FY2001 Progress Report for Automotive Lightweighting

FY2001 Progress Report for Automotive Lightweighting
• Office of Advanced Automotive
Technologies FY 2001 Program Highlights
• Vehicle Propulsion and Ancillary Subsystems
• Automotive Lightweighting Materials
• Automotive Propulsion Materials
• Fuels for Advanced CIDI Engines and Fuel Cells
• Combustion and Emission Control for
Advanced CIDI Engines
• Fuel Cells for Transportation
• Advanced Technology Development
(High-Power Battery)
• Batteries for Advanced Transportation
Technologies (High-Energy Battery)
• Vehicle Power Electronics and Electric Machines
• Vehicle High-Power Energy Storage
AUTOMOTIVE L IGHTWEIGHTING M ATERIALS
• Spark Ignition, Direct Injection Engine R&D
2001 A NNUAL P ROGRESS R EPOR T
Office of Transportation Technologies
Series of 2001 Annual Progress Reports
AUTOMOTIVE
LIGHTWEIGHTING
MATERIALS
2001
ANNUAL
PROGRESS
REPORT
• Electric Vehicle Batteries R&D
U.S. Department of Energy
Energy Efficiency and Renewable Energy
Office of Transportation Technologies
www.cartech.doe.gov
DOE/EERE/OTT/OAAT - 2001/003
ACKNOWLEDGEMENT
We would like to express our sincere appreciation to Argonne
National Laboratory and Computer Systems Management, Inc.,
for their artistic contributions in preparing the cover of this
report and to Oak Ridge National Laboratory for its technical
contributions in preparing and publishing this report.
In addition, we would like to thank all our program participants
for their contributions to the programs and all the authors who
prepared the project abstracts that comprise this report.
This document highlights work sponsored by agencies of the U.S. Government. Neither the U.S.
Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied,
or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately
owned rights. Reference herein to any specific commercial product, process, or service by trade name,
trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the U.S. Government or any agency thereof. The views and opinions of authors
expressed herein do not necessarily state or reflect those of the U.S. Government or any agency thereof.
Printed on recycled paper
U.S. Department of Energy
Office of Advanced Automotive Technologies
1000 Independence Avenue S.W.
Washington, DC 20585-0121
FY 2001
Progress Report for Automotive Lightweighting Materials
Energy Efficiency and Renewable Energy
Office of Transportation Technologies
Office of Advanced Automotive Technologies
Vehicle Systems Team
Robert Kost
January 2002
Vehicle Systems Team Leader
This report was prepared as an account of work sponsored by an agency of the
United States Government. Neither the United States Government nor any
agency thereof, nor any of their employees, makes any warranty, express or
implied, or assumes any legal liability or responsibility for the accuracy,
completeness, or usefulness of any information, apparatus, product, or process
disclosed, or represents that its use would not infringe privately owned rights.
Reference herein to any specific commercial product, process, or service by trade
name, trademark, manufacturer, or otherwise, does not necessarily constitute or
imply its endorsement, recommendation, or favoring by the United States
Government or any agency thereof. The views and opinions of authors expressed
herein do not necessarily state or reflect those of the United States Government
or any agency thereof.
Automotive Lightweighting Materials
FY 2001 Progress Report
CONTENTS
1.
INTRODUCTION................................................................................................................................
1
2.
AUTOMOTIVE ALUMINUM R&D .................................................................................................
9
A.
B.
C.
D.
E.
F.
G.
H.
3.
Forming Advances for Aluminum Automotive Components and Applications ............................
Active Flexible Binder Control System for Robust Stamping .......................................................
Warm Forming of Aluminum—Phase 2 ........................................................................................
Hydroforming of Aluminum Tubes................................................................................................
Electromagnetic Forming of Aluminum Sheet...............................................................................
Optimization of Extrusion Shaping: Aluminum Tubular Hydroforming .......................................
Improved A206 Alloy for Automotive Suspension Components...................................................
Die-Casting Die Life Extension .....................................................................................................
ADVANCED MATERIALS DEVELOPMENT................................................................................ 51
A.
B.
C.
D.
E.
F.
Low-Cost Powder Metallurgy for Particle-Reinforced Aluminum Composites ............................
Low-Cost Cast Aluminum Metal Matrix Composites....................................................................
Structural Cast Magnesium Development ......................................................................................
Magnesium Powertrain Cast Components .....................................................................................
Advanced Magnetherm Process for Production of Primary Magnesium .......................................
Solid Oxygen-Ion-Conducting Membrane Technology for Direct Reduction
of Magnesium from Its Oxide at High Temperatures.....................................................................
G. Understanding the Economics of Emerging Titanium Production Processes ................................
H. Structural Reliability of Lightweight Glazing Alternatives............................................................
4.
9
13
21
25
29
33
39
45
51
55
61
67
71
73
81
85
POLYMER COMPOSITES R&D...................................................................................................... 93
A. Development of Manufacturing Methods for Fiber Preforms ........................................................ 93
B. Composite-Intensive Body Structure Development for Focal Project 3 ........................................ 97
C. Study of Thermoplastic Powder-Impregrated Composite Manufacturing Technology
for Automotive Applications .......................................................................................................... 101
5.
LOW-COST CARBON FIBER .......................................................................................................... 107
A.
B.
C.
D.
E.
Low-Cost Carbon Fibers from Renewable Resources....................................................................
Low-Cost Carbon Fiber Development Program.............................................................................
Low-Cost Carbon Fiber for Automotive Composite Materials......................................................
Economical Carbon Fiber and Tape Development from Anthracite Coal
Powder and Development of Polymer Composites Filled with
Exfoliated Graphitic Nanostructures ..............................................................................................
Microwave-Assisted Manufacturing of Carbon Fibers ..................................................................
iii
107
111
123
127
133
FY 2001 Progress Report
6.
Automotive Lightweighting Materials
RECYCLING ....................................................................................................................................... 137
A. Sorting Mixed Alloys from Shredded Automobiles ......................................................................
B. Recycling of Polymer Matrix Composites.....................................................................................
C. Investigation of the Cost and Properties of Recycled Magnesium Die Casting for
Employment in Automotive Applications .....................................................................................
D. Recycling Assessments and Planning ............................................................................................
7.
147
153
ENABLING TECHNOLOGIES......................................................................................................... 157
A. Durability of Carbon-Fiber Composites.........................................................................................
B. Creep, Creep Rupture, and Environment-Induced Degradation of Carbon and
Glass-Reinforced Automotive Composites....................................................................................
C. NDE Tools for Evaluation of Laser-Welded Metals (Steel and Aluminum) .................................
D. Modeling of Composite Materials for Energy Absorption ............................................................
E. Composite Crash Energy Management..........................................................................................
F. Intermediate-Rate Crush Response of Crash Energy Management Structures..............................
G. Long-Life Electrodes for Resistance Spot Welding of Aluminum Sheet Alloys
and Coated High-Strength Steel Sheets .........................................................................................
H. Plasma Arc Welding of Lightweight Materials .............................................................................
I. Performance Evaluation and Durability Prediction of Dissimilar Material Hybrid Joints ............
J. Joining of Dissimilar Metals for Automotive Applications: From Process
to Performance ...............................................................................................................................
K. Technical Cost Modeling ...............................................................................................................
L. Nondestructive Evaluation Techniques for On-Line Inspection of Automotive Structures ..........
8.
137
143
157
163
169
173
177
181
185
189
191
197
203
207
HIGH-STRENGTH STEELS ............................................................................................................. 215
A.
B.
C.
D.
E.
F.
G.
H.
I.
Enhanced Forming Limit Diagrams...............................................................................................
High-Strength Steel Stamping Project ...........................................................................................
Hydroform Materials and Lubricants Project ................................................................................
Sheet Steel Joining Technologies ..................................................................................................
Sheet Steel Fatigue Characteristics ................................................................................................
Strain Rate Characterization ..........................................................................................................
High-Strength Steel Tailor-Welded Blanks ...................................................................................
Tribology........................................................................................................................................
Modeling of High-Strain-Rate Deformation of Steel Structures ...................................................
215
217
219
221
223
227
229
233
235
Appendix A: ACRONYMS AND ABBREVIATIONS............................................................................. 247
iv
Automotive Lightweighting Materials
FY 2001 Progress Report
1. INTRODUCTION
Automotive Lightweighting Materials R&D
As a major component of the Office of Advanced Automotive Technologies (OAAT) in the U.S. Department
of Energy’s (DOE’s) Office of Transportation Technologies (OTT), the Automotive Lightweighting Materials
(ALM) Program focuses on the development and validation of advanced lightweight materials technologies to
significantly reduce automotive vehicle body and chassis weight without compromising other attributes such
as safety, performance, recyclability, and cost. Through many of its technology research programs, OAAT
has supported the government/industry Partnership for a New Generation of Vehicles (PNGV) since its
inception. The PNGV leadership is now reevaluating the partnership goals to identify changes that will
maximize the potential national petroleum-savings benefit of the emerging PNGV technologies. When these
PNGV goal changes have been defined, the OAAT will adjust the focus of its technology research programs
accordingly.
The specific goals of the ALM program are (1) by FY 2000, develop and validate materials technologies that,
if implemented, will enable a reduction in automobile body and chassis weight by 50% relative to a 1994
baseline automobile, at 1.5 times the cost of using the baseline materials, while maintaining safety, reliability,
and recyclability; and (2) by 2004, improve materials technology to enable a 50% reduction in the weight of
automobile body and chassis components, and a 40% reduction in the overall vehicle weight (PNGV target),
while achieving cost competitiveness with conventional materials.
The program is pursuing five areas of research: cost reduction, manufacturability, design data and test
methodologies, joining, and recycling and repair. The single greatest barrier to use of lightweight materials is
their high cost; therefore, priority is given to activities aimed at reducing costs through development of new
materials, forming technologies, and manufacturing processes. Priority lightweight materials include
aluminum, magnesium, titanium, and composites such as metal-matrix materials and glass- and carbon-fiberreinforced thermosets and thermoplastics.
Collaboration and Cooperation
The ALM Program collaborates and cooperates extensively in order to identify and select its research and
development (R&D) activities and to leverage those activities with others. The primary interfaces have been
and still are with the Big Three domestic automotive manufacturers, namely the PNGV Materials Technical
Team, the Automotive Composites Consortium (ACC) and the United States Automotive Materials Partnership (USAMP). This collaboration provides the means to determine critical needs, to identify technical barriers, and to select and prioritize projects. Other prominent partners include such organizations as the Aluminum Association, the American Iron and Steel Institute, the American Plastics Council, the Vehicle Recycling
Partnership, the Society for the Advancement of Materials and Process Engineering, the International Magnesium Association, the International Titanium Association, and the Auto Parts Rebuilders Association. The
program also coordinates its R&D activities with entities of other U.S. and Canadian federal agencies. Interaction with the DOE Office of Industrial Technologies (OIT) and Office of Heavy Vehicle Technologies
(OHVT) and with the Department of Natural Resources of Canada is especially important by virtue of
overlaps of interests in lightweight materials.
Once selected, R&D projects are pursued through a variety of mechanisms, including cooperative research
and development agreements (CRADAs), cooperative agreements, university grants, R&D subcontracts, and
directed research. This flexibility allows the program to select the most appropriate partners to perform
critical tasks. The ALM efforts are conducted in partnership with automobile manufacturers, materials
suppliers, national laboratories, universities, and other nonprofit technology organizations. These interactions
1
FY 2001 Progress Report
Automotive Lightweighting Materials
provide a direct route for implementing newly developed materials and technologies. Laboratories include
Albany (Oregon) Research Laboratory, Ames Laboratory (Ames), Argonne National Laboratory (ANL),
Lawrence Berkley National Laboratory (LBNL), Lawrence Livermore National Laboratory (LLNL), Los
Alamos National Laboratory (LANL), Oak Ridge National Laboratory (ORNL), Pacific Northwest National
Laboratory (PNNL), and Sandia National Laboratories (SNL). PNNL manages the Northwest Alliance for
Transportation Technologies, drawing on the expertise and developments in the Northwest. ANL oversees
recycling efforts, and ORNL provides overall technical management, including management for the DOE
cooperative agreement with USAMP.
Research areas and responsible organizations
Coordinated area
Organization
Production and fabrication of aluminum
The Aluminum Association, OHVT, OIT, Natural
Resources of Canada (NRCAN)
International Magnesium Association, NRCAN
Production and fabrication of magnesium
Recycling, reuse, repair of automotive parts and
materials
Fabrication of steel and cast iron
Electric and hybrid vehicle technology program
Auto Parts Rebuilders Association, OIT, Vehicle
Recycling Partnership
American Iron and Steel Institute, the Auto/Steel
Partnership
DOE Office of Energy Research, National Science
Foundation
Department of Commerce—National Institute of
Standards and Technology’s Advanced Technology
Program
Department of Transportation
Materials research for defense applications
Department of Defense
Materials research for space applications
National Aeronautics and Space Administration
Crashworthiness
Department of Transportation
International vehicle material R&D
International Energy Association
Production and fabrication of titanium
The International Titanium Association
Production and fabrication of composites
American Plastics Council
Fundamental materials research
High-volume composite processing
FY 2001 Accomplishments
To meet the goals set forth for the program, the ALM Program is developing materials and material processing technologies; validating these technologies through fabrication and evaluation of representative, nonproprietary test components; and developing adequate design data to facilitate their beneficial application.
The research is balanced between nearer-term objectives and longer-term, higher-risk research. As the technical barriers are removed, the technology is made available to industry. Because of the broad area of research
and the limited resources, projects have been selected to overcome the most significant barriers within the
technical areas that the materials community considers higher-risk but that, if successfully developed, would
result in significant progress toward program goals.
2
Automotive Lightweighting Materials
FY 2001 Progress Report
Composite Materials
Low-Cost Carbon Fiber
One of the highest-priority technical needs is the development of low-cost carbon fibers. Carbon fiber
composites are the lightest material available for making primary automotive structures. Their use in
automotive structural composites could reduce the body and chassis weight of vehicle components by up to
67%. Currently, the use of these advanced lightweight composites in primary automotive structures is limited
because carbon fibers are much more expensive than traditional automotive materials. To address this
challenge, research is being conducted to reduce the cost of materials being converted into carbon fiber and
the cost of the production processes used for making carbon fiber.
Originally, the program had four projects aimed at developing low-cost carbon fiber precursors. During the
last year, that field was down-selected to the two most promising technologies. The two choices were a
nearer-term, medium-risk project using commodity-grade polyacrylonitrile (PAN) precursor that could reduce
finished fiber costs by $1.60 per pound; and a longer-term, higher-risk project using lignin-based precursors
that could reduce the price of carbon fiber to near $3.00 per pound.
Textile-Grade PAN Precursors
During 2001 a method was successfully developed to produce lower-cost carbon fiber precursor using
commodity-grade PAN. Commodity textile-grade PAN is commercially available in large quantities at about
half the cost of large-tow PAN precursors. Pretreatment and processing methods were developed using
textile-grade PAN precursors. Appropriate mechanical and chemical properties were determined, and a
laboratory-scale proof of concept was demonstrated. The next step will be to scale this technology up for
industrial applications. These new carbon fiber precursors should complete full-scale development within
18 months and be ready for implementation into production facilities shortly afterward. The resulting
reduction in the price of carbon fiber is projected to be $1.60 per pound.
Textile PAN being processed into carbon fiber.
Before- and after-processing views of carbon fiber
made from textile-grade PAN.
Lignin-Blend Carbon Fiber Precursors
It is expected that commodity-grade textile precursor will allow carbon-fiber–based composites to penetrate
the automotive market at significant levels by allowing the $5.00 per pound goal for carbon fiber to be met.
However, for carbon fiber composites to become more prevalent as structural materials, prices of $3.00 per
pound must be approached. The second precursor project seeks to reach that goal by using lignin as the
feedstock. Lignin is a by-product of papermaking and is often burned in paper mills to recover its energy
value. It can be purchased on the open market for $40 per ton. During this year, very inexpensive feedstock
3
FY 2001 Progress Report
Automotive Lightweighting Materials
materials were desalted and purified, and processing schedules were determined. Lignin blends were
successfully spun, stabilized, carbonized, and graphitized in small batches. It was demonstrated that goodquality carbon fiber can be made from lignin feedstocks and that the process is economically viable.
Carbonized fibers from lignin blends.
As-carbonized surface of a polyethylene oxide/
lignin-blend fiber.
Microwave-Assisted Plasma Processing of
Carbon Fiber Precursors
In addition to developing lower-cost precursors, the
program has been working on developing lower-cost
production methods for converting precursors into carbon
fiber. In previous years, a process was demonstrated that
used microwave energy to produce carbon fibers; it
required processing times that were only a fraction of those
necessary using conventional methods. The tow size of
carbon fibers was small; and the process, after initial
development, was small and the output line speed still
much slower than desired. During the past year, the
Microwave-assisted plasma carbon fiber
microwave-assisted plasma process for manufacturing
production unit.
carbon fiber using microwave energy was greatly
improved. Scale-up of the process was conducted this year
using large-tow (48,000-filament) precursors. Line speeds were increased from 4 inches to 48 inches per
minute, and fiber quality was improved to match that of fiber commercially available from carbon fiber
vendors. Development of this technology should be completed in the next 18 months.
Commercialization and Advancement of the P4 Process
The programmable powder preform process (P4) is a low-waste, high-speed method of making fiber performs
that was developed under this program. (Preforms are the fiber mats that are infiltrated with a resin to make a
composite.) In earlier work, the economic viability of P4 was demonstrated using glass fiber; this year two
variations of this technology were commercialized on vehicle platforms. In addition, this year the process for
making high-quality composite performs using carbon fiber was advanced past the proof-of-concept stage.
The project identified the best fibers and made the first carbon fiber composite performs. The preforms were
then molded into composite test panels.
4
Automotive Lightweighting Materials
FY 2001 Progress Report
Aston Martin Vanquish with front quarter panels
made using P4.
Chevy 1500 with pickup truck bed made using P4.
First Durability Guidelines Delivered for Designing with Carbon Fiber Composites
An ongoing program effort is to develop the durability-based design guidelines necessary to enable
automotive engineers to design automotive structures made from composites and be assured that the
structures will last for the life of the vehicle. To do so, it is necessary to account for the effects of all the
various load histories, temperature extremes, and environmental stressors that a vehicle may see and then
develop an understanding of their synergistic effects. In previous years, design guidelines were developed for
glass-fiber composites. During this year, the program completed the first durability design guideline for
carbon fiber composites. The first design guidance manual was distributed to all three domestic automakers; it
includes the synergistic effects of multiple-load, environmental, and temperature histories on carbon-fiber–
based composite materials used for primary automotive structures.
Completed Phase 1 of Focal Project 3
The program integrates all development projects into large demonstration projects that demonstrate that
structures can be manufactured and assembled within required cycle times and at required costs while
meeting performance targets acceptable to industry. The third focal project is the development of an allcarbon-fiber composite body-in-white. During this year, phase 1 of Focal Project 3 was completed. The
resulting structure met or exceeded all design requirements and can be cost-effectively manufactured from
carbon fiber composites. It is more than 60% lighter than the baseline steel structure. The figure on page 6
shows the structure and the amount of material used in it.
Metals and Alloys
Low-Cost Powder Metallurgy for Particle-Reinforced Aluminum Composites
Powder metallurgy particle-reinforced aluminum (PMPRA) composites exhibit a number of benefits for
automotive applications. Unfortunately, several barriers, including cost, durability, and manufacturability,
have limited their widespread implementation. A project led by USAMP has focused on developing the
manufacturing methods, machining technology, and design methodologies necessary to establish the
commercialization potential of these materials. The PMPRA project is nearing completion. Major process and
material development milestones were achieved in FY 2001. The press and sinter sub-team has fabricated an
aluminum-silicon-alloy PMPRA transmission oil pump gear set with new dimensionally accurate tooling to
demonstrate the newly developed technology. Parts have passed two major durability hurdles, including wear
tests. In addition, a set of sintering process conditions based on correlations to wear-resistant alloy
microstructures and calorimetry measurements have been developed. The direct powder forging sub-team has
established the feasibility of making connecting rods from PMPRA composites by demonstrating the
fabrication of components in small lots using existing (non-optimized) tooling with an aluminum-SiC powder
5
FY 2001 Progress Report
Automotive Lightweighting Materials
Phase 1 results:
67% mass savings over baseline
Bending stiffness exceeded 20%
Torsional stiffness exceeded 140%
Durability and abuse load cases satisfied
Manufacturing strategy developed
Materials/mass distribution:
Chopped carbon—54.8 kg
Carbon fabric—17.7 kg
Core–3.2 kg
Adhesive—1.6 kg
Inserts—8.8 kg
blend and cold-isostatically-pressed preforms. The results of roomtemperature tensile and fatigue tests and microstructural characterization
indicate a high probability of achieving design and performance strengths.
These activities have demonstrated a high potential for using powder
metallurgy aluminum composite materials in both moderate-strength and
high-strength/high-temperature applications in high-volume production.
Low-Cost Cast Aluminum Metal Matrix Composites
In addition to the activity focused on powder-metallurgy–based metal
matrix composites (MMCs), efforts have been supported in the
development of low-cost cast aluminum MMCs. This project seeks to
make aluminum MMC materials more attractive for widespread use in
automotive applications by reducing the costs of the raw materials, the
processing methods, and the machining and finishing operations. Initial
activities have resulted in the selection and validation of lower-cost SiC
reinforcement materials and the scale-up of the mixing and holding
system for producing low-cost castings from 60 to 600 kg. Other
activities focused on the development of innovative shape-casting and
Prototype 600-kg modular
molding technologies—including centrifugal casting, pressure-infiltrated
mixing system built by MC-21:
melter (rear), mixer (middle),
preforms, and ceramic-based composite samples—to produce aluminum
and holding furnace
MMC automotive braking systems. With the promise of a lower-cost
(foreground).
aluminum MMC material, a less expensive compositing process, and
supporting work on downstream processing, future efforts will focus on
innovative brake system designs that take advantage of the unique
properties of MMC materials to produce cost-effective, lightweight replacements for cast iron systems.
Improved A206 Alloy for Automotive Suspension Components
Aluminum alloy A206 is significantly stronger than the aluminum casting alloys normally used at present and
has mechanical properties approaching those of some grades of ductile iron. It also has excellent hightemperature tensile and low-cycle fatigue strength. Consequently, this material has potential for use in a
number of applications to reduce vehicle weight. Cost savings may also result because less material would be
required to provide the strength needed for the application. In spite of its excellent properties, A206 alloy has
not been used in significant quantities because of its propensity for hot cracking. The ALM Program is
supporting a project team, under the direction of USAMP, to investigate the potential for improving the
castability of A206 using a better method of grain refinement and a new ultrasonic inspection technique to
6
Automotive Lightweighting Materials
FY 2001 Progress Report
Lower control arm castings produced in Phase 1 trials. (Left) A front lower control arm for a
GM Grand Am. (Right) A rear lower control arm for a GM Cadillac.
test for the presence of hot cracks. Suspension components have been successfully cast using a new
experimental version of A206. Initial results indicate higher resistance to hot cracking and mechanical
properties superior to those of conventional A356 alloy parts, even though the casting process had not been
optimized. Further experiments were conducted to establish optimum process parameters for the production
of more than 100 castings with little or no porosity under production conditions. Continued testing is
expected to result in a database that the automotive industry can use to design and cast lighter-weight, higherstrength components.
Solid Oxygen-Ion-Conducting Membrane Technology
Because of their low densities, magnesium alloys have the
potential for application in many automotive components. One
of the barriers to the increased use of magnesium in automotive
applications is the relatively high cost of the primary metal. The
current methods for producing magnesium are either electrolysis
from a halide electrolyte bath, which requires extensive and
expensive feed material preparation, or metallothermic reduction
at high temperatures, which involves expensive metal reductant.
The ALM Program supports efforts aimed at developing a solid
oxygen-ion-conducting membrane (SOM) process for producing
magnesium from its oxide. The SOM process has the potential to
be more economical, less energy-intensive, and more
environmentally sound. In work to date, high magnesium
production rates have been achieved in laboratory-scale
demonstrations without damaging the yittria-stabilized zirconia
membrane or other components of the cell. Steady state
operation of the laboratory-scale cell has indicated superior
performance compared with state-of-the-art processes for
magnesium production and with Hall Cell aluminum production.
Experiments to determine parameters critical to the potential
scale-up of this technology for commercial magnesium
production will continue.
7
Apparatus for the laboratory-scale
electrolysis and collection of magnesium
metal using the SOM technology.
FY 2001 Progress Report
Automotive Lightweighting Materials
Roadmapping and R&D Planning
A report on initial planning was completed with the American
Plastics Council (APC). The report was begun at a May 2000
workshop. A report on recycling, based on a September 2000
workshop, was completed as a joint effort of several interested
organizations, such as the APC, the Aluminum Association, the
American Iron and Steel Institute and the Auto Parts Rebuilders
Association. The planning report is particularly notable in that it
seems to have stimulated further planning on the part of the plasticssupplier industry in this area of plastics for the automotive market.
The first comprehensive program review by an independent panel of
outside experts (outside the program) was completed. The panel
consisted of four retired materials and manufacturing
researcher/manager experts from the auto industry, one from the
aluminum industry, one from the steel industry and one from the U.S.
Department of Defense.
Future Direction
FY 2001 marked the completion of a ramp-down of R&D thrusts
begun around FY 1994–1998, and completion of a ramp-up of new
thrusts begun around FY 1999. Most projects—with the exception of
one effort on casting of unreinforced aluminum and all projects on
glass-fiber-reinforced polymer-matrix composites (PMCs)—were
completed or began winding down toward final wrap-up in FY 2002.
New projects on magnesium casting and on carbon-fiber-reinforced
PMCs came up to full funding levels. These new efforts will demand
most of the funds available in FY 2002 and most funds likely to be
available in FY 2003. FY 2001 claims to have developed titanium
production processes with potential costs compatible with some
(higher-value) automotive applications were confirmed. Some of the
titanium being produced will be assessed for suitability for
fabrication in FY 2002 in anticipation of further escalation of the
titanium activities in late FY 2002 or FY 2003. These events
represent notable shifts away from aluminum—the prime candidate
for meeting the original PNGV goal of a cost-competitive 40%
weight reduction—to less-explored and higher-risk materials thought
capable of enabling even greater weight reductions over the longer
term.
Planning report completed with
American Plastics Council.
Roadmapping report for recycling of
vehicle materials.
Seven new projects on forming of components out of sheet steel or aluminum began in FY 2001, as did five
on joining of steel, aluminum, and magnesium to one another and to PMCs. This effort signifies another
programmatic trend toward downstream manufacturing, as opposed to the emphasis on upstream aluminum
sheet production and alloy development during the earlier years of the ALM Program. The multi-material
nature of the joining projects highlights still another trend that arises from recent thinking that various
materials will compete for use in future vehicles, without domination by just one, steel, as in the past.
FY 2002 will be a year for planning future new efforts in the areas of polymers and recycling, based on the
FY 2000 and FY 2001 roadmapping efforts (see Roadmapping and R&D Planning), that could be addressed
as the new thrusts mentioned wind down. Further R&D efforts in the areas of MMCs, glass, and
nondestructive evaluation will be considered as the present efforts in these areas wind down in FY 2002.
8
Automotive Lightweighting Materials
FY 2001 Progress Report
2. AUTOMOTIVE ALUMINUM R&D
A. Forming Advances for Aluminum Automotive Components and Applications
Robert P. Evert
Alcoa, Inc., Alcoa Technical Center
100 Technical Drive, Alcoa Center, PA 15069
(724) 337-2410; fax: (724) 337-2255; e-mail: [email protected]
DOE Program Manager: Joseph A. Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: Alcoa, Inc.
Contract No.: XSU544C
Objectives
•
Develop and demonstrate advanced forming processes to improve the formability of aluminum sheet for
automotive panel applications.
•
Improve the aluminum sheet metal stamping process through the development of a method for controlling the
flow of metal into the tooling cavity during the stamping process through the application of variable binder
loads.
OAAT R&D Plan: Task 2; Barriers A, B
Accomplishments
•
Conducted single-cylinder testing and development at Erie Press to resolve a technical difficulty in controlling
the binder cylinders that was uncovered during the full-scale press demonstration trials.
•
Successfully reprogrammed the binder cylinder control system to obtain satisfactory results under actual press
conditions.
Future Direction
•
Ship unit to Troy Design and Manufacturing (TDM) for next phase press evaluations.
•
Complete a final report on the technical and economic evaluations.
2000. This report details progress on the
process/press optimization project.
Introduction
More extensive use of aluminum on vehicles has
been identified as an important means of reducing
tailpipe emissions and fuel consumption. This
contract comprised two distinct projects,
process/press optimization and warm forming. Work
on the warm forming project was completed in
Research Results
The full-scale press demonstration trials at Troy
Design and Manufacturing (TDM) in 2000
uncovered a technical difficulty in controlling the
9
FY 2001 Progress Report
Automotive Lightweighting Materials
binder cylinders in the last 1–2 inches of the press
stroke. The project team determined that the control
system required re-programming using the control
valves’ actual pressure drop at various flow rates.
The optimal procedure to accomplish this was to
pull the full-size binder control system from the
press at TDM and work on a single cylinder at Erie
Press.
One complete binder cylinder assembly was
removed from the fixture and placed into a test
stand. The test stand consisted of a larger hydraulic
cylinder directly opposed to the binder cylinder. The
binder cylinder received its electrical commands and
hydraulic fluid from the existing binder control
system. The control of the opposed cylinder was
derived from an external source. This control
allowed the velocity of the opposed cylinder to be
programmed.
A servo valve opening\flow\pressure drop
spreadsheet was acquired from the valve
manufacturer. These data allowed modification of
the control algorithm to predict the valve opening
required while maintaining binder cylinder force
during the slower press velocities.
Initial tests were performed with the opposed
cylinder traveling at constant velocities. The results
were favorable, especially at the slower velocities
that were not under control during the in-forming
trial at TDM.
For the next test, we programmed the opposing
cylinder to mimic the velocity curve of the press at
TDM. It was observed that the binder cylinder
pressure fell further below the set point as the
velocity was reduced. The problem was traced to
binder cylinder speed reference calculation. The
speed reference was derived by examining the
feedback from the binder fixture analog position
sensor and calculating the rate of change during
every controller scan. The position sensor’s
feedback contained a substantial amount of electrical
noise that required us to incorporate digital filtering
to produce a stable result. This filtering also
dampened the speed reference signal’s response to a
change in velocity. Then a digital encoder and PC
card were added to allow acquisition of digital
(noise-free) position information at a 20- V rate.
Testing resumed with the new encoder and PC
card installed and the opposing cylinder pushing the
binder cylinder down at the press velocity. It was
observed that when the binder cylinder was first
contacted, pressure control was satisfactory.
However, as the stroke increased, the cylinder
pressure became unstable. It was surmised that the
sensitivity of the new speed sensor allowed the
binder cylinder control system to respond much
faster than did the previous method, and that it was
compensating for the velocity changes (increasing
and decreasing) as the opposed cylinder tried to
maintain the press velocity curve. At that time, it
was concluded that the simple hydraulic cylinder
setup was not capable of simulating the inertial
characteristics and velocity curve of a large
mechanical press.
Next, the single binder cylinder assembly was
installed in a 4000-T forging press. This test would
eliminate the possibility that the binder cylinder
pressure changes would affect the velocity of press
and thus induce system instability. The press was
operated at about 1/3 its normal speed in an attempt
to match the TDM press velocity curve. The initial
tests were encouraging, but the feed forward
algorithm (derived from the valve manufacturer’s
data) was causing the pressure to rise significantly
above the set point as the press reached bottom dead
center. The data gathered from this test allowed the
team to derive the proper equation needed to
improve the feed forward algorithm.
With the new equation in place, testing resumed
with acceptable results. Successful pressure profiles
were obtained throughout the entire press stroke for
the three pressure trajectories required from the
finite element modeling of the process. Figure 1
shows the three prescribed pressure curve
trajectories for groups of cylinders. Figures 2–4
show the actual single-cylinder pressure output
under actual press conditions. (Note that 14.22 psi =
1kN.) This test completed the single-cylinder testing
and development at Erie Press, and the unit was
prepared to ship back to TDM for press trial
reevaluations. At this time, the program
management transferred to Mahmoud Demeri of
Ford Motor Company.
10
Automotive Lightweighting Materials
FY 2001 Progress Report
Figure 3. Actual cylinder pressure for pin group 2.
Figure 1. Recommended load trajectories from finite
element modeling.
Figure 2. Actual cylinder pressure for pin group 1.
Figure 4. Actual cylinder pressure for pin group 3.
11
Automotive Lightweighting Materials
FY 2001 Progress Report
B. Active Flexible Binder Control System for Robust Stamping
Project Leader: Mahmoud Y. Demeri
Sr. Technical Specialist, Manufacturing Systems Department
Ford Research Labs, Ford Motor Company
2101 Village Road
Dearborn, MI 48124
(313) 845-6092; fax: (313) 390-0514; e-mail: [email protected]
DOE Program Manager: Joseph Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: U.S. Automotive Materials Partnership
Contract No.: DE-FC05-95OR22363
Objective
•
Develop and demonstrate, on an industrial scale, an optimized closed-loop flexible binder control system that
can be installed in presses to improve the quality, reduce the variability and maintain the accuracy of stampings
made from aluminum alloys and ultra-high-strength and stainless steels. The system will also reduce the cost
for developing and setting production tools.
OAAT R&D Plan: Task 2, 13; Barrier B
Approach
•
Conduct open-loop control demonstration of flexible binder technology.
•
Develop methodology and guidelines for designing and building flexible binders.
•
Develop computer simulation and process optimization capabilities for flexible binders.
•
Develop a closed-loop flexible binder control system with appropriate sensors.
•
Demonstrate closed-loop control of the flexible binder system on an industrial part.
•
Evaluate technical and economic feasibility of flexible binder technology.
Accomplishments
•
Completed project team formation and selection of academic and industrial partners.
•
Initiated the project on March 20, 2001.
•
Conducted monthly project, steering committee, and task meetings.
•
Received from the Institute for Metal Forming Technology the design guidelines for segmentation of lower
flexible binders and structural designs for rigid lightweight upper binders.
•
Developed an original equipment manufacturer–approved 5-phase project plan for development of the
computer simulation and process optimization capabilities. (Development work done by Ohio State University).
•
Built tooling for testing single-point binder force control for axi-symmetric parts.
13
FY 2001 Progress Report
Automotive Lightweighting Materials
•
Conducted computer simulation on blank holder force (BHF) optimization of the conical cup.
•
Evaluated geometry- and energy-based wrinkling detection methods.
•
Investigated ductile fracture and thinning as criteria for splits.
control in the stamping industry to extend the
forming window of difficult-to-form materials,
thereby increasing the robustness of the forming
process and improving the quality and consistency
of stampings. This technology will use computer
simulation and process optimization to predict the
optimum binder force trajectory on hydraulic
cushions with programmable multi-pins on
mechanical transfer presses.
Introduction
Significant weight savings can be achieved by
replacing parts made from mild steel with those
made from lightweight materials (aluminum and
magnesium alloys) and high-specific-strength
materials (ultra-high-strength and stainless steels).
Such materials are less formable than mild steel, and
parts made from them lack dimensional control
because of the significant amount of springback that
they produce after forming.
Traditional stamping leaves no flexibility in the
stamping process for using difficult-to-form
materials and for responding to process variations
(e.g., lubrication, material, die wear, blank
placement) that can lead to stamping inconsistencies
or even failure. It has been found that failure by
wrinkling or tearing is highly dependent on the
magnitude and trajectory of the binder force.
Recently, dynamic variation of the binder force
during the forming stroke has been shown to affect
formability, strain distribution, and springback.
Optimal forming trajectories can be obtained under
constant and variable binder force conditions, but
there is no guarantee that process variables will
remain constant during the stamping process.
Specifying a binder force trajectory is not easy
because the part shape changes during forming.
Also, stresses in the part cannot be determined
because the coefficient of friction is not a
controllable quantity and it varies from location to
location. Therefore, the forming process must be
controlled and a closed loop system with an
appropriate local control parameter (friction, drawin) must be used to track a predetermined optimum
control parameter trajectory.
Details of Project and Team Formation
Significant effort was spent in detailing project
tasks and in selecting team members. A steering
committee for the project was formed, task leaders
were identified, a project administrator was hired,
and academic and industrial partners were selected.
The project was initiated on March 20, 2001.
Monthly project and steering committee meetings
were conducted and a number of task meetings were
convened. The meetings reflected a high level of
participation and an increased interest in binder
control technology. Tasks 1, 2 and 3 are being
conducted simultaneously.
Task 1: Conduct Open-Loop Trials Using
Flexible Binder and Liftgate Tooling
Task 1 was inherited from an earlier U.S.
Automotive Materials Partnership/DOE project that
ended in December 2000. Proceeding with Task1
was contingent upon receiving a functional binder
control unit from that project. The control unit was
designed and built by Erie Press Systems, and
demonstration trials were planned at Troy Design
and Manufacturing (TDM). Initial tests showed that
the control unit did not function properly; therefore,
it was sent back to Erie Press to find and fix the
problem. Erie Press reprogrammed, reassembled,
and retested the control unit and returned it to TDM
for installation in a mechanical press to conduct
tryouts on the liftgate inner tooling. Tests conducted
in March 2001 showed a serious problem with
controlling individual hydraulic cylinders in the
binder control unit. The problem has been identified
Scope
The project scope involves using flexible binder
control technology in conjunction with innovative
tool designs and closed-loop control to produce
robust processes for stamping aluminum and highstrength-steel automotive panels. The focus of this
project is to implement binder and feedback process
14
Automotive Lightweighting Materials
FY 2001 Progress Report
•
•
as a control problem: pressure within the hydraulic
cylinders could not be achieved in a mechanical
press with a simple proportional integral derivative
(PID) control because the punch velocity is not
constant as it is in hydraulic presses. A number of
options are being considered to surmount the
problem. It is important to realize that Task 1 cannot
proceed until a functioning, reliable binder control
unit becomes available.
•
Allowable stresses and strains in flexible binders
Computer-assisted designs for flexible binder
tools
Hydraulic cylinder and pin sizes, capacity and
characteristics
Task 2 is proceeding according to plan. Some of
the accomplishments include design guidelines for
convex, straight, and concave segments of lower
flexible binders (Figures 2–4) and honeycomb,
triangular, and circle designs for rigid lightweight
upper binders (Figure 5).
Task 2: Develop Guidelines for Designing
and Building Flexible Binders
The Institute for Metal Forming Technology of
the University of Stuttgart in Germany pioneered the
development of flexible binder technology. The
Institute was selected to work on Task 2. Flexible
binder technology uses innovative tooling designs
(Figure 1). The lower binder has cone-shaped
segments to provide local binder force control, and
the upper binder employs structures to provide
rigidity and light weight. Building flexible binders
requires optimization of the tool designs to achieve
flexibility in the lower binder and lightweight
rigidity in the upper binder.
Figure 2. Design for lower binder.
Figure 1. Flexible binder technology with flexible lower
and rigid upper binders.
Task 2 objectives are to develop guidelines for
designing and building flexible binders. Specific
areas of development include these:
•
•
•
•
Figure 3. Designs for lower binder segments.
Binder segmentation
Cone shape, numbers, and size
Lower and upper binder material
Effect of stress concentration and fatigue on
flexible binders
15
FY 2001 Progress Report
Automotive Lightweighting Materials
related to temporal binder force control, such as
fracture prediction, wrinkling detection, and
temporal binder force control strategies. Several
fracture prediction and wrinkle detection methods,
based on part geometry and energy, were analyzed.
A conical cup-drawing tool was designed and
manufactured for use in developing temporal binder
force optimization algorithms and in verifying the
wrinkle-detection and fracture-prediction methods.
The accomplishments include
1. Design and manufacture tooling for conical cup
drawing
2. Preliminary simulations on BHF optimization of
the conical cup drawing
3. Investigation of wrinkling detection methods
(geometry-based and energy-based)
4. Investigation of fracture prediction criteria
(ductile fracture and thinning)
Figure 4. Designs for lower binder segments.
Results from the conical cup drawing
simulations conducted to determine the optimal
linear variable BHF profiles are shown in Figure 6.
The types of wrinkle detection methods
evaluated in this investigation fall into two
categories: (1) geometry-based and (2) energy
based. Wrinkle types obtained in a conical cup are
flange wrinkles and sidewall wrinkles. One of the
geometry-based methods used is measuring wrinkle
amplitudes at different cross-sections (Figure 7).
The evolution of wrinkle severity of a drawn
cup is shown in Figure 8. Results show that the
average wrinkle amplitude varies with the draw
depth of the conical cup.
Computer simulations were conducted on the
conical cup to determine the optimal constant BHF
profiles and to verify wrinkle and fracture
predictions. These simulations were conducted with
several different constant BHF values to determine
how the applied constant BHF values affect the
maximum obtainable drawing depths. Figure 9
shows plots of maximum obtainable cup depths at
different constant BHF levels for two wrinkle
amplitudes, recorded when failures were detected.
Flange and sidewall wrinkles occur at lower BHF
levels (<20 KN). Sidewall wrinkle is the only type
of failure that occurs at medium BHF levels
(20 KN~140 KN); fracture is the only failure type
when higher BHF levels (>140 KN) are applied.
From Figure 9, it is seen that at a constant BHF of
Figure 5. Designs for upper binder construction.
Task 3: Develop Computer Simulation and
Process Optimization for Flexible Binders
The Engineering Research Center for Net Shape
Manufacturing of the Ohio State University was
selected to work on Task 3 because of its expertise
and the availability of equipment and facilities
needed to complete the work. Ohio State developed
a 5-phase project plan for Task 3 that was approved
by the original equipment manufacturers:
Phase 1: State-of-art review in stamping simulation
for binder force control
Phase 2: Single-point binder force control on axisymmetric parts
Phase 3: Single-point binder force control on nonsymmetric parts
Phase 4: Multipoint flexible binder force control
Phase 5: Technology transfer
Work on the task is proceeding according to
plan. A literature survey was conducted of issues
16
Automotive Lightweighting Materials
FY 2001 Progress Report
Figure 6. Linear variable blank holder force profiles used in finite element model simulations.
Figure 7. Wrinkle amplitude method.
Figure 9. Effect of constant blank holder force on
drawing depth of cup.
Figure 8. Wrinkle amplitude variation with draw depth
in a conical cup.
Figure 10. Thinning distribution on the tensile test
sample (simulation).
140 KN, the maximum obtainable cup depth is
42 mm. In order to obtain a deeper cup (>42 mm), a
non-constant BHF (i.e., varying BHF curve) must be
used.
Using the stresses and strains obtained from the
tensile test simulations and the ductile fracture
constants, the ductile fracture criteria values were
calculated for each element and each time step.
Figure 10 shows the thinning distribution on a
quarter mesh of the tensile test simulation when the
ductile fracture criteria values become unity at
element 1.
In order to investigate the effectiveness of the
ductile fracture criteria, the fracture values
(calculated at element 1 using the four fracture
criteria) are plotted with respect to the engineering
strain (along the longitudinal direction) in Figure 11.
The thinning value (calculated at
17
FY 2001 Progress Report
Automotive Lightweighting Materials
Figure 11. Fracture criterion values at element 1.
•
element 1) is also plotted. In Figure 11, ductile
fracture criteria values become unity when the
engineering strain is 0.43–0.45. This corresponds to
thinning of 25–30%. It seems that these ductile
fracture criteria can be used to predict fracture with
reasonable accuracy.
•
Accomplishments for Task 3
•
•
•
•
•
•
Literature survey on temporal BHF control and
wrinkling and fracture detection methods were
conducted and summarized.
Conical cup tooling has been designed and built.
Initial conical cup drawing simulations were
conducted.
Simulation results show that a linearly
increasing BHF profile can increase the
maximum drawing depth of the conical cup,
compared with the cup drawing with a constant
BHF.
Four wrinkling detection methods have been
introduced.
Wrinkle amplitude and distance to ideal surface
methods were found to be good in reflecting
both (1) the detailed information of individual
wall wrinkles and (2) the global severity of wall
wrinkles of drawn parts. The surface area ratio
method can only describe the global severity of
a wrinkled part. The Nordlund’s energy method
can only reflect the location of individual
wrinkles.
Four ductile fracture criteria have been
evaluated by conducting simulations of the
tensile test and the conical-cup drawing
operations.
The use of the ductile fracture criteria in
predicting splits in BHF optimization seems
promising. However, because of the lack of
experimental data to determine damage values
for different materials (aluminum alloys and
high-strength steels), a thinning limit will be
used, for now, in the development of BHF
optimization algorithms.
The following is the status of subtasks
performed by Ohio State University:
•
•
18
Development of wrinkle detection
Geometry-based—100% complete
Energy-based—100% complete
Verify wrinkle detection with the conical cup
drawing—50% complete
Development of fracture detection
Thinning/thinning rate—100% complete
Forming limit diagram/forming limit stress
diagram—50% complete
Ductile fracture criteria—100% complete
Verify fracture detection with the conical cup
drawing—50% complete
Automotive Lightweighting Materials
•
FY 2001 Progress Report
•
Determination of temporal binder force (conical
cup)
Optimization techniques—10% complete
Closed-loop control techniques—10% complete
19
Variable binder force experiments with the
conical cup—10% complete
Automotive Lightweighting Materials
FY 2001 Progress Report
C. Warm Forming of Aluminum—Phase 2
Project Leader: Suresh Rama
DaimlerChrysler Corporation
2730 Research Drive
CIMS 463-00-00
Rochester Hills, MI 48309
(248) 838-5356; fax: (248) 838-5338; e-mail: [email protected]
Project Administrator: Jack McCabe
National Center for Manufacturing Sciences
3025 Boardwalk
Ann Arbor, MI 48108-3266
(734) 995-4919; fax: (734) 995-1150; e-mail: [email protected]
DOE Program Manager: Joseph Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: U.S. Automotive Materials Partnership
Contract No.: DE-FC05-95OR22363
Objective
•
Prove the manufacturing and economic feasibility of the warm forming process for aluminum for industrial
applications in a mass production environment:
Improve the formability of the aluminum alloy.
Develop a washable lubricant for use with heated blanks and dies.
Optimize the temperature distribution of the die.
Develop a rapid heating system for blanks.
Optimize the process design and layout.
OAAT R&D Plan: Task 2, 13; Barriers A and B
Approach
•
•
Develop alloys:
Select candidate alloy compositions.
Determine microstructures after heating.
Perform tensile tests.
Conduct preliminary forming tests.
Analyze test results.
Develop lubricant system:
Select lubricant candidates.
Perform screening tests.
Design and lab test lubricant application system.
Install and test system in forming facility.
21
FY 2001 Progress Report
•
•
•
•
•
•
Automotive Lightweighting Materials
Conduct thermal analysis:
Develop die and sheet models and equations for thermal analysis.
Develop and verify predictions of die temperatures and distortions.
Forecast sheet temperatures and distortions after stamping.
Prepare for blank heating:
Verify heater performance in University of Michigan Laboratory.
Design, build, and validate production heater.
Install and test infrared heating in stamping facility.
Fabricate alloys:
Prepare candidate material for laboratory testing at University of Michigan.
Fabricate full-scale material for production stamping tests.
Demonstrate full-scale process:
Develop blank transfer mechanism.
Install infrared heating system.
Perform system tests.
Perform production tests.
Complete panel inspection and material analysis:
Evaluate material strength and corrosion properties.
Inspect panels.
Evaluate the process:
Develop the business case for warm forming.
Evaluate the technical and economic feasibility for warm forming.
Accomplishments
•
Held a project kick-off meeting at U.S. Automotive Materials Partnership on June 5, 2001.
•
Met to gain consensus on process cost model on September 28, 2001.
•
Formed a team to prepare a cost model straw man for team review in November.
•
Initiated the lubricant selection process.
•
Resolved intellectual property issues satisfactorily.
•
Selected Camanoe Associates to perform technical cost modeling.
•
Received a proposal from Oakland University to perform thermal analysis and requested a second proposal from
the University of Michigan.
Future Direction
•
Specify and produce four alloys for testing.
•
Hot-test lubricant samples produced by Fuchs and continue evaluations to develop the optimum lubricant.
•
Build a straw man technical cost model and present it to the team for evaluation.
•
Build and deliver a lab heater for the University of Michigan.
•
Select either Oakland University or the University of Michigan to conduct the thermal analysis.
22
Automotive Lightweighting Materials
FY 2001 Progress Report
Introduction
Phase 2 Issues and Plans
This project focuses on developing the
materials, equipment, and processes to costeffectively produce automotive panels from
aluminum alloys using a technology called warm
forming.
Aluminum alloys generally have limited
formability, a characteristic that has limited their
extensive use in the industry. To relax this
limitation, DaimlerChrysler, General Motors, and
Ford collaborated to investigate the feasibility of
increasing the formability of aluminum using warm
forming technology.
Phase 1 focused on demonstrating the gains in
formability due to the warm forming process. Since
that milestone has been passed successfully, the next
steps need to focus on the technical and economic
feasibility of applying warm forming technology to
the rigors of a production process.
Five critical issues must be addressed
successfully to pass the Phase 2 milestone: alloy
formability improvement, lubrication system
development, blank heating, full-scale production
rate demonstration, and cost analysis.
The first issue is improving the formability of
the alloy to increase its applicability to parts
requiring a material with more extreme drawing
capability.
In Phase 1, graphite was used as the die
lubricant. Although graphite performs wells as a
lubricant at warm forming temperatures, it is not
easily removed and does not satisfy painting
requirements. The second Phase 2 issue is
developing a lubricant that can be applied in a costeffective manner and can be satisfactorily washed
away to meet painting requirements.
The third issue is demonstrating that blanks can
be uniformly heated and fed to a press at a speed
commensurate with standard production rates.
The fourth issue is demonstrating that the
heater, lubrication system, press, and feeder
mechanism can function as a unit at production
rates.
The fifth issue is building a technical process
cost model of the warm forming process that would
enable a cost comparison between the currently used
process and warm forming and would flag areas
where the product cost could be minimized.
Phase 1 Accomplishments
The first phase of the project involved
developing a modified 5000-series aluminum alloy,
producing blanks, and performing warm forming
trials on a Dodge Neon door inner, a door panel
with deep-drawn profiles.
Stamping trials were performed at Sekely
Industries, Inc., an industrial stamping supplier and
member of the Phase 2 project. The principal result
of Phase 1 was a stamping demonstration of Dodge
Neon inner door panels that were formed in a single
die stage with one hit. A typical panel from these
tests is shown in Figure 1.
Phase 2 Accomplishments
The project was initiated on June 5, 2001, with
a kick-off meeting attended by the project team. All
participants reviewed their roles in the project and
summarized their deliverables. Team assignments
were made, one of which was to seek contractors for
performing the cost analysis and the thermal
analysis.
A search was made to identify an appropriate
contractor to perform the critical task of cost
modeling. Camanoe Associates was selected based
Figure 1. Dodge Neon door inner panel.
Producing these panels in a single stage was an
encouraging feat because the equivalent steel panel
required two stages. The newly developed alloys
also showed significant promise for use as Class A
automotive panels. Reaching this goal would have a
significant payback because it would simplify
procurement, production, and recycling of
aluminum-intensive vehicles in the future.
23
FY 2001 Progress Report
Automotive Lightweighting Materials
on its reputation for delivering quality products on
other automotive industry processes.
Camanoe submitted a proposal, which was
presented at a team meeting on September 28, 2001.
The Camanoe proposal was accepted, and a
purchase order is being processed.
The key issue raised at the September 28
meeting was the definition of the scope and
parameters for the process cost model. A model
team was formed that will be led by Camanoe and
will include representation from DaimlerChrysler,
Tower Automotive, and the University of Michigan.
Oakland University submitted a proposal to
execute the thermal analysis task. A second proposal
is expected from the University of Michigan. The
proposals will be evaluated, and an award should be
made before the end of November 2001.
Future Direction
•
•
•
•
•
24
Work is planned in five areas:
Alloy selection: Four alloys will be specified by
the University of Michigan and produced by
Pechiney for testing at the University.
Lubricant: GM will hot-test the lubricant
samples produced by Fuchs, and Fuchs will
continue evaluations to develop the optimum
lubricant.
Cost model: Camanoe will build a straw man
technical cost model and present it to the team
for evaluation.
Infrared heater: IR Technologies will build and
deliver the lab heater for the University of
Michigan.
Thermal analysis: The University of Michigan
will be asked to submit a proposal for
conducting the thermal analysis, and the project
team will select either Oakland University or the
University of Michigan as the contractor.
Automotive Lightweighting Materials
FY 2001 Progress Report
D. Hydroforming of Aluminum Tubes
Principal Investigator: Michael L. Wenner
GM R&D Center
Mail Code 480-106-359
30500 Mound Road
Warren, MI 48090-9055
(810) 986-1108; fax: (810) 986-0574; e-mail: [email protected]
DOE Program Manager: Joseph A. Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: U.S. Automotive Materials Partnership
Contract No.: DE-FC05-95OR22363
Objective
•
Develop material data, design rules, and computer modeling capability to enable the design and manufacture of
hydroformed aluminum tubes for the reduction of mass in automotive bodies.
OAAT R&D Plan: Task 2, 13; Barrier B
Approach
•
Undertake an experimental program to determine the basic material and frictional properties of the tube
material.
•
Design and carry out simple laboratory experiments to determine the response of the tubes to bending and
hydroforming.
•
Use the basic material and frictional properties measured in the laboratory to complete computer simulations in
order to evaluate and improve our ability to model the bending and hydroforming of aluminum tubes.
•
Use analysis of the laboratory data to determine simple design rules that product and process designers could
use to ensure that hydroformed designs are manufacturable.
•
Instrument a rotary draw bender to determine the exact forces imparted on the tubes and to explore the effects
of process parameters on bending behavior and on subsequent formability.
Accomplishments
•
Obtained tensile data for all materials and measured frictional data.
•
Completed and documented free tube expansion tests.
•
Completed and documented straight-tube corner fill tests.
•
Began rotary draw bend tests.
•
Began computer modeling of the tests that have been conducted.
25
FY 2001 Progress Report
Automotive Lightweighting Materials
Future Direction
•
Install and fully instrument a bender on the University of Waterloo campus to bend tubes under a wide range of
end feed and bend radii.
•
Test these bent tubes at Industrial Research and Development Institute (IRDI) to determine their residual
formability.
•
Participate in model validation studies with the automobile companies and develop formability criteria suitable
for finite element modeling.
•
Characterize failure and damage metallographically.
•
Complete work on IRDI’s experimentally determined forming limit diagrams, correlate friction measurements
and achievable corner radii in straight-tube corner fill tests, and determine an empirical prediction of burst
pressure and corner radius.
This project brings together engineers from
DaimlerChrysler, Ford, General Motors, Alcoa,
Alcan, VAW, IRDI, the University of Waterloo, and
several manufacturing concerns. The expertise of the
group includes material science, production
methods, computational mechanics and
experimental methods. This group meets monthly to
monitor progress and to plan future project work.
Introduction
Automotive structural parts can be made from
hydroformed tubes. These are tubes, not necessarily
circular, which are first bent to shape in a rotary
draw bender and then put into a die and expanded by
fluid pressure to the die configuration. These parts
have many advantages over the welded sheet metal
parts they replace: They help reduce manufacturing
costs (by decreasing part count, welding, and scrap)
and improve performance (by reducing mass and
increasing stiffness). Steel hydroformed tubes are
finding increased application in the automotive
industry, but the behavior of aluminum in the
bending and hydroforming operations are currently
poorly understood; therefore, aluminum
hydroforming is not yet used in the U.S. automobile
industry. Figure 1 shows a hydroformed engine
cradle.
Accomplishments
Materials used in this project included welded
Al3.5Mg alloy tubes from VAW and extruded tubes
from Alcoa in alloys 6061-T4 and 6061-T6. Alcan’s
Kingston laboratories obtained tensile data for all
materials, and IRDI measured frictional data.
Free-tube expansion tests— tests carried out
without a die—were completed and fully
documented. Data collected during these tests
included the magnitude of expansion, the fluid
pressure, the axial force, and the burst pressures.
This information enables us to begin to estimate the
basic formability of the tube material, and it
provides experimental data that we can use to
validate computer models. Figure 2 shows Al3.5Mg
tubes, 3.5 mm in thickness, at maximum, medium,
and minimum end feeds and as received (left to
right).
Straight-tube corner fill tests were also
completed and documented. These tests involve
expanding a tube into a square die. In such a test,
frictional forces between the tube and the die
become important. Three different lubricants were
used. Measurements included internal pressure, axial
force, and corner displacement. It was found that the
lubricants had a significant effect on the amount of
Figure 1. A hydroformed engine cradle.
26
Automotive Lightweighting Materials
FY 2001 Progress Report
Data from these tests also provide information
on the basic formability of the tube materials.
Rotary draw bend tests were begun during
FY 2001. These tests are not yet complete. When
they have been completed, the bent tubes will be
tested in hydroforming dies to evaluate their residual
formability.
Computer modeling of the tests described has
begun. Models of the rotary draw bending have been
completed and compared with data from
experiments, with quite good agreement. The freetube expansion test has also been modeled, and
comparisons with the experimental data are under
way. A benchmark document has been prepared to
assist in comparing the various computer models
with each other and with the experimental data.
Much of the time in our monthly meetings for
this time period was spent in laying out a course of
work for the next phase of this project. In view of
the high strains created in the bender, the lack of
knowledge of subsequent formability, and the fact
that the bending operation is responsible for most of
the process variability in hydroforming, we decided
to concentrate on the bender. The University of
Waterloo will install and fully instrument a bender
on campus. It will be used to bend tubes under a
wide range of end feed and bend radii. These bent
tubes will then be tested at IRDI to determine their
residual formability.
Waterloo will also participate in model
validation studies with the automobile companies
and will develop formability criteria suitable for
finite element modeling. Waterloo will also
characterize failure and damage metallographically.
IRDI will complete work on its experimentally
determined forming limit diagrams. IRDI will also
correlate friction measurements and achievable
corner radii in straight-tube corner fill tests and will
determine an empirical prediction of burst pressure
and corner radius. The latter two projects are aimed
at providing validated design rules for hydroforming
systems.
Figure 2. Al3.5Mg tubes, 3.5 mm in thickness, at
maximum, medium, and minimum end feeds
and as received (left to right).
corner expansion that could be achieved, and that
this effect correlated quite well with the data taken
in the basic friction tests carried out by IRDI. These
results suggest that the friction tests are valid and
that they might be enlarged in scope to provide
guidelines for designing hydroforming systems.
Figure 3 shows one of the tubes that did not burst.
The variation in thickness is clearly evident.
Figure 3. A test tube that did not burst.
27
Automotive Lightweighting Materials
FY 2001 Progress Report
E. Electromagnetic Forming of Aluminum Sheet
Principal Investigator: Richard W. Davies
Pacific Northwest National Laboratory
P.O. Box 999, Richland, WA 99352
(509) 376-5035; fax: (509) 376-6034; e-mail: [email protected]
Technical Coordinator: Sergey Golovashchenko
Ford Motor Company
Ford Research Laboratory, Dearborn, MI 48121-2053
(313) 337-3738; fax: (313) 390-0514; email: [email protected]
DOE Program Manager: Joseph A. Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Participants
Dwight Rickel, EMF System Development, Los Alamos National Laboratory
James Sims, Coil Durability, Materials, and Design, Los Alamos National Laboratory
Sergey Golovashchenko, Project Coordinator, Ford Motor Company
Jeffrey Johnson, EMF System and Control, Pacific Northwest National Laboratory
Gary Van Arsdale, EMF System Design and Metal Forming, Pacific Northwest National Laboratory
Nick Klymyshyn, Computational Mechanics, Pacific Northwest National Laboratory
Dick Klimisch, Project Coordinator and Material Supply, Aluminum Association
Oxford Automotive Company, EMF Industrial Embodiment
Contractor: Pacific Northwest National Laboratory
Contract No.: DE-AC06-76RL0 1830
Objective
•
Develop electromagnetic forming (EMF) technology that will enable the economical manufacture of
automotive parts from aluminum sheet. EMF is a desirable process because the dynamic nature of the
deformation results in benefits including increased forming limits and reduced springback. These benefits will
result in increased use of aluminum and, therefore, vehicles that are more fuel-efficient because their mass is
reduced.
OAAT R&D Plan: Task 2, 13; Barriers A, B
Approach
•
Address three main technical areas:
Analysis methods for forming system designs.
Development of durable actuators (coils).
Industrial embodiment of the EMF process.
29
FY 2001 Progress Report
Automotive Lightweighting Materials
Accomplishments
•
Completed a literature search for information on EMF, coil materials, and coil design/durability.
•
Completed design and assembly of a 150-kJ pulsed power unit (Figure 1) at Los Alamos National Laboratory
(LANL).
•
Completed a C4 survey of the laboratories.
•
Requested and received a proposal from LANL for a paper study on the design of durable coils.
•
Installed the 150-kJ pulsed power supply at Pacific Northwest National Laboratory (PNNL).
Future Direction
•
Establish an operating procedure for the 150-kJ pulsed power unit.
•
Develop design concepts for durable coils for aluminum sheet.
•
Design and build a test rig for coil durability testing.
•
Develop concepts/designs for test systems that are suitable to generate hybrid forming (combined static and
dynamic deformation) biaxial formability data.
•
Establish project partners that can contribute to the development of technology for the industrial embodiment
and formability areas.
•
Explore and develop modeling capabilities that can assist in the design of EMF systems.
involved either the expansion or compression of
cylinders (tubes). The forming of sheet materials is
considerably more complex and has received
relatively little attention.
Introduction
In the EMF process, a transient current pulse of
high magnitude is generated in a coil by a lowinductance electric circuit. During the current pulse,
the coil is surrounded by a strong transient magnetic
field. The transient nature of the magnetic field
induces current in a nearby conductive workpiece
that flows in the opposite direction to the current in
the coil. The coil and the workpiece are set up as
adjacent magnets with poles oriented in opposite
directions to repel each other. The force of repulsion
can be very high, equivalent to surface pressures on
the order of tens of thousands of pounds per square
inch. Thin sheets of material can be accelerated to a
high velocity in a fraction of a millisecond.
A recent interest in understanding the EMF of
metals has been stimulated by the desire to use more
aluminum in automobiles. The high workpiece
velocities achievable using this forming method
enhance the formability of materials such as
aluminum. Also, the dynamics of contact with the
forming die can eliminate springback, an undesired
effect that cannot be avoided in other forming
techniques such as stamping. The commercial
application of this process has existed since the
1960s. The large majority of applications have
Project Deliverables
At the end of this program, methods and data to
assist in the economical design of EMF sheet
forming systems will be documented. These will
include materials information and design methods
for durable coils, coil durability test data for selected
materials and design concepts, dynamic and hybrid
formability data, methods for modeling the forming
process, and concepts for the industrial
implementation of the technology in an automotive
manufacturing environment.
Planned Approach
This project will address three main technical
areas. The first technical area involves establishing
analysis methods for designing forming systems.
These methods will be based on developed
knowledge of forming limits and of relations
between electrical system characteristics and
deformation responses for specific aluminum alloys
of interest. The second area of technical challenge is
30
Automotive Lightweighting Materials
FY 2001 Progress Report
in coil durability. Existing knowledge of EMF and
relevant knowledge from pulsed power physics
studies will be combined with thermo-mechanical
analyses to develop durable coil designs that will be
tested experimentally. Until a more thorough
understanding is achieved of economic factors
determining the required durability, a nominal level
of a 100,000-cycle coil life will be a project goal.
The third technical area involves the industrial
embodiment of the EMF process. In this project,
EMF is expected to be hybridized with conventional
sheet metal stamping. Different approaches to
hybridization will be analyzed for issues affecting
economical implementation in a modern stamping
plant. Different system concepts will be developed
and studied. Existing knowledge of the EMF process
and technical achievements in this project will be
combined to establish a methodology for designing
hybrid forming systems that can be readily
integrated into modern manufacturing facilities for
the economical production of automotive sheet
aluminum components.
Figure 1. The capacitor bank and load cables connected
to the forming coil.
EMF System Commissioning
The initial testing and trials of the new EMF
system at PNNL were completed in September
2001. The trials consisted of assembling the new
EMF power supply system, load cables, and
inductive load coil. The apparatus used to conduct
the experiments is illustrated in Figure 1. This figure
shows the four parallel coaxial conductors connected
between the power supply and the EMF coil.
Figure 2 is an enlarged view of the single-turn coil.
The coil used was a single-turn, low-inductance
aluminum alloy coil made from AA6061-T6. Also
shown in Figure 2 are multiple sheets of Mylar
sheeting and G10 (glass-fiber composite) insulating
materials. Not shown in the figure are the coil
containment shroud and associated supports.
The experiments involved multiple cycles of
charging and discharging of the capacitor bank
through the load coil at various known energy
levels. The capacitor bank was controlled via the
custom system developed for the unit. The sequence
of testing consisted of charging the capacitor bank,
isolating the charging power supply, triggering
(releasing) the capacitor charge, and monitoring the
response of the system. The system response was
recorded using a high-speed digital oscilloscope.
Figure 2. The single-turn forming coil and the load cable
connections.
Figure 3 illustrates the typical response of the
system during a 15-kJ discharge of the capacitor
bank. This figure shows that the half-current of the
system is approximately 86 kA, so that a total
current of 172 kA passed through the load coil. The
system rise time was shown to be approximately
26 microseconds. This power supply system
performed as predicted during its design phase, and
follow-on work is currently under way.
Coil Design Concepts and Durability
In EMF, the large forces that deform the
workpiece are mirrored in the coil. The coil,
insulators, and support structure must resist these
forces, as well as related thermal cycles, without
significant permanent deformation or material
failure. In contrast to typical cylindrical coils, sheet
forming will require coils with general threedimensional shapes that are inherently less resistant
31
FY 2001 Progress Report
Automotive Lightweighting Materials
modular, and highly durable (nominally
100,000 cycles) if they are to be relevant to
automotive manufacture.
Current Waveform - Initial Trials of the PNNL Capacitor Bank
System Charged to ~15kJ, Single turn coil, No ring suppression, Half current plotted
100
80
Conclusions
Current (kA)
60
The technical feasibility of EMF for aluminum
sheet in an automotive application has been
demonstrated in a prior project conducted by the
U.S. Council for Automotive Research. However,
the durability of relevant coils systems and methods
for the economical design, construction, and
implementation of forming systems is yet to be
demonstrated. There is also a need for dynamic
formability data for relevant aluminum alloys. This
project targets these issues. Progress has been made
in assessing the current state of knowledge for
materials, coil design, formability, and system
design. Also, a pulsed power system has been
designed and fabricated to serve in experimental
testing of coil systems. As this project progresses, a
balanced combination of analysis and experiment
will be applied to develop durable coil systems that
meet the performance requirements of automotive
manufacturing. Formability data will be produced
and combined with system design analyses to
develop methods for designing EMF systems.
Project partner(s) will be recruited to study the
issues of industrial implementation of EMF sheet
forming systems.
40
20
0
-50
0
50
100
150
200
-20
-40
Time (microseconds)
Figure 3. The current waveform that resulted during the
initial system trials at PNNL.
to forces induced during forming. The key issues
involve materials selection and design. Materials
must be selected for both electromagnetic and
mechanical properties. Materials must be compatible
for bonding, presence of coolants, manufacture, and
system integration for hybrid forming systems that
combine conventional stamping and EMF. The
design must integrate these elements while
delivering the primary function of a spatial and
temporal load distribution that achieves the desired
deformations. Coil systems will have to be low-cost,
32
Automotive Lightweighting Materials
FY 2001 Progress Report
F. Optimization of Extrusion Shaping: Aluminum Tubular Hydroforming
Principal Investigator: Richard W. Davies
Pacific Northwest National Laboratory
P.O. Box 999, Richland, WA 99352
(509) 376-5035; fax: (509) 376-6034; e-mail: [email protected]
Project Coordinator: David Williams
Alcoa Inc.
100 Technical Drive, Alcoa Center, PA 15069-0001
(724) 337-2861; fax: (724) 337-2209; e-mail: [email protected]
DOE Program Manager: Joseph A. Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Participants
Glenn Grant, Laboratory Hydroforming and Microscopy, Pacific Northwest National Laboratory
Kenneth Johnson, Computational Mechanics, Pacific Northwest National Laboratory
Kirit Shah, Numerical Modeling and Project Coordination, Alcoa Inc.
Edmund Chu, Numerical Modeling and Formability, Alcoa Inc.
Robert Evert, Extrusion, Manufacturing Process, and Formability, Alcoa Inc.
Frederic Barlat, Constitutive Modeling, Alcoa Inc.
Ming Li, Constitutive Modeling and Fracture, Alcoa Inc.
Valery Guertsman, Electron Microscopy, Pacific Northwest National Laboratory
Contractor: Pacific Northwest National Laboratory
Contract No.: DE-AC06-76RL01830
Objective
•
Develop and experimentally verify materials and forming models that can be used to optimize the hydroforming
of complex aluminum tubular components. The project also focuses on developing predictive methods to
determine the forming limits of extruded aluminum alloys during the hydroforming process.
OAAT R&D Plan: Task 2; Barrier B
Approach
•
Evaluate existing finite element (FE) modeling codes and baseline capabilities.
•
Select a representative complex three-dimensional hydroformed component to demonstrate integrated stretch
bending and hydroforming process steps.
•
Develop laboratory experimental hydroforming test equipment and evaluate candidate extruded aluminum
alloys.
33
FY 2001 Progress Report
Automotive Lightweighting Materials
•
Develop and verify mathematical-based methods for predicting material forming limits under hydroforming
process conditions.
•
Evaluate candidate aluminum alloys and tubular products under simulated hydroforming conditions.
Accomplishments
•
Developed forming-limit diagrams (FLDs) for several aluminum alloys and compared the material formability
with theoretical FLD calculation methods.
•
Established the effects of non-proportional loading on the forming limits of aluminum alloy extrusions.
•
Established the effects of tubing dimensions and strain hardening on forming limits of aluminum alloy
extrusions.
•
Determined the mechanical property gradients within structural aluminum alloy tubing that affect the overall
formability of these tubes. Structural extrusions are normally made using the cost-saving manufacturing method
known as “portal die extrusion.”
Future Direction
•
Carry out additional work to improve understanding of the differences between structural and seamless
extruded tubing.
•
Characterize seam-welded aluminum alloy tubing for a quality and cost comparison with extruded aluminum
alloy tubing.
•
Develop numerical algorithms coupled to FE codes that optimize or refine hydroforming process control prior
to part prototyping phases, which would advance use of aluminum hydroforming.
Introduction
Aluminum is currently a principal lightweight
material candidate for the Partnership for a New
Generation of Vehicles (PNGV) initiative. However,
some technical challenges remain that inhibit the
economical and efficient use of aluminum in highvolume automotive manufacturing. A major
challenge for many aluminum-intensive vehicle
designs is the ability to predict the dimensional
variations that develop during the forming of
individual parts and structural elements, and to use
these predictions to control and optimize the
forming process that is used. The objective of this
project is to develop analytical and experimental
tools that can be used to model and predict the
shaping and forming processes for individual
aluminum structural elements. This project is mainly
focused on numerical simulation of extrusion
shaping processes and experimental evaluation of
aluminum alloy formability.
Project Deliverables
Designing a vehicle by selecting among the
multitude of different aluminum alloys, heat
34
treatments, and manufacturing techniques for
optimum performance in an automotive structure is
a formidable challenge. Several manufacturing
technologies have been identified that appear to
have significant influence on the formability of
aluminum extrusions. There are also major issues
that will predict the manufacturability and
dimensions of automotive components made from
aluminum extrusions. The scope of these projects
includes evaluating the limits of formability of
different types of aluminum alloy tubing under
hydroforming conditions investigating the capability
to perform multiple sequential forming operations
using FE analysis to accurately simulate the
bending, die closure, and hydroforming of aluminum
alloy tubing
Planned Approach
The approach and work plan of the project
includes (1) evaluating material formability under
laboratory hydroforming conditions, (2) performing
prototype-scale commercial hydroforming trials, and
(3) implementing FE codes and other analytical
methods.
Automotive Lightweighting Materials
FY 2001 Progress Report
both axial load and internal pressurization. The load
frame actuator controlled the axial position and
force imparted to the tube, and it had a maximum
travel of 152 mm. The specimen grips consisted of a
locking collet, an internal mandrel, and a seal
assembly that are typically used for hydrostatic
testing of tubular products. The internal
pressurization system consisted of an air-driven
water pump that was capable of achieving pressures
up to 69 MPa. A programmable valve regulated the
internal pressure during testing. The hydroforming
control system consisted of a multi-axis digital
servohydraulic controller, which controlled forming
process parameters and data acquisition.
Many different populations of aluminum alloy
extrusions were evaluated. This report presents two
of the subject materials, which are AA6063-T4 and
AA6061-T4. The AA6063-T4 population
dimensions were an outside diameter (OD) of 76
mm with a wall thickness of 3.5 mm. The AA6061T4 population dimensions were OD 50 mm with a
wall thickness of 2.0 mm. The overall lengths of the
tubes tested during experimentation were 560 mm
for AA6063 and 460 mm for AA6061. The
unsupported lengths of tube between grips were
420 mm and 320 mm, respectively, with 70 mm of
tube inserted into each grip. Each subject tube
specimen had a 2.54 mm2 grid of fine lines applied
to its surface prior to deformation, and the
development of plastic strain was quantified using
strain grid analysis. The strain grid analysis results
were compiled to create an FLD for each tube
population.
The experiments included uniaxial tensile tests
that determined the strain hardening coefficients of
n = 0.26 and n = 0.27 for AA6061-T4 and AA6063T4, respectively. After the uniaxial tensile properties
were determined, each sample population was
subjected to a series of experiments to determine the
level of plastic strain that developed under various
proportional loading ratios of internal pressure and
axial load. Figure 2 illustrates the typical results
obtained for a single population of samples tested
under free hydroforming conditions. The specimens
are shown so that the bottom specimen received the
highest amount of axial compression during testing,
and the top specimen was the uniaxial specimen.
The strain-gridded regions surrounding the fractures
were characterized using a strain-grid analysis
system. This system determined the magnitude of
strains that developed in the specimen in the region
Experimental Characterization of Forming
Limits
Tube hydroforming is a manufacturing process
receiving considerable attention as a method to
produce complex and precise structural automotive
components. The hydroforming process typically
applies internal pressure and axial compression to a
tube material to manipulate its expansion to fill a
precise die cavity. Figure 1 schematically presents
Figure 1. Schematic of the tube hydroforming process
before and after forming.
the concept of tube hydroforming. Currently, the
automotive industry is applying this process
predominately with steel. However, in an effort to
reduce vehicle weight and vehicle emissions,
considerable attention is focused on applying
aluminum alloys in applications formerly dominated
by steel. One significant focus of this project was to
evaluate the formability of various aluminum
extrusions. This section presents the results of an
experimental investigation of the formability of
aluminum alloy extrusions, which is representative
of much of the project’s experimental focus. The
experimental characterization of tubular materials
under simulative hydroforming conditions requires a
special apparatus. The system required the
simultaneous application of internal pressure and
axial load. The system also required unique
specimen grips and process control capabilities. The
mechanical apparatus used to perform the
experiments is shown in Figure 2. The base system
for the free hydroforming tool was a 300-kN
uniaxial hydraulic load frame. The system was fitted
with a 222-kN load cell and unique grips to support
35
FY 2001 Progress Report
Automotive Lightweighting Materials
Figure 2. The apparatus used to perform free hydroforming experiments (left). Photograph showing the results of
applying increasing magnitudes of axial compression during internal pressurization on the AA6063-T4
population (right).
around the failure. Each of the originally 2.54-mm2
grids characterized in the area surrounding the
fracture was identified as necked if the nodes near
the failure showed elevated levels of local strain
(necking) in the grid, or the fracture had penetrated
under (undercut) the grid surface. Figure 3 shows
the results of strain-grid analysis for both
populations plotted on a circumferential-versusaxial-strain FLD.
These laboratory hydroforming experiments
have shown that the forming limits of aluminum
alloy tubing may be determined using a laboratory
apparatus that incorporates both axial end feed and
internal pressurization. For the AA6061-T4 and
AA6063-T4 materials presented, the limits of the
material formability have proved to be significantly
different even though they have a similar n-value
(strain hardening coefficient).
In addition to the seamless 6000-series
extrusions presented, this project evaluated the
formability of 5000-series materials, 6000-series
alloys subjected to retrogression heat-treatment, and
various structural extrusions produced using portal
die extrusion techniques. Additional work was
conducted to improve understanding of the
differences in formability of various AA6061-T4
extrusions with different grain sizes and
microstructures.
36
Laboratory Hydroforming with Dies—
Hydrocone Experiments
Commercial hydroforming is significantly
different from the free hydroforming experiments
described. Free hydroforming is a valuable
technique to quantitatively determine and
distinguish the performance of different materials,
but this technique does not offer all the insight
needed for commercial hydroforming predictions.
Therefore, a laboratory apparatus was devised that
was more representative of commercial
hydroforming. Figure 4 is a schematic representation
of the hydrocone assembly developed and fabricated
at Pacific Northwest National Laboratory. This
system was designed to simulate the friction, end
feed and sealing, and buckling mode constraints that
often exist in commercial hydroforming. This
arrangement allows tests to be conducted under
more severe axial compression than free
hydroforming. Figure 5 is a photograph of an
AA6061-T4 tube with an OD of 50 mm and a wall
thickness of 2 mm tested in the hydrocone
apparatus. This tube achieved increased
circumferential expansion compared with same
materials under free hydroforming, which is
consistent was commercial experience. Figure 6
illustrates an FLD showing the relative change in
plastic strain achieved using the hydrocone
compared with free hydroforming. This apparatus
Automotive Lightweighting Materials
FY 2001 Progress Report
Figure 3. Form-limiting diagrams for the two sample populations determined via free hydroforming.
Figure 6. The friction and reduced buckling in the
hydrocone dies increases the axial
(compressive) end feed of the tube prior to
buckling and increases the circumferential
expansion.
Figure 4. Schematic illustration of the Pacific Northwest
National Laboratory hydrocone assembly.
Figure 5. Photograph of an AA6061-T4 tube tested in
the hydrocone apparatus.
37
FY 2001 Progress Report
Automotive Lightweighting Materials
enables the investigation of more severe nonproportional loading with respect to axial end feed,
which is more representative of commercial
hydroforming.
Sequential Forming Numerical Simulation
Modeling of extrusion shaping was performed to
develop reliable methods for simulating
3-dimensional extrusion forming operations so that
the forming process path and final dimensional
characteristics of the part can be predicted. This type
of work requires accurate material models, realistic
failure criteria, and control methodologies that allow
modeling and prediction of a successful forming
process. This report describes a FE modeling and
simulation study performed for sequential stretchbending and hydroforming of an aluminum
extrusion. The purpose of the study was to
investigate and develop procedures for threedimensional FE modeling of sequential forming
operations and to assess both the viability of
implicit-versus-explicit solution formulations and
the utility of 3-dimensional versus two-dimensional
models. The complete sequence of forming
operations was simulated, from initial clamping of
the ends of the extrusion for stretch-bending, to
depressurization and release of the part from the
hydroforming dies.
Figure 7 illustrates a representative
3-dimensional FE model of a complex hydroforming
problem. This figure illustrates the final step in a
sequence of stretch forming and hydroforming of an
initially non-circular 6000-series extrusion. Results
of the simulations, including springback, residual
stresses, and deformed shape, showed the value of
3-dimensional models compared with 2-dimensional
models. Figure 8 shows the predicted final shape
and contours of plastic strain from a sequence
simulation encompassing both stretch bending and
forming.
38
Figure 7. A finite element simulation of a complex
hydroform component. This was the final step
in a sequence of numerical models used to
predict the formability, failure, and final
dimensions of the manufacturing process.
Figure 8. An automotive hydroformed component as
predicted by the sequential numerical
simulations.
Automotive Lightweighting Materials
FY 2001 Progress Report
G. Improved A206 Alloy for Automotive Suspension Components
Principal Investigator: Geoffrey K. Sigworth
GKS Engineering Services
1710 Douglas Avenue, Dunedin, FL 34698-3703
(727) 733-2179; fax: (727) 733-2179; e-mail: [email protected]
Project Manager: Paul A. Bujalski
Daimler Chrysler Corp., CIMS 481-01-41
800 Chrysler Drive, Auburn Hills, MI 48326
(248) 576-6994; fax: (248) 576-7288; e-mail: [email protected]
DOE Program Manager: Joseph Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: U.S. Automotive Materials Partnership
Contract No.: DE-FC05-95OR22363
Objectives
•
Demonstrate the capability to produce suspension parts in 206 alloy without hot cracks.
•
Establish the quality of 206 control arms.
•
Provide design data and guidelines needed to produce other parts in high-strength 206 alloy.
•
Determine if a quick and inexpensive ultrasonic test can replace dye penetrant inspection.
OAAT R&D Plan: Task 12; Barriers A, B
Approach
•
Establish that it is feasible to produce 206 alloy control arm castings by employing a combination of improved
grain refinement and correct casting process parameters.
•
Optimize the alloy chemistry and casting parameters for a GMX-130 control arm and make a production
quantity of these castings.
•
Provide detailed mechanical property data for the resulting control arms, at room temperature and at 250°F, and
provide engineering and manufacturing design guidelines for production of other castings in 206 alloy.
•
Establish whether the quick, inexpensive ultrasonic test method, developed at Stevens Institute of Technology,
is a reliable indicator for the presence of cracks.
Accomplishments
•
Completed Task 1 and manufactured two control arm castings in 206 alloy.
•
Completed Task 2 and made more than 100 production-quality castings.
•
Partially completed Task 3 (some preliminary mechanical property data are available.
39
FY 2001 Progress Report
Automotive Lightweighting Materials
Future Direction
•
Complete testing of the phase 2 castings.
•
Prepare design guidelines.
•
Complete a final report on the project by February 2002.
linked to improved CAFE and vehicle performance.
There is also a potential for cost savings.
Introduction
A206 alloy is significantly stronger than
aluminum casting alloys normally used currently
and has mechanical properties approaching those of
some grades of ductile iron. It also has excellent
high-temperature tensile and low-cycle fatigue
strength. Consequently, this material could be used
in a number of applications to reduce vehicle
weight. Cost savings may also result, because less
material would be required to provide the strength
needed for the application.
In spite of its excellent properties, 206 alloy is
seldom used because of its propensity for hot
cracking. A better method to grain-refine this alloy
has been discovered, which has the potential to
eliminate the hot-cracking problem. In addition, a
new ultrasonic inspection technique to test for the
presence of hot cracks has been developed at
Stevens Institute of Technology. Hence, it is time to
reconsider the commercial feasibility of 206 alloy.
The high-strength 206 alloy has a number of
potential applications, but its high strength and
excellent ductility make it an ideal candidate for
suspension components. Consequently, control arms
will be poured to establish the viability of this
material for suspension components.
An extensive material properties database has
been developed for aluminum alloys under the
current U.S. Council for Automotive Research
(USCAR) project titled “Design and Product
Optimization for Cast Light Metals”. However, no
information on 206 alloy is presently available in
this database.
Program and Deliverables
This project will take approximately 12 months
to complete and will proceed in four stages. The
total project cost is $232,000, 56% of which is
provided by in-kind contributions from the
participants and the three automobile companies.
DOE funds provide the balance.
Phase 1
The viability of casting the 206 alloy as
suspension components is to be established. Two
General Motors (GM) control arms will be poured at
Stahl Specialty Company. The tooling for these
castings is available at Stahl and will be used to
make A206 (conventional) and A206E
(experimental) versions of 206 alloy. These castings
will be examined for cracks by die penetrant tests,
and mechanical properties will be determined for
test bars cut from the castings.
Phase 2
Assuming the results in phase 1 confirm the
viability of casting the A206E alloy, the second
stage of the program will begin. Further casting
trials will be conducted at Stahl Specialty with the
object of “fine tuning” or optimizing the casting
process for the GMX-130 casting. When the “sweet
spot” for the casting has been found, a significant
number of production castings will be made. Dr.
Donskoi from Stevens Institute of Technology will
participate in the evaluation of crack inspection
techniques at Stahl Specialty during these trials.
Justification
Automakers will be under increasing pressure to
reduce CO2 emissions through increased Corporate
Average Fuel Economy (CAFE) standards. Because
of its higher strength, 206 alloy structures have the
potential to reduce vehicle mass, which is directly
Phase 3
The third stage of the program includes a full
spectrum of mechanical properties and other
inspection tests. Most of these will be conducted at
40
Automotive Lightweighting Materials
FY 2001 Progress Report
Westmoreland Mechanical Testing Laboratories.
The results will give USCAR a good comparison of
properties between A356 and A206 alloys poured in
the same mold. In addition to the mechanical
property tests, nondestructive tests will be
conducted at Stevens Institute of Technology.
Results Obtained
Phase 1
Two control arms were selected for these tests.
Both parts have been produced commercially in the
past by using tilt-pour molds in A356 alloy, heated,
treated, and aged to maximum strength (T6
condition). Two versions of A206 alloy were
poured, using virtually the same melt and process
parameters used for A356 alloy. No changes were
made in the mold. Minor changes in tilt rate and
ejection times were made for the A206 alloy trials,
but no attempt was made to fully optimize the
process. The intention was to establish the viability
of pouring automotive suspension components in
A206 alloy. The two castings are both GM parts.
They are shown in Figure 1. The casting on the right
hand side of Figure 1 is a rear lower control arm for
GM’s Cadillac. This part is a substitute for a steel
stamping and was not designed primarily to be an
aluminum casting. The geometry is complicated, the
draft angle on the deep ribs is minimal, and it is a
large casting (weighing 6.7 lb or 3 kg). This casting
was selected as a worst-case scenario. Contrary to
our expectations, we were successful in producing
this part. The most serious problem was cracking
during ejection. The tool used in our trials was old
and showed a significant amount of erosion at the
in-gates. Thus the part tended to hang up in the
mold. Only with careful maintenance of the mold
Phase 4
In the fourth and final stage of the project, GKS
Engineering will prepare a detailed analysis and
report of important experimental results obtained
during the program. The report will
• Determine if the hot cracking problem can be
eliminated by a combination of improved grain
refinement, proper mold design, correct casting
parameters, and better on-line inspection
techniques.
• Provide detailed mechanical properties data for
the A206 alloy control arm, at room temperature
and at 250°F.
• Provide engineering and manufacturing design
guidelines for production of castings in 206
alloy
• Establish whether the quick and inexpensive
ultrasonic test method, developed at Stevens
Institute of Technology, is a reliable indicator
for presence of cracks
Figure 1. Lower control arm castings produced in phase 1 trials. (Left) A front lower control arm for a GM
Grand Am. (Right) A rear lower control arm for a GM Cadillac.
41
FY 2001 Progress Report
Automotive Lightweighting Materials
casting process has not been optimized for 206
alloy.
coating in these areas could the part be ejected
without cracking.
The casting on the left hand side of Figure 1 is a
front lower control arm for GM’s Grand Am. This
part was originally designed to be an aluminum
casting.
The GMX-130 castings were subjected to dye
penetrant tests. For the parts produced with the
regular version of A206 alloy, 4 of 24 castings or
16.6% exhibited cracks. For castings poured with
the improved A206E alloy, 5 of 52 castings or 9.6%
exhibited cracks. The other control arm was also
examined for cracks, but these data are less
meaningful, because the cracks were caused by a die
ejection problem. The results consequently reflect
the state of the die coating.
The castings were heat treated, and tensile bars
were cut from primary load-bearing parts of the
casting, where mechanical properties are critical.
The average values of the tensile properties
measured are summarized in Table 1. Data for the
same castings made with A356 alloy are also
included for comparison.
Phase 2
A week-long casting trial was held in August at
Stahl Specialty Company in Kingsville, Missouri. A
set of statistically designed experiments was used to
establish optimum process parameters for the
GMX-130 casting. When the optimized process was
used, X-ray inspection of parts showed the casting
to have excellent soundness (no porosity). Then one
and a half shifts of production castings were made
using the optimized process.
During these trials, the chemistries of the alloy
and of the heat treatment parameters were adjusted
slightly to increase the strength of the castings.
In addition to the GMX-130 control arm
castings, separately cast test bars and step-mold
castings were poured for comparison. The step mold
has five sections of different thickness. Tests of
tensile bars cut from each section will give us an
idea of how the section thicknesses, and the
resulting solidification rates, affect the strength of
castings made in 206E alloy.
Table 1. Average values of the properties measured
for the tensile bars cut from heat-treated
castings
Phase 3
The extensive series of destructive and
nondestructive tests of the castings made in phase 2
was just beginning as this report was being
prepared. Only preliminary test data on separately
cast test bars are available.
One of the problems with the use of 206 alloy is
that little technical information about the material is
available in current handbooks. As a consequence,
an extensive literature search was made regarding
the Al-Cu–based and Al-Cu-Mg–based alloys.
Studies published on other Al-Cu-Mg alloys
nearly 50 years ago suggest that 206 alloy should be
subject to a natural aging process—holding castings
at room temperature after quenching from the hightemperature solution heat treatment. We therefore
decided to study this strengthening process by
measuring the Brinell hardness of step-mold
castings and the tensile strength of separately cast
(ASTM B-108) test bars. Tests were made on the ascast samples and on castings at times of 1 hour,
8 hours, and 1,2 3, 4, and 7 days after the water
Mechanical properties of RLCA castings
Yield
Ultimate
Elongation
Alloy
(ksi)
(ksi)
(%)
206-T4
36.6
54.3
12.6
206E-T4
33.5
51.7
15.8
356-T6
31.7
40.1
9.1
Mechanical properties of GMX-130 castings
Yield
Ultimate
Elongation
Alloy
(ksi)
(ksi)
(%)
206-T4
33
51.2
17.0
206E-T4
34.3
53.4
17.8
356-T6
28.6
41.6
9.6
The phase 1 results were encouraging. We were
successful in making both control arm castings. The
new experimental (E) version of 206 alloy appears
to offer better resistance to hot cracking, and the
mechanical properties are comparable to those of
the conventional version of the alloy. Also, the
206 control arm castings are stronger than
conventional 356 alloy parts, even though the
42
Automotive Lightweighting Materials
FY 2001 Progress Report
First, the average elongations shown appear to
be controlled largely by the presence of small oxides
in the test bar castings. These are sometimes
observed on the fracture surface of the bar, and they
lower the elongation to fracture. When the metal is
completely free of oxides, the elongation varies
from 25% at 1 hour after the water quench, to 22%
at 3 or more days after the solution treatment.
To a lesser extent, the ultimate tensile strength
(UTS) is also influenced by the presence of oxides.
When oxides are present, the UTS is usually
lowered by about 1000 to 2000 psi (7–14 MPa).
The first test results, obtained on control arms in
phase 1 of this study and presented earlier in
Table 1, were made within 24 hours after the hightemperature solution treatment. Thus full strength
was not achieved in these castings at the time tests
were made.
The complete test results from the phase 2
castings should be available by the end of January
2002. The design guidelines should also be prepared
by then.
The final report for this project should be
complete by the end of February 2002.
quench at the end of the solution treatment. The
averages of test results for five castings are
presented in Figure 2. The natural aging process is
seen to be complete after 3 days of holding. At this
time, the average values for the tensile properties (in
separately cast permanent mold test bars) are
Yield strength
37,800 psi (261 MPa)
Ultimate
62,200 psi (429 MPa)
Elongation
20%
The Brinell hardness is 106.
Brinell Hardness Yield and Ultimate Tensile Strength (ksi)
Several points should be considered regarding
the results shown in Figure 2.
70
450
60
400
350
50
MPa
300
40
250
30
200
110
100
90
Elongation
80
20
10
0
as
cast
0
1
2
3
4
5
6
7
time of natural age (days)
Figure 2. Mechanical properties of 206 alloy during
natural aging.
43
Automotive Lightweighting Materials
FY 2001 Progress Report
H. Die-Casting Die Life Extension
Principal Investigator and Lab Coordinator: Edward L. Courtright
Pacific Northwest National Laboratory
P.O. Box 999, Richland, WA 99352
(509) 882-4754; fax: (509) 882-5335; e-mail: [email protected]
Industry Coordinator: Dave Brooks
DaimlerChrysler Corporation
800 Chrysler Drive East, Auburn Hills MI 48236-2757
(248) 576-4992; fax: (248) 576-2182; e-mail: [email protected]
DOE Program Manager: Joseph A. Carpenter
(202) 586-1022; fax: (202) 586-2182; e-mail: [email protected]
Participants
George Fenske, ANL Coordinator, Argonne National Laboratory
James Richardson, Intense Pulsed Neutron Source, Argonne National Laboratory
Claude Reed, Laser Applications Laboratory, Argonne National Laboratory
Srinath Viswanathan, ORNL Coordinator, Oak Ridge National Laboratory
Craig Blue, Infrared Processing, Oak Ridge National Laboratory
Thomas Watkins, X-ray Diffraction Analysis, Oak Ridge National Laboratory
Narendra B. Dahotre, Laser Materials Research, University of Tennessee
John Deibler, Computational Mechanics, Pacific Northwest National Laboratory
Nick Klymyshyn, Computational Mechanics, Pacific Northwest National Laboratory,
Ed Flynn, Project Coordinator, General Motors, Corp
Rod Whitbeck, Project Coordinator, Ford Motor Company
Contractor: Pacific Northwest National Laboratory
Contract No.: DE-AC06-76RL01830
Objective
•
Extend the useful life of the dies that are used in die casting aluminum for automobile components. Extending
the life of dies will make it more economically attractive to replace steel components with aluminum
components.
OAAT R&D PLAN: Task 12; Barriers A, B
Approach
•
Identify failure mechanisms.
•
Develop an analytical modeling capability to improve understanding of the relationships between the thermal
cycle, stress, and the amount of accumulated damage in a die.
•
Identify potential solutions, with emphasis on surface modifications.
•
Develop a predictive modeling capability to help identify solutions.
•
Perform in-plant demonstrations of enabling technologies.
45
FY 2001 Progress Report
Automotive Lightweighting Materials
Accomplishments
•
Completed dry-lube and gas-cooling demonstration project. Results showed that this process has commercial
potential.
•
Completed commercial demonstrations of laser-heat-treated die inserts. The heat-treatment process reduced
cracking in sensitive areas.
•
Completed commercial demonstrations of laser-surface-coated die inserts. Two coatings were tested, both of
which showed good initial soldering resistance but were eventually consumed.
•
Completed residual stress study on a four-cavity diode plate die. An unexpected reversal in stress was identified
after an extended service life, and analytical modeling was used to help identify the cause.
Future Direction
•
Complete a few tasks remaining after completion of four major project milestones during the last 3 months of
FY 2001.
•
Document the project results in a final report.
of a dry powder was evaluated for commercial
potential. A pilot-scale demonstration project,
funded by the automobile company partners, was
successfully performed at a commercial die-casting
facility. The castings produced by this method met
all the quality requirements of a production part.
Surface heat-treating was identified as a method
to increase the resistance of dies to heat-checking
and extend the life of dies overheated in service.
Several die inserts were heat-treated on the surface
by a laser and run successfully in commercial
production at a General Motors (GM) plant.
Laser surface-coating processes were developed
at the University of Tennessee and Argonne
National Laboratory. Coatings were applied to parts
of the same dies that had been heat-treated. These
coatings also performed satisfactorily in commercial
operation.
A four-cavity diode plate die and matching pairs
of die sets made from H-13 and KDA-1 alloy steel
were removed from service at intervals over an
extended period of time and tested at Oak Ridge
National Laboratory (ORNL) for residual stress
levels. It was discovered that the residual stress at
the die surfaces cycled between tension and
compression over time.
Introduction
In the first phase of the die life extension
program, the national laboratory participants
identified and categorized the fundamental failure
mechanisms associated with reduced die life. The
study included metallurgical analysis of die
components furnished by industrial partners,
literature reviews, interviews with field experts, and
analytical modeling. Environmentally-assisted
cracking, which had not been considered before, was
identified as a potential contributor to limited die
life. Crack growth studies were performed with
several parting agents, i.e., lubricants; and it was
determined that crack growth rates were accelerated
in the presence of some of them. Localized
overheating was also identified as a major factor
contributing to heat-checking. Tests showed that
infrared heating and laser surface processing had the
potential to reharden temper-softened die surfaces
by selectively heat-treating only the affected areas.
Analytical modeling was used to confirm the
benefits of these approaches and to help determine
the role of residual stress in die performance.
Project Deliverables
Demonstration projects were organized to
evaluate new technologies and approaches that
showed promise for extending die life.
To reduce the debilitating effects of splashing
cold water on hot dies, a process involving gas
cooling and the application of a lubricant in the form
Planned Approach
Summaries of the completed demonstration
projects follow.
46
Automotive Lightweighting Materials
FY 2001 Progress Report
experience in Japan has reportedly shown a
reduction in porosity defects when dry powder lube
systems are used.
The degree of heat-checking was undetectable in
both processes. This result was not unexpected,
because the runs were not of sufficient duration to
generate surface cracks. Japanese studies have also
reported a 20-X increase in die life when heatchecking is the primary failure mode.
Soldering was observed on the barrel insert for
the throttle body casting (see Figure 1), but the
amount of solder build-up was actually less for the
powdered lubricant process than for the
conventional water-cooled cycle. However,
soldering increased when the dry-lube process was
shortened to reduce cycle time. These results
correlated with an increase in the die insert surface
temperature as recorded by thermographs and an
imbedded thermocouple.
Pilot Testing of Gas Cooling and Dry Powder
Lubricant
Prior modeling work and experimental studies
performed on the thermal fatigue apparatus at Case
Western Reserve University suggested that
resistance to heat-checking could be improved by
substituting gas for water during the cooldown
portion of the die-casting cycle. Eliminating water,
which is also used to apply the lubricant (i.e., parting
agent), required a dry lubricant application process.
A commercially available delivery system for
powder lubricant was custom-fit to an existing
production die-casting press, and over 6000 castings
were produced at the MagTec casting facility in
Jackson, Michigan. The results of the quality
comparison showed that castings produced with
liquid lubricant and those produced using the dry
lubricant were comparable, and both were free of
porosity defects. Large-volume production
Figure 1. Soldering patterns on large barrel inserts for experiments comparing wet and dry lube cycles.
47
FY 2001 Progress Report
Automotive Lightweighting Materials
Laser Surface Heat Treating
Experimental studies performed earlier on the
Case Western Reserve University thermal fatigue
testing system had shown that surface heat-treating
significantly reduced heat-checking. Figure 2 shows
a die insert designated as GM detail 504 and heattreated on the surface with a 1.6-kW pulsed
Nd:YAG laser equipped with a fiber-optic beam
delivery system by Argonne National Laboratory.
Figure 3. Photograph of a used GM insert, detail 504,
showing heat-checking in area 1.
Figure 2. Photograph showing laser surface heat-treated
areas on a GM die insert. The heat-treating
was performed by the Laser Applications
Laboratory at Argonne.
Figure 4. Photograph of laser heat-treated and lasercoated GM die insert after 33,285 shots.
Vertical laser coating bands are still visible on
one surface.
The laser treatment leaves overlapping bands on
the surface of the die that can readily be discerned in
the photograph. This insert and another one heattreated by the University of Tennessee were placed
into commercial production at the GM Bedford
facility. The two dies collectively produced more
than 50,000 castings.
Another view of a standard 504 insert that was
not surface heat-treated is shown in Figure 3.
Extensive heat-checking can be observed in the
region designated as Area 1 in the photograph. For
comparison, Figure 4 shows the identical area on the
insert that was surface heat-treated with a
continuous-wave multi-kilowatt industrial Nd:YAG
laser equipped with a 600-micron-diameter fiber
optic for beam delivery at the University of
Tennessee. Virtually no evidence of cracking can be
seen on the laser surface-treated insert in the
affected area. However, photomicrographs have
revealed that a few small cracks do exist in the
curved area adjacent to the heat-treated region.
Further improvements in controlling the laser heattreating process along concave and convex surfaces
will be required to enhance the efficacy of this
process.
Laser Surface Coating
Originally developed as a cladding process, laser
surface processing has been modified to apply
cermet coatings consisting of a solder-resistant
ceramic that is cemented to the die surface with a
metallic binder. A coating developed at the
University of Tennessee, consisting of titanium
diboride particles in a ferrous alloy binder, was
applied to the post or pillar portion of the die insert
in Figure 4. No soldering can be observed on this
surface after 33,000 shots; however, very little
coating remains. The opposite surface—which
cannot be seen in the photograph—was heavily
48
Automotive Lightweighting Materials
FY 2001 Progress Report
soldered, and no remnants of the coating could be
detected on it. Castings that were examined at two
prior service intervals showed little evidence of
soldering. These results suggest that the coating,
while resistant to soldering, was slowly consumed.
Heavy soldering apparently occurred very rapidly
once the coating was gone. Similar observations
were made on another insert coated with tungsten
carbide by Argonne National Laboratory.
then go immediately into a state of residual tension
as a result of the compressive plastic strains imposed
during the injection of hot molten aluminum into the
die cavity. The active stress in the die is reversed
when cold water is applied to cool the hot surface;
but the temperatures are low enough, and the yield
strength of the steel high enough, to preclude plastic
tensile strains from developing. Thus the residual
stress remains tensile. Prior observations reported by
Case Western Reserve University on the
redevelopment of compressive residual stresses were
not widely published because they were believed to
be erroneous. However, these new measurements,
some of which were made at extended service life,
show clearly that compressive residual stresses can
be reestablished.
The potential consequence of this residual stress
reversal is reduced die life, because the stresses are a
result of plastic tensile strains being imparted into
the die. These strains will promote crack nucleation
and accelerate the growth of existing cracks.
Analytical modeling has been used to create a
scenario that helps explain why this reversal is
occurring. Initially, an insulating layer of baked-on
die lube builds on the surface. This layer retards
solidification of the casting and extends the casting
cycle, maybe by as much as 7 seconds in some
cases. To compensate, additional cold water is
sprayed onto the die while it is hotter than normal.
Tensile stresses develop rapidly, and the
temperature-dependent yield strength of the die is
exceeded. Plastic tensile strains form in the surface
layers of the die, and these force the residual stress
back into compression.
Residual Stress in Production Dies
Ford Motor Company provided a set of female
die halves from four cavity diode plate die sets. Two
of these were made with premium-grade H-13 die
steel and two with KDA-1 alloy. Residual stresses
were measured on these dies at ORNL using X-ray
diffraction methods over a 2-year period. The initial
measurements were made when the dies were new—
that is, at zero service life—and then after 25,000,
113,000, and 235,000 shots. The residual stress
results are plotted as a function of service life in
Figure 5, together with other data reported by
Iwanaga1 and Case Western Reserve.2
80
60
40
20
0
-20 0
-40
-60
-80
50
100
150
200
25 0
O R N L /H -1 3 -E
O R N L / K D A -E
CW RU
IW A N A G A
O R N L /H -1 3 -C
O R N L / K D A -C
-100
-120
References
Figure 5. Comparison of the residual stresses measured
in various commercial dies as a function of
cycle time multiplied by 1000.
1. S. Iwanaga, “Initiation and Propagation of
Heat Checking and Variation of Residual Stress in
Aluminum Die Casting Dies,” Transactions, 19th
International Die Casting Congress and Exposition,
paper T97-081, 1997.
2. Dave Schwam, Case Western Reserve
University, private communication with Edward
Courtright, Pacific Northwest National Laboratory,
March 2001.
It was previously shown that dies begin life in a
state of residual compression as a result of the
surface polishing that is performed as the last step
prior to the die’s being placed into service. The dies
49
Automotive Lightweighting Materials
FY 2001 Progress Report
3. ADVANCED MATERIALS DEVELOPMENT
A. Low-Cost Powder Metallurgy for Particle-Reinforced Aluminum Composites
Project Leader: Dr. Jean C. Lynn
DaimlerChrysler Corporation
800 Chrysler Drive
CIMS 484-01-13
Auburn Hills, MI 48326-2757
(248) 576-3192; fax: (248) 576-2176; e-mail: [email protected]
Project Administrator: Dr. Manish Mehta
National Center for Manufacturing Sciences
3025 Boardwalk
Ann Arbor, MI 48108-3266
(734) 995-4938; fax: (734) 995-1150; e-mail: [email protected]
DOE Program Manager: Joseph Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: U.S. Automotive Materials Partnership
Contract No.: DE-FC05-95OR22363
DOE CRADA No.: 96-Mult-AMP-0444(EE-07-02)
Objectives
•
Develop low-cost powder metallurgy (PM) manufacturing methods for particle-reinforced aluminum (PRA)
composite components.
•
Advance PRA machining technology and PRA composite design methodologies.
•
Provide sufficient information that U.S. Automotive Materials Partnership (USAMP) members and suppliers
can make sound commercialization decisions regarding PM PRA technologies.
•
Through participation by leading suppliers, establish (enhance) technical and manufacturing capabilities in
North America to provide low-cost PM PRA components to the automotive industry through the USAMP
partners.
OAAT R&D Plan: Task 12; Barriers A, B
Approach
During this 6.5-year project, the following tasks will be executed in two sub-projects, press and sintering
(P&S) and direct powder forging (DPF):
•
Establish property requirements and cost targets.
•
Develop and assess low-cost PM processes for PRA components.
•
Develop and apply process models for PM PRA technologies.
51
FY 2001 Progress Report
Automotive Lightweighting Materials
•
Develop machining technology for PM PRA components.
•
Develop mechanical behavior and wear database for PM PRA materials.
•
Develop PRA composite design optimization methodologies.
•
Combine new process, machining technology, and composite design methodologies to optimize and fabricate
PRA materials for at least two components.
•
Bench and engine test at least three PM PRA components.
•
Develop and continuously update cost models for PM PRA materials, fabrication processes, and PRA
machining processes.
Accomplishments
P&S sub-project to develop moderate-strength aluminum composites:
•
Fabricated Al-Si alloy PM PRA transmission oil pump gear sets with new dimensionally accurate tooling,
demonstrating team-developed technology. Parts have passed two major durability hurdles, including wear tests.
Supplier has nearly completed evaluation of all candidate new coatings and the testing of compacted gears to
meet more stringent wear and durability test regime defined with the original equipment manufacturer (OEM)
pump test engineers. P&S material/process design database is being finalized.
•
Tested new wear coatings for further evaluation based on cost performance.
DPF sub-project to develop high-strength, high-temperature aluminum composite applications:
•
Established the feasibility of PM PRA connecting rod by demonstrating component fabricated in small lots
using existing (non-optimized) tooling with Al-SiC powder blend and preforms.
•
Selected the coarse/coarse powder blend for process optimization. Additional commercial powder blending
methods are being evaluated with suppliers to optimize processing and handling.
•
Conducted tensile and fatigue property tests at room temperature and conducted microstructure studies. The
results indicate a high probability of achieving design and performance strengths. Elevated temperature testing
is needed on the optimum blend.
Future Direction
•
Complete all wear/durability/dimensional conformance studies to recommend optimized P&S technology for
high-volume gear set production (prior to OEM evaluations of supplied gear sets).
•
Complete evaluation trials of commercial preforming, blending, and agglomeration technologies for DPF
technology; fabricate/test 150 connecting rods from optimum parameters for OEM evaluation.
•
Complete composite material/process design guidelines and database for both P&S and DPF projects.
•
Identify and explore additional opportunities for cost reduction and new powertrain applications of Al powder
blends.
Major Project Deliverables
In FY 2001, the P&S and DPF teams delivered
the following items to USAMP:
•
•
A set of sintering process conditions (correlated
to a wear-resistant alloy microstructure) for the
52
P&S process, based on microstructural analysis
and calorimetry measurements.
For the DPF program, an alternative powder
shape consolidation process (that does not
require a binder or lubricant additive) for
fabrication of forging preforms was developed
and recommended for the model test shape.
Automotive Lightweighting Materials
•
•
•
•
FY 2000 Progress Report
development, processing, and handling experience
gathered thus far suggest that design rules and
tooling (presently applicable to ferrous-based
materials) may need to be revised in order to
completely exploit the benefits of using aluminum
PM composites.
The PM PRA program expects to achieve
several key objectives in FY 2002. Note that the
P&S effort with the supplier team will formally end
after the first quarter of FY 2002. The original
equipment manufacturers (OEMs) will evaluate the
results and components for bench testing and
dynamometer testing during the reminder of
FY 2002.
The emphasis of the team will shift in FY 2002
to supporting the DPF team, led by Metaldyne, in
successfully demonstrating the DPF process for
manufacturing optimized high-strength Al-SiC
composite bearing caps for preforms and
demonstrating Metaldyne’s sinter-forge processing
to permit evaluation of base material property and
dimensional and thermal fatigue. Another objective
in FY 2002 is the optimization of powder flow
technology. Next, prototype aluminum connecting
rods and tooling will be designed and optimized to
meet the OEM engine design specification. The
original objective to produce up to 150 connecting
rods for evaluation in bench test devices still stands.
Analyses from fabricated connecting rod samples
will be used to optimize processes and tooling and
perform machining evaluations. The DPF program
is scheduled to conclude at the end of the first
quarter of FY 2003.
The USAMP team stands committed to
achieving its major P&S and DPF technology
milestones in FY 2002 for component fabrication,
testing, and validation.
Investigation of the commercial potential for
incorporating the process in a connecting-rod
manufacturing line.
Specimen fatigue testing, mechanical property,
and modeling data for the quantitative
assessment of material alloy and blend
characteristics in DPF PM PRA.
Identification and test results from investigation
of commercial powder blending technologies
required for the agglomeration of new
aluminum alloys and composite materials using
PM processes and best practices.
The PM PRA project management team directed
an effort led by Aluminum Consultants Group to
develop a project database for the P&S and DPF
projects. Its purpose is to capture key technical
information from the project for use by
powertrain product designers and process
engineers. The project website, maintained by
Technologies Research Corporation–National
Center for Manufacturing Sciences, is being
used for real-time development of the P&S
database, with the final version to be produced
on CD for archival and distribution purposes.
The team has also finalized an alloy
designation/nomenclature system for the DPF
project.
Conclusions
This DOE-funded USAMP program is nearing
completion with major process and material
development milestones achieved during FY 2001.
The P&S project has demonstrated high potential
for using PM aluminum composite materials in both
moderate strength and high-strength/hightemperature applications in high-volume production.
The insights gained from the DPF project materials
53
Automotive Lightweighting Materials
FY 2001 Progress Report
B. Low-Cost Cast Aluminum Metal Matrix Composites
PNNL Contract Manager: Mark T. Smith
(509) 376-2847; fax: (509) 376-6034; e-mail: [email protected]
PNNL Principal Investigator: Darrell R. Herling
(509) 376-3892; fax: (509) 376-6034; e-mail: [email protected]
DOE Program Manager: Joseph A. Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: Pacific Northwest National Laboratory
Contract No.: DE-AC06-76RL01830
Objectives
•
Develop low-cost aluminum (Al) metal matrix composite (MMC) materials that are cost-competitive with
typical Al alloys used in the automotive industry.
•
Develop a modular mixing and holding system for low-cost Al MMC materials.
•
Develop innovative technologies to produce Al MMC automotive braking systems and powertrain components.
OAAT R&D Plan: Task 12; Barriers A, B
Approach
•
Develop rapid MMC mixing process and lower-cost, castable Al MMC material.
•
Conduct a foundry evaluation of castability.
•
Evaluate mechanical and physical properties.
•
Conduct friction and wear testing of sample products.
•
Explore innovative casting and finishing options.
•
Develop innovative brake system designs to utilize Al MMC materials.
Accomplishments
•
Completed scale-up of rapid MMC mixing process from 60-kg batch sizes to 600 kg. The cost of developing
the larger equipment was shared by MC-21, Inc.
•
Completed the processing of innovative cast and pressure-infiltrated samples for wear testing.
•
Completed wear and friction testing of seven candidate Al-based MMC materials at Rockwell Science Center
using an instrumented test system.
•
Provided sample materials for brake wear simulation testing by Ford Research Laboratory.
•
Initiated collaboration with Visteon Chassis System for the development and testing of Al MMC brake system
components.
55
FY 2001 Progress Report
Automotive Lightweighting Materials
Future Direction
•
Continue development work in the area of evaluating innovative casting and machining concepts, using the
low-cost MMC material approach. Development of a brake system design task involving automotive industry
Tier I suppliers will occur in FY 2002. This project will be focused around the use of Al MMC materials in
automotive braking systems.
Introduction
Modular Rapid Melting/Mixing/Holding
MMC System
The initial goal of the project is to develop the
next-generation mixing process for producing lowcost Al MMC materials for casting applications. The
concept is to use a modular melting-mixing-holding
unit, designed to be placed on the foundry floor, in
which the MMC material can be mixed as needed
for casting various products. The rapid mixing
process, combined with the use of lower-cost SiC
reinforcement feedstock, was selected to reduce the
overall cost of particulate-reinforced Al MMC
materials. The target material cost, which was
established by the domestic automotive industry, is
approximately $1/lb for production of a range of Al
MMC products.
As part of the development of the modular
mixing concept and low-cost Al MMC material, it is
necessary to evaluate the materials that are generated
from the new process and equipment and compare
them with a currently available commercial product.
The Al MMC material produced by Duralcan, Ltd.,
is considered to be the industry standard for stir-cast
Al MMC material. Therefore, mechanical and
physical property evaluation trials were conducted to
compare the properties of the new low-cost
materials against those of the equivalent Duralcan
material.
The ultimate goal of the project is to
demonstrate the ability to reduce the cost of cast
MMC powertrain and brake system components to
that of a comparable system consisting of cast iron
components in an automobile application. In
addition to reducing the cost of the feedstock Al
MMC material, manufacturing costs must be
reduced to produce a component at a cost-effective
price. Therefore, to further reduce the cost of Al
MMC components, innovative casting technology
projects have been initiated to address the issues and
the low-cost requirements of casting/manufacturing
of MMC brake system components.
To make Al MMC materials more attractive for
widespread use in automotive applications, rawmaterial costs will need to be reduced. The current
cost for castable Al MMC material is $2–3 per
pound, depending on the quantities ordered from the
supplier. The domestic automotive industry has
estimated that it would require a feedstock cost of
approximately $1 per pound to make MMCs more
attractive for use in powertrain and brake system
components. Therefore, the primary goal of this
project is to develop the processing methods to
produce Al matrix composites at a lower cost, with
no detrimental effect on the mechanical and physical
property performance. In addition to reducing the
compositing time through the application of a rapid
mixing process, a further cost savings can be
realized by reducing the raw material cost of the SiC
particulate used.
MC-21 in Carson City, Nevada, developed an
evolutionary compositing process to mix MMC
materials. The design of the mixer incorporates a
rapid mixing process that significantly shortens
compositing time and, therefore, reduces labor costs
associated with production of the MMC material.
The concept of the rapid mixing process calls for inplant placement of the system so that the MMC
material can be produced as needed. This
arrangement can eliminate the need for remelting
prior to casting, which can result in added expense
and detrimental effects to the material quality. The
system is modular in design, so it can be scaled
according to the casting foundries’ business needs,
thus providing for a cost-effective flexible
compositing method. In addition, the ability to
produce smaller batch sizes potentially allows rapid
changeover between matrix alloy/reinforcement
product combinations, at a reduced cost. A U.S.
patent (6,106,588) issued in 2000 covers the rapid
mixing concept and the details of the process.
Along with a less labor-intensive compositing
process, a lower-cost SiC reinforcement can be
56
Automotive Lightweighting Materials
FY 2001 Progress Report
mixed while controlling pour temperature. All three
units were demonstrated during two 600-kg material
runs. The resulting cast ingot material was delivered
to Pacific Northwest National Laboratory (PNNL)
for microstructural evaluation. Additional materials
will be provided to selected foundries and casting
houses for evaluation. The estimated production rate
for the 600-kg system is two batches per 8-hour shift
with an electric melter unit, and four to six batches
per 8-hour shift with an auxiliary reverb melter. The
system is designed in a modular fashion in order to
promote the installation on a foundry floor, and the
capacity can be sized to meet specific needs.
utilized for further cost savings. For most stir-cast
Al MMC materials, a specific grade of SiC
particulate is used as reinforcement. The designation
for this grade of SiC is F500, which indicates that a
narrow size distribution of particle sizes is used—
approximately 15–25 µm. The added separation
required to achieve the narrow size distribution
increases the cost of the raw material and adds to the
overall cost of the MMC material. For the production of an Al MMC material with the MC-21
process, a lower-cost SiC material was selected that
had a wider size distribution of approximately
3–30 µm. MC-21 estimates the price for a
359/SiC/20p composite material, using the lowercost SiC and the rapid mixing process, at $1 per
pound.
During FY 2001, MC-21 completed the scale-up
of the modular mixing process from the previous
60-kg capacity to 600 kg (Figure 1). The larger
equipment consists of a separate melting furnace, a
modular mixing unit, and a separate holding unit
capable of keeping the reinforcement phase properly
Low-Cost Al-MMC Material Testing and
Evaluation
During FY 2001, MMC testing and evaluation
has focused on microstructural, mechanical property
and thermal property evaluations of the material
systems that are included in the innovative casting
task. Included in the materials evaluations are the
baseline Duralcan F3S.20S and the new MC-21 lowcost SiC materials, both statically cast at 20-volume
fraction reinforcement. Innovative shape casting and
molding concepts include the centrifugally cast
Duralcan and MC-21 materials and the pressureinfiltrated preform materials produced by Metal
Matrix Cast Composites (MMCC) in Waltham,
Massachusetts. In addition, ceramic-based
composite samples were evaluated that were made
by BFD, Inc., in Columbus, Ohio, using BFD’s
ONNEX process. Microstructural evaluations have
focused on characterization of reinforcement volume
fraction, reinforcement distribution, and porosity
levels in feedstock and cast materials. Oak Ridge
National Laboratory (ORNL) is performing thermal
physical property measurements, including thermal
conductivity, thermal diffusivity, and coefficient of
thermal expansion. Sample materials for the thermal
property measurements were delivered to ORNL
near the end of the second quarter. Updated property
data are provided in Table 1 on tensile properties for
the MC-21 material, with comparable data on the
Duralcan F3S.20S material. As indicated by the
data, the mechanical properties are in line with the
benchmark Duralcan data.
Figure 1. Prototype 600-kg modular mixing
system built by MC-21: melter
(rear), mixer (middle), and holding
furnace (foreground).
57
FY 2001 Progress Report
Table 1.
Automotive Lightweighting Materials
Tensile data for MC-21 and Duralcan materials
Composite material
Modulusa
(GPa)
Elongationb
(%)
MC-21 359/SiC/20p w/F500SiC
100.5
MC-21 359/SiC/20p w/low cost SiC
102.3
Duralcan F3S.20S ingot tested
100.4
Duralcan F3S.20S product data
98.6
a
Results obtained by ultrasonic testing.
b
T6 condition values.
c
UTS = ultimate tensile strength.
d
0.2% off set not achieved.
e
Fracture strength.
0.38
0.62
–
0.40
Because of the unique structures that were
generated during centrifugal casting and preform
infiltration, the evaluation of mechanical properties
from representative reinforced areas of the samples
focuses primarily on compression testing using
cylindrical samples. Compressive properties are felt
to be representative of the primary loading
conditions that occur in brake rotor applications, and
they provide a basis for comparison among the
various candidate materials. Room-temperature
compression tests were performed on the materials
made by the innovative shape-casting methods; the
results are tabulated in Table 2. The data represent
Table 2.
Duralcan F3S20 reinforced region
MC21 new pig ingot reinforced
region
MMCC Al-Mg/Al2O3/40p
reinforced region
MMCC Al-Si/SiC/40p
reinforced region
BFD ONNEX cermet rotor
UTSc
T6
T71
(MPa) (MPa)
310
NA
288
240
NA
290e
217
262
Innovative Casting and Finishing Technology
The goals of the Partnership for a New
Generation of Vehicles (PNGV) project depend
significantly on the ability to achieve weight
reduction cost-effectively. One area that has been
targeted for weight reduction is the brake system, in
which replacing cast iron with an AL MMC is
desirable. Although the current prototype PNGV
vehicles have Al MMC brake rotors, they are
considered too expensive for high-volume use. The
target is to achieve the weight reduction at cost
parity with cast iron.
It is recognized that meeting the goal of a lowcost brake rotor will require more than just a lowercost Al MMC material. There are opportunities to be
realized in reducing costs for brake system
components through the application of innovative
manufacturing methods and design. Therefore, work
was initiated on innovative casting and machining
technologies to address downstream processing
methods and costs related to the production of an
automotive brake rotor.
During FY 2001, contracts were placed with
BFD and MMCC to produce Al MMC test disks for
brake wear testing and for use in characterizing
materials properties and microstructures. In addition
to the BFD and MMCC test materials, PNNL
prepared a series of test disks using conventional
static gravity casting and centrifugal casting to
produce a segregated, selectively reinforced disk.
Samples of each material set were then machined
into 200-mm-diameter, 12.5-mm-thick test
specimens and delivered to Rockwell Science Center
for wear testing. Wear and friction testing were
conducted at two initial operating temperatures,
150ºC and 400ºC. The initial brake pad materials
Compression data
Material
Yield strength
T6
T71
(MPa)
(MPa)
290
NA
274
214
d
NA
310
214
0.2% offset yield
(MPa)
352
349
204
269
248
the compressive strength of the reinforced areas,
with the centrifugal cast materials showing the
greatest strength. The volume fraction in the
reinforced region of the centrifugal cast components
is approximately 36–38%, while the pressureinfiltrated MMCC materials have a reinforcement
volume fraction of 40%. The compressive strengths
are consistent with the expected values for Al MMC
materials with this level of reinforcement.
58
Automotive Lightweighting Materials
FY 2001 Progress Report
were the Plymouth Prowler rear pads (which operate
on an aluminum MMC brake rotor) and a standard
cast iron pad for the baseline iron rotor tests. Results
of the testing are shown in Figure 1. Note that using
the Prowler MMC formulated brake pad material
with the MMC rotor specimens resulted in very high
wear rates for the friction pad. Additional work
performed by Rockwell Science Center investigated
the use of more aggressive friction pad materials,
resulting in more favorable pad wear rates.
During the last half of FY 2001, a collaboration
was initiated with Visteon Chassis Systems
(Dearborn, Michigan). Visteon is now in the process
of signing a cooperative research and development
agreement with PNNL and will participate in the
design and testing of brake systems that incorporate
Al MMC brake rotors. During the first half of
FY 2002, Visteon will develop the design for the test
rotors and provide it to PNNL and the selected
MMC producers. It is expected that test rotors will
be produced by PNNL, BFD, MMCC, and THT
Presses in Dayton, Ohio (which uses the squeeze
casting process). PNNL, in conjunction with ORNL,
will characterize the mechanical and thermal
properties of the scaled-up test rotors and supply
material properties data to Visteon for evaluation
and incorporation into its brake system models.
PNNL and the Aluminum Consultants Group will
update cost models for the new rotor and brake
system designs and compare them with the baseline
cast iron system. In addition to developing the MMC
brake rotors, the team will examine additional Al
MMC applications in conjunction with the Original
Equipment Manufacturers Steering Committee and
Visteon. As required, additional friction and wear
evaluations will be performed at the Rockwell
Science Center.
Future Work
With the promise of a lower-cost Al MMC
material, along with supporting work on
downstream processing, it is important to consider
the opportunities for innovative brake system
design. Designs must take advantage of the different
physical properties of Al MMC materials, as well as
their unique manufacturing requirements compared
with cast iron, rather than simply substituting them
in existing designs. Therefore, as part of the
FY 2002 activities, the project will secure the
participation of a Tier 1 brake system supplier in a
cost-shared project to develop innovative brake rotor
designs that take full advantage of the light weight
of Al MMCs as a cost-effective replacement for cast
iron. It is envisioned that this project would proceed
through the steps of design, prototype
manufacturing, and testing.
Gray Iron
Gray Iron
0.5
F3S20S
0.3
MC21LC20
0.3
BAT01A
-0.5
BAT01B
-0.5
ONNEXW1561
-0.2
Rotor
0.37
93.70
9.5
58.20
304.00
49.40
241.00
166.70
288.30
BBA
2004
0.3
PA533
0.29
PA533
0.28
PA533
0.26
PA533
0.26
PA533
0.29
PA533
Pad
Friction Coefficient
Normalized Wear Rate (10-8 in/ft)
Figure 1. Aluminum metal matrix composite rotors tested against PA533 at 400°C .
59
Automotive Lightweighting Materials
FY 2001 Progress Report
C. Structural Cast Magnesium Development
SCMD Project Chairman: Richard J. Osborne
General Motors Corporation
Mail Code 480-205-314
30007 Van Dyke Road
Warren MI, 81033-57039
(810) 575-7039; fax: (810) 492-5115; e-mail: [email protected]
SCMD Project Administrator: D. E. Penrod
Manufacturing Services and Development Inc
4655 Arlington Drive
Cape Haze, Florida 33946
(941) 697-5764; fax: (941) 697-5764; e-mail: [email protected]
DOE Program Manager: Joseph A. Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: U.S. Automotive Materials Partnership
Contract No.: DE-FC05-95OR22363
CRADA No.: 00-Mult-AMP-0596
Objectives
•
Develop the science and technology necessary to implement a front structural cradle (and two other magnesium
castings) that will interface with other concurrent magnesium programs proposed for the U.S. automotive
industry.
OAAT R&D Plan: Task 11, 12; Barriers B, C
Approach
•
Investigate various methods of casting and techniques used.
•
Develop mold fill parameters.
•
Develop/investigate corrosion factors and methods of protection.
•
Develop nondestructive evaluation (NDE) and sensor applications.
•
Enhance the current design guideline.
•
Develop and continue the existing database.
•
Develop joining and best practices.
•
Develop a business case, cost model, and actual casting of a Corvette front magnesium cradle.
•
Transfer technology for all project items to industry.
61
FY 2001 Progress Report
Automotive Lightweighting Materials
Accomplishments
•
Initiated the development of a magnesium material design property database.
•
Started collecting magnesium production castings for the development of the material property database.
•
Selected a Corvette front cradle (currently a production aluminum production casting) for conversion to
magnesium.
•
Selected two additional magnesium production castings for project development.
•
Established nine steering committees to work with the core team on solving the project tasks.
•
Collected commercial software from industry suppliers for evaluation by Oak Ridge National Laboratory
(ORNL). This software will be used for future project modeling work.
•
Started corrosion evaluation and casting protection work.
•
Started quantitative characterization of cast microstructures on the magnesium production castings.
•
Developed a magnesium metal material design property database.
Future Direction
•
Award a tooling order (for a magnesium cradle) to one of the project participants.
•
Follow the tasks listed in the project statement of work to complete the project on time.
•
Continuously solicit several additional industrial (equipment) participants.
•
•
•
Introduction
The structural cast magnesium development
(SCMD) project objectives focus on resolving
critical issues that now limit the large-scale
application of magnesium castings in automotive
components. This project will develop the science
and technology necessary to implement a front
structural cradle (and two other magnesium castings)
that will interface with other concurrent magnesium
programs proposed for the U.S. automotive industry.
Such components offer all of the difficult
manufacturing issues—including casting processes
(e.g., high-pressure die casting and semisolid, lowpressure, and squeeze casting) and joining—along
with harsh service environment challenges, such as
corrosion, fatigue, and the stress relaxation
associated with fasteners.
The project team includes personnel from
• The Big Three automotive companies
• More than 34 companies from the casting supply
base
• Academic personnel
• Independent testing and research laboratories
• American Foundrymen’s Society
• Technical associations
ORNL
Sandia National Laboratories (SNL)
Lawrence Livermore National Laboratory
(LLNL)
Figure 1 illustrates the technology we are trying
to develop to improve the properties of cast
components. Figure 2 illustrates a generated
Property Influence
Reduced material property
variation with increased = Lower Cost
design strength leads to …... & Weight
•
•
•
•
Material
Properties
YS
UTS
Ductility
Fat.Str.
Corvette Front Cradle
mean
Figure 1. Technology will be developed that will help
engineers and product designers reduce cast
component variation, thereby yielding lower
component cost and weight.
62
Automotive Lightweighting Materials
FY 2001 Progress Report
has begun evaluating quality assurance
methodologies/equipment and will further
develop a new radioscopic standard for
magnesium castings based on improving
American Society for Testing and Materials
(ASTM) standards.
Magnesium Alloy Database
Users
• Product Designers
• Release Engineers
• Structural Analysts
Contents
Technology
Technology
• Material Data
• Engineering Properties
• Microstructure / Property
Linkage
Features
• Data Comparison
• Searches/plotting
• Statistical Information
•
Benefits
• Enhance product design
• Improved simulation parameters
• Reduce component weight
Evaluation and development of numerical
modeling techniques to predict cast
microstructure and subsequent mechanical
properties throughout cast component sections.
Figure 2. Comprehensive materials database.
A major goal of the project is to develop mathbased models that predict the microstructure and
subsequently the mechanical (tensile and cyclic)
properties of cast components from the mold,
casting, and component function criteria
(Figures 4 and 5). The results of this work will
be a joint effort of industry participants, national
laboratories, and academia.
database that uses an entirely different architecture
for comparing magnesium. It includes historical
literature data and comprehensive mechanical
property data derived from samples excised from
actual production magnesium castings. Casting
processes examined include high-pressure diecasting and gravity, permanent-mold, semisolid,
low-pressure and squeeze casting.
Microstructure
Monitoring Technologies
•
The study of variability in ductility and ultimate
tensile strength of production casting of magnesium
alloys has started. These samples have been
successfully correlated to the presence of defects
through careful qualitative and quantitative
fractographic observations and quantitative
microstructural data obtained through digital image
analysis and stereological techniques. Quantitative
correlation between tensile elongation and total area
Development of on-line process monitoring,
feedback control, and non-destructive evaluation
technique(s) to ensure cast component
consistency and quality.
LLNL previously adapted and applied rapidresponse temperature optical sensors for process
monitoring of two production casting facilities
in a previous project. Similar applications will
be used for the SCMD project (Figure 3). LLNL
High Pressure Cavity Fill Analysis
Sensors and NDE Technology
Sensor Production Trials
1. Tilt Pour Mold - Stahl Specialty
2. HPDC - Gibbs Die Cast
Infrared Fiber Optic Sensor
Initial shot sleeve/ plunger position
Position after 0.16 sec
Project Support
New Foil Design
ASTM Level 3 (1/4”)
1. NDE equipment & procedures
2. Sample evaluations
3. Support model validation
Manufacturable Rep. Quality Indicator (RQI)
Rapid cavity fill - 2X
Figure 3. Process sensors, radioscopic quality indicators
and nondestructive evaluation project support.
Medium cavity fill - 1X
Figure 4. Math-based simulation model that predicts
cast microstructure.
63
FY 2001 Progress Report
Automotive Lightweighting Materials
Mechanical Property Simulation Model
Multi-Scale Fracture Mechanics
ISV Model
Experiment
Fracture of Silicon
Growth of voids
Part Design
Optimize Geometry
Predict Failure
IVS Model
Void
VoidNucleation
Growth
Void/Void
CrackCoalescence
Growth
Void/Particle Coalescence
Void coalescence
FEM Analysis
Component Geometry
Monotonic/ Cyclic Loads
Structural Analysis
Development Activities
1. Sandia constitutive damage
model
2. Material testing
3. Linkage of material properties
to microstructure features
4. Development of multi-scale
fracture theory
Sandia’s model uses 27 material parameters to
predict time and location of static and cyclic failure.
•
Figure 5. Material property simulation model.
fraction of defects (e.g., oxides, pores, flux residues)
of the fracture surface has already been
demonstrated for various groups of production
castings on the previous project. Fractographic
analysis has already been completed of the tensile
test specimens of AM60 Mg-alloy specimens from
an instrument panel. Georgia Tech completed this
analysis, which will be useful for the modeling work
performed by ORNL and SNL.
Project Benefits
Milestones
The expected benefits to DOE and U.S. industry
from successful completion of this research would
include the following:
• Reductions in mass for ground and air
transportation vehicles will lead to reduced fuel
consumption, emissions, and dependence on
foreign oil.
• Automakers are under increasing pressure to
reduce CO2 emissions and increase Corporate
Average Fuel Economy (CAFE) standards. The
North American auto industry currently uses
approximately 70,000 MT of magnesium/year,
which is equivalent to some 3.5 Kg per vehicle.
• The ability to significantly increase magnesium
usage will help the auto industry meet future
Federal CAFE targets and reduce exposure to
CAFE penalties. Cast magnesium structures
have the potential to reduce vehicle mass by an
over 100 Kg, which could reduce emissions by
5% and reduce fuel consumption by
approximately 1.0 mpg (ignoring secondary
mass savings).
• Light metal alloys have greater recycling value
and reduce the energy required for recycling
compared with plastics (including melting,
Cooperative Agreement
•
•
•
•
A front cradle has been chosen for conversion
from aluminum to magnesium.
Two additional magnesium castings have been
chosen for evaluation by the project team.
Detailed quantitative microstructural
characterization has been completed for
production components cast with AM 60 and
AZ 91 magnesium alloys.
Nine project steering committees have been
established and are actively working with the
project team and all industrial participants on
major task issues.
Cooperative Research and Development
Agreement
•
•
tomography. Comparisons are encouraging,
considering that no coalescence is included yet
in the magnesium model, which only considers
void growth. We expect to include more about
pore-pore coalescence as the project continues.
We have also quantified some microstructureproperty relations for fatigue based upon some
microstructural evaluations of AM60B and
AZ91D magnesium alloys. This information
will be used for our multi-scale fatigue model.
LLNL—Radiographic analysis of validation
mule castings has started to determine
discontinuity types and grades at predicted high
stress locations. Implementation testing of a
fiber optics in-mold thermal monitoring system
for high-pressure die-casting will be initiated in
several high-volume production facilities.
Laboratory personnel will start computed
tomography of notched tensile bar magnesium
samples to measure porosity void nucleation and
growth for constitutive equation modeling
parameters.
ORNL—ORNL has received modeling software
from three of the supply teams and started the
evaluation review. The resulting ORNL models
will be used in current commercial software.
SNL—Finite element simulations of AM60B
notch tensile tests were performed to corroborate
experimental load-displacements curves with
porosity data from the interrupted X-ray
64
Automotive Lightweighting Materials
•
•
•
•
FY 2001 Progress Report
•
machining, handling, and transportation energy
requirements).
The competitive global posture of the Big Three
U.S. automakers will increase if they can design
and manufacture vehicles that offer greater
consumer value. This could improve the nation’s
trade balance with countries that market vehicles
that are more fuel-efficient than those produced
in North America.
Health and environmental issues for workers are
reduced during light metal casting operations,
compared with such problems in ferrous
foundries and polymer molding operations.
The national laboratories will gain valuable
manufacturing development and product
application experience.
The national laboratories will gain an
opportunity to develop math-based simulation
models and NDE technologies that benefit both
the auto industry and federal technology
programs.
the delay and requirement by the U.S. Council
for Automotive Research Business and Policy
Committee that all existing 34 project
participants re-sign and comply with the
recently designed generic work agreement.
Two important elements are associated with the
SCMD project:
• The research institutions participating in the
project—universities, the national laboratories,
industry, and CAMNET—must work together to
carry the scientific research that is required (e.g.,
microstructure; effects of modeling; corrosion
and fastening properties, NDE methods).
• Industry needs a fast-track effort to convert an
existing aluminum cradle to magnesium and
have a part ready for testing on an actual vehicle
in 3 years. Therefore, the industrial participants
cannot wait for all of the intricate scientific
aspect to be understood before a process is
chosen, the tooling built, parts cast and the
validation tests are run.
The project scope of work will be followed to
complete the project in time. However, the project
has been somewhat curtailed by
• the delay by the Big Three automakers in
signing the cooperative research and
development agreement
Each “side” of the project, researchers and
industry, must remain constantly aware of the needs
and progress of the other.
65
Automotive Lightweighting Materials
FY 2001 Progress Report
D. Magnesium Powertrain Cast Components
Project Manager: Bob R. Powell
GM Research & Development Center
MC 480-106-212, 30500 Mound Road, Warren, MI 48090-9055
(586) 986-1293; fax: (586) 986-9204; e-mail: [email protected]
Project Administrator: Peter Ried
Ried & Associates, LLC
6381 Village Green Circle, Suite 10, Portage, MI 49024
(616) 327-3097; fax: (616) 321-0904; e-mail: [email protected]
DOE Program Manager: Joseph Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: U.S. Automotive Materials Partnership
Contract No.: DE-FC-05-95OR22363
Objectives
•
Demonstrate and enhance the feasibility and benefits of using magnesium alloys in engine components.
•
Design an ultra-low-weight, cost-effective-performance engine block, bedplate, and structural oil pan using the
best low-cost, recyclable, creep- and corrosion-resistant magnesium alloys.
•
Compare and select the alloys on the basis of common-protocol casting and testing. (The resulting cast specimen
database will provide the necessary design data for the components.)
•
Design and build the component dies.
•
Cast and test the components in operating engines (dynamometer or vehicle).
•
Validate the performance benefits, component durability, and system costs.
•
Create a material specification for magnesium powertrain alloys common to original equipment manufacturers.
•
Identify critical challenges for future magnesium alloy and component developments and use this information to
promote scientific research in North America.
OAAT R&D Plan: Task 11, 12; Barriers A, B
Approach
•
Conduct the Magnesium Powertrain Cast Components (MPCC) project in two phases, each containing three
tasks. A decision gate separates the two phases.
Phase I: Produce the finite element analysis (FEA) engine design, the alloy property requirements for the
engine, the comparison of the alloys in a cast-specimen material property design database, a cost model of
the engine, and a set of scientific challenges to serve as the basis for Requests for Proposals for fundamental
research programs to be conducted in North America.
67
FY 2001 Progress Report
Automotive Lightweighting Materials
Phase II: Produce the magnesium-intensive-engine test results using cast components, assembled and tested
engines to validate Phase I predictions, an extensive design database using alloy specimens excised from
cast components, the common magnesium materials specification, and the results of the fundamental
research programs, the need for which were identified in Phase I.
Accomplishments
•
Launched the MPCC project and achieved participation of the key magnesium alloy producers and die casters,
as well as a full complement of industrial partners.
•
Selected the engine to serve as the baseline for the design of the magnesium-intensive version.
•
Defined the mechanical property database content and format to support the FEA design needs for the engine
components and to compare and select the best magnesium alloys for the components.
•
Completed alloy literature review and compilation of producers’ alloys properties to enable the initial screening
and downselection of the alloys for inclusion in the cast-specimen mechanical property database.
•
Defined the alloy casting protocol and evaluation criteria.
oil supplies, DOE and the U.S. automotive industry
undertook the MPCC project to extend the weightreduction potential of magnesium to the powertrain.
The project will increase interest in magnesium
powertrains and promote greater competition for
and development of magnesium sources,
manufacturing capability, and engineering expertise.
It will also encourage scientific research into the
properties, processing, and behavior of magnesium
by universities and national laboratories in North
America. The project was started in 2001. It is being
conducted in two phases separated by a decision
gate, at which point the predicted feasibility and
cost-effective performance benefits will be
evaluated prior to entering Phase II.
The engine chosen for the MPCC project is the
Ford Duratec, as shown in Figure 1. The
components of this V-6 engine to be designed for
magnesium are the block, the bedplate, and the
structural oil pan. The bedplate and oil pan will be
high-pressure die cast, and the block will be sand
cast.
Introduction
The MPCC project will provide comprehensive
answers to the questions of the feasibility and costbenefit of using magnesium in powertrain
components. Although magnesium has been
demonstrated to significantly reduce weight at
acceptable automotive costs in many areas of the
vehicle, structural powertrain components have not
benefited from this material. The reasons for this are
as follows:
• high cost of alloys that are able to withstand the
operating temperatures of the engine without
deforming under load (creeping)
• limited powertrain design experience with
magnesium alloys
• no long-term field validation or controlled-fleet
testing data of magnesium powertrain
components
• limited scientific infrastructure in the United
States that is directed toward acquiring a
fundamental understanding of magnesium alloys
and casting processes.
Details of Phase I
With the recent development of several
potentially low-cost, creep-resistant alloys, the
growth of non-powertrain magnesium alloy design
experience—including the launch of the Structural
Cast Magnesium Database (SCMD) project by DOE
(see article 3C in this progress report), and the
renewed recognition of U.S. dependence on foreign
The first-phase goals of the project comprise
three tasks: (1) evaluation of the alloys, (2) FEA
design of the magnesium engine components, and
(3) identification of the critical scientific knowledge
necessary for future magnesium powertrain
materials.
68
Automotive Lightweighting Materials
FY 2001 Progress Report
model will be refined to achieve ultra-lightweight,
cost-effective performance. The design and cost
model results using the best alloys from Task 1 will
form the basis for the decision gate for entry into
Phase II of the project.
Task 3
Task 3 identifies the fundamental scientific
challenges of using magnesium alloys and casting
processes in powertrain components. While alloy
development is not within the scope of this project
directly, the purpose of Task 3 is to promote new
and strengthen existing scientific research in North
America, specifically for future magnesium alloy
and component developments (e.g., determining
detailed mechanisms of creep and fatigue of
magnesium alloys at elevated temperatures and
establishing thermodynamic and phase equilibrium
databases). Several such needs have already been
identified in the project, and these will be presented
in appropriate forums in the near future.
Figure 1. The Ford 2.5L Duratec engine.
Task 1
Task 1 is the evaluation of the new creepresistant alloys based on cast specimens using a
common die and standardized mechanical property
and corrosion tests to yield a short list of suitable
alloys. Evaluation will be based on tensile and
fatigue properties, creep and corrosion resistance,
castability, recyclability, and estimated alloy costs.
On the basis of already published data and
properties data provided by alloy suppliers, the
initial set of candidate alloys has been identified,
their producers have agreed to participate in the
program, and the test matrix has been defined. The
alloys will be cast into the test specimens for
evaluation. The documentation of the test results
will be based on the same database structure and
format as that developed in the project “Design and
Production Optimization for Cast Light Metals” that
was completed for aluminum and launched for
magnesium (SCMD) in 2001. MPCC and SCMD
will have a common database of non-powertrain
magnesium alloys, benefiting both groups of
potential magnesium alloy users.
Details of Phase II
The second-phase goals of the project also
comprise three tasks: (4) casting the engine
components, (5) completing the magnesium alloy
property database, and (6) conducting validation
tests of the assembled engines. Upon entry into this
phase of the project, its specific goals and objectives
will be reconsidered and altered as necessary to best
serve the overall goals of the MPCC project. In
addition, the identification and promotion of
scientific research projects will continue.
Task 4
Task 4 will accomplish the casting of the engine
components, which will ultimately be validationtested in Task 6. The dies for the bedplate and
structural oil pan and patterns for the engine block
will be designed from fill and solidification models
and fabricated. Components will be cast, inspected,
and approved for assembly. Allowances for up to
three iterations of casting trials and subsequent die
and pattern modification have been included in the
project schedule.
Task 2
Task 2 is the FEA design and cost model of the
engine and its magnesium components. The initial
designs will be based on the property data for the
alloys that came from the literature evaluation in
Task 1. As casting and test results are acquired, the
69
FY 2001 Progress Report
Automotive Lightweighting Materials
durability of the components. All tested engines will
be torn down for inspection; and the final report will
contain the test and teardown results, the FEA
designs, the fill and solidification models, the alloy
property databases, the final cost model, a common
magnesium materials specification for powertrain
components, and recommendations for future work.
Task 5
Task 5 will complete the magnesium alloy
database and will use alloy specimens excised from
cast components obtained in Task 4. This database
will not comprise all the initially considered alloys;
rather, only those selected at the end of Tasks 1 and
2 will be included. Validation of “actual” properties
may result in identifying additional areas of
technical challenge and the need for further
scientific research, thereby furthering the objective
of Task 3.
Summary
The MPCC project is an aggressive attempt to
address key concerns regarding the future prospects
for a magnesium-intensive powertrain. Upon
completion of the project, in addition to addressing
these concerns, additional scientific and economic
implementation barriers, if any, will have been
identified and programs to overcome the scientific
barriers will have been undertaken by qualified
North American universities and laboratories.
Task 6
Task 6 will achieve the completion of the
project: the cast components assembled into
complete engines and tested on dynamometers or
vehicles to validate the design, performance, and
70
Automotive Lightweighting Materials
FY 2001 Progress Report
E. Advanced Magnetherm Process for Production of Primary Magnesium
Principal Investigator: Robert J. Fondell
Northwest Alloys, Inc.
1560 Marble Valley Road, Addy, WA 99101
(509) 935-3415; fax: (509) 935-3414; e-mail:[email protected]
PNNL Project Manager: Russell H. Jones
(509) 376-4276; fax: (309) 376-0418; e-mail: [email protected]
DOE Program Manager: Joseph Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: ALCOA, Inc., Alcoa Center, Pennsylvania
Contract No.: 306420-AN7
Objective
•
Design, fabricate, construct, and demonstrate the process, equipment, and operating parameters required to
produce magnesium economically using plasma torch technology and atmospheric condensation of magnesium.
OAAT R&D Plan: Task 11; Barriers A, B
Accomplishments
•
Completed an economic model and demonstrated the successful operation of the plasma torch furnace at
Mintek (Randburg, South Africa) in the submerged arc mode.
•
Made a decision not to pursue the development of a liquid condenser, based on difficulties encountered in the
operation of the prototype condenser.
Mintek portion of the work) and not the productivity
increases associated with operation of the liquid
condenser that would allow continuous tapping of
liquid magnesium.
An engineering study is in progress to retrofit an
existing furnace with a graphite electrode for the
plasma torch operation. Testing at Mintek revealed
that it may be feasible to operate with a submerged
arc environment powered by the Northwest Alloys’
AC transformers. However, a decision by ALCOA
to shut down its Addy, Washington, magnesium
reduction facility resulted in a decision to terminate
this project.
Summary
The prototype liquid condenser experienced two
difficulties: (1) durability of the high-temperature
seal, and (2) stresses that exceeded the design limits
for the condenser material because of the
temperature needed to operate a liquid condenser. A
decision was made to hold off on further
development of the liquid condenser since a
satisfactory solution has not been found for these
issues. This decision was also based on the
economic model identifying potential major cost
savings from changes in slag chemistry and
associated feed materials that are feasible with the
increased heat from the plasma torch (i.e., from
71
Automotive Lightweighting Materials
FY 2001 Progress Report
F. Solid Oxygen-Ion-Conducting Membrane Technology for Direct Reduction of
Magnesium from Its Oxide at High Temperatures
Principal Investigator: Uday Pal
Department of Manufacturing Engineering
Boston University
Brookline, MA 02446
(617) 353-7708; fax: (617) 353-5548; e-mail: [email protected]
DOE Program Manager: Joseph Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Participants
Timothy Keenan, Masters Candidate, Boston University
George B. Kenney, ELMEx, Medford, Massachusetts
Ajay Krishnan, Masters Candidate, Boston University
Christopher Manning, Research Associate, Boston University
Contractor: Boston University
Contract No.: C#407200-AN8
Objective
•
Demonstrate the technical and commercial viability of the solid oxygen-ion-conducting membrane (SOM)
process for producing magnesium directly from its oxide. The SOM process employs yttria-stabilized zirconia
(YSZ) as a solid-oxide, oxygen-ion-conducting membrane to separate the anode and cathode of a hightemperature electrolytic cell. This process exhibits several advantages over existing magnesium production
routes, including improved economics and reduced environmental impact.
OAAT R&D Plan: Task 11; Barriers A, B
Approach
•
Identify molten flux systems to operate the SOM process at 1300–1400(C.
•
Examine the long-term stability of the YSZ SOM membrane in the candidate flux systems.
•
Determine key physical properties of the candidate flux systems: volatilization rate, ionic conductivity, and
mass transfer characteristics.
•
Conduct laboratory-scale experiments with various anode-electrolyte-cathode configurations in order to
determine the dissociation potential and the I-V characteristics for producing magnesium from its oxide in the
selected flux systems.
•
Evaluate various methods of magnesium collection during the SOM process.
•
Operate the SOM cell between 1300 and 1400°C at current density on the order of 1 A/cm2 and maintain low
overall ohmic and polarization losses in the system so that the applied electric potential does not exceed 5V.
This arrangement will keep the power consumption at less than 13 KWh/kg of magnesium produced.
73
FY 2001 Progress Report
Automotive Lightweighting Materials
Accomplishments
•
Conducted experiments to examine potential interactions between the YSZ membrane material and two
candidate flux compositions: 29 wt% SiO2–51 wt% MgF2–20 wt% MgO and MgF2–5.3 wt% MgO. Based upon
these results, the binary eutectic MgF2–5.3 wt% MgO flux was selected as the candidate flux for the SOM
process.
•
Conducted thermogravimetic experiments to determine the volatility rate of the MgF2–5.3 wt% MgO flux at
temperatures of interest for the SOM process.
•
Assembled the experimental apparatus for the in situ investigation of possible molten flux/YSZ interactions
using impedance spectroscopy.
•
Assembled an SOM cell incorporating the MgF2–5.3 wt% MgO flux, the YSZ membrane, and a magnesium
vapor collection apparatus for the laboratory-scale electrolysis of magnesium between 1300 to 14000C.
•
Achieved high magnesium production rates (Faradaic current densities in excess of 1 amp/cm2) without
damaging the YSZ membrane or other components of the electrolytic cell. The applied electric potential was
below 5 V.
•
Operated the SOM cell successfully at 1300°C for more than an hour with a steady-state current density of
800 mA/cm2. The applied electric potential was 4.0 V, and the power consumption was estimated to be
10 KWh/kg of magnesium produced; these figures are clearly much better than the state-of-the-art processes for
magnesium production and also the Hall Cell for aluminum production. Magnesium metal vapor that was
produced was successfully condensed as crystals in the condenser.
Future Direction
•
Complete the impedance measurement experiments and model the long-term stability of the YSZ membrane in
contact with the MgF2–5.3 wt% MgO flux.
•
Determine the ionic conductivity of the MgF2–5.3 wt% MgO flux.
•
Continue the laboratory-scale magnesium electrolysis experiments to identify the rate-limiting process steps,
and optimize the SOM cell configuration and the magnesium vapor condenser.
•
Use the results obtained to verify the energy requirements and the environmental and economic benefits of
magnesium production by the SOM process. The data will be used to model the process for scale-up in order to
evaluate commercial viability.
Introduction
The SOM process is a generic environmentally
sound and energy-efficient alternative for oxide
electrolysis using an inert-oxygen-ion-conducting
membrane-based anode. The process can be an
attractive alternative for the synthesis of highenergy-content metals such as aluminum,
magnesium, titanium, high-purity silicon, chromium,
and value-added ferroalloys. This project focuses on
the synthesis of magnesium metal using this
technology.
The current production methods for magnesium
are either electrolysis from a halide electrolyte bath,
which requires extensive and expensive feedmaterial preparation, or metallothermic reduction
(the magnetherm process) at high temperatures
74
(1600°C) involving expensive metal reductant
(FeSi). The electrolytic magnesium process is based
on an anhydrous chloride feed material that is
recovered from brines via an elaborate and
expensive front-end dehydration process. This feed
preparation process step may represent 80% of the
plant footprint and 30% of its capital cost. This
process also generates a chlorine by-product at the
graphite anode that may represent an environmental
concern and requires additional capital and operating
resources. The alternative metallothermic process
for producing magnesium is based on the thermal
reduction of calcined dolomite or magnesite with
ferrosilicon to form a magnesium vapor that is
collected in an attached condenser. The ferrosilicon
reductant represents 12 KWh/Kg, or roughly 40% of
Automotive Lightweighting Materials
FY 2001 Progress Report
the total energy required to produce magnesium; it
also represents roughly a third of its total production
cost. Furthermore, this batch process generates about
4–5 tons of slag per ton of magnesium that must be
properly disposed of.
In the SOM process, the oxide reduction is
electrochemical, but it dramatically alters the
existing electrolytic magnesium process flowsheet.
It replaces the magnesium chloride dehydration
process with a simple magnesium oxide calcining
operation, thereby significantly reducing front-end
capital and operating costs. It also reduces the
overall energy consumption of magnesium
production by roughly 50%. In addition, the
elimination of the magnesium chloride cell feed
eliminates the production of chlorine and chlorinated
hydrocarbons. Compared with the current
metallothermic and the electrolytic processes, the
SOM process has the potential to be more
economical, less energy-intensive, and more
environmentally sound.
The fundamental configuration of the SOM
process, shown in Figure 1, consists of a solidoxygen-ion-conducting stabilized-zirconia
electrolyte (membrane) that separates the anode
from the melt containing the oxide of the metal to be
reduced.
H (g)
O -(YSZ ) --> H
2
(g) + 2e
A no de
olid-oxide ox ygen-io n-conducting
embra ne
DC
2- (melt ) --> O
- (YSZ )
Ionic melt with dissolved MeO
M e n+ (mel t) + ne - --> Me
atho de (inert)
Elec tron
flow
be increased as long as the potential at the meltzirconia interface does not exceed the dissociation
potential of the solid zirconia and undesired oxides
are not reduced at the cathode. Therefore, larger
potentials can be applied between the electrodes in
order to increase the rate of production of the
desired metal. The full benefit of the SOM process
can be realized if the process is conducted at
temperatures between 1200 and 1400°C.1,2 At these
temperatures, the overall resistance drop across the
stabilized zirconia membrane and the melt are
expected to be sufficiently low to allow high current
densities on the order of 1 A/cm2 or greater to be
obtained. In addition, at these temperatures, the
process efficiency can be further increased by
directly reforming hydrocarbon fuel over the anode.
It should be noted that several attempts have been
made to employ SOMs at temperatures below
1000(C. However, these efforts have not been
successful in developing a commercially viable
process mainly because sufficient high-current
densities could not be obtained through the
membrane.3–8
Work in Progress
Keeping commercial viability in mind, the
present research is aimed at developing a SOM cell
capable of producing and collecting magnesium with
nearly 100% current efficiency at steady-state
current density on the order of 1 A/cm2 or greater
with applied electric potentials not exceeding 5V.
This development would result in a magnesium
production rate that is at least 1.33 times the current
aluminum mass production rate in Hall cells at
comparable power consumption. The following
sections summarize the various components of
current research efforts toward this goal.
Molten Flux Selection
Figure 1. Solid oxygen-ion-conducting membrane
process for metal (Me) production by direct
reduction of its oxide.
An inert cathode is placed in the melt. When the
applied electric potential between the anode and the
cathode exceeds the dissociation potential of the
oxide to be reduced, the desired metal cations are
reduced at the cathode; oxygen ions migrate through
the membrane and are oxidized at the anode. The
applied electric potential between the electrodes can
75
Two melt systems were selected as candidate
fluxes for the SOM process for magnesium
production. Selection criteria include low volatility
at the temperatures of interest, viscosity of less than
10 poise, electrical conductivity of greater than
1 (ohm-cm)í, and MgO solubility of at least 5 wt%,
at temperatures from 1300 to 1400°C. These
requirements are expected to provide kinetically
favorable conditions suitable for scale-up.7, 8, 9–11 The
first flux considered was the eutectic composition of
the binary MgF2–MgO system (MgF2–8mol%
FY 2001 Progress Report
Automotive Lightweight Materials
MgO.) The second flux investigated was based on
the eutectic composition of the ternary MgO–MgF2–
SiO2 system; the exact composition was 29 wt%
SiO2–5 1wt% CaF2–20 wt% MgO.
For the proposed SOM technology to be
technically feasible on a commercial scale, it is
critical that the YSZ membrane be reasonably stable
in the molten flux. Therefore, to identify a single
best candidate flux composition for the SOM
electrolytic cell, preliminary experiments were
conducted to evaluate the stability of the SOM
material, YSZ, in the two melts described earlier. In
these experiments, YSZ crucibles containing the
respective candidate fluxes were heated to the
proposed operating temperature of the cell (1300°C).
The flux-containing crucibles were held at
temperature for periods of 10 and 30 hours and then
cooled to room temperature. After cooling, the
interfaces between the YSZ crucible and candidate
fluxes were examined using optical and scanning
electron microscopy for signs of dissolution,
erosion, or general reaction. Chemical analysis of
the region near the slag-crucible interface was
performed using an electron microprobe equipped
with a wavelength-dispersive X-ray spectrometer.
Examination of the solidified samples indicated
that YSZ is relatively stable in contact with both
slags. For both slag compositions, the reacted zone
between the slag and crucible wall was less than
500 µm after 30 hours. For all experiments,
microprobe analysis of the crucible wall near the
solid-liquid interface indicated a deficiency in yttria.
This suggests that a selective dissolution of yttria
occurs at the liquid-solid interface. It may be
possible to alleviate the problem by adding yttria to
the initial flux composition to reduce the driving
force for yttria dissolution from the YSZ membrane.
It was found that the degree of reaction appeared to
be slightly higher for the ternary slag containing
silica. In addition, it was found that some pitting of
the YSZ crucible occurred in the vicinity of the free
surface of the silica containing ternary slag. This
phenomenon was not observed for the binary MgO–
MgF2 slag. Based on these observations, it was
concluded that YSZ is more stable in contact with
the binary MgO–MgF2 slag, and this flux
composition was selected for further
experimentation.
76
Characterization of MgF2–8mol%MgO Flux
An investigation is in progress to accurately
determine certain key properties of the binary
magnesium oxy-fluoride slag, which has been
selected for use in the SOM electrolytic cell at
temperatures between 1300 and 1400°C. These
properties include
• precise erosion/corrosion rate of YSZ by this
flux
• its volatility
• its ionic conductivity
• its mass transfer properties
The first parameter in the list is critical. The
feasibility and design of a large-scale SOM-type
reactor for the electrolysis of magnesium are highly
dependent upon the long-term stability of YSZ in the
proposed flux. Therefore, laboratory experiments are
planned to precisely determine the rate of attack of
YSZ by the magnesium oxy-fluoride slag. For these
experiments, the plan is to use an impedance
spectroscopy technique that has been successfully
employed in the past to investigate refractory-melt
interactions in other systems.11,12 The scientific
principles of this technique will be only briefly
explained here. The impedance of an
electrochemical cell is a function of several
parameters, including the wetted area and geometric
configuration of the electrodes and the conductivity
of the electrolyte. For carefully designed
experiments, very small changes in the wetted area
of the electrodes can be detected by monitoring
changes in the impedance of the cell. In the present
study, two cylindrical electrodes with a known and
constant separation distance will be inserted into the
molten MgO–MgF2 flux contained in a YSZ crucible
and the impedance measured as a function of time.
Interaction between the MgO–MgF2 flux and the
YSZ will result in a change in the measured
impedance.
Thus far, the experimental apparatus for these
experiments has been assembled. The cell constant
and the dependence of the cell impedance upon the
wetted area of the electrodes has been determined
for the proposed apparatus using a standard solution
of KCl in water with a precisely known
conductivity. Impedance data regarding the stability
of YSZ in the MgO–MgF2 flux as a function of
temperature is being gathered and will be presented
in subsequent reports. These experiments will also
Automotive Lightweighting Materials
FY 2001 Progress Report
allow for the precise determination of the ionic
conductivity of the flux.
The specific rate of volatilization of the MgO–
MgF2 flux is also an important process parameter.
Significant flux losses due to volatilization will not
only complicate the laboratory experiments in
progress, but also decrease the attractiveness of a
commercial-scale process based on this flux.
Therefore, thermogravimetric (TGA) experiments
are in progress to accurately determine the rate of
volatilization of this slag at the proposed operating
temperatures for the SOM process. Experiments
have been conducted using a high-temperature
D101-02 AT Cahn Balance.
For these experiments, approximately 20 grams
of slag were contained in a graphite crucible having
an internal diameter of 2.2 cm and suspended from
the microbalance using a mullite rod. All
experiments were conducted in an inert atmosphere
with a purging flow of argon gas. Volatilization
experiments have been conducted at various
temperatures from 1300 to 1550(C. At these
temperatures, the rate of volatilization has been
found to be in the range of 10í7 to 10í6 g/cm2sec.
This rate is relatively low and will not be a problem
for either the laboratory experiments or the eventual
commercial process. Based upon these experiments,
the temperature dependence of the volatilization
rate, R, was calculated in g/cm2-sec as
R = − 3.96 × 10 −4 exp
−14281
T Figure 2. Apparatus for the laboratory-scale
electrolysis and collection of magnesium
metal using the SOM technology.
Laboratory-Scale Electrolysis and Collection
of Magnesium
An apparatus for the electrolysis and collection
of magnesium on a laboratory scale using the SOM
technology is schematically shown in Figure 2. The
major features of this cell include an iron cathode; a
magnesia collection and condenser tube to capture
the magnesium metal vapor that is produced at the
cathode; a YSZ crucible containing the magnesium
oxy-fluoride flux, serving as the SOM; and a liquid
copper bath maintained at a very low oxygen
chemical potential, serving as the anode. As
experiments continue, the apparatus shown in
Figure 2 is likely to evolve.
Potentiodynamic and potentiostatic experiments
have been conducted to evaluate the performance of
the apparatus and the behavior of the magnesium
77
oxy-fluoride flux at 1300(C. Results of these
experiments are extremely promising. Figure 3
shows the experimentally determined currentpotential behavior of the cell up to a current density
of 1 amp/cm2. The current density is calculated
using the active YSZ membrane area because it was
found that the membrane area was rate-limiting.
Power consumption at 1 amp/cm2 is estimated to be
12 KWh/kg of magnesium produced. The results of
this and other potentiodynamic experiments indicate
that the MgO dissociation potential under the
imposed experimental conditions is around 1 V, and
the cell resistance is primarily ohmic in nature; no
diffusion-limited current develops as long as the flux
contains an adequate source of MgO.
FY 2001 Progress Report
Automotive Lightweight Materials
Condenser tube
Current Density (amps/cm2)
1.2
1
0.8
0.6
0.4
0.2
0
0
1
2
3
4
5
6
Condensed Mg metal
Applied Voltage (volts)
Figure 5. Condensed magnesium metal
vapors produced in the SOM cell
at 1300(C.
Figure 3. Potentiodynamic sweep conducted on the
magnesium oxy-fluoride flux using the SOM
cell shown in Figure 2 at 1300(C.
250
200
150
100
50
0
0
0.2 0.4 0.6 0.8
1
1.2 1.4 1.6 1.8
0.8
0.6
0.4
Figure 6. Energy-dispersive X-ray spectra scan of the
magnesium metal deposit in the condenser
tube.
Summary
0.2
References
0
0
10
20
30
40
Time (minutes)
50
60
70
Figure 4. Steady-state current in the SOM cell under
potentiostatic conditions (4.0 V) at 1300(C.
78
2
Energy (keV)
Significant progress has been made in this multifaceted research program to evaluate the use of the
SOM technology for the direct electrolysis of
magnesium from its oxide. Laboratory experiments
to determine parameters critical to the potential
scale-up of this technology for commercial
magnesium production will continue.
1
Current Density (amps/cm2)
cps
20 kV Scan (2-4 micron penetration)
In addition, the cell has been operated under
potentiostatic conditions at a current density of
around 0.8 A/cm2 for more than an hour (Figure 4).
The applied electric potential was 4 V, and the
corresponding power consumption was 37 kW/g or
10.3 kWh/kg of magnesium produced. It is notable
that the electrical efficiency achieved is better than
that in the highly efficient modern aluminum Hall
cell. A significant quantity of magnesium metal was
deposited in the crystalline form in the cooler
regions of the condenser tube (Figure 5). The
magnesium metal analysis was confirmed using a
scanning electron microscope equipped with an
energy-dispersive X-ray spectrometer (Figure 6).
1. U. B. Pal and S. C. Britten, U.S. Patent No.
5,976,345, Boston University, Boston,
Massachusetts, November 2, 1999.
2. U. B. Pal and S. C. Britten, U.S. Patent No.
6,299,742, Boston University, Boston,
Massachusetts, October 9, 2001.
Automotive Lightweighting Materials
FY 2001 Progress Report
3. R. W. Minck, U.S. Patent No. 4,108,743,
Ford Motor Company, Detroit, Michigan,
August 22, 1978.
4. D. S. Poa, L. Burris, R. K. Steunenberg, and
Z. Tomczuk, U.S. Patent No. 4,995,948, United
States Department of Energy, Washington, D.C.,
February 26, 1991.
5. A. F. Sammuels, U.S. Patent No. 4,804,448,
Eltron Research, Inc., Aurora, Illinois, February 14,
1989.
6. B. Marincek, U.S. Patent No. 3,692,645,
Swiss Aluminum Ltd., Chippis, Switzerland,
September 19, 1972.
7. D. E. Woolley, U. B. Pal, and G. B. Kenney,
in Proc. of the International Symposium on Molten
Salts, Slags and Fluxes, S. Seetharaman, ed.,
Stockholm, Sweden, ISS-TMS Publication, 2000,
CD-ROM. (Accepted for publication in High
Temperature Materials and Processes 2001).
79
8. D. E. Woolley, U. B. Pal, and G. B. Kenney,
Journal of Metals, October 2001: 32–35.
9. S. Yuan, U. B. Pal, and K. C. Chou, J. of
American Ceramic Society, 79(3) 641–650 (1996).
10. P. Soral, U. B. Pal, and H. R. Larson,
Metallurgical and Materials Transactions, 30B(2),
307–321 (1999).
11. S. B. Britten and U. B. Pal, Metallurgical
and Materials Transactions, 31B(4), 733–753
(2000).
12. U. B. Pal, J. C. MacDonald, E. Chiang,
W. C. Chernicoff, K. C. Chou, J. Van Den Avyle,
M. A. Molecke, and D. Melgaard, “Behavior of
Ceria as an Actinide Surrogate in Electro-Slag
Remelting and Refining Slags,” to be published in
Metallurgical and Materials Transactions, 32B
(2001).
Automotive Lightweighting Materials
FY 2001 Progress Report
G. Understanding the Economics of Emerging Titanium Production Processes
Co-Principal Investigator: Randolph E. Kirchain
Camanoe Associates
P.O. Box 425242, Cambridge, MA 02142
(617) 253-4258; fax: (617) 258-7471; e-mail: [email protected]
Co-Principal Investigator: Richard Roth
Camanoe Associates
P.O. Box 425242, Cambridge, MA 02142
(617) 253-6487; fax: (617) 258-7471; e-mail: [email protected]
Jacqueline Isaacs
Northeastern University
305 Snell Engineering Center
Boston, MA 02115
(617) 373-3989; fax: (617) 373-2921; e-mail: [email protected]
PNNL Contract Manager: Russell H. Jones
(509) 376-4276; fax: (509) 376-0418; e-mail: [email protected]
Participants
Professor Joel P. Clark, Massachusetts Institute of Technology
Frank R. Field III, Massachusetts Institute of Technology
Contractor: Pacific Northwest National Laboratory
Contract No.: DE-AC06-76RLO 1830
Objective
•
Explore the potential for the widespread adoption of titanium in automobile structures by understanding the
manufacturing economics of emerging processes for its production and for forming it into parts.
OAAT R&D Plan: Task 11; Barrier A
Approach
•
Identify promising titanium production processes. These include
Plasma quench process being developed by Plasma Quench Titanium Incorporated (PQTI)
Armstrong process being developed by International Titanium Powders (ITP)
FFC Cambridge process being developed by British Titanium
•
Catalog technical fundamentals of processes of interest and their potential for technological development.
•
Develop process-based cost models for each of the titanium production processes.
•
Use the cost models to explore the behavior of each process’s manufacturing economics.
•
Identify a probable potential lowest production cost for each process.
•
Develop a process-based cost model of automotive parts production using titanium.
•
Use the cost model to reveal the cost drivers associated with titanium parts production.
81
FY 2001 Progress Report
Automotive Lightweighting Materials
Accomplishments
•
Constructed cost models of PQTI and ITP processes.
•
Analyzed the economic behavior of each process, including quantifying parametric sensitivity and identifying
key cost drivers.
•
Assessed titanium production cost under likely current, pessimistic, and optimistic development of each
technology.
•
Constructed cost model of hot isostatic pressing.
•
Analyzed the sensitivity of part production cost to key cost drivers.
•
Generated and delivered a report summarizing these results.
Future Direction
•
Develop an analogous cost model of the FFC process.
•
Use this model to generate analyses of FFC process economics.
that have the potential for significantly reducing the
cost of producing titanium. They are
• plasma quench (PQTI)
• Armstrong
• FFC
Introduction
An automobile design trend that has received
much attention is the reduction of vehicle mass.
Reducing mass can improve both performance and
fuel economy. Although design changes can play a
large role in reducing mass, ultimately, large
reductions will require the substitution of higherspecific-strength/stiffness materials for the now
universal carbon steel. Primary contenders in this
race are high-strength steels, aluminum, magnesium,
and fiber-reinforced polymer composites. One
material that is not on this short list but that could
provide reductions in selected applications is
titanium. Although titanium is light and strong, its
role in the automobile has been almost nonexistent
because of its exorbitant price. This high price is a
direct result of the current production route, the
Kroll process, which is time-consuming; energy-,
capital- and labor-intensive; and batch-based.
However, new technologies are emerging that
may change the characteristics of the titanium
market. In particular, these technologies may reduce
the price of titanium sufficiently to allow it to
compete in high-volume markets, possibly even
automotive markets. This project examines the
production costs for three of these technologies in
detail.
To understand the economics of these processes,
process-based cost models were created for the
PQTI and the Armstrong processes. (A model for
the FFC process is under development.)
Furthermore, to understand how changes in raw
material price ultimately impact part cost, a model
of a part forming process, hot isostatic pressing, was
developed.
Planned Approach
Because all three processes represent
technologies that are not yet implemented at full
manufacturing scale, traditional accounting-based
approaches to cost are ineffective. However, a set of
methods, collectively referred to as “technical cost
modeling,” has been developed to address such
questions. Technical cost models (TCMs) derive
manufacturing costs by building up from the
engineering realities of a process. Specifically,
TCMs combine engineering process models,
operational models, and an economic framework to
map from the details of product and process to
operating costs. Because TCMs are built around
technical details, they allow an exploration of how
cost evolves as a technology changes. Because of
Project Deliverables
Three processes were identified by the
Northwest Alliance for Transportation Technology
82
Automotive Lightweighting Materials
FY 2001 Progress Report
rapidly enough to prevent significant back reaction.
The actual output from the PQTI process is, in fact,
titanium hydride (TiH1–2), which contains ~97%
titanium by weight, and hydrochloric acid (HCl).
The principal costs for the PQTI process are
• TiCl4
• electricity
• quench gas
• capital expenditures
the obvious applicability, the TCM approach was
adopted for this project.
Analysis Summary
Production Cost ($/kg)
The cost analyses of both of these processes
proved to be very promising. Figures 1 and 2 give a
summary look at these costs for the baseline
scenarios considered. These charts indicate the costs
for producing briquettes by lightly sintering
powders. As the diagrams show, both processes can
deliver titanium powder at a price well below the
prevailing market price. In fact, under a range of
operating conditions, these processes could
manufacture powder for less than the selling price of
sponge today.
In addition to the sale of Ti powder, HCl can
provide a significant source of revenue.
In addition to a number of specific analyses, the
economic performance of the PQTI process was
investigated under three development scenarios:
(1) optimistic development, (2) likely current
implementation, and (3) pessimistic development.
Figure 3 shows the production costs associated
with each of these scenarios. Notably, using
optimistic assumptions, the titanium powder
production cost falls nearly to half of the current
sponge price. Furthermore, although the range from
“optimistic” to “pessimistic” cost covers nearly a
factor of three, the “pessimistic” result is still well
below prevailing prices for titanium powder.
$15
Mat’l
Labor
Energy
Fixed
$10
$5
$0
Without Densification
With Densification
$15
Production Cost ($/kg)
Production Cost ($/kg)
Figure 1. Baseline cost breakdown by major elements
for PQTI process.
$15
Mat’l
Labor
Energy
Fixed
$10
Without Dens.
With Dens.
$10
$5
$0
$5
Optimistic
Likely Current
Pessimistic
Figure 3. PQTI costs across development scenarios.
$0
Without Densification
With Densification
Armstrong Process
Figure 2. Baseline cost breakdown by major elements
for Armstrong process.
The process used at ITP, referred to as the
Armstrong process, involves the reaction of gaseous
TiCl4 with sodium to produce titanium. In realizing
this process, ITP’s clear breakthroughs are
permitting continuous operation of the process as
well as the ability to produce powder directly.
Because their product is in powder form, they can
much more readily and effectively extract the
titanium from the byproduct NaCl and any
Plasma Quench Process
The plasma quench process involves the thermal
dissociation and reduction of titanium tetrachloride
(TiCl4). This is accomplished by passing these
reactants through an electric arc. The resultant
plasma is drawn through a Delaval nozzle. The
nozzle accelerates and expands the gas, quenching it
83
FY 2001 Progress Report
Automotive Lightweighting Materials
remaining sodium. Because of the strong exothermic
nature of the reaction, the primary energy demands
derive not from initiating the process, but rather
from cooling its products.
The principal costs for the ITP process are
• TiCl4
• sodium
• capital expenditures
Figure 4 shows production costs for the
Armstrong process across the scenarios investigated.
Again, all cases represent notable decreases in
titanium production cost, with the optimistic
scenario coming in below current production costs
of Kroll sponge.
Production Cost ($/kg)
$15
In addition to the sale of Ti powder, NaCl can
provide a source of revenue, albeit a small one.
Without Dens.
With Dens.
$10
$5
$0
Optimistic
Likely Current
Pessimistic
Figure 4. Armstrong costs across development
scenarios.
84
Automotive Lightweighting Materials
FY 2001 Progress Report
H. Structural Reliability of Lightweight Glazing Alternatives
PPG Program Manager: Rick Rueter
(412) 820-8731; fax: (412) 820-8705; e-mail: [email protected]
Visteon Program Manager: Mike Brennan
(313) 755-1879; fax: (313) 755-7485; e-mail: [email protected]
DOE Program Manager: Joe Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
PNNL Program Manager: Moe Khaleel
(509) 375-2438; fax: (509) 375-6605; e-mail: [email protected]
Contractor: Pacific Northwest National Laboratory
Contract No.: DE-AC06-76RL01830
Objective
•
Create lightweight glazing products for automotive applications. The technical focus of this project includes the
following tasks:
Develop lightweight laminated windshield and body glass by using thin glass layers and asymmetric
construction.
Characterize the contribution of thin glazing to the structural rigidity of automobiles through
instrumentation and modeling.
Characterize the strength of new and used typical automotive glass and thin lightweight glass.
Characterize stresses imparted on automotive glazing systems that are due to external factors such as stone
impacts, handling and mounting, and aerodynamic loads.
Develop numerical models to study the behavior of automotive glass (typical and new innovative
windshields) under stone impacts.
Develop numerical modeling and simulation tools to design and assess the structural behavior and
reliability of new, thinner glazing systems.
OAAT R&D Plan: Task 8; Barriers B, C
Accomplishments
•
Fabricated new lightweight windshield and side-body glasses that are 30% lighter than conventional glazing
systems.
•
Fabricated new glass with polyvinyl butyral (PVB) thicknesses ranging from 0.5 to 0.76mm.
•
Determined the contribution of the glazing system to overall car rigidity.
•
Developed new tests for measuring bending and edge strength of finished windshield specimens and for
measuring stone impacts.
•
Determined the tin/air side strength and the effect of temperature on glazing strength.
•
Developed models to simulate the through-thickness damage evolution in monolithic and layered glass
subjected to blunt and sharp indentors.
•
Designed a new computational tool for design of lightweight glass.
85
FY 2001 Progress Report
Automotive Lightweighting Materials
Future Direction
•
Develop lightweight side-door glass.
•
Study asymmetric glass construction and/or the use of coatings to harden the outer surfaces of glass for weight
reduction.
•
Investigate the effect of strength and thickness of PVB on overall windshield/side-door strength and
performance.
•
Investigate the overall thermal behavior based on the new lightweight glass, including the effect of the
laminated side-door glass.
•
Characterize the strength of the new lightweight glass.
•
Verify and extend the predictive stone impact models to investigate the effect of the impact of sharp indenters
on glass.
•
Enhance and extend the numerical tool for design of lightweight glass to include stresses encountered during
fabrication of glass.
•
Develop a methodology for modeling acoustical response of glass.
Introduction
Lightweight Glass Manufacturing
This project is a cooperative research and
development agreement (CRADA) between DOE,
Pacific Northwest National Laboratory (PNNL),
Visteon Automotive Systems (Glass Division) and
PPG, Inc. PPG joined this project at the beginning of
CY 2000. The project started in June 1998.
Reduction of weight is a major concern for all
future cars because weight can contribute
significantly to fuel economy. The 100–150 lb of
glass in current cars can potentially be cut by a third
or more by creating tempered thin glass, asymmetric
glass construction, and glass/plastic bi-layers. The
development of experimental procedures and of
accurate and efficient numerical methods to predict
the impact of stresses on the structural reliability of
new, lightweight glazing products—such as
glass/plastic bi-layers and tempered thin glass—is
the key for ensuring widespread use of these
products in PNGV. With tools developed in this
project, glass will also potentially begin playing
more of a structural role so that metal weight can be
reduced. Although 50 lb is the minimum savings,
there may be a cascading impact on weight savings
in other areas. For example, it may be possible to
decrease weight in other vehicle parts as a result of
increased body stiffness imparted by stronger,
thinner glass. The contribution of improved thinner
glass properties to enhancing thermal management
may also make it possible to reduce weight in
thermal management components.
New lightweight windshield glass (Figure 1)
was fabricated using 1.6-mm, 1.8-mm, 2.0-mm, and
2.1-mm glass plies (a conventional automotive glass
ply has a thickness of between 2.4 and 2.6 mm; total
windshield thickness is usually over 5.2 mm). The
new glass was formed and laminated at the Visteon
Tulsa Plant. The new glass constructions were
•
•
•
2.1 mm glass, 0.76 mm PVB, 1.6 mm glass
(29% weight reduction)
1.8 mm glass, 0.76 mm PVB, 1.8 mm glass
(31% weight reduction)
2.0 mm glass, 0.76 mm PVB, 1.8 mm glass
(27% weight reduction)
Figure 1. Lightweight glass
materials windshield.
New lightweight laminated side glass (Figure 2)
was fabricated using 1.6-mm, 2.1-mm, and 2.6-mm
plies. The new side-glass constructions were
86
Automotive Lightweighting Materials
•
•
•
FY 2001 Progress Report
16.44 kNm/deg without glass (i.e., the glass
contributes about 30% of the overall car rigidity).
Reducing the glass thickness by 40% resulted in a
minor reduction in the rigidity to 23.26kNm/deg.1
Additional analyses were conducted to
investigate the effect of glazing molding stiffness on
the contribution of the glazing system to the P2000’s
torsional rigidity. The results show that increasing
the glazing molding stiffness results in a marginal
gain in torsional rigidity, but decreasing the glazing
molding stiffness results in a significant loss in
torsional rigidity.
1.6 mm glass, 0.76 mm PVB, 1.6 mm glass
2.1 mm glass, 0.76 mm PVB, 2.1 mm glass
2.6 mm glass, 0.76 mm PVB, 2.6 mm glass
Figure 2. Lightweight glass
materials sidelight.
Strength Characterization
PNNL built unique experimental tools for
measuring (1) the bending strength of finished
windshield specimens, (2) the edge strength of
windshields, and (3) the effects of precision impacts
with blunt and sharp indentors. The bending strength
apparatus is similar to a standard ring-on-ring
apparatus that produces a uniform bending stress in
flat glass; however, the new device accommodates
the radii of finished manufactured windshields.
Extensive strain gage data and experimental
windshield strength results have been obtained to
validate the apparatus. We measured the strength
and flaw measurements for conventional automotive
windshields in new and used conditions. We also
examined the effect of high-speed impact damage on
windshield strength. The strength of single glass
plies, notably in the edge region, was also examined.
The data were used to design lightweight glazing
systems for future vehicles. These are some of the
conclusions from the experiments on conventional
used and new windshields:
Studies on other asymmetric glass constructions,
on coatings to harden the outer surface, and on
tempering will continue during FY 2001. We will
continue to seek more weight reduction while
maintaining or enhancing glass performance.
Contribution of Glass to Structural Rigidity
Experiments and modeling were conducted to
determine the contribution of glass to overall car
rigidity, and the stresses transferred to the glass. The
work focused on the 200 CW 170 Focus and the
P2000 vehicles. Visteon Glass Division investigated
the strain induced in a windshield for a 200 CW 170
Focus subject to road load vibration, body flex, and
mechanical loading. The vehicle was instrumented
for acquisition of strain and acceleration data. Three
tests were performed—road load data acquisition,
direct mechanical loading of windshield, and
torsional body flexure. Based on the observed strain,
it does not appear that the loading methods produced
significant strains in the windshield.
A finite element model of the P2000 body in
white was obtained from the Vehicle Safety
Research Department at Ford. It was used to study
the contribution of the front windshield and backglass to the overall rigidity of the P2000, as well as
the effect of the glass thickness and molding
stiffness on both the contribution of the glass to
overall rigidity and the loads transferred to the glass.
Analyses have been conducted using the P2000
model to determine the contribution of the front
windshield and back-glass to the overall rigidity of
the P2000 [ref. 1]. The analyses showed that the
torsional rigidity of the P2000 is 24.29 kNm/deg
with windshield and backlight glass, and
•
•
•
•
87
The maximum strength of carefully handled new
windshields approaches 30,000 psi.
The minimum strength of production-quality
new windshields is 7500 psi, or 25% of the
maximum strength. This is comparable to the
strength of used windshields with 53,000 road
miles, indicating the severe handling damage
during manufacturing.
The adhesion of a PVB layer is critical, not only
for containment of glass fragments during an
accident, but also for maximizing the loadbearing capacity of laminated windshields.
The well-bonded PVB layer transmits a
membrane stress of as much as 21% of the
FY 2001 Progress Report
Automotive Lightweighting Materials
maximum tensile stress in each of the glass plies
during ring-on-ring flexure testing.
Figure 3 shows the flexible ring-on-ring
apparatus and Figure 4 shows the edge strength test.
(See ref. 2 for more details.) Figure 5 shows the
edge strength of single plies. We obtained minimum
strength estimates of 29, 23, and 19 MPa at failure
probabilities of 1 × 10 , 1 × 10 and 1 × 10 ,
respectively. These are only 20 to 30% of the
maximum edge strength of 86 MPa, indicating the
need for better control of edge quality during
finishing and handling. The combination of edge
flaws, inner band residual tension, and bending
stress induced by lamination can initiate premature
cracking from the edge and undermine the long-term
durability of the windshield. Such premature
cracking is generally minimized by improving edge
quality through diamond grinding and/or by
introducing moderate levels of compression in the
edge region by convective cooling following sagbending of individual plies.
Figure 6 shows the bending strength of square
plate specimens measuring 70 × 70 mm and 2.0 mm
thick. These were prepared from float glass sheets
and tested in the concentric ring fixture (b = 9.5 mm,
c = 28.6 mm) in ambient conditions at a crosshead
speed of 1.27 mm/min, the bending strength of
single plies. The key message conveyed by Figure 6
is that the air side is consistently stronger than the
tin side. This may be helpful for manufacturing sidedoor glass because the optical requirements for it are
less restrictive.
Edge
S p e c im e n
Figure 4. Setup for edge strength test.
Figure 5. Weibull distribution of edge strength of
single glass plies.
Failure Probability
99.5
95
80
60
40
20
Tin side
Air side
10
5
2
1
50
150
250
Strength (MPa)
350 450
Figure 6. Weibull plot of strength distribution
of tin versus air side of single plies.
Future windshield designs are required to be
safer and more economical without compromising
performance. In this regard, it is critical to
understand the strength behavior of conventional
laminated windshields under various operating
Figure 3. Photograph of the ring-on-ring apparatus
during a test.
88
Automotive Lightweighting Materials
FY 2001 Progress Report
conditions—notably temperature and relative
humidity. All of the strength measurements were
carried out in controlled-temperature environments
and ambient humidity. The data in Figure 7, as
might be expected, show that the biaxial strength is
28% lower at 50ºC and 21% higher at ºC than
that at room temperature. In addition, the scatter in
strength data at 50ºC is significantly higher than that
at 25ºC and ºC (see ref. 3 for more details).
Figure 8 shows the experimental set-up for a
precision impact of glass (up to 70 miles/hour), and
Figure 9 is a representative result. PNNL built the
impact apparatus to experimentally investigate the
response of automotive glazing to impact events.
The apparatus uses a spring-loaded projectile launch
system, capable of firing symmetric or asymmetric
projectiles. The system measures the speed of the
projectile immediately prior to impact via two pairs
of infrared beam sensors. The system precisely
positions a strain-gage-instrumented single-ply or
laminated glass sheet into the impact system and
measures the dynamic response of the glass to the
impact. The preliminary results show that a singleply sheet of 3.6-mm glass develops 4000 psi of
tensile stress on the glass sheet immediately
opposite to the impact event.
Figure 8. Diagram of an impact apparatus
to measure the response of
automotive glazing to impacts.
Residual Stress Characterization
Measurements of the surface stress distribution
were obtained for three automotive windshields.
Measurements were performed using two
techniques. For the first technique, we used a laser
Figure 9. Strain data from a test on the impact apparatus.
grazing angle surface polarimeter (GASP)
instrument manufactured by StrainOptics, Inc.4 This
instrument is capable of measuring surface stresses
only. For the second technique, an instrument was
used to measure the polarization change after
double-passing light through the glass. The
innovation here was to make this measurement
simultaneously over a whole windshield with 2-in.diameter optics.5 Figures 10 and 11 show the
maximum principal stresses on the windshield’s
Figure 7. Weibull plot of strength distribution of new
windshields.
89
FY 2001 Progress Report
Automotive Lightweighting Materials
glazing to abrasion and particle impact, while at the
same time minimizing weight.6
We are performing stone impact experiments for
conventional and lightweight glass. The windshields
are impacted with ¼-in.-diameter steel spheres to
simulate windshield damage due to stone impacts.
These experiments are used to validate the models.
We are also doing staircase (threshold-seeking)
stone impact experiments to better determine the
values of parameters used in the models. Our models
are being validated using controlled experiments
done at PNNL. Figure 12 shows a half-symmetry
model, and Figure 13 shows the contour plots for
web-shape damage for the tri-layered glass
constructions under a ball bearing impact of 60 mph.
It was predicted that visible damage would be
produced on surface no. 2 for all of the four trilaminated cases. It is important to note that the least
damage results when the thin piece of glass is used
as the inside layer.
Figure 10. Maximum principal stress (in MPa) on outer
surface of windshield.
Steel ball
Glass/PVB interface
glass
Figure 11. Maximum principal stress (in MPa) on inner
surface of windshield.
outer and inner surfaces, respectively. We are using
the residual stresses as input to the stone impact
models and the design tools. The measured residual
stresses are also guiding the manufacturing of the
lightweight glass to determine suitable forming
times and tempering.
PVB
glass
Figure 12. Half-symmetry representation of the finite
element model.
Windshield Resistance to Stone Impact
A Numerical Tool for Design of Lightweight
Glass
We developed axisymmetric finite element
models to simulate the glass damage evolution when
a windshield glass is subjected to impact loading of
stones. The windshield glass consists of two glass
layers laminated with a thin PVB layer. A model is
used to predict and examine through-thickness
damage evolution patterns on different glass
surfaces. It also examines cracking patterns for
windshield design factors such as symmetric or
asymmetric glass construction, thickness of glass
plies, and curvature. These models are used to
determine appropriate glass construction schemes
that meet requirements such as strong resistance of
We have developed a method for predicting
structural failure probabilities for automotive
windshields (see ref. 7 for more details). The
predictive model is supported by the data from
strength tests performed on specimens of automotive
glass. Inputs for stresses are based on finite element
calculations, or from measurements of the residual
stresses that arise from fabrication. The failure
probability for each surface or edge subregion of a
windshield is calculated from the local state of
stress, the surface area or length of the subregion,
90
Automotive Lightweighting Materials
FY 2001 Progress Report
Case A(2.9/1.6)
Case G(2.3/2.2)
Case H(2.2/2.3)
Case N(1.6/2.9)
Figure 13. Contour plots of predicted damage for total thickness of 5.2mm (web shape).
and the statistical distribution of glass strengths.
Figure 14 shows failure probabilities for an entire
windshield due to residual stresses. This new tool
will be used to design lightweight glass with
reasonable reliability levels as compared with
conventional automotive glass. Recent work has
resolved concerns with the original probabilistic
formulation and has validated the numerical results
for cases of high failure probabilities. The work will
now focus on applications of the failure prediction
model. We have obtained data from PPG for
measured edge strengths of laminated side-light
glass and for measured stresses induced by door
slam tests. This information will be used as input to
predict failure probabilities for the loading condition
of concern. Results will be compared with results of
similar calculations to be performed by PPG
research staff.
Side Door
New lightweight laminated side glass using
1.6-mm, 2.1-mm, and 2.6-mm plies was fabricated.
The new side glass constructions were
•
•
•
1.6 mm glass and 0.76 mm PVB or 0.5 mm PVB
and 1.6 mm glass
2.1 mm glass and 0.76 mm PVB or 0.5 mm PVB
and 2.1 mm glass
2.6 mm glass and 0.76 mm PVB or 0.5 mm PVB
and 2.6 mm glass
•
Side door impact has been conducted at PPG.
Figure 15 shows typical strain measurements.
Figure 14. Predicted failure probabilities due to residual
stresses (failure probability due to edge flaws
is 4.2E-05, and due to surface flaws is
2.0E-07).
91
FY 2001 Progress Report
Automotive Lightweighting Materials
2. S. T. Gulati, J. D. Helfinstine, T. A. Roe,
M. A. Khaleel, R. W. Davies, K. K. Koram, and
V. Henry, “Measurement of Biaxial Strength of New
vs. Used Windshields,” Proceedings of the 2000
International Body Engineering Conference, Detroit,
October 3–5, 2000, Society of Automotive
Engineers Technical Paper Series 2000-01-2721,
SAE International, Warrendale, Pennsylvania.
3. S. T. Gulati, J. D. Helfinstine, T. A. Roe,
M. A. Khaleel, R. W. Davies, K. K. Koram, and
V. Henry, “Effect of Temperature on Biaxial
Strength of Automotive Windshields,” Proceedings
of the 2000 International Body Engineering
Conference, Detroit, October 3–5, 2000, Society of
Automotive Engineers Technical Paper Series 200001-2722, SAE International, Warrendale,
Pennsylvania.
4. M. A. Khaleel, J. L. Woods and C. L.
Shepard, “Surface Stress Measurement on
Automobile Windshields”, Glass Technology, 42(2)
49–53 (2001).
5. B. D. Cannon, C. L. Shepard and M. A.
Khaleel, “Stress Measurements in Glass by Use of
Double Thermal Gratings,” Applied Optics, 40(30),
5354–5369 (2001).
6. X. Sun, M. A. Khaleel, R. W. Davies, and
S. T. Gulati, “Effect of Windshield Design on High
Speed Impact Resistance,” Proceedings of the 2000
International Body Engineering Conference, Detroit,
October 3–5, 2000, Society of Automotive
Engineers Technical Paper Series 2000-01-2723,
SAE International, Warrendale, Pennsylvania.
7. F. A. Simonen, M. A. Khaleel, R. W. Davies,
K. K. Koram, and S. T. Gulati, “Probabilistic Failure
Prediction of Automotive Windshields Based on
Strength and Flaw Distributions,” Proceedings of
the 2000 International Body Engineering
Conference, Detroit, October 3–5, 2000, Society of
Automotive Engineers Technical Paper Series 200001-2720, SAE International, Warrendale,
Pennsylvania.
20000
15000
[1] Strain
[2] Strain
[3] Strain
[4] Strain
10000
[5] Strain
[6] Strain
[7] Strain
[8] Strain
5000
[9] Strain
[10] Strain
[11] Strain
0
0.125
[12] Strain
0.145
0.165
0.185
0.205
0.225
0.245
0.265
[13] Strain
[14] Strain
[15] Strain
-5000
-10000
Figure 15. Measured strains from slam test
Future Work
Future work will focus on fabrication of
lightweight glass using cost-effective manufacturing
technologies. We will continue to seek more weight
reduction while maintaining or enhancing glass
performance. The static and dynamic strength of the
glass will be characterized, and the computational
models will be verified. Modeling efforts will
continue to focus on stone impact studies and
acoustics. Work will also be conducted to enhance
and extend the numerical tool for design of
lightweight glass.
References
1. M. A. Khaleel, K. I Johnson, J. E. Deibler,
R. W. Davies, K. Morman, K. K. Koram, and V.
Henry, “Effect of Glazing System Parameters on
Glazing System Contribution to a Lightweight
Vehicle’s Torsional Stiffness and Weight,”
Proceedings of the 2000 International Body
Engineering Conference, Detroit, October 3–5,
2000, Society of Automotive Engineers Technical
Paper Series 2000-01-2719, SAE International,
Warrendale, Pennsylvania.
92
Automotive Lightweighting Materials
FY 2001 Progress Report
4. POLYMER COMPOSITES R&D
A. Development of Manufacturing Methods for Fiber Preforms
Program Manager: Norm Chavka
Automotive Composites Consortium Principal Investigator: Norman G. Chavka
(313) 322-7814; fax: (313) 390-0514; e-mail: [email protected]
Project Manager: C. David Warren
Oak Ridge National Laboratory
P.O. Box 2009, Oak Ridge, TN 37831-8050
(865) 574-9693; fax: (865) 574-0740; e-mail: [email protected]
DOE Program Manager: Joseph A. Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: DOE Cooperative Agreement
Contract No.: DE-FC05-95OR22545
Objectives
•
Develop and demonstrate new fiber preforming processes to decrease cost, increase manufacturing rates, and
improve reproducibility of large preforms for composite molding.
•
Provide process development support, mainly development of a suitable carbon-fiber roving for use in the
programmable, powdered preforming process (P4), in support of Automotive Composite Consortium (ACC)
Focal Project 3.
OAAT R&D Plan: Task 3, 10; Barriers A, B
Accomplishments
•
Tested a variety of carbon-fiber rovings supplied by various manufacturers in P4. All rovings supplied were
essentially “off-the-shelf” and unfortunately did not exhibit characteristics suitable for chopped-fiber preform
processing.
•
Fabricated flat-panel preforms from carbon-fiber rovings supplied by several carbon-fiber manufacturers.
Subsequently used these performs to support structural reaction injection molding trials and material
characterization efforts within the ACC.
•
Designed a modified fiber-handling system and fabricated a prototype system.
•
Performed several upgrades on the ACC’s P4 preforming machine during the first half of FY 2001.
•
Tested prototype thermoplastic string binder materials (supplied by PPG) in P4.
•
Successfully tested a choppable version of the Vetrotex Twintex roving (commingled polypropylene/glass
fiber) in P4.
93
FY 2001 Progress Report
Automotive Lightweighting Materials
Future Direction
•
Continue to develop and evaluate new carbon rovings.
Carbon-Fiber Roving Development
Currently available off-the-shelf carbon-fiber
materials are not suitable for use in chopped-fiber
preforming processes. These materials were
originally designed for use in filament winding,
pultrusion, woven/engineered fabrics, and
preimpregnated material manufacture. Due to this
fundamental difference in processing methodology,
current materials cannot successfully be processed in
high-speed chopped-fiber applications. In order to
achieve low-cost, random carbon-fiber composites
that are suitable for use in the automotive industry, a
carbon-fiber roving suitable for use in a choppedfiber preforming application must be developed.
Based on the lack of availability of suitable carbonfiber materials, a carbon-fiber roving specification
was developed and distributed to several carbonfiber manufacturers. Carbon-fiber suppliers then
developed several experimental carbon-fiber rovings
for use in chopped-fiber preforming processes.
The ACC’s P4 preforming machine located at
the National Composite Center in Dayton, Ohio, was
used to perform all material evaluations. All
materials were processed through an Aplicator
SMART chopper gun system. During the carbonfiber testing program, two separate preform tools
were utilized to examine the processing
characteristics of carbon rovings and to manufacture
panels for destructive evaluations. The two tools are
the shape preforming tool (Figure 1) and the flatpanel preforming tool (Figure 2).
Figure 2. Flat-panel preforming tool.
The shape tool is approximately 195 × 600 ×
325 mm (l × w × h) and contains five separate twodimensional surfaces. These five surfaces are (from
the uppermost surface downward) as follows
(angle/l × w): 0°/95 × 600 mm, 90°/115 × 600 mm,
0°/60 × 600 mm, 45°/110 × 600 mm, and
0°/90 × 600 mm. Each surface is blended to another
with either a concave or convex radius (6 mm). This
allows for a variety of surfaces to be examined for
material deposition and conformability
characteristics. Shape preforms are also
destructively evaluated using a cut-and-weigh
procedure to determine areal density distribution.
The flat-panel preforming tool is 700 × 700 mm, and
the consolidation thickness can be varied with the
use of shims along the perimeter of the tool. Flatpanel preforms fabricated on this tool are used for
destructive evaluation to determine areal density
distribution. In addition, flat-panel preforms were
supplied to support ACC molding development and
material characterization efforts.
All of the carbon-fiber rovings tested failed to
meet the ACC carbon-fiber roving specification and
were also determined to be unsuitable for highvolume applications because of preform processing
issues. Unfortunately, the material form has not
improved significantly, and processing of all the
available materials remains somewhat of an issue.
Successful development of carbon-fiber rovings will
depend upon extensive research and development at
the carbon-fiber suppliers to produce rovings to
specifications. Currently, carbon-fiber roving
Figure 1. Shape preforming tool.
94
Automotive Lightweighting Materials
FY 2001 Progress Report
development at suppliers is minimal, and off-theshelf materials not suitable for chopped-fiber
processing are being supplied. Unless funding of
development programs at carbon-fiber
manufacturers can be realized, advanced carbon
fibers for chopped-fiber applications will not be
developed in a timely manner.
Because of lack of carbon-fiber supplier support
in manufacturing carbon-fiber rovings according to
the ACC’s specifications, new programs are being
developed to fund programs for roving development
at carbon-fiber manufacturers. It is still unclear
whether these programs will be funded directly or
via the cooperative agreement. This change in
programmatic approach is still under investigation
and is dependent upon supplier willingness to
perform the proposed research. P4 carbon-fiber
preforming on the ACC’s preforming machine is
currently an ongoing effort, with new materials
being developed and tested. The ACC will continue
to develop and evaluate new carbon rovings in an
effort to create a low-cost carbon-fiber gun roving
suitable for use in the P4 process.
roving contact surface area within the fiber-delivery
system. Based on the experimental test results, a
second-generation fiber-handling system is currently
being designed to further reduce friction by utilizing
eyelets with less contact surface, which will improve
carbon-fiber roving processibility. In addition, the
new handling system will improve processing of
both thermoset and thermoplastic string binder
materials due to friction reduction.
PPG String Binder
PPG string binder material is a thermoplastic
binder strand comingled with the glass-fiber rovings.
The thermoplastic strands are composed of either
polystyrene and polypropylene or a co-polyester and
polyester. Preliminary materials could not be
successfully processed due to issues related to fiber
strength; however, development of a more suitable
material is ongoing. In addition, equipment
modifications, including redesigned ejector tubes
and a delivery system to reduce friction, are being
investigated to process this material more
efficiently.
P4 Machine Upgrades
Vetrotex Twintex
P4 machine upgrades included relocation of the
binder hoppers, new surface veil choppers with
electric servo control, quick electrical connects and
disconnects to facilitate moving of the P4
equipment, improvements to the chopper gun SLC
program, resolution of hydraulic system oil-leaking
issues, replacement of damaged electrical cables,
and an improved robot hose and chopper gun offset
beam design. Many of the modifications were due to
wear and tear of the equipment, and the remaining
modifications are the current state of the art in P4
technology.
Vetrotex Twintex material was processed
favorably by P4. Several issues emerged; however,
these are not deemed to be severe. A major issue
identified was the high preform loft immediately
following chopping and also following preform
consolidation. Preform loft following chopping was
approximately 10 times that of the desired molded
part thickness. In the same way, preform loft
following consolidation was approximately 5 times
that of the desired molded part thickness. The high
loft may cause problems on component surfaces
with relatively small draft angles (i.e., less than 7°),
leading to “wiping” when both the preforming and
molding tools are closed. Additionally, the relative
lack of stiffness within the thermoplastic strands can
lead to fiber jamming within the output tubes. Based
on the preliminary test results, a more detailed
investigation is currently being planned.
Prototype Fiber-Delivery System
The prototype fiber-handling equipment reduced
friction when processing carbon-fiber rovings, thus
allowing improved processing characteristics,
including a reduction of fiber fuzzing. The main
focus of the prototype system was a reduction in
95
Automotive Lightweighting Materials
FY 2001 Progress Report
B. Composite-Intensive Body Structure Development for Focal Project 3
Principal Investigator: Nancy Johnson
GM Research & Development Center
MC 480-106-256, 30500 Mound Road, Warren, MI 48090-9055
(810) 986-0468; fax: (810) 986-0446; e-mail: [email protected]
Project Manager, Composites: C. David Warren
Oak Ridge National Laboratory
P.O. Box 2009, Oak Ridge, TN 37831-8050
(865) 574-9693; fax: (865) 574-0740; e-mail: [email protected]
DOE Program Manager: Joseph Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: U.S. Automotive Materials Partnership
Contract No.: DE-FC-59OR22545
Objectives
•
Support the goal of Focal Project 3 (FP3) of the Automotive Composites Consortium (ACC) to design, analyze,
and build a composite-intensive body-in-white (BIW), while meeting structural and production objectives.
•
Achieve BIW structure—compared with steel—(1) achieves cost parity, (2) yields 60% mass reduction,
(3) provides equivalent or better structural performance, and (4) affords equivalent better dimensional tolerance.
•
Develop high-volume production techniques (>100,000 units per year).
•
Provide a focus for bringing together technology developed by each of the ACC working groups through
emphasis on carbon-fiber-reinforced composites and the use of hybrid materials, faster manufacturing processes,
design optimization, and rapid joining methods.
OAAT R&D Plan: Task 10; Barriers A, B, C
Approach
•
Optimize the design and complete the finite element analysis (Phase 1).
•
Construct one part of the BIW to demonstrate high-volume processing methods for both component and
assembly fixtures (Phase 2). The body component will be tested before continuing with the complete BIW build.
•
Build the complete body-in-white (BIW), ensuring that the properties of each part are consistent with those that
will be obtained from production tools (Phase 3).
Accomplishments
•
Completed Phase 1 of the program, including design optimization and finite element analysis of the selected
BIW structure. (The structure meets or exceeds all design requirements.)
•
Identified critical manufacturing issues for Phase 2 of the project.
97
FY 2001 Progress Report
•
Automotive Lightweighting Materials
Began work to build a section of the body side, which will be used as a learning tool to conduct further
preforming and molding research.
Future Direction
•
Develop the programmable powdered performing process (P4).
•
Produce a learning tool to address manufacturing issues.
were not exceeded. The finite element model of
the final concept for the BIW is shown in
Figure 1.
Introduction
Much of FY 2000 was devoted to completing
detailed design and analysis of the chosen BIW
concept. The analysis predicts the mass of the
BIW to be 86.2 kg, a 67% reduction over that for
the steel. The actual mass will be slightly higher
because of manufacturing constraints on the
variations in thickness; however, we still expect to
exceed the target of 60% reduction.
Details of Phase 1
The package was determined to be roughly
equivalent to the JA; therefore, extensive concept
and styling changes were not possible. The first
stage was to optimize the geometric structure to be
efficient for composite materials. The philosophy
was to produce a BIW with a low panel count,
good connectivity, efficient load paths, stability,
and—above all—maximized section properties.
Performance, cost, processibility, and
minimum thickness were key issues. The structure
is based on three sets of carbon materials: random
chopped fibers for the majority of panels; stitched
or woven material (with core) for the flat,
regularly shaped floor and roof structures; and
either braided or helically wound lower rails.
From the sound platform of optimized
geometry with efficient part integration, the
process of optimizing the material content against
the prescribed load cases was undertaken.
Topological and topographical optimization
techniques were developed and utilized on the
random material. The considerations for the sheet
materials utilized ply-management techniques for
local patch and insert reinforcement.
The load cases assessed were bending and
torsion stiffness (which were exceeded by 80 and
200%, respectively) and durability and abuse
loads to ensure that material fatigue allowables
Figure 1. Finite element model of body-in-white.
Crash performance of the structure was
developed with respect to two key criteria:
controllable crush of the front rails and stability
and integrity of the backup safety structure.
Nonlinear crash analysis capability is not yet fully
developed and therefore was not used for this
study. High safety margins are placed on the
backup safety structure throughout the stable
crush of the rails. The loads from the predicted
crash pulse show that a factor of safety of 3 exists
for material allowables.
The predicted mass of the complete structure
is 86 kg, a 67% reduction over that for the steel
BIW mass and comfortably below the 60% target
reduction.
The feasibility study undertaken demonstrates
that the broad project objectives are feasible, and
the group is continuing with Phase 2.
Details of Phase 2
Processing technology is being developed in
conjunction with the materials and processing
groups. These include
98
Automotive Lightweighting Materials
•
•
•
FY 2001 Progress Report
The materials and processing groups are
developing P4 for carbon fiber. P4 development
plans needed for FP3 have been discussed, and
preliminary work has begun. Plans are in place to
produce a learning tool to address the necessary
manufacturing issues to be able to successfully
build the demonstration part. The tool
incorporates challenging issues requiring
development, such as preforming and molding
sections of variable thickness. The demonstration
part will also be used to address issues related to
bonding.
high-volume processes and assembly
− expansion of the programmable,
powdered preforming process (P4) for
carbon
− evaluation of other processes (as needed)
− identification of methods to reduce
molding time
− adhesive bonding
material property characterization of carbonreinforced composites
sandwich construction.
.
99
Automotive Lightweighting Materials
FY 2001 Progress Report
C. Study of Thermoplastic Powder-Impregnated Composite Manufacturing
Technology for Automotive Applications
Principal Investigators: James Grutta and Larry Stanley
Delphi Automotive Systems
8385 South Allen Street #140
Sandy, Utah 84070-6434
(801) 568-0170; e-mail: [email protected]
Program Manager: Mark T. Smith
Pacific Northwest National Laboratory
P.O. Box 999, Richland, WA 99352
(509) 376-2847; fax: (509) 376-6034; e-mail: [email protected]
Program Manager, Composites: C. David Warren
Oak Ridge National Laboratory
P.O. Box 2009, Oak Ridge, TN 37831-8050
(865) 574-9693; fax: (865) 574-0740; e-mail: [email protected]
DOE Program Manager: Joseph A. Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: Pacific Northwest National Laboratory
Contract No.: DE-AC06-76RL01830
Objectives
•
Demonstrate and develop the manufacture of glass- and carbon-fiber-reinforced thermoplastic matrix
composites at production rates consistent with those of high-volume automotive structural applications.
•
Investigate and develop methods for shortening the process cycle times and improving materials performance.
•
Develop methods for achieving a Class A surface finish requirement, consistent with outer body panels for
automotive use.
•
Investigate the potential for cored (sandwich) structure composites with thin-walled thermoplastic matrix
composite skins.
•
Determine the cost-effectiveness of thermoplastic matrix composites in production volumes consistent with
those required for automotive applications, and the potential to achieve cost targets in the future.
OAAT R&D Plan: Task 10; Barriers A, B
Approach
•
Use data from previously completed parametric studies of time, temperature, and pressure profiles required to
achieve properties for structural automotive composites and determine combinations of processing factors that
will accelerate the process cycle to under one minute.
•
Develop methods of heating and cooling and process operations to achieve consistent cycle times.
101
FY 2001 Progress Report
Automotive Lightweighting Materials
•
Analyze alternative methods of melt processing for the various operations required for consolidation, including
experimental evaluation of cored structures.
•
Apply a multi-scale materials modeling approach to develop predictive numerical models for thermal and
structural responses for constituent materials and architectures; apply those models to processing conditions for
cycle optimization, material properties predictions, and end-use performance predictions.
•
Use data to influence materials suppliers in modification of base material feeds in ways that support increased
cycle times.
•
Use complex-shaped components to determine process limitations and factors for economic evaluation.
•
Evaluate methods of achieving a Class A surface finish with reinforced thermoplastic matrices and analyze
them in production cycles on complex tooling.
Accomplishments
•
Developed a two-press system and mold clamping/shuttle arrangement to integrate novel process parameters for
accelerating heating and cooling cycles.
•
Used the heated/cooled cone mold for process and formability trials to fabricate test components.
•
Developed a method for tabbing and analyzing thin, curved compressive test samples from the conical parts
fabricated in the formability mold.
•
Investigated all compatible core materials for processing and properties and determined two acceptable core
materials, based on structural and processing factors. (Core processing appears to be mainly useful with highertemperature matrices if a two-step process is used. Lower-temperature thermoplastics will limit automotive
applications.)
•
Developed a new processing method for continuous processing with multiple tooling sets and with optimized
heat/cooling methods. Delphi has filed a record of invention for the process.
•
Analyzed surface films with in situ and post-processing methods for achieving a Class A surface on highly
loaded structural components. Methods with films and veils were generally unsuccessful, but this work has led
to a novel method developed by Pacific Northwest National Laboratory (PNNL) of process and material
modifications that have promise for achieving Class A structural thermoplastic composites. PNNL has filed
invention disclosures on the system.
•
Developed a method of materials processing at PNNL to integrate with the traditional heating/cooling cycle,
which has achieved greatly reduced cycle times in small panel bench tests. The method is being further
developed to determine whether scaling is a factor and whether it will be applicable to production cycles.
•
Analyzed materials production and materials forms with the suppliers to determine cost and volume potential
and material supply limitations.
•
Determined cost/volume scenarios for thermoplastic TP components for different processing methods.
•
Developed initial macro-model results indicating the potential to predict important process parameters.
•
Refined meso-scale models of the fabric structure to predict formability relationships for stamp forming of
2 × 2 twill-weave fabrics.
•
Developed a family of micro-scale models for unidirectional composite (fiber bundle) property predictions.
Future Direction
•
Refine process models to predict formability limits under various process conditions.
•
Continue meso-scale (fabric) model developments to investigate effects of weave architecture on woven lamina
properties.
•
Fully develop, integrate, and document the models in the multi-scale approach.
102
Automotive Lightweighting Materials
FY 2001 Progress Report
•
Develop process and test equipment for a production cycle demonstration of thin cored structures.
•
Develop a methodology for control of “print through” to achieve a Class A surface and adapt to higher-rate
processing.
•
Develop tooling and production cycle for demonstration of continuous process methods for achieving cycle
times of less than 1 minute.
•
Develop novel process methods to determine their capability for implementation into higher-volume
production.
•
Work with materials supplier to adapt to other materials forms that are more cost-effective and that can achieve
increases in production cycle times.
Introduction
Project Deliverables
Polymer matrix composites are a class of
materials identified as having a combination of
properties that are required to achieve significant
reductions in vehicle mass compared with
conventional materials. Although they are widely
regarded as being the material with the most
potential for future weight savings, their use is
severely restricted by fundamental issues with
achieving the necessary production volumes and
with the cost of basic materials. These issues have
been studied and analyzed based primarily on liquid
thermosetting methods for matrix composites, such
as structural reaction injection molding (SRIM) as
used in the Automotive Composites Consortium
(ACC) focal projects. The cost and availability of
fibers and materials, especially carbon fibers, are
being addressed by DOE under the low-cost carbon
fiber projects. This project has identified carbon
fiber thermoplastic matrix composites as having
significant potential to meet structural property
requirements and to achieve rapid processing cycles
similar to the cycles for stamping of metal
components. To recognize the potential for
widespread use of polymer matrix composites in
automotive applications, two other fundamental
issues must be addressed—cored, thin-walled
structures (for optimum structural performance and
minimum weight) and Class A surfaces in semistructural components. Developing thermoplastic
composites to meet these requirements will require a
combination of experimental approaches, materials
and process modeling, and modification of materials
forms and processing methods to achieve an
optimum manufacturing process.
This multi-year program will develop and
demonstrate knowledge concerning four technology
areas associated with the thermoplastic composite
forming process: (1) results of process cycle times
and materials properties achievements at high
forming rates; (2) results of Class A surface finish
trials; (3) methods for the manufacture of thin facesheet cored structures; and (4) modeling tools for
predicting and optimizing the thermal and structural
materials behavior for fiber-reinforced
thermoplastics.
Planned Approach
Experimental and analytical methods are being
used and developed for high-rate forming of
thermoplastic composite sheet and cored structures.
The project team is focusing on achieving
automotive stamping production rates and on
materials modifications to achieve rapid flow and
consolidation. By integrating these two approaches,
the project is able to determine where difficulties in
the process and materials form occur and suggest
methods to overcome these barriers. Several
significant invention disclosures have resulted, and a
further focus on the process and materials kinetics
will allow the bench scale demonstration of highrate stamping.
Experimental Methods
The initial experimentation used flat plate molds and
developed a process diagram based on achieving a
pre-defined “acceptable” quality of consolidation, as
measured by compressive strength of the coupons.
These process conditions were used as a basis for
103
FY 2001 Progress Report
Automotive Lightweighting Materials
trials, and a two-press system was developed over
the past year to simulate a newly developed
continuous cycle process. The approach will be
developed and refined in the coming year with novel
methods of integrating the heat and cooling cycles to
reduce overall cycle time and to address thermal
load issues.
Class A surface trials were run with films and
veil materials that have been employed in the liquid
molding industry for improving surface finish.
These trials were conducted with the flat and cone
mold tooling. Under the higher pressures and
thermal cycle conditions for this process, they
proved ineffective for approaching Class A surface
conditions, except at unacceptably large thicknesses.
The development and study of conditions have led to
a novel method of controlling surface finish that is
under experimental and theoretical investigation
now, which could allow a significant reduction in
the thickness requirement of the surface film/veil.
An experimental study was made of potential
core materials for one-stage and two-stage forming
of cored structures. The analysis proved valuable in
separating manufacturer claims for capabilities of
the core materials from the actual behavior
demonstrated in rapid process cycle conditions
(Figure 3). The materials that seem to have the most
potential for a one-step process are the phenolic and
end-grain balsa cores, since both can withstand the
temperatures and pressures of processing. Cored
panels are very sensitive to requirements for facesheet processing; therefore, the selection of
materials and processing parameters has to be based
on part requirements. A process cost study was done
to determine the effectiveness of the different
approaches (Figure 4).
the processing of a conical mold that had varying
angles of fabric shear (deformation) and varying
forming radii (both convex and concave) to provide
an understanding of the forming limits under
different process conditions (Figures 1 and 2).
Coupons have been taken from the conical mold and
tested in compression to match the consolidation
results to the results from the basis flat plate mold.
Areas showing varying degrees of resin content and
fiber deformation were selected for study.
Compressive strength data collected so far on cone
mold formed parts are summarized in Table 1.
Methods for achieving rapid process cycles were
developed as an ongoing part of the conical mold
Figure 1. Heated/cooled cone mold for formability
trials.
Analytical Methods
A multi-scale modeling approach is being
applied to develop modeling tools with long-term
value to the technology of fabric-reinforced
thermoplastic composites. As illustrated in Figure 5,
three important scales have been addressed. At the
micro-scale, the fiber and matrix properties are
captured, along with their interactions that depend
on microstructure. At the meso-scale, the fiber
bundle properties from the micro-scale models are
used to simulate the composite material response for
particular weave architectures. Finally, properties
predicted at the meso-scale are applied in macro-
Figure 2. Forming trial parts.
104
Automotive Lightweighting Materials
FY 2001 Progress Report
Table 1. Maximum failure stress of formed sections (MPa)
P
(MPa)
1
1.7
2.4
1 minute
3 minutes
6 minutes
25 mm
25 mm 37.5 mm
25 mm
12.5 mm
12.5 mm
37.5 mm
12.5 mm
Center
Center
from
from
from
from
from center
from center
from center
from center
center
center
center
center
Specimen
location
Center
Left side
259
130
Right side
325
Left side
289
Right side
Left side
318
214
279
165
187
225
257
382
298
304
298
92
251
345
172
252
221
319
333
349
251
262
224
190
341
354
314
255
181
302
310
251
214
189
126
93
402
376
342
255
241
303
244
303
147
310
162
206
262
276
195
207
286
329
214
258
Right side
163
122
230
127
263
259
179
213
280
200
216
248
Average
233
218
192
180
316
267
245
229
261
308
272
269
scale models that predict performance in the
manufacturing and application environments. The
implicit goal of any multi-scale modeling approach
is to create ties between properties applied at the
macro-scale and the features of micro- and mesoscale models. This is particularly desirable for
composite materials because the macro-scale
properties are difficult to characterize completely by
experimental testing alone.
The finite element method is the fundamental
general-purpose modeling tool applied at all three
scales. The unit of construction for a fabric
architecture is a bundle of fibers that, at any point
along the bundle, consists of unidirectional fibers
adhesively bound together by a thermoplastic
polymer. Therefore, the properties of the bundle can
be represented by the unidirectional fibers with
interstitial polymer. Here, it is assumed that the
fibers are arranged in a periodic array so that a small
periodic unit is representative of the entire bundle.
Several variations of this micro-scale model have
been developed and used to predict thermal and
structural properties of consolidated fiber bundles.
At the meso-scale, the fiber bundles are arranged
in woven patterns so that the fiber bundles are
intertwined in coherent form to facilitate the
formation of thin planar structures. By nature, the
weaves have a periodic structure that facilitates a
small representative volume of the fabric as the basis
for a meso-scale model that can predict the detailed
behavior of the fabric, which in turn can be
homogenized to estimate macro properties of the
composite. Thermal and structural models of a plain
weave have been constructed, and thermal models
Average Results From Flexural Testing
25000
Average Maximum Stress (psi)
In-Situ Balsa
In-Situ Rohacell 1
20000
In-Situ Divinacell
2 step balsa
15000
In-Situ Aluminium Honey comb
10000
5000
0
In-Situ Balsa
In-Situ
Rohacell 1
In-Situ
Divinacell
2 step balsa
In-Situ
Aluminium
Honey comb
Sandwich type
Figure 3. Sandwich panel mechanical testing evaluation.
Cost Comparison
8 ft2 thermoplastic sandwich panel
$140.00
In-Situ
Processing
Two-step
Processing
$120.00
Cost per part
$100.00
$80.00
$60.00
$40.00
$20.00
$0
37.5 mm
from
center
20,000
40,000
60,000
80,000
100,000
120,000
Number of parts
Figure 4. Cost evaluation of core panel manufacturing.
105
FY 2001 Progress Report
Fiber/Matrix
Interactions
Automotive Lightweighting Materials
Process
Modeling
Fiber Bundle
Interactions
MICRO-……… …….……..…MESO-………………..………
……MACRO-
Figure 5. Multi-scale modeling of fabric-reinforced composites. Shown are a hexagonal fiber pack unit
cell for micro-scale modeling, a 2×2 twill fabric unit cell for meso-scale modeling, and cone
forming process model results at the macro-scale.
have been developed for a 2 × 2 twill weave
material. For further understanding of the effects of
fiber architecture, work has been initiated on a
4-harness satin weave meso-scale model. At the
macro-scale, it is desirable to predict performance in
the manufacturing and application environments. In
this project, the immediate interest is in modeling of
the forming process as it relates to the formability of
particular weaves and heat transfer properties under
forming cycle conditions. Process models have been
developed to initially study the effect of in-plane
shear stiffness on formability and drape
characteristics for the cone-shaped mold. In
addition, macro-models are being applied to predict
in-mold pressure distributions during forming. These
pressure distributions and their evolution relate to
the time required at pressure for complete
consolidation of the composite to occur.
Infrared tomography experiments have been
conducted to collect data for verification of thermal
model predictions. However, the matrix materials
under consideration are strongly transparent to
infrared radiation. This infrared transparency has the
effect of “seeing” the fiber surfaces internally, rather
than an overall thermal response of the composite.
This effect has complicated the interpretation of the
experimental data for comparison with the models.
Future work will be directed at better numerical
representations of the die surfaces and material
properties. Improved properties will result from
predictions of the meso- and micro-scale models. In
particular, meso-scale models will predict in-plane
shear stress-strain histories that ultimately relate to
formability at the macro level. In addition, alternate
element formulations will be employed to permit
through-thickness stresses currently unavailable
from the shell element used in the forming process
models. This will be useful for predicting final
thickness distributions and transient pressure
distributions. As the models are refined, more
comparisons will be made with experimental data
associated with cone mold forming trials.
Conclusion
The technical feasibility of stamp forming of
thermoplastic carbon/nylon composites for
automotive applications has been demonstrated for
simple flat panel shapes and is being further
developed for complex-shape forming. The
investigation and modeling results have led to the
development of some novel processes and
technologies that have the potential to further
advance the forming science, surface appearance,
and applicability of these material forms. Significant
challenges still lie in optimization of the process and
its development into a commercially viable robust
process, especially given the requirements for cored
thin-wall structures and Class A surfaces for
structural components. Materials costs and
production volumes are within acceptable limits for
large-scale automotive applications.
106
Automotive Lightweighting Materials
FY 2001 Progress Report
5. LOW-COST CARBON FIBER
A. Low-Cost Carbon Fibers from Renewable Resources
C. F. Leitten, Jr. (Project Contact), W. L. Griffith, A. L. Compere, J. T. Shaffer
Oak Ridge National Laboratory
Post Office Box 2009
Oak Ridge, TN 37831-8063
(865) 576-3785; fax: (865) 574-8257; e-mail: [email protected]
Program Manager: C. David Warren
Oak Ridge National Laboratory
Post Office Box 2009
Oak Ridge, TN 37831-8050
(865) 574-9693; fax: (865) 574-0740; e-mail: [email protected]
DOE Program Manager: Joseph Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: Oak Ridge National Laboratory
Contract No.: DE-AC05-00OR22725
Objective
•
Investigate and establish proof-of-principle processing conditions for several types of low-cost carbon fiber
precursor materials that are suitable—in properties, volume, and cost—for use by the automotive industry.
OAAT R&D Plan: Task 5; Barriers A, B
Approach
•
Evaluate the spinning of fiber from high-volume renewable and recycled fiber streams.
•
Evaluate carbonization and graphitization of renewable and recyclable fibers using a variety of bench techniques
(e.g., thermogravimetric analysis, micrography, resistance/conductance).
•
Evaluate the effects of modern carbon fiber production technologies (e.g., hot stretching, controlled thermal and
atmospheric environment) on selected fibers to improve properties.
•
Focus on factors (feedstock availability and cost, yield) critical to economic feasibility.
•
Evaluate mechanical and composite properties of graphitized melt-spun fibers.
Accomplishments
•
Found effective stabilization, carbonization, and hot-stretching techniques for single-strand, melt-spun, ligninbased fiber.
107
FY 2001 Progress Report
Automotive Lightweighting Materials
•
Developed compositions that decrease water uptake during storage and shipment of lignin-blend fibers.
•
Prepared and evaluated scanning electron micrographs of a wide variety of raw, carbonized, and graphitized
fibers.
•
Obtained preliminary data that indicate that yields of 50%, consistent with those obtained commercially for
lignin-based activated carbon, are feasible.
•
Working with Westvaco, prepared and provided large amounts of hardwood kraft lignin for project use.
•
Successfully prepared lignin-blend fibers with a dominant graphitic content at a firing temperature consistent
with intended use.
Future Direction
•
Spin, characterize, and oxidize/carbonize/graphitize a wide variety of single fibers.
•
Evaluate data on the physical properties of several recyclable petrochemical polymer and lignin-blend fibers
using bench-scale oxidation and carbonization facilities.
•
Scale up fiber spinning and treatment to permit evaluation of mechanical properties of several experimental
fibers.
•
Systematically evaluate fiber properties and economics to downselect the number of fiber formulation systems
under active investigation.
•
Attempt melt spinning of multiple fibers to create a small tow (20–40 fibers) to (1) provide a better basis for
understanding downstream spinning issues and (2) provide material for small composite tests of successful fiber
systems. Contracts for this activity are being placed with North Carolina State University and the University of
Tennessee.
Introduction
Project Deliverables
This project focuses on development of carbon
fibers from high-volume, low-cost, renewable or
recycled fiber sources to reduce precursor and
processing costs. Use of these materials also
decreases the sensitivity of cost to changes in
petroleum production and in energy costs. Earlier
literature reports indicate that a wide variety of
fibers—kevlar, nylon, polyacetylene, polyethylene,
polystyrene, and pitch—as well as natural and
renewable products such as wool, rayon, and
lignosulfonate were also used to make lowperformance carbon fibers in the 1960s and 1970s.
In quality, cost, and volume, kraft lignins and
recycled petrochemical polymers are attractive in
the United States. This project will evaluate the
feasibility of preparing carbon fiber precursors from
a variety of high-volume renewable and recycled
polymers.
By the end of this multi-year program, the
production of one or more environmentally friendly,
economically feasible carbon fiber precursors will
be demonstrated.
Planned Approach
Producing low-cost, high-volume carbon fiber
precursors requires the simultaneous development
of processes for making the feedstock fibers and of
downstream processing techniques that improve the
properties of each type of fiber. Because of high
levels of emissions and costs typically associated
with spinning of fiber from liquids, first priority was
placed on development of melt-spinning techniques
for fiber. The use of lignin and other nonnitrogenous feedstocks would eliminate cyanide
emissions during furnacing. The use of modern
furnacing techniques, such as hot-stretching and
controlled-atmosphere processing, is being
evaluated to improve the properties and yield of
carbon fiber precursors from feedstock. Promising
108
Automotive Lightweighting Materials
FY 2001 Progress Report
mixing extruder have varying diameters that are
typically larger than those of fibers typically used in
composites. These problems can be solved with
more uniform stretching techniques and larger draw
ratios.
Since results with single fibers are favorable, a
decision was made to spin amounts of multiple
precision fibers for graphitization and mechanical
property evaluation. Subcontracts for multiple fiber
spinning are being placed.
fiber blends will be prepared as precision fibers and
will, additionally, be graphitized to permit the
evaluation of finished fiber structure and
mechanical properties of individual fibers and of
composites.
Carbon fiber for use in aerospace or recreational
products is typically produced in three steps:
(1) spinning, (2) stabilization/
oxidation [with hot stretching, in the case of
polyacrylonitrile (PAN], and (3) carbonization/
graphitization under reduced oxygen or inert
atmosphere to increase strength and stiffness.
Systematic evaluation will be required to find
appropriate conditions for spinning and furnacing
melt-spun lignin-based fibers.
Stabilization, Carbonization, and
Graphitization
Conditions that provide good stabilization of
lignin fiber blends containing polyesters,
polyolefins, and polyethers have been profiled.
Computer-controlled furnacing and a novel
technique for stretching single fibers are producing
dense, solid fibers with little or no inter-fiber
adhesion or melting.
After stabilization, the lignin-blend fibers are
carbonized in a reducing atmosphere, again with hot
stretching. Carbonized lignin-blend single fibers are
dense and compact with relatively few visible
defects.
Lignin-blend fibers have been graphitized to
temperatures ranging between 1600 and 2400(C
(Figure 1). Yields remain near 50%.
Fiber Spinning
Melt-spun polymer blends of kraft lignin, which
is commercially available as a byproduct, received
significant attention. Lignin is a ubiquitous polymer
that typically constitutes 20–30% of wood and
woody biomass. In pulping, lignin and lignin
compounds are separated from cellulose and either
burned or recovered and sold as a byproduct.
Detailed estimates place the volume of lignin
produced and burned by the U.S. paper industry at
around 1000 times the present volume of worldwide
carbon fiber production. Thus kraft lignin could
support production of the amounts of fiber needed
by the automotive industry. Gasification, which is
being evaluated as an alternative to current lignin
combustion processes, could provide significant
amounts of lignin as well as both high-temperature
process heat and low-cost electricity. Development
of this process could greatly increase availability of
carbon fiber and decrease the price.
Studies of melt-spun single lignin-blend fibers
have continued during this year. Because of its
narrow melting point range, Westvaco hardwood
kraft lignin has been used. As reported earlier, it
was desalted prior to use to prevent void formation
during graphitization.
Single fibers were spun to our specifications
from prepared lignin mixtures by North Carolina
State University staff and returned on cardboard
spools for our evaluation. Although the current
fibers are strong enough to withstand high-speed
spooling and to permit hand lay-up on quartz for
furnacing, fibers produced by the Atlas Laboratory
Figure 1. A scanning electron micrograph showing the
end of a graphitized lignin-blend fiber.
Development of Graphitic Structure
Raw, stabilized, carbonized, and graphitized
lignin-polyester blend fibers have been evaluated.
Crystalline structure was not apparent in raw,
109
FY 2001 Progress Report
Automotive Lightweighting Materials
stabilized, or carbonized materials. As shown in
Figure 2, graphitic structure is increasingly
pronounced as firing temperature increases, as
indicated by powder X-ray diffraction analysis.
Future Directions
Proof-of-concept production of single ligninblend fibers indicates the potential for use of this
material as a feedstock for industrial-grade carbon
fibers. To determine whether the technology is
feasible, melt-spinning and processing of multiple
fibers will be evaluated. Hot-stretching and
furnacing conditions for multiple fiber tows will be
evaluated. Surface treatments and sizing to permit
the production of small composites for properties
tests will be evaluated.
Conclusions
Several lignin-based blend fibers that can be
oxidized, carbonized, and graphitized have been
demonstrated using equipment and techniques
developed during this project. Preliminary
evaluations indicate that production of carbon fiber
precursor from renewable and recycled materials is
likely to be feasible. The yield of fiber is ~50%.
Further studies are under way to (1) extend the
range of evaluated fibers, (2) continue evaluation
and improve processing conditions, (3) spin and
process multiple small-diameter fibers, and
(4) begin evaluation of multiple fibers and
composite properties.
Figure 2. Powder diffraction data on graphitized ligninblend fiber showing the predominance of
graphitic structure.
110
Automotive Lightweighting Materials
FY 2001 Progress Report
B. Low-Cost Carbon Fiber Development Program
Program Manager: Mohamed G. Abdallah
Hexcel Corporation
P.O. Box 18748, Salt Lake City, UT 84118
(801) 508-8083; fax: (801) 508-8090; e-mail: [email protected]
Project Manager, Composites: C. David Warren
Oak Ridge National Laboratory
P.O. Box 2009, Oak Ridge, TN 37831-8050
(865) 574-9693; fax: (865) 574-0740; e-mail: [email protected]
DOE Program Manager: Joseph A. Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Principal Investigators:
Harini Dasarathy, Carlos Leon y Leon, John McHugh, Rick O’Brien, Stephen Smith, and Brent Hansen
Contractor: Hexcel Corporation
Contract No.: 450001675
Objective
•
To define technologies needed to produce a low-cost carbon fiber (LCCF) for automotive applications at a cost
of $3.00 to $5.00 per pound in quantities greater than one million pounds per year. The required carbon fiber
properties are tensile strength of greater than 400 ksi, modulus of greater than 25 Msi, and strain at failure
greater than 1%.
OAAT R&D Plan: Task 5; Barriers A, B
Approach
•
Develop new precursors that can be converted into carbon fiber at costs below the costs of current processes.
•
Explore processing by other methods than thermal pyrolysis.
•
Develop technologies for significant improvement in current production methods and equipment.
•
Develop alternative methods for producing carbon fiber from pitch, polyacrylnitrile (PAN), or other precursors.
•
Reduce precursor cost by the use of commercially available energy-efficient precursors and high conversion
yields.
•
Improve precursor production economics of scale and throughput.
•
Introduce novel low-cost carbon fiber production methods.
111
FY 2001 Progress Report
Automotive Lightweighting Materials
Accomplishments
•
Evaluated proposed research areas through laboratory trials and refinement of manufacturing cost analyses.
These focused on
PAN-based precursors: large-tow benchmark, commodity textile acrylic tow, chemical modifications,
acrylic fibers spun without solvents, and radiation and nitrogen pretreatment of PAN-based materials.
Precursors other than PAN: polyolefins—polypropylene (PP), linear low-density polyethylene (LLDPE)
and high-density polyethylene (HDPE); polystyrene; and polyvinyl chloride (PVC) pitch.
•
Scaled up proposed research areas to pilot line trials.
•
Assessed the technical and economic feasibility of the proposed research areas.
•
Down-selected the most promising technologies to meet the program objectives:
Commodity textile acrylic tow with chemical modification and/or radiation pretreatment.
Polyolefins (LLDPE and PP).
•
Developed detailed manufacturing cost models for the down-selected technologies.
Future Direction
•
Continue pilot-scale and scale-up experimental trials and assessment of the manufacturing cost of the downselected technologies.
•
Down-select the LCCF technology(ies) to be developed for large-scale engineering and commercial
manufacturing feasibility.
•
Develop manufacturing project plans that include the configuration of facilities, sites, supply of raw materials,
and economic feasibility studies for the selected technology(ies) for the large-scale production of LCCF for
automotive applications.
Introduction
The goal of this program is to define and
demonstrate technologies needed for the
commercialization of LCCFs to be used in
automotive applications. Lighter-weight automotive
composites made with carbon fibers can improve the
fuel efficiency of vehicles and reduce pollution. In
order for carbon fibers to compete more effectively
with other materials in future vehicles, their cost
must be reduced.1–3 Specifically, this program
targets the production of carbon fibers with adequate
mechanical properties, in sufficiently large
quantities, at a sustainable and competitive cost of
$3 to $5 per pound.
Project Deliverables
At the end of this multi-year program,
technologies for LCCF production will be defined.
This definition will include the required materials
and facilities and will be supported by detailed
manufacturing cost analyses and processing cost
models. Laboratory trials and pilot-scale
demonstrations will be performed to support the
defined technologies.
Planned Approach
This program was divided into two phases
(Figure 1). Phase I (15 months): Critical review of
existing and emerging technologies, divided into
two tasks:
Task 1.1. Literature review and market analysis.
Task 1.2. Laboratory-scale trials and
preliminary LCCF manufacturing cost
assessments of the proposed
technologies. Phase 1 led to further
refinement and down-selection of the
most promising technologies for
Phase 2 (Table 1).
Phase 2 (27 months): Evaluation of selected
technologies using pilot-scale equipment and cost
models. Phase 2 was divided into three tasks:
112
Automotive Lightweighting Materials
FY 2001 Progress Report
Table 1. Phase 1 low-cost carbon fiber (LCCF) development program technology downselection
Primary LCCF
Continued in
development issues
Precursor
Technology
Phase 2?
Technical
Cost
PAN-based
Not PAN-based
Large-tow benchmark
Textile acrylic
Chemically modified
Acrylic fibers spun without solvents
Radiation-treated
N2-prestabilized
Polyethylene
Polypropylene
Polystyrene
Polyvinyl chloride
X
X
X
X
X
X
X
X
X
X
X
X
X
–
X
X
–
X
X
X
X
–
–
such as the programmable powdered performing
process (P4) (ref. 1) use continuous tows as their
feedstock. Large tows (>24,000 filaments) can be
produced more economically than typical aerospacegrade tows (≤24,000 filaments). However, the
manufacture of carbon fibers from large tows
presents a variety of technical challenges. These
challenges were addressed in three ways. First, small
tows were combined to investigate the effects of
various parameters (tow count, filament diameter,
tow width, air flow velocity and orientation) on
large-tow processability. We then carried out the
conversion of a large-tow PAN precursor regarded
as an industrial benchmark (Acordis large-tow
precursor). The equipment modifications and
parameters needed to safely process the precursor in
a carbon fiber pilot line were established. As a
result, large-tow precursors were processed and
converted to carbon fibers with mechanical
properties in excess of LCCF program targets
(tensile strength of up to 516 ksi, modulus of up to
30 Msi). However, the high cost of this precursor
precludes its use for LCCF purposes. We then
contacted alternative large-tow PAN manufacturers
and obtained less expensive developmental
precursors for testing. Work on large-tow
developmental precursors again allowed us to
exceed LCCF targets (tensile strength of up to
463 ksi, modulus of up to 30 Msi).
Task 2.1. Pilot-scale design for the evaluation of
selected LCCF technologies. This
included modifications of a PAN
spinning pilot line and two different
carbon fiber conversion lines (a
single-tow research line and a multitow pilot line) and the construction of
continuous sulfonation processing
equipment for polyolefin fibers.
Task 2.2. Experimental evaluation of downselected LCCF technologies,
including commodity textile tow PAN
(with chemical modification and
radiation and/or nitrogen
pretreatment) and polyolefins
(LLDPE and PP).
Task 2.3. Large-scale feasibility study of
selected LCCF technologies.
The planned approach included decision points
within each of the tasks in order to further downselect and define the most promising technology(ies)
that will meet the program objectives, as indicated in
the overall LCCF program timeline (Figure 1).
Polyacrylonitrile
Large-Tow Benchmark
Carbon fibers from PAN precursors can be made in
a variety of product formats (e.g., tows, mats). Highrate automotive composite manufacturing processes
113
FY 2001 Progress Report
Task Name
Automotive Lightweighting Materials
99
Qtr 3
Qtr 4
Qtr 1
2000
Qtr 2 Qtr 3
Qtr 4
Qtr 1
2001
Qtr 2 Qtr 3
Qtr 4
Qtr 1
2002
Qtr 2 Qtr 3
Qtr 4
LCCF Development
Project Plans & Review
Phase I: Critical Review
I.1 Existing & Emerging Technologies
Go/No-Go
1/27
I.2 Lab Trials of Proposed Technologies
Go/No-Go
8/10
Go/No-Go
1/30
Phase II: Evaluation of Selected Technologies
II.1 Pilot-scale Equipment Design
Go/No-Go
8/20
Go/No-Go
12/28
II.2 Experimental Evaluation
Go/No-Go
7/30
II.3 Large Scale Feasibility Study
Figure 1. Low-Cost Carbon Fiber Development Program timeline.
Extensive discussions with large-tow PAN
suppliers also allowed us to address factors that may
lead to lower LCCF costs. These include PAN
chemistry, spinning parameters, PAN filament
characteristics, tow band uniformity and splittability,
and tow packaging issues. For instance, supplying
tows in compact bales instead of boxes could reduce
packaging and shipping costs, and hence PAN fiber
costs, by as much as 15%. However, additional trials
with developmental precursors indicate that costsaving measures can compromise their LCCF
performance. The knowledge gained in processing
large-tow PAN precursors is being applied to
support our work with more economical large-tow
PAN derived from commodity-textile-grade PAN
polymers.
Commodity Textile Acrylics
Commodity-textile-tow PAN is commercially
available in large quantities at about half the cost of
large-tow PAN precursors. Its use as a LCCF
precursor could thus lower the carbon fiber
manufacturing costs to LCCF program target levels.
However, the properties of commodity textile PAN
are less than optimum from a carbon fiber
conversion perspective. Textile-grade PAN
polymers are usually made with relatively large
amounts of vinyl acetate or methyl acrylate
114
comonomers, and without acidic comonomers or
other stabilization enhancers. These polymers are
generally wet-spun or dry-spun in large production
lines. The large tows are combined, sized, crimped,
and compacted in bales. The resulting tows are
unsplittable and are difficult to process in
conventional carbon fiber lines.
To address those difficulties, raw materials
(fibers and polymers) were obtained from various
textile PAN suppliers. Initially, wet-spun or dryspun textile PAN tows made with methyl acetate or
vinyl acetate comonomers, and commercially
available in small filament deniers, were tested. Lab
trials followed by pilot line conversion trials of
selected materials yielded conditions at which
thermal runaway reactions, filament fusion, and
other processing difficulties could be overcome.
However, the mechanical properties of the resulting
carbon fibers fell below the LCCF program targets
(see Table 2). In addition, the relatively long
oxidation time and low carbon yield of textile PAN
(Table 2) can be expected to increase LCCF
manufacturing costs. Laboratory-scale development
of novel PAN modifications and conversion
technologies led to several ways to improve the
potential of commodity textile PAN as an LCCF
precursor. These include chemical modification,
radiation treatment, and N2 prestabilization.
Automotive Lightweighting Materials
FY 2001 Progress Report
Table 2. Comparison of properties of carbon fibers made from large-tow polyacrylonitrile (PAN)
Large-tow PAN
Developmental PAN
Commodity textile
Parameter
benchmark
precursor
tow PAN
Filaments per tow
Tensile strength, ksi
Modulus, Msi
322,000
332,000
666,000
516
463
121
30
30
11
Strain to failure, %
1.6
1.5
1.1
Carbon fiber density, g/cm3
1.79
1.77
1.67
Carbon fiber yield, % wt.
54
Relative oxidation time
43
1
1.36
Chemical modification of acrylics is one of
several methods reported in the literature to
accelerate the rate-limiting stabilization stage in the
carbonization process. Lab-scale investigation
identified an inorganic oxidizing agent (hydrogen
peroxide) in an alkaline medium as the most
effective modifier of PAN polymer/fiber. Evaluation
of modified polymers and fibers by various
analytical methods indicated that partial oxidation of
PAN had been accomplished. Pilot-line trials
demonstrated that adding a chemical modification
step in the PAN spinning line prior to drying would
induce a similar degree of partial oxidation in the
spooled precursor fiber. Several factors—such as
modifier composition, fiber morphology, and
application process—were optimized. As expected,
optimally modified PAN stabilized in approximately
half the time required for its virgin counterparts.
Properties of carbon fiber made from chemically
modified precursor are shown in Table 3.
Work for year 3 of the program is planned as
follows:
1. Spin textile polymer (control and chemically
modified) on the PAN pilot line.
2. Prepare several tow sizes (3, 12, 48, 84 K,
etc.) by combining smaller tows.
3. Carbonize various tow sizes to model the
effect of tow mass on various stages in the
conversion process.
4. Optimize the oxidation/carbonization
process to maximize mass throughput and
mechanical properties and minimize carbon
fiber cost.
5. Identify process parameters for scaling up to
standard textile tow counts (>300K).
Table 3. Properties of carbon fibers produced from
chemically modified specialty
polyacrylonitrile (PAN)
Tensile
modulus
(Msi)
Strain at
failure
(%)
Density
(g/cm3)
Mass per
unit length
(g/m)
420
24.5
1.71
1.746
0.215
1.61
process to more economical textile PAN fibers.
Exploratory lab work on chemical modification of
textile PAN polymers indicated partial oxidation
similar to that in the precursor polymer. Adaptation
to an in-line modification process for textile fiber,
similar to that used in the PAN pilot-line trials,
began in the fourth quarter of the second year of this
project. Textile polymer with vinyl acetate (vinyl
acetate is more cost-effective than methyl acetate)
was purchased from a textile fiber manufacturer. It
was successfully spun on the PAN pilot line and
enough textile polymer–based fiber was made for
further work. In parallel, a cost analysis of carbon
fiber made from chemically modified textile PAN
fiber showed that reduced stabilization time
translates to lower capital investment and ultimately
to lower carbon fiber production cost.
Chemically Modified Acrylics
Tensile
strength
(Ksi)
45
Following these encouraging initial results, the
focus turned toward adapting the modification
115
FY 2001 Progress Report
Automotive Lightweighting Materials
6. Collaborate with textile fiber manufacturers
to implement the process on a production
line.
Acrylic Fibers Spun Without Solvents
Work on this topic was substantially reduced in
the fourth quarter of 2000 and was halted by March
2001 for the following reasons:
1. The costs of a melt-spun acrylic fiber were
estimated using other melt-spinning systems
as models. The results showed that the
expected lowest cost, for state-of-the-art
large-scale manufacture, could not fall
below $0.50–0.55/lb, with the actual selling
price probably needing to be much higher
(current polyester filament prices give some
guidance in this regard, at $ 0.65 to 0.70/lb).
Set against significant uncertainties related
to technology implementation, and capital
risk, we see only a small economic
justification for pursuing melt-acrylic
technology, possibly $0.20/lb as carbon
fiber. The only reason for prioritizing this
work would be a compelling technical
advantage over textile acrylic fibers. These
are available at ~$0.75/lb, and the range of
acrylic polymer variants that might be made
available at this price is limited only by the
scale of the market and by customer/supplier
interests.
2. The problem of extruding a melt filament is
not the primary technical challenge. In the
first and second quarters of 2001, it became
apparent that an intractable problem with
melt-acrylics was stabilization without
fusion in a cost-effective time frame.
Radiation treatment was found to be helpful
in inhibiting the degree of fusion
experienced and to permit thermal oxidation
to be applied, but the required time frames
were simply too long for carbon fiber
operations to be practical.
Based on these findings, work on this subtask
was halted. In subsequent work during 2001, it has
become even more apparent that textile acrylic fibers
have more LCCF potential than had been previously
realized.
116
Radiation Pretreatment
Radiation pretreatment of PAN precursors to
reduce carbon fiber manufacturing costs by
decreasing stabilization times was technically
demonstrated and has been an option since the late
1970s. Recent advances in radiation treatment
technology and reduced processing costs have
rekindled interest in this approach. For LCCF
development, this route is especially attractive if it
could be applied to low-cost textile PAN fiber with
the same effect as those observed for PAN
precursors. As noted under “Commodity Textile
Acrylics,” commodity textile acrylic fibers can be
purchased for approximately half the cost of largetow precursors currently available in the market.
A significant technical problem associated with
the conversion of textile fibers to carbon fiber is the
long stabilization time necessary before
carbonization. The low reactivity of textile PAN
fibers relative to PAN precursors demands high
temperatures to initiate oxidation (compared with
those required for precursors). Since the stabilization
process is accompanied by a high residual heat of
reaction, the use of high temperatures leads to an
uncontrollable exothermic process, resulting in fiber
fusion and/or burning. Preliminary lab-scale
evaluation of textile fiber samples irradiated with
either gamma or E-beam sources indicated a
reduction in heat evolution and the initiation of
oxidation at lower temperatures. Analysis of the two
radiation treatments, from both cost and operability
points of view, favored the E-beam radiation
treatment as the process of choice. The treatment
dose level was optimized to obtain maximum benefit
in terms of reducing both heat evolution and
treatment cost. Figure 2 shows cost model
calculations derived from optimization trials. The
optimum dose level for textile acrylic fibers was
defined as 30 Mrad.
Commercial textile PAN fiber is crimped to
maintain the tow integrity of the >300K filaments
and is packaged in bales. Radiation facilities in
general are not equipped to handle large tows or the
packaging formats of standard textile PAN
shipments. For these reasons, textile fiber was spun
on Hexcel’s PAN pilot line using commercially
available textile polymer (~90–92% acrylonitrile +
8–10% vinyl acetate) to demonstrate the technical
feasibility of radiation treatment on textile PAN
tows. A 12-K tow spool was radiated in a continuous
line operation with E-beam equipment at an off-site
Automotive Lightweighting Materials
FY 2001 Progress Report
0.40
overall stabilization time of textile PAN fibers. The
temperature/ time profile in oxidation was also
optimized, without increasing the total heating time,
in order to obtain a higher carbon yield (profile 2 in
Figure 3). The radiated fiber was oxidized and
carbonized on the research line in a continuous
process. The control fiber could not be carbonized
continuously because of its poor strength. The
properties of the carbon fiber obtained are shown in
Table 4.
1.00
0.30
0.90
0.20
0.80
0.10
0.70
Ð
0.00
0
10
20
Relative Heat Evolved
Treatment Cost [$/lb CF]
Selected
30 Mrad
0.60
30
40
50
60
70
80
90
Table 4. Carbon fibers properties after E-beam
pretreatment of textile polyacrylonitrile
fiber @ 30 Mrad
100
Applied Dose [Mrad]
Figure 2. Effect of applied E-beam radiation dose on
treatment cost and relative heat evolution of
commodity textile tow PAN.
facility. The line speed was set to treat the fibers at
30 Mrad. Radiated fibers and as-spun virgin fibers
were then subjected to batch oxidation/stabilization
studies. The results from these studies (Figure 3)
showed that a total oxidation time of at least
75 minutes was necessary to stabilize the preirradiated fibers (to an oxidized fiber density of
1.324 g/cm3) to enable the fibers to withstand typical
carbonization temperatures. By comparison, the
oxidized fiber density of control fibers reached only
1.329 g/cm3 after 135 minutes.
Virgin textile fibers
Residence Time (min)
Radiated @ 30 Mrad -Profile 1
1.329 g/cm3
135 min
Radiated @ 30 Mrad -Profile 2
80
60
40
1.324 g/cm
75 min
3
3
1.347 g/cm
75 min
20
0
220
225
230
235
240
245
250
Tensile
modulus
(Msi)
Strain at
failure
(%)
Density
(g/cm3)
MPUL
(g/m)
256
22.7
1.10
1.71
0.776
These trials demonstrated the concept of using
radiation as a pretreatment for reducing carbon fiber
conversion costs through a reduction in overall
stabilization time. In this first attempt to carbonize
the sample, the carbon fiber modulus obtained
almost met the LCCF program’s target of 25 Msi.
While the tensile strength is still below the LCCF
target of 400 ksi, the reported strength is very
encouraging for fibers derived from a textile PAN
polymer. In fact, many options are available to
improve the mechanical properties of these textile
PAN fibers. Higher strengths can be expected when
an optimum degree of stretch is applied in the
oxidation and tar removal stages of the process. In
addition, the property envelope could be pushed
further if a commercial source of PAN polymer with
compositions of less than 8% vinyl acetate and more
than 92% acrylonitrile could be identified. This
source should be able to provide the material at a
cost similar cost to that of the textile PAN polymer
used for the trials.
Based on the experimental results obtained,
economic analyses performed to date, and practical
options available for improvement in carbon fiber
properties, textile PAN radiation pretreatment is the
most promising route to achieve the LCCF
development goals. As a result, work has been
planned for Year 3 of the program as follows:
120
100
Tensile
strength
(Ksi)
255
Temperature (°C)
Figure 3. Effect of air stabilization profile on ox-density
of textile PAN fibers with and without E-beam
radiation pretreatments.
These results clearly showed that, in these trials,
E-beam irradiation led to a 45 % reduction in the
117
FY 2001 Progress Report
Automotive Lightweighting Materials
1. Carbonize radiated Hexcel-spun 48K textile
acrylic fiber.
2. Explore the effects of small changes in vinyl
acetate content on carbon fiber properties.
3. Design hardware (with vendors) to be able
to radiate 83K and >300K tows uniformly.
4. Carbonize various tow sizes to model the
effect of tow mass on various stages of the
conversion process.
5. Optimize the oxidation/carbonization
process to maximize mass throughput and
mechanical properties and minimize carbon
fiber cost.
6. Identify process parameters needed for
scaling up the technology to handle standard
textile tows (>300K).
Nitrogen Prestabilization Treatment
Work within the LCCF program confirmed
earlier observations that N2 prestabilization can help
to reduce the overall processing time of precursor
and textile-grade PAN. Heating PAN in N2 promotes
cyclization reactions without parallel oxidation
reactions. This greatly reduces the reaction onset
temperature and the amount of exothermic heat
released upon oxidation. Hence, N2-prestabilized
PAN can be processed more quickly and safely
through conventional oxidation ovens. Bench-scale
LCCF work showed that N2 prestabilization
treatments were most effective when they were
carried out (a) at relatively high temperatures and
(b) for short times. For instance, large tows of textile
PAN can ignite if oxidized directly at temperatures
as low as 240°C. In contrast, subjecting textile PAN
to a single N2 prestabilization step at 280°C for
5–30 minutes showed the potential to cut down its
overall stabilization time by as much as 30%. The
elimination of any possible filament fusion by
employing a two-stage N2 prestabilization treatment
(including a low-temperature N2 pretreatment step)
was also investigated. However, preliminary LCCF
manufacturing cost models indicated that other
LCCF technologies are more promising than N2
prestabilization in terms of overall cost reduction.
For that reason, the decision was made to halt work
on this subtask and to focus on the most promising
LCCF technologies.
118
Non-Polyacrylonitrile
Polyolefins
Polyolefin fibers are appealing as LCCF
precursors because of their commercial availability,
high carbon content, and relatively low cost.
Recycled polyolefin resins are also available at low
cost, and their use would reduce the impact of
industrial polyolefins on the environment. However,
commercial polyolefin fibers melt before they can
be thermally stabilized. Their melting can be
suppressed through stabilization treatments that
promote chain cross-linking while minimizing
fusion or destructive chain scission. Such treatments,
when effective, tend to involve the use of hazardous
and/or costly materials. A few open literature reports
show that some of those treatments have been
employed in the past to successfully convert
polyethylene (PE), but not PP to carbon fibers. In
those few instances, short lengths of noncommercial, expensive PE fibers were subjected to a
variety of batch stabilization treatments. In one
experiment, batch-stabilized HDPE fibers were
subsequently converted to carbon fibers with
mechanical properties that approach LCCF program
targets (tensile strength of up to 369 ksi, modulus of
up to 22 Msi).4
In contrast to the processes discussed in
literature reports, our approach involved taking
commercially available polyolefin tows and
developing an industrially scalable method for their
conversion to low-cost carbon fibers. Accordingly,
polyolefin tows with 1000 to 3000 filaments and 2
to 3 denier per filament were obtained from various
commercial suppliers. Portions of the tows were
subjected to sulfonation, radiation, or combined
treatments with or without cross-linking additives.
Preliminary results and manufacturing cost model
calculations indicated that polyolefin fiber
treatments in hot, concentrated sulfuric acid could
provide adequate stabilization within LCCF program
targets. A multi-stage sulfonation reactor was built
to allow the continuous feeding, prewetting,
sulfonation, washing, and winding of polyolefin
tows. Line speeds were set to allow the direct
carbonization of sulfonated and washed polyolefin
tows if desired. LLDPE was selected over HDPE
based on its lower cost, easier spinnability, and more
open structure. Initial trials allowed the
Automotive Lightweighting Materials
FY 2001 Progress Report
determination of the experimental conditions at
which either LLDPE or PP tows could withstand
continuous immersion in hot sulfuric acid without
breaking. These trials confirmed that both PE and
PP tows can be successfully stabilized in a
continuous process.5 However, tests with the more
economical PP indicated that the conditions needed
to fully stabilize it led to poor tow processability and
quality. The trials carried out to date also confirmed
the importance of careful control of temperature
stability, line tension, acid concentration, and drying
conditions.
estimates, work on polystyrene was halted to focus
on the most promising LCCF technologies.
Polyvinyl Chloride Pitch
The strategy of producing carbon fiber from
PVC pitch was abandoned as a sub-task in early
2001. Several factors contributed to this decision.
The negative factors that were known prior to
experimental work included a very low theoretical
carbon yield (resulting from the presence of chlorine
in the polymer backbone); the necessity to discard or
recycle hydrogen chloride), a relatively low-value
pollutant and byproduct of PVC pitch pyrolysis; and
the rather poor carbon fiber made via this route as
reported in the literature. Based on optimistic
assumptions about the ability to develop a process
that achieves the LCCF program target properties at
a reasonable stabilization time, this sub-task
remained a marginal candidate on cost potential
through the year 2000. This was primarily because
PVC is the cheapest commercial polymer available,
and the conversion of pitch fiber to carbon fiber is
reasonably well understood.
Developing a process for extruding pitch fiber
with good conversion potential brought additional
technical problems, some of which were eventually
solved. However, little evidence was developed that
suggested a pitch could be made via this route that
would be superior to commercially available pitches
as a carbon fiber precursor. This conclusion was due
to the rather narrow window of PVC-to-pitch
processing conditions available to make a pitch with
a reasonable softening point.
It was shown, however, that extrudable pitches
could be made by heat-treating a standard grade of
PVC powder in the range of 25 to 45 minutes of
reaction time at 400°C. Extrusion experiments with
these samples and a control, a commercial coal tarbased pitch, were carried out with the aid of a
capillary rheometer. Although the rheometer is not
an ideal tool for melt-spinning and fiber collection,
the experiments allowed some qualitative judgments
to be made. It was not possible to wind fiber from
the rheometer extrude from PVC-pitch samples that
possessed a reasonable softening point, and the
relatively low softening point that resulted from the
longest stabilization time would indicate great
difficulty in stabilization. The only recourse to carry
this work further would have been to attempt a
scale-up of the PVC-pitch pyrolysis to generate
Polystyrene
Like polyolefins, polystyrene resins are also
appealing as LCCF precursor materials because of
their commercial availability, high carbon content,
and relatively low cost. But unlike polyolefins,
polystyrene is not available commercially in the
form of fiber tows. After contacting a series of
polystyrene resin suppliers and commercial pilotline melt spinners, it was possible to produce multifilament tows of polystyrene fibers and to acquire
hands-on experience with their manufacture. Owing
to their molecular characteristics, polystyrene fibers
can be melt-spun; but the resulting fibers are so
brittle that they cannot be drawn at reasonable line
speeds. Higher spinning temperatures improved their
spinnability but enhanced their degradation rate.
Literature reports provide potential pathways to
improve the manufacture and properties of meltspun polystyrene fibers (e.g., through multi-stage
draw processes). Some evidence also indicated that
polystyrene fibers can be cross-linked much more
quickly and effectively than polyolefin fibers. In
fact, patent literature reports indicate that largediameter carbon fibers with low mechanical
properties (tensile strengths of up to 71 ksi and
modulus of up to 9 Msi) have been obtained by
carbonization of polystyrene fibers treated with
warm sulfuric acid for a few minutes.
Based on these observations, and on kinetic
modeling parameters derived from successful benchscale stabilization and carbonization trials, a detailed
LCCF manufacturing cost model was developed.
The conclusion from the cost model calculations
was that polystyrene fibers, if made available, could
not offer any significant cost advantages as LCCF
precursors compared with PAN fibers. In view of
the technical difficulties and unappealing cost model
119
FY 2001 Progress Report
Automotive Lightweighting Materials
sufficient quantities for a true melt-spinning study.
Given the lack of encouragement from the benchscale work, this did not seem to be worth the effort
and expense, relative to other, more promising subtasks.
Cost Modeling of LCCF Technologies
8.0
Without SA/RD/ROI
With 10% SA/RD
With 10% SA/RD, 20% ROI
7.0
CF
Co 6.0
st
[$/
lb]
5.0
120
4.0
3.0
Large Tow
Benchmark
Commodity
Textile PAN
Chemically
Modified CTPAN
RadiationTreated CTPAN
Sulfonated
Polyethylene
Figure 4. Predicted carbon fiber manufacturing costs for
selected technologies.
to fall well below the $5.00/lb upper LCCF program
boundary. If 20% ROI is included for these LCCF
technologies, the minimum attainable carbon fiber
costs would be on the order of ~$5.60 to 5.90/lb. A
distribution of costs into their various components
(Figure 5) shows that the savings relative to the
large-tow benchmark (baseline) technology would
mostly be attained by using less expensive materials
as precursors, and to a lesser extent by reducing
capital costs.
3.0
Materials
Capital
Utilities
Labor
2.5
CF Cost [$/lb]
Manufacturing cost modeling has been an
essential component of all the LCCF technologies
investigated. Based on an earlier model by Cohn and
Das,6 derived from conventional PAN-based carbon
fiber manufacturing practices, costs of potential
LCCF technologies were initially divided into four
main categories: precursor, stabilization,
carbonization, and other.2, 3, 6 The most significant
cost contributor to the conventional carbon fiber
manufacturing process is the precursor.3 This is
followed by the precursor stabilization step and to a
lesser extent by the carbonization step. Hence, initial
LCCF efforts were geared toward uncovering less
expensive precursors and stabilization methods. The
cost model incorporates fixed capital, maintenance,
taxes, insurance, labor, and utility costs, as well as
variables such as sales and administrative/research
and development (SA/RD) costs and return on
investment (ROI). Technical factors such as carbon
fiber yield, line speed, and productivity were also
incorporated into the model. In each case, capital
investment costs were estimated for a production
volume of 4 million lb/year of carbon fiber in two
parallel manufacturing lines. This resulted in
estimated capital expenses on the order of
$36 million for the baseline case using a commercial
large-tow PAN precursor, while required capital
expenses for the candidate LCCF technologies
ranged from $30 to $40 million. The projected
carbon fiber manufacturing costs for the LCCF
technologies deemed most promising at the end of
Year 2 of the LCCF program are shown in Figure 4.
Figure 4 shows that the $5.00/lb upper LCCF
program cost target cannot be met in a sustainable
manner (that is, without allowing for overhead or
profit margins) based on the use of commercially
available large-tow PAN precursor. In fact, the
present cost model suggests that to meet the ~$6.00
to $6.50/lb sales price currently claimed by some
large-tow carbon fiber manufacturers, the ROI
margin could only be within ~5.5 to 11.1%. The
predicted costs of the other four technologies,
including 10% SA/RD but excluding ROI, are seen
2.0
1.5
1.0
0.5
0.0
Large Tow
Benchmark
Commodity
Textile PAN
Chemically
Modified CTPAN
RadiationTreated CTPAN
Sulfonated
Polyethylene
Figure 5. Distribution of carbon fiber manufacturing
costs for selected technologies.
Conclusions
The results of the laboratory trials of proposed
technologies) were as follows.
1. We halted further work on acrylic fibers
spun without solvents, plasticized PAN,
Automotive Lightweighting Materials
FY 2001 Progress Report
PVC, and polystyrene because of technical,
environmental and cost issues (Table 5).
2. We evaluated and selected the following
most promising LCCF technologies for
Phase 2: commodity textile PAN-based
precursors (as-received and with
pretreatment using chemical modification,
radiation, and nitrogen prestabilization
technologies) and polyolefin precursors
(LLDPE and PP).
3. We used a large-tow PAN precursor
technology benchmark as a metric to
evaluate the proposed technologies in terms
of their potential to meet the LCCF
program’s cost targets. The difference
between commodity textile PAN and largetow precursor, based on carbon fiber cost, is
approximately $1.80 vs $3.10/lb.
while producing carbon fiber properties that
meet the LCCF program requirements using
standard Hexcel’s precursor PAN.
3. Evaluated and selected E-beam radiation
technology for the prestabilization of
commodity textile acrylic tow to achieve the
cost and mechanical properties targeted by
the LCCF program. This was demonstrated
by working with acrylic textile producers
and E-beam and gamma radiation equipment
manufacturers and by acquiring and
processing acrylic polymer materials.
4. Developed a continuous two-stage
sulfonation process for the stabilization of
LLDPE and PP fibers. This was
demonstrated by acquiring polyolefin fibers
and producing sulfonated samples at various
processing conditions and carbon fiber
conversion levels.
5. Developed cost models for the selected
technologies: large-tow benchmark,
commodity textile acrylic fibers with
chemical modification and radiation
pretreatments, and LLDPE sulfonation. The
cost models incorporate fixed capital,
maintenance, taxes, insurance, labor, and
utility costs, as well as variables such as
S/RD and ROI costs. Technical factors such
as carbon fiber yield, line speed, and
productivity were also incorporated into the
model. For comparison purposes, capital
Summary
These are highlights of the progress during
FY 2001.
1. Processed and evaluated various state-ofthe-art PAN and developmental PAN
precursors that meet the targeted properties
for the LCCF program using a single-tow
research line and a multiple-tow pilot line.
2. Developed the continuous in-line chemical
modification of PAN precursor on researchline and pilot-line scales and demonstrated a
reduction in stabilization times of ~50%
Table 5. Technologies eliminated and the reasons for their elimination
Eliminated technologies
Reasons for elimination
Comments
Technical
Cost
Acrylic fibers without
solvents
X
X
Fusion
Insufficient cost incentive to solve
technical problems
Plasticized PAN
X
X
Capital-intensive
Needs non-aqueous solvent
Polyvinyl chloride
X
X
Low yield
Cost of environmental compliance
Polystyrene
X
X
Spinning issues
Environmental issues
No cost advantages
121
FY 2001 Progress Report
Automotive Lightweighting Materials
6. investment costs were estimated for a
production volume of 4 million lb/year of
carbon fiber in two parallel manufacturing
lines. This resulted in estimated capital
expenses on the order of $36 million when
using a commercial large-tow PAN
precursor (baseline) and $30 to $40 million
for other LCCF technologies. The predicted
costs of the other four technologies,
including 10% SA/RD but excluding ROI,
were found to fall well below the $5.00/lb
upper LCCF program boundary. If 20% ROI
is included for these LCCF technologies, the
minimum attainable carbon fiber costs
would be on the order of ~$5.60 to 5.90/lb.
It was also concluded that the lowest cost
for polymeric precursor fibers cannot be less
than $0.50/lb. Submitted two records of
invention: “Polyolefin Sulfonation and
Carbonization” and “Chemical Modification
of PAN” and a public disclosure on
“Chemical Modification”
8. Produced three papers [two published during
Society for the Advancement of Materials
and Process Engineering (SAMPE) 2000,
and one accepted for SAMPE 2001].
Commodity textile acrylic fiber costs are ~$0.70
to $0.90/lb. At current carbon fiber consumption
levels, one acrylic facility could supply the world
demand for carbon fiber precursor. Also, very large
carbon fiber lines are not decisive in achieving low
carbon fiber costs. Individual units at 2 million
lb/year can show favorable economicsA significant
cost benefit would be realized by moving from
large-tow precursor to commodity acrylic, with
savings of about $1.30/lb of carbon fiber. However,
manufacturing a carbon fiber with a tensile strength
of 400 ksi from commodity acrylic is challenging.
Enhanced stabilization rates using radiation
and/or chemical modification should lower carbon
122
fiber costs to ~$4/lb (without ROI, SA/RD). The
best results will be achieved by working closely with
acrylic fiber manufacturers to improve the carbon
fiber conversion potential of currently available
material.
Developmental “precursors” have met the LCCF
program’s targets, but only within a very narrow
cost window. Also, high-volume melt-processed
polyolefins present a competitive alternative to
PAN. Polyolefins have been shown to approach the
LCCF program’s targets, but a continuous largescale process must be developed.
References
1. C. D. Warren, “Carbon Fiber in Future
Vehicles,” SAMPE Journal 37(2), 7–15 (2001).
2. C. A. Leon y Leon, R. A. O’Brien, H.
Dasarathy, J. J. McHugh, and W.C. Schimpf,
“Development of Low-Cost Carbon Fibers for
Automotive and Non-Aerospace Applications. Part
I: Review of Existing and Emerging Technologies,”
pp. 1–11 in M. Rokosz, R. B. Boeman, D. T.
Buckley, and J. Jaranson, eds., SAMPE-ACCEDOE-SPE Midwest Advanced Materials and
Processing Conference, Dearborn, Michigan, 2000.
3. S. Smith, “Development of Low-Cost Carbon
Fibers for Automotive and Non-Aerospace
Applications. Part II: Precursor Processing
Technologies,” pp. 12–23 in M. Rokosz, R. B.
Boeman, D. T. Buckley, and J. Jaranson, eds.,
SAMPE-ACCE-DOE-SPE Midwest Advanced
Materials and Processing Conference, Dearborn,
Michigan, 2000.
4. S. Horikiri, J. Iseki, and M. Minobe, U.S.
Patent 4,070,446 (Sumitomo 1978).
5. C. A. Leon y Leon, R. A. O’Brien, J. J.
McHugh, H. Dasarathy, and W. C. Schimpf, 33rd
International SAMPE Technical Conference, Seattle,
Washington, 2001 (accepted).
6. S. M. Cohn and S. Das, “A Cost Assessment
of Conventional PAN Carbon Fiber Production
Technology,” Oak Ridge National Laboratory, 1998
(unpublished draft).
Automotive Lightweighting Materials
FY 2001 Progress Report
C. Low-Cost Carbon Fiber for Automotive Composite Materials
Principal Investigator: Donald G. Baird
Department of Chemical Engineering
Virginia Tech, Blacksburg, VA 24061
(540) 231-5998; fax: (540) 231-2732; e-mail: [email protected]
Co–Principal Investigators, Virginia Tech
James E. McGrath, Department of Chemistry, (540) 231-5976; e-mail: [email protected]
Garth L. Wilkes, Department of Chemical Engineering, (540) 231-5498; e-mail: [email protected]
Principal Investigator: Amod A. Ogale, Clemson University
Co-Principal Investigators: Dan D. Edie and John M. Kennedy, Clemson University
Project Manager: C. David Warren, Composites
Oak Ridge National Laboratory
P.O. Box 2009, Oak Ridge, TN 37831-8050
(865) 574-9693; fax: (865) 574-0740; e-mail: [email protected]
DOE Program Manager: Joseph A. Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Participants, Virginia Tech
Priya Rangarajan, Post-doctoral associate
Michael Bortner, Ph.D. student
V. Bhanu, Post-doctoral associate
David Godshall, Ph.D. student
Kent Wiles, M.S. student
Technical Support, Clemson University
Chuan Lin, Ph.D., Visiting assistant professor/research associate
Conceicao Paiva, Ph.D., Visiting assistant professor/research associate
Tom Haynie, Graduate research assistant
William Poole, Undergraduate research assistant
Contractor: Virginia Tech
Contract No.: 4500011036
Objective
•
Develop a cost-effective route to producing low-cost carbon fiber (<$5/lb) with a tensile strength of greater than
2.7 GPa, a modulus of >170 GPa, and a strain-to-failure rate of >1%. This range of carbon fiber properties has
been deemed necessary for generating composite materials suitable for automotive applications.
123
FY 2001 Progress Report
Automotive Lightweighting Materials
OAAT R&D Plan: Task 5; Barriers A, B
Approach
•
Develop melt-spinnable polyacrylonitrile (PAN) copolymers by controlling the copolymer ratio, molecular
weight, and end-groups.
•
Identify a benign solvent or plasticizer for PAN copolymers, allowing precursor fibers to be generated by meltspinning techniques.
•
Identify an alternate melt-spinnable precursor system containing a higher carbon content.
•
Develop a thermal treatment for oxidative stabilization and carbonization of fibers made from different
precursors to produce low-cost carbon fibers with desired mechanical properties.
•
Develop a new stabilization process based on ultraviolet (UV) irradiation to cross-link the fibers.
Accomplishments
•
Successfully prepared melt-processable acrylonitrile (AN) copolymers by copolymerizing methylacrylate (MA)
with AN to produce a series of copolymers of controlled molecular weight but of variable MA content.
•
Found that compositions in the range of 88/12 to 90/10 AN/MA with small amounts of chemical stabilizers
(e.g., 1 wt% boric acid) exhibited sufficient stability for spinning on a laboratory system.
•
Treated fibers with a UV-sensitive coating, which allowed the surface to be stabilized by UV radiation. (The
fibers could not be stabilized by standard thermal oxidation treatment.) The fibers could then be subsequently
stabilized (i.e., cross-linked and cyclized) in an oxidative step.
•
Successfully carbonized the stabilized fibers.
•
Successfully saturated copolymers of AN and MA, with levels of AN as high as 90 mole%, with CO2 and
extruded them at temperatures low enough to prevent significant cross-linking. Carbon dioxide is an
environmentally benign processing aid that will diffuse out and leave the fiber in a state that will allow it to be
stabilized by thermal oxidation.
•
Discovered and developed an environmentally attractive, economical, and solvent-free polymer system that has
the potential to serve as a carbon fiber precursor.
Future Direction
•
Fine-tune the AN/MA melt-spinnable copolymer composition so that it is sufficiently stable for spinning in
larger-size equipment, but yet has the potential to be stabilized using a reasonable oxidation schedule. This will
be accomplished by selecting an appropriate AN/MA ratio (between 85 and 90 mole% AN), the amount of
stabilizer (i.e., boric acid), and the polymerization mechanism.
•
Develop fundamental information on the copolymers concerning rheological properties as a function of
temperature, time, and shear rate needed to design a large-scale spinning system.
•
Scale up the reactor for producing the AN/MA copolymer precursor for pilot-scale melt spinning.
•
Fine-tune the UV stabilization process and scale it up for handling multiple-filament bundles.
•
Design a spinning pack to handle marginally stable precursor copolymers.
•
Reduce the cost of the carbon fibers further by synthesizing AN/MA/epoxy terepolymers that can be
photochemically stabilized in minutes rather than hours.
124
Automotive Lightweighting Materials
FY 2001 Progress Report
•
Continue to develop alternate methods for reducing oxidative stabilization time by introducing sites in the
AN/MA copolymers that provide sites for cross-linking at lower temperatures.
•
Continue to search for a lower-cost, higher-carbon-content precursor that can be photochemically rather than
oxidatively stabilized.
Introduction
Melt-Spinnable PAN Copolymers
To use plastics in the manufacture of
automobiles to reduce weight and hence improve
energy consumption, it is necessary to reinforce
them with glass or carbon fibers. Glass is
inexpensive (less $1.00 per pound), but its
reinforcing capabilities and density make it less
desirable than carbon fiber. At present, carbon fiber
suitable for use in automotive composites costs more
than $7.00 per pound, a cost prohibitive for general
use. It has been determined that in order to
economically use carbon fiber in the manufacture of
automobiles, the price must be less than $5.00 per
pound. Approximately half of the present cost for
carbon fiber is associated with the production of the
carbon fiber precursor. For automotive composites,
the precursors at present are generated from PAN
copolymers that are solution-spun using highly toxic
solvents. Considerable savings could be recognized
if these copolymers could be melt-spun without the
use of the highly toxic solvents. Furthermore, as the
PAN copolymers contain only about 50 wt% carbon,
new precursors with higher carbon levels would
reduce the cost even more.
Melt-processable AN copolymers were
successfully prepared by copolymerizing MA with
AN to produce a series of copolymers of controlled
molecular weight but of variable MA content. In a
significant departure from past procedures,
copolymers were prepared with an azo-initiator,
which produced non-associative end groups and
thereby led to a significant reduction in melt
viscosity. These materials were successfully
extruded on a laboratory scale (see Figure 1) to
Project Deliverables
Figure 1. Micro-spinline for producing carbon fiber
precursors from small quantities of polymer
(<50 g).
At the end of this multi-year program, a meltspinnable polymer system will be developed and
scaled up that is suitable for producing carbon fiber
precursors that can be stabilized more rapidly by
means of radiation. It will lead to a significant
reduction in the cost of producing carbon fiber.
produce carbon fiber precursors. Efforts must be
continued to obtain the highest AN-to-MA ratio
possible while keeping the molecular weight at the
highest level possible in order to provide adequate
melt strength for melt-spinning. The approach we
are taking requires that a delicate balance be
maintained between a copolymer that is sufficiently
stable to be melt-spun but eventually will become
unstable, so that cross-linking and cyclization can
take place in a subsequent step. Therefore, we are
investigating other avenues for stabilizing the fiber
precursor besides direct thermal oxidative methods.
One approach employs a UV-sensitive coating
applied to the fibers which, when subjected to UV
radiation, will initiate the cross-linking of the
Planned Approach
To meet the target of generating carbon fiber
costing less than $5 per pound, the present approach
using PAN copolymers must be significantly
modified. These modifications include the
development of melt-spinnable PAN copolymers
that can be stabilized and carbonized efficiently and
economically, and the identification of a precursor
system that contains higher carbon levels.
125
FY 2001 Progress Report
Automotive Lightweighting Materials
the carbon fiber manufacturing process. Another
approach, which is available as a backup in case we
cannot produce a system that is sufficiently
thermally stable, is the use of CO2 as a plasticizer for
PAN copolymers. Carbon dioxide is a benign
solvent that is obtained from air and can be released
back to air. PAN copolymers that are currently
solution-spun may be melt-spun in the presence of
small amounts of CO2 (~5 wt%). Although we will
not actively pursue this approach in the coming year,
it could be an important technique for handling the
UV-sensitive terepolymers.
copolymer at the surface. The fiber can then be
subjected to heating to promote thermal oxidative
stabilization of the remainder of the fiber (see
Figure 2). It is desirable to make the whole fiber
Mitsubishi
Amlon
VaTech/Clemson
Figure 2. Scanning electron microscope micro-graphs of
UV–oxidatively stabilized fibers (UV 2.5 h–
150°C, 180–220°C).
Higher-Carbon-Level Precursors
Another factor that must be considered in
reducing the cost of carbon fiber is the use of a
precursor material that contains higher carbon
content. We have tested an entirely different
polymer, syndiotactic polybutadiene, as an alternate
precursor because it has a theoretically higher yield
of carbon. However, thus far it has been found to be
inferior to PAN-based precursors in hightemperature processing steps involved in the
preparation. Future efforts will be concerned with
adding a photosensitive group that would allow the
polymer to be cross-linked by means of UV
radiation.
UV-sensitive by synthesizing an AN/MA/epoxy
terepolymer. Significant cost reductions can result
from eliminating or reducing the time required for
oxidative stabilization. At present, we have shown
that fibers stabilized in this fashion can be converted
to carbon fiber (Figure 3). This is the slowest step in
Amlon
Mitsubishi
VaTech/Clemson
Figure 3. Scanning electron microscope micro-graphs of
carbon fibers produced by UV–oxidative
stabilization (2.5 h–150°C, 180–220°C)
followed by carbonization (1500°C–30 min).
126
Automotive Lightweighting Materials
FY 2001 Progress Report
D. Economical Carbon Fiber and Tape Development from Anthracite Coal
Powder and Development of Polymer Composites Filled with Exfoliated
Graphitic Nanostructures
Vice-President of Research and Development: Richard M. Pell
Cornerstone Technologies
350 Second Street, Plains Township, Wilkes-Barre, PA 18702
(570) 825-2799; fax: (570) 825-5744; e-mail: [email protected]
Project Manager, Transportation Composites: C. David Warren
Oak Ridge National Laboratory
P.O. Box 2009, Oak Ridge, TN 37831-8050
(865) 574-9693; fax: (865) 574-0740; e-mail: [email protected]
DOE Program Manager: Joseph A. Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Participants
Pennsylvania College of Technology: Michael E. Starsinic, C. H. (Hank) White
Contractor: Cornerstone Technologies, LLC
Contract No.: 4500011695
Objectives
•
Develop low-cost fiber spinning and tape casting/extrusion processes from low-cost, carbon-based raw
materials.
•
Produce high-stiffness, high-strength carbon-fiber-like particles for automotive composite structural
applications.
OAAT R&D Plan: Task 5; Barriers A, B
Accomplishments
•
Produced a unique new structural/conductive graphite platelet material, exfoliated graphite nanostructures
(EGN), with the promise of breakthrough performance as an injection-molded graphite-polymer composite
component.
•
Conducted trials of EGN-filled thermoplastic resin with successful injection molding at virgin-resin conditions.
•
Evaluated compounding and resin processing technologies that generate unique EGN-filled polymeric
composites.
•
Identified additives for EGN-polymer formulations to enhance its compatibility and to improve the performance
of EGN-polymeric composite components.
127
FY 2001 Progress Report
Automotive Lightweighting Materials
Future Direction
•
The project concludes at the end of FY 2001.
characteristics at substantially reduced molding
pressures and shearing forces. The resulting
extrudate can be compression-molded, injectionmolded, extruded, cast, blow-molded, vacuumformed, poltruded, and formed with other
processes common to the plastics, preimpregnated textile, and polymeric composites
industries. The primary advantage of a low-cost,
high-performance powder component for polymer
composites such as EGN is to drive polymer
performance in automotive applications closer and
closer to that of metals such as aluminum and
high-cost engineered fiber reinforcements
currently used for body, under-the-hood, and
interior components.
Introduction
The use of polymeric composites in materialsbased industries has almost quadrupled during the
last two to three decades, especially in automotive
applications. The benefits of polymerics include
lower cost; weight reduction; styling potential;
functional design; and new effects, including
superior acoustic characteristics, reduced
maintenance, and corrosion resistance. In most
applications, however, the resin system must be
reinforced by fibers and/or particulates to yield the
required stiffness and strength. The resins are
mixed with various additives for processing and
durability requirements and sometimes with fillers
for further cost reduction. These modifications
cause major reductions in toughness and
processability and difficulties in controlling
surface properties of the final product.
Expanded graphite is used along with carbon
black, various types of carbon fiber, metal fibers,
and ceramic nitrides as fillers to improve the
electrical and thermal conductivity of normally
non-conducting or poorly conducting polymeric
materials. Generally, these are used in
combination with other fillers such as glass, talcs,
aramid fibers, mica, and the like to obtain
additional property enhancements for overall heat
transfer solutions in polymer applications.
Exfoliated graphite nanostructure (EGN) is a
graphitic layer in platelet form, consisting of
multi-aromatic carbon ring nanostructures (see
Figure 1). EGNs are identical to the graphite
ribbon component of carbon fiber and are capable
of performing in thermoplastics as a low-cost
carbon reinforcement. Current EGN platelets have
high aspect ratios of 100:1 to 1000:1. In the fourth
quarter of 2001, EGN particles haves been
reduced in to 5 µm in width by 20 nm in thickness.
EGN micro-particles should act as agents to
inhibit micro-crack propagation, the dominant
damage mechanism producing composite and
polymer failures.
EGN-filled polymer resins can be
conventionally compounded using twin-screw
extrusion techniques. Compounding trials
demonstrate excellent compounding
Anticipated Property Benefits of Exfoliated
Graphite Nanostructures
EGN has the advantages of
• electrical conductivity
• thermal stability
• heat transfer
• flexural modulus
• tensile modulus
• ductility
• heat distortion temperature
• abrasion resistance
• surface quality
Table 1 lists some common uses of polymeric
composites in automobile manufacturing. Table 2
lists some of the disadvantages of those materials
in automotive applications.
Core Process Technology
Cornerstone’s process technology works by
focusing the kinetic energy of a high-pressure
liquid so that concentrated destructive forces are
applied to raw material feedstocks. It attacks
particles along natural cleave planes and defects
and high-energy grain boundaries, and applies
hydro-wedging attacks on particle grain
boundaries. The process generates extremely
intensive turbulence, shearing, high-velocity
128
Automotive Lightweighting Materials
FY 2001 Progress Report
Exfoliated graphitic nanostructures
Magnification 3000X
Note: Aspect ratio exceeds 100:1.
Crystalline graphite precursor (non-expanded)
Magnification 90X
Figure 1. Graphite precursor and EGN platelets.
Table 1. Typical polymers used in automotive applications
Resin system
Automotive application
Glass-reinforced polypropylene
Battery trays, brackets
Mineral-filled polypropylene
Heating, ventilating, air-conditioning, instrument panel, structural, under-hood
Thermoplastic olefins
Fascias, sound abatement, moldings, step pads
Unfilled polypropylene
Interior trim, splash shields, wiring conduit
Mineral-filled nylons (6 and 6,6)
Covers, gears, connectors
Acetal
Interior (knobs, levers), structural
Acrylonitrile butadiene styrene
Fascias
(ABS)
Urethanes
Instrument panel, interior trim
Polyvinyl chlorides
Interior trim
Polyphenylene sulfide
Gears, connectors, fuel pump components
Polyethlene terephthalate (highStructural, exterior
impact)
Table 2. Issues with current automotive resins
Resin system
Issues
Thermoplastic olefins/polypropylene/olefin Stiffness, heat stability
Nylons
Impact strength, cost
Polyvinyl chloride, urethanes, acrylonitrile
Cost (ABS, urethanes), recycling (all)
butadiene styrene (ABS), acetal
Thermosets
Impact, recycling
Fiber-filled resins
Surface/aesthetics/paintability cost of chopping fibers and blending
All polymer-based systems
Electrostatic paintability
collision/abrasion, and destructive cavitation. Its
combined hydraulic forces result in the
penetration, pressurization, and ultimate
exfoliation of layered structures. It successfully
comminutes a diverse range of materials including
graphite, coal, silica, wollastonite, zirconia,
alumina, ferrochrome, chromium metal, cordierite,
and other difficult-to-process materials. High-
aspect ratio particles are produced at the submicron and nano-scales at relatively low labor and
energy costs.
EGN Product and Process Performance
To date, uncoated and unmodified EGN
(18-mm D50 particle size, 50–100 nm in thickness)
129
FY 2001 Progress Report
Automotive Lightweighting Materials
•
has been successfully compounded with nylon 6
and polypropylene copolymer from 5% through
70% by weight. EGN-filled polypropylene and
nylon 6 pellets were injection-molded into formed
articles using virgin resin conditions.
EGN has been produced in a coated
concentrate (dust-free powder) format of 80%–
90% EGN with a coating of functionalized
polymer or polymeric-compatibilizer. A batch of
1000 lb has been produced, and production tooling
is in place for volume output. EGN production
capacity is 1500 lb per hour for a single
production system.
•
Exfoliated graphite nanostructure
advantages in nylon 6
•
•
•
•
•
Experimental Results
•
Table 3 and Figures 2–4 show the mechanical,
electrical, and thermal properties of polypropylene
copolymer composites with various levels of EGN
filler.
•
•
Exfoliated graphite nanostructure
advantages in polyolefins
•
•
•
•
•
•
•
Processability comparable to that of unfilled
(injection molding) at loadings of up to 55%
Substantial improvement of ductility and
tensile with property compatibilizers
•
Stiffness increase of up to 700% in initial
study of EGN advantages in polyolefins
Tensile strength increase of up to 30%
Impact decrease of only 50% at maximum
mechanical property improvements
Thermal performance surpassing that of
nylons
Electrical properties modified from insulating
to dissipative/conductive
Excellent surface quality at loadings of up to
50–55% in polypropylene copolymer
Potential unique combinations of mechanical,
electrical, and thermal properties
Tensile strength increase of up to 55% Tensile
modulus increased by 450%
Flexural modulus increased by 500%
Heat deflection temperature raised by 300%
Impact decrease of only 55% at maximum
strength increase
Thermal performance approaching that of
high-performance liquid crystal polymers
Electrical properties modified from insulating
to dissipative/conductive
Excellent surface quality at up to 50% EGN
loading
Potential unique combinations of mechanical,
electrical, and thermal properties
Processability comparable to that of unfilled
(injection molding) at up to 50% loading
Nanotechnology Breakthrough
Cornerstone’s core technology focuses highpressure liquid in a manner that applies
concentrated destructive forces on raw material
feedstocks, creating hydro-wedging attacks on
particle grain boundaries, extremely intensive
turbulence, shearing, high-velocity
collision/abrasion, and destructive cavitation.
These combined hydraulic forces result in the
penetration, pressurization, and ultimate
exfoliation of layered structures. In the case of
Table 3. Mechanical properties of EGN-filled polypropylene copolymer
Polymer/filler
Fina 7825
5% EGN
10% EGN
20% EGN
36% EGN
53% EGN
Yield
stress
(PSI)
3563
3920
3896
3916
4100
4516
Modulus
Tensile
(PSI)
47400
56500
59100
63700
85400
94600
Flexural
(PSI)
115000
173700
211000
281000
514000
912000
Elongation
Yield
(%)
5
4
4
6
3
5
130
Break
(%)
715
592
490
71
13
6
Izod
impact
Heat deflection
temperature
(ft*lb/in.)
(qC)
0.76
0.68
0.67
0.69
0.40
0.40
46
48
49
55
65
82
Automotive Lightweighting Materials
FY 2001 Progress Report
FINA 7825 with EGN
TENSILE
5000
4500
4000
Stress (psi)
3500
3000
2500
2000
1500
1000
500
0
0
5
10
15
20
25
30
35
40
45
50
55
60
65
70
Strain (%)
Virgin FINA 7825
FINA 7825 w/ 10% EGN
FINA 7825 w/ 36% EGN
FI
FI
FI
Figure 2. Polypropylene/exfoliated graphite nanostructure stress-strain curve.
FINA 7825 with EGN
Flexural
7000
6000
Stress (psi)
5000
4000
3000
2000
1000
0
0
1
2
3
4
5
6
Strain (%)
Virgin FINA 7825
FINA 7825 w/ 5% EGN
FINA 7825 w/ 10% EGN
FINA 7825 w/ 20% EGN
FINA 7825 w/ 36% EGN
FINA 7825 w/ 53% EGN
Figure 3. Conductivity in low-density polyethylene, injection-molded, and roll-milled commercial.
graphite, Cornerstone has created a unique, highaspect ratio morphology, with platelets exhibiting
nano-scale thickness.
feasibility of commercializing carbon-reinforced
polymer products that are not polyacrylonitrile
(PAN) based, use low-energy processing, and
reduce costs compared with PAN-based
alternatives.
It is anticipated that Cornerstone’s technology
will help the U.S. carbon and automotive
industries develop advanced processes for
manufacturing new products at reduced costs that
can be passed on to the consumer. In addition, a
new process-reactor technology will be introduced
Conclusions
From extensive preliminary studies and
literature searches, it was concluded that
dissolving anthracite coal in chemical agents was
neither feasible nor cost-competitive.
Cornerstone’s generation of a new exfoliated
graphite particle has permitted a modified
approach. The approach is to demonstrate the
131
FY 2001 Progress Report
Automotive Lightweighting Materials
Conductivity Comparison
0
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
-2
log Conductivity (S/cm)
-4
50 micron graphite/LDPE (Roll
Mill)
-6
Conductex/LDPE (Compression
Mold)
-8
EGN/LDPE (Compression Mold)
-10
EGN/PP (Injection Mold)
-12
Conductex/PP (Injection Mold)
-14
-16
-18
Volume Fraction Particles
Figure 4. Comparison of mechanical properties of polypropylene resin and polypropylene with EGN
filler.
for the efficient production of carbon/graphite
products.
The development of enhanced EGN-filled
polymeric composites would fulfill a number of
automotive and truck project goals and objectives.
Low-cost, conductive, engineered plastic
composites could enable the commercialization of
paintable, abrasion-resistant, high-performance
body components; functional under-the-hood
parts; and lightweight injection-moldable energy
management materials.
132
Automotive Lightweighting Materials
FY 2001 Progress Report
E. Microwave-Assisted Manufacturing of Carbon Fibers
Principal Investigator: Felix L. Paulauskas
Oak Ridge National Laboratory
Oak Ridge, TN 37831-8048
(865) 576-3785; fax: (865) 574-8257; e-mail: [email protected]
Project Manager, Composites: C. David Warren
Oak Ridge National Laboratory
P.O. Box 2009, Oak Ridge, TN 37831-8050
(865) 574-9693; fax: (865) 574-0740; e-mail: [email protected]
DOE Program Manager: Joseph A. Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5068; fax: (865) 576-4963; e-mail: [email protected]
Technical Support
Terry L. White, Oak Ridge National Laboratory
Kenneth D. Yarborough, Oak Ridge National Laboratory
Professor Roberto Benson, Matthew J. Williams, University of Tennessee, 434 Dougherty Engineering
Building, Knoxville, TN 37996-2200
Contractor: Oak Ridge National Laboratory
Contract No.: DE-AC05-00OR22725
Objectives
•
Investigate and develop a microwave-assisted technical alternative to carbonize and graphitize
polyacrylonitrile- (PAN-) based precursor.
•
Prove that carbon fibers with properties that make them suitable for use by the automotive industry can be
produced inexpensively using microwave-assisted processing.
•
Demonstrate that microwave-assisted processing can produce acceptably uniform properties over the length of
the fiber tow.
•
Show that, for specified microwave input parameters, fibers with specific properties may be controllably and
predictably manufactured using microwave furnaces.
•
Demonstrate the economic feasibility of producing fibers of approximately 30-Msi modulus at a significantly
lower cost than fibers produced conventionally.
OAAT R&D Plan: Task 5; Barriers A, B
Approach
•
Conduct parametric studies with the existing single-tow, pilot continuous carbon fiber processing unit to
provide data about unit performance, energy balance, limitations, and thresholds.
•
Characterize the properties of carbon fibers produced by microwave-assisted plasma (MAP) processing.
133
FY 2001 Progress Report
Automotive Lightweighting Materials
•
Use results of the parametric and fiber characterization studies to develop and incorporate modifications
necessary to transform the single-tow processing unit from a low-production-speed line to a high-productionspeed line.
•
Continuously evaluate the underlying technology, hardware, and ideas required to transform this laboratory unitow, high-production-speed line pilot unit to a multi-tow, moderate- to high-production-speed line.
Accomplishments
•
Completed the construction and basic evaluations of the MAP continuous carbon fiber processing prototype
laboratory unit. (This is a single carbon fiber tow processing facility.)
•
Demonstrated the processing of PAN-based stabilized precursor to carbon fiber in a continuous mode.
•
Acquired, modified, and installed two modular microwave units for use as an independently controlled
preheating section. Each unit has a nominal output power of 3 kW. The units are part of a hardware upgrade
leading toward a pilot processing unit, with a high-production-speed line, for single-tow carbon fiber.
•
Completed development of the required methodologies for morphological (crystallographic) evaluation of the
MAP-produced and conventionally produced carbon fibers. For these crystallographic evaluations, wide-angle
X-ray diffractometry was used. This permitted a fundamental comparison at the level of atomic crystal
unit/configuration of the MAP fiber and conventionally produced carbon fibers.
•
Developed methodologies for characterizing elemental chemical composition and functional groups present on
the surface of the carbon fiber. This characterization was undertaken with conventionally produced PAN-based
carbon fibers. It is performed with X-ray photoelectron spectroscopy and secondary ion mass spectroscopy.
•
Continued the effort to put in place the required underlying methodologies and resources for other types of
characterizations of the MAP-manufactured continuous carbon fibers on a routine basis.
•
Designed, constructed, installed, and evaluated a unique vacuum eyelet system in the processing unit. These
new devices will guarantee proper vacuum during processing and reduce frictional forces in the fiber tow,
thereby reducing the amount of filament breakage.
•
Filed a new patent application for a spin-off technology from this carbon fiber manufacturing program: No. IR0841: “Microwave and Plasma-Assisted Modification of Composite Fiber Surface Topography.”
Future Direction
•
Study process parameters, limitations, and threshold in this continuous pilot facility and incorporate
modifications.
•
Incorporate required new modifications and advancements into this continuous unit with the intent to achieve
higher production-line speed and multiple-tow capability.
•
Complete the required characterization for the MAP-processed carbon fiber and compare it with conventionally
processed, low-cost, PAN-based carbon fibers.
•
Initiate preliminary work in demonstrating related technologies in the area of carbon fiber manufacturing (e.g.,
surface treatment) as resources are available.
processed carbon fibers indicated that this
technology produces fibers with high-density values,
acceptable electrical resistivities, and standard fiber
diameters. The weight loss in the fibers ranged from
45% to 55%, which is equivalent to the values
obtained from conventional processes. The
mechanical properties were slightly lower than but
reasonably close to those of conventionally
Introduction and Progress
In the earlier stages of this program, microwaveassisted processing of PAN precursor for the
manufacturing of carbon fibers was proved to be a
feasible technical alternative to conventional oven
processing. Proof of this concept had been limited to
batch or quasi-batch processing. Test results on
134
Automotive Lightweighting Materials
FY 2001 Progress Report
processed carbon fibers. Required processing times
showed good potential for this process to produce a
significant cost savings.
With this knowledge, a more advanced
continuous processing unit was conceived,
developed, and built. Additional technical issues
were resolved, such as the need for a suitable
vacuum device/system to maintain an adequate
processing vacuum during operation in this unit.
Fiber-tow-handling devices (e.g., feeding,
tensioning, adjustment, take-up) were also
developed or bought to suit this project. The core of
the unit is the 18-foot-long reactor, which contains
the carbon fiber processing discharge tube. Nitrogen
gas is leaked into the discharge tube under
controlled conditions to maintain the required
(vacuum level) operating pressure. A section of this
reactor features a rectangular waveguide with
periodic openings in two opposite sides to enable
monitoring of the temperature of the tube via a twocolor infrared system during operation.
MAP carbon fiber was produced at speeds of up
to 3 in/min. All these MAP fibers were analyzed for
mechanical and electrical properties and for their
densities, tow cross-sectional area (calculated), and
single filament diameter (calculated). Mechanical
evaluations of these runs in the manufacturing of
carbon fiber in continuous mode indicated that the
product made steady progress toward meeting the
PNGV requirements during the first half of the
reporting period. More recent evaluations of these
manufactured carbon fibers indicate that they now
surpass all technical requirements of the PNGV
program. (The minimal requirements of the PNGV
program are tensile strength of 300 ksi, modulus of
25 Msi, elongation at break of 1%.)
Table 1 shows a comparison between
conventionally manufactured carbon fiber and
carbon fiber manufactured by the Oak Ridge
National Laboratory MAP process. The MAP
process was characterized at different line high
speeds (7.5, 20.5 and 34.6 in./min). Both
conventional and MAP processes were evaluated for
50-K tows. In general, these results indicate that the
MAP carbon fibers are comparable to and, in the
case of some specific properties, even better than the
corresponding conventionally processed, low-cost
carbon fibers. Furthermore, the results of
morphological analysis (X-ray diffractograms)
indicate that these MAP fibers possess a degree of
carbon fiber graphitization that is comparable to or
higher than that of conventionally processed carbon
fibers (see Figure 1). It was further demonstrated
that the morphological properties of the fiber
produced by the MAP process do not vary
significantly as a function of position along the tow
length. Figure 1 also indicates the production
flexibility in the MAP process to manufacture
carbon fiber to the required (average) level of
graphitization.
Table 1. Testing data for PAN-precursor-based fibers, 50-K tows, MAP carbon fibers,
immediately after processing
Physical testing data
MAP
Production line speed
Density/pycnometer
Calculated tow area
Calculated filament diameter
Tow linear electric resistance
Electrical resistivity
(in./min)
7.5
20.5
34.6
(gr/cm3)
1.83
1.82
1.77
(× 10–2 cm2)
1.90
2.03
2.04
(µ m)
6.97
7.19
7.20
(× 10–2 Ω/cm) 8.37
9.80
8.90
(× 10–3 Ω/cm) 1.57
1.89
1.82
Mechanical testing data
Panex 33
–
1.81
1.96
7.37
8.23
1.62
Conventional
Fortafil F3 (C)
–
1.77
2.17
7.43
7.80
1.67
Conventional
PNGV Program
Fortafil
Goals
Panex 33
F3 (C)
Production line speed
(in./min)
7.5
20.5
34.6
–
–
–
Modulus
(× 106 lb/in.2)
31.7
30.5
29.0*
26.1
31.1
25
Ultimate strength
(× 103 lb/in.2) 424.1 384
341.9*
407.8
485.3
300
Elongation at break
(%)
1.3
1.22
1.25*
1.5
1.5
1.0
*These fibers were neither surface treated nor sized and tested 4 months after the manufacturing date.
MAP
135
FY 2001 Progress Report
Automotive Lightweighting Materials
3000
Project Deliverables
At the end of this multi-year program, a
protoype high-production-speed line for continuous
MAP processing of carbon fiber will have been
developed, constructed, and tested. A final report
will be issued with the results of tests of the carbon
fibers processed with this unit. Appropriate industry
briefings will be conducted to facilitate
commercialization of this economically enabling
technology.
Intensity (counts)
2500
a) Commercial
(AKZO F3(C) 50K)
b) High energy MAP
c) Low energy MAP
2000
1500
1000
500
0
10
15
20
25
30
35
40
45
50
2-Theta (degrees)
Figure 1. Comparison of the X-ray diffractograms for
commercial and MAP carbon fiber produced
under different microwave energy processing
levels (curves not normalized).
136
Automotive Lightweighting Materials
FY 2001 Progress Report
6. RECYCLING
A. Sorting Mixed Alloys from Shredded Automobiles
Project Contacts: Jim Quinn, General Motors
(248) 680-4732; fax: (248) 680-2843; e-mail: [email protected]
Dick Klimisch, The Aluminum Association
(248) 784-3007; fax: (248) 784-3006; e-mail: [email protected]
John Green, The Aluminum Association consultant
(410) 465-6354; e-mail:[email protected]
Adam Gesing, Huron Valley Steel Corporation
(734) 697-6313; e-mail:[email protected]
DOE Program Manager: Joseph Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Participants
Mike Thomas, Alcan
Mike Wheeler, Alcan consultant
Paul Schultz, Alcoa Aluminum
Andy Sherman, Ford Motors
Contractor: U. S. Automotive Materials Partnership
Contract No.: DE-FC05-95OR22363
Objective
•
Identify, evaluate, and validate methods for separating and sorting aluminum in shredded automotive scrap.
OAAT R&D Plan: Task 9; Barrier E
Approach
•
Demonstrate the capability to separate wrought from cast materials.
•
Demonstrate existing and developed scrap sorting capability using color and laser- induced breakdown
spectroscopy (LIBS).
•
Integrate the various elements (scrap preparation, decoating, sorting, recycling) required for commercial
recycling of automotive wrought alloy scrap.
Accomplishments
•
Optimized Alcoa’s proprietary caustic etching alloy-coloring process to provide color tints to the alloy mix in
order to facilitate color sorting.
•
Conducted melt tests to establish a composition for each color grouping.
•
Used the LIBS technology to sort 5xxx alloys from 6xxx alloys.
137
FY 2001 Progress Report
Automotive Lightweighting Materials
•
Conducted trials evaluate bulk cleaning procedures for the alloy samples.
•
Demonstrated a clear, consistent separation of 5182 and 6111 alloy samples into separate bins.
Future Direction
•
Conduct sorting tests of additional alloys to establish that the specific separations are possible and establish
procedures to group materials by major alloying elements.
Project Rationale
The fraction of the wrought aluminum used in
the automobile grows, the automotive aluminum
alloys in which use of aluminum in automobiles is
becoming increasingly widespread. Previously, this
aluminum was predominantly cast material for
applications in the power train. Now much more
wrought aluminum, both sheet and extrusions, is
being used in the bodies of vehicles.
In the past, all aluminum that was recycled from
auto shredders was remelted into casting alloys and
then reused in cast components. As the amount of
wrought material increases, this procedure will
eventually become unacceptable because the
wrought materials are compositionally distinct from
and metallurgically incompatible with the cast
materials.
As the fraction of wrought aluminum used in
automobiles grows, the scrap-consuming alloys
currently used—mainly foundry 380 and 319—will
no longer be able to use the entire volume of
automotive aluminum scrap. The recycled content of
some other alloys will have to be increased to keep
the total aluminum recycling system functioning
efficiently. In the long term, this change will require
the capability to batch desired “hardener”"
compositions by sorting mixed alloy scrap shred.
While dismantling of autos will clearly be used
to some extent to recover the aluminum, the
overwhelming portion of aluminum will be
recovered from shredded automobiles. Accordingly,
it is envisaged that there will be a modification to
the current recycle process. Figure 1 illustrates
schematically the additional processing of alloy
sorting that will be inserted into the overall recycle
loop (see the hatched area). Alloy sorting will
require decoating and pre-cleaning of the shredded
particles plus separation of aluminum from other
metals. Technologies to accomplish these tasks are
already all commercially practiced.
Figure 1. Future automotive aluminum recycling
process. Source: “The Aluminum Industry
Roadmap for the Automotive Market,” The
Aluminum Association, May 1999.
There are intermediate aluminum sorting steps
that may open some other markets, both automotive
and non-automotive, for the shredded aluminum
scrap and delay the date when the recycling stream
will need composition-based alloy sorting
technology. These intermediate steps include
separation of cast from wrought particles and further
property-based grouping within the wrought and cast
alloy families. Accordingly, a more comprehensive
picture of the future recycling loop incorporating
composition-based alloy sorting is shown in
Figure 2. The additional elements are shown in the
bottom left side of the figure.
Using alloy-sorting technology to batch tailored
hardener compositions will enable keeping the
various loops of the total aluminum recycling
system closed without requiring that each alloy in a
car have substantial recycled content. Thus
designers will be able to keep optimizing component
138
Automotive Lightweighting Materials
FY 2001 Progress Report
The materials collected were designed to
anticipate those wrought alloys that will be added to
the recycling stream in 10–15 years. The following
wrought alloys were already collected in ~100 kg
quantity: 2036, 3003, 4147, 5182, 6061, 6111, 7003,
7016, and 7129. Additional samples of 5052, 5754,
6016, 6022 and 6063 have been sought. These
known alloys will be seeded into the current
unknown alloy mix of both wrought and cast shred
pieces. Additional alloy material is continuously
being sought. Recently, Ford Motor Company
donated three AIV hulks to be shredded and added
to the material mix for separation.
The approximate composition of samples of
unknown shred pieces was determined using a handheld optical emission spark spectrometer. HVSC has
used this technique to characterize the current
generation of the aluminum alloy shred mix,
analyzing and identifying ~6,000 individual shred
pieces. These pieces are ready for LIBS
composition-based sorting.
The HVSC pilot plant was used for the sorting
tests. This pilot line simulates full-scale plant
operation and operates at a production rate of
7000 lb per hour. Plant tests on separation of cast
from wrought alloys were done at the HVSC
Belleville and Overpelt, Belgium facilities on a scale
of 100–200 tons of aluminum mix input.
Figure 2. Future aluminum recycling system using alloy
sorting technology.
performance and minimizing vehicle weight to make
the maximum reductions in greenhouse emissions.
Sorting Test Results
Cast from Wrought Alloys
Collection, Identification, and Preparation of
Scrap for Sorting Demonstration
Based on its prior experience, HVSC developed
and implemented a proprietary wrought–cast
separation method. This method has been tested in
plants at both Belleville and Overpelt. The test at
Belleville involved 90 tons of material; the resulting
products of these tests have been sold commercially,
indicating the beginnings of a market for sorted
aluminum material.
Table 1 gives the results of the initial Belleville
test, expressed in terms of the percentage of the
component recovered. The starting material
contained some 60% wrought material. The final
wrought component contained 85% of the original
wrought material, and the final cast component
contained 88% of the original cast component.
The demonstration of the viable wrought–
cast separation is a major achievement in the
The Huron Valley Steel Corporation (HVSC)
location in Belleville, Michigan, was the site for the
sorting tests. HVSC has developed significant
capabilities for the separation of scrap materials and
is in an advanced phase of the development of
technology for the sorting of aluminum alloys.
HVSC has provided mixed-alloy aluminum shred
from its current production in quantities of 100–
200 tons for plant testing and 1–2 tons for pilot plant
testing.
All participants were asked to contribute
materials of known composition for the sorting tests.
Materials were either as-received from auto
shredders or cut to < 4 in. in the largest dimension,
and were then hammer-milled to simulate the
shredding operation.
139
FY 2001 Progress Report
Automotive Lightweighting Materials
a sample of ~400 kg of shred. Color sorting
separated the unknown mix into three distinct
categories: “bright”—pure aluminum and binary
AlMg alloys, “gray”—AlMn, AlSi, and AlMgSi
alloys, and “dark”—alloys containing zinc and
copper (see Figure 4).
Table 1. Recovery of component in
product in the 90-ton test at
Belleville
Products
Components
Wrought
Cast
Wrought
84.6%
14.3%
Cast
6.3%
88.0%
development of an overall sorting strategy for
shredded automobile material, and the development
of commercial sales is a further positive sign for the
use of aluminum in automotive applications.
Alloy Family Grouping by Selective Etching
and Color Sorting
A proprietary alloy coloring process developed
by Alcoa was used to provide a color tint to the alloy
mix to facilitate color sorting. In this process,
exposure to a caustic etching process colors various
alloy families differently. With the support of Paul
Schultz of Alcoa, the caustic etching process was
optimized. The specific conditions chosen following
extensive testing were exposure to 10% NaOH
solution at 50°C for 60 seconds. These conditions
enabled good control and tint reproducibility and
minimized metal loss, environmental fumes, and
cost.
In Figure 3, material to be sorted passes through
the etching process before being placed on the belt.
The conveyor belt takes the material under a camera
that then enables the sorting process.
Figure 4. Scrap pieces can be sorted into three major
groupings following etching.
Subsequent melt tests were conducted to
establish a composition for each color grouping.
Compositional and mass balance analysis indicated
the following results:
•
•
•
•
There is a small metal loss of ~0.5% from the
etching process.
The contrast developed during etching can be
increased by higher temperatures, concentrations
and times; however, this also increases the metal
loss.
The color sorting method is well suited to
separating mixtures of known alloys with good
color contrast. Unfortunately, however, there is
no direct correlation of color to alloy family
group. The sorting method collects unknown
alloy mixes into color groups that roughly
correlate to the composition of individual pieces.
The family grouping will be sufficient to open
up some new market for auto shred material.
Figure 3. Color sorting process.
Scrap Sorting Using LIBS
Batch samples of alloyed shred were subjected
to coloring by selective etching under previously
described conditions. Samples of both known and
unknown alloys were etched and sorted. The final
sort of the HVSC unknown alloy mix was based on
To batch a tailored hardener for a particular
alloy from shredded scrap, a sorting method based
on chemical composition, rather than physical
properties, will be required. The LIBS technology
permits quick, precise, non-contact, elemental
140
Automotive Lightweighting Materials
FY 2001 Progress Report
chemical analysis of the surfaces of randomly
shaped scrap particles. This technology is based on
optical emission spectroscopy, is more complex and
costly than other technologies, and likely will be
used only as a final sorting step in combination with
other techniques.
Recent proprietary work at HVSC has been
successful in integrating LIBS technology to full
production scale on a pilot line. Specifically, the
integration of laser operation, particle tracking,
analysis, and separation has been achieved at
production rates of up to 50 shred particles per
second.
While some laser ablation of the surface is
possible during the LIBS analysis, for optimum
analysis results, the surfaces need to be cleaned of
all paint and protective coatings. The LIBS
technique is schematically illustrated in Figure 5.
concentration in the two alloys, as opposed to silicon
or other alloying additions.
Additional trials were conducted to evaluate
bulk cleaning procedures for the alloy samples.
Specifically, in these experiments, the LIBS Mg/Al
and Si/Al spectral intensity ratios, measured for
freshly ground surfaces, were compared for other
methods of cleaning. On freshly ground surfaces, it
was found that that a stable intensity ratio was
achieved from the first laser shot (note: there are 4
energy pulses within each laser shot). For other
cleaning procedures, ranging from water washing to
caustic etching, more than one multi-pulse laser is
necessary to achieve the stable intensity ratio, which
is assumed to be representative of the bulk alloy
composition.
When the experimental conditions had been
optimized, the 5182 / 6111 alloy sorting experiment
was conducted. It was found that 5182 samples were
diverted consistently to bin 1 and 6111 samples were
collected in bin 2. A clear separation of 5xxx and
6xxx alloys had been demonstrated, which is highly
significant from the viewpoint of enhancing the
economics and efficiency of automotive recycling.
Given this successful demonstration, additional
alloy sorting tests have been scheduled. For
example, the separation of the following alloy
groups is under way or is planned:
Figure 5. Schematic of the LIBS sorting process.
•
•
•
•
Sorting Alloy by Alloy
Additional work on the LIBS technology has
been conducted during the past period of the
contract. The sorting of 5xxx alloys from 6xxx
alloys has been conducted. This is of especial
interest to the automakers because it is anticipated
that these alloy families will compose the bulk of
wrought material used in future vehicles.
For this sorting demonstration, attempts were
made to separate the alloys 5182 from 6111. Both
these alloys are being used in current vehicle
production. For the purposes of this demonstration,
the pilot plant sorter at HVSC was used. Sheet
samples of known 6111 alloy (containing ~1%Mg)
and 5182 alloy (containing ~5%Mg) were placed
randomly on the sorter belt moving at 300 ft per
minute. Following calibration experiments, it was
determined that it was best to attempt the sorting
based on a distinction of the magnesium
3003/3004
3003/6111/5182/2036/7129
5182/5754/5052
6111/6061/6063/6082
In addition to establishing that these specific
separations are possible, these tests will also
establish procedures to group materials by major
alloying elements. Alternatively, these tests could
enable sorting of particles compatible with a
particular alloy or batching of materials to achieve a
particular alloy composition. All this should make it
possible to batch a product with the particular
selected proportions of alloying elements to provide
a “tailored hardener,” as was discussed in the
context of the future aluminum recycling scheme
shown in Figure 2.
Process Improvements
Since the start of the project, several process
improvements have been incorporated into the
141
FY 2001 Progress Report
Automotive Lightweighting Materials
HVSC pilot line. For example, a faster and more
powerful laser system has been acquired, and a more
efficient spectrometer has been received and
incorporated into the line. Also, new laser beam
delivery optics and light collection optics are due to
be installed during the October–November
timeframe. The latter equipment is designed to
minimize variations in the LIBS spectral line
intensity resulting from variations in particle height
or position on the moving belt. These improvements
will further improve the effectiveness of the LIBS
sorting system.
Finally, the pilot line has been modified to
enable “multi-stream diversion.” Initial test results
showed 95% diversion of the targeted particles into
the first two product streams and ~80% into the third
stream at industrial-level belt speeds and loadings.
However, it is already clear from testing on the
HVSC pilot plant that the process of LIBS is
sufficiently robust to sort particulate material from
shredded autos on a particle-by-particle basis. In
fact, HVSC is moving forward with plans to develop
a commercial facility in anticipation of the
developing market for the recycling of aluminumcontaining material from shredded automobiles.
Specifically, it has been shown that it is possible to
readily separate 5xxx from 6xxx alloys. This is
especially significant to automakers because
wrought material from these alloy families will be
used extensively in future vehicle production. The
project has demonstrated that technology will be
available to sort auto shred material when the market
need arises. Accordingly, there should be no concern
that the growing quantities of aluminum going into
the automotive cannot be recycled in the future.
The large market for recycling wrought
aluminum from automotive is probably still at least
10 years away, since the current vehicles with high
aluminum content will be used by the consumer for
the next decade. Exactly how the technology will be
deployed in the future will depend on many factors,
such as the specific alloys used, the cost of process
sensors, the amounts of aluminum used in future
vehicles, and the overall economics of the sorting
process. However, there is little doubt, based on the
demonstrations conducted to date on this project,
that scrap sorting technology will be available when
it is needed.
Additional Reading
A. Gesing et al., “Separation of Wrought
Fraction of Aluminum Recovered from Automobile
Shredder Scrap,” Aluminum 2001, Das, Kaufman,
and Lienert, eds., TMS and the Aluminum
Association, 2001, pp. 31–42.
Aluminum Industry Roadmap for the Automotive
Market, The Aluminum Association, Washington,
D.C., May 1999.
P. B. Schultz and R. K. Wyss, “Color Sorting of
Aluminum Alloys for Recycling,” Plating and
Surface Finishing, 87(4) (April 2000), 10–12 and
87(6) (June 2000), 62–65.
U. S. Patent 6,100,487, August 8, 2000.
“Automotive Aluminum Scrap Sorting,” DOE
Office of Industrial Technologies, available at
www.oit.doe.gov/aluminum.
Automotive Engineering, July 2001.
American Environmental Review, to be aired in
the fall of 2001.
Summary and Outlook
Test materials from shredded automobiles have
been assembled and prepared for sorting tests.
Etching procedures have been developed to provide
a color tint to a mix of shredded aluminum auto
material, and techniques have been developed to
enable this material to be sorted by color. Additional
work will be needed to refine and verify the process
economics.
The separation of cast from wrought shredded
auto material has been demonstrated in two largescale tests in the United States and Europe. This
process has already been industrially implemented in
the HVSC plants, opening new markets for the
wrought aluminum shred from auto shredders.
The sorting of alloys using the LIBS process has
also been shown to be technically feasible and can
clearly separate material on an alloy-by-alloy basis.
142
Automotive Lightweighting Materials
FY 2001 Progress Report
B. Recycling of Polymer Matrix Composites
Principal Investigator: Bassam J. Jody
Argonne National Laboratory
9700 South Cass Avenue, Argonne, IL 60439
(630) 252-4206; fax: (630) 252-1342; e-mail: [email protected]
Program Manager: Edward Daniels
Argonne National Laboratory
9700 South Cass Avenue, Argonne, IL 60439
(630) 252-4206; fax: (630) 252-1342; e-mail: [email protected]
DOE Program Manager: Joseph A. Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: Argonne National Laboratory
Contract No.: W-31-109-Eng-38
Objective
•
Develop efficient and cost-effective processes for recovering carbon fibers from polymer matrix composites
(PMCs).
OAAT R&D Plan: Task 9; Barrier E
Approach
•
Conduct process improvement studies to determine the optimum operating conditions.
•
Determine disposal and treatment requirements of the effluent stream.
•
Develop a conceptual design for a full-scale plant.
Accomplishments
•
Evaluated three processes to separate carbon fibers from PMC scrap: thermal treatment, chemical treatment,
and thermal shock. The thermal treatment method, a single-step process, recovered carbon fibers with
properties equivalent to those of their virgin counterparts. Economic analysis of this method also showed it to
be potentially economical. Therefore, its development was continued in FY 2001.
Future Direction
•
Work with a carbon fiber manufacturer to produce and test PMC panels made with recovered fibers and
compare the properties of these panels with those of similar panels made with the same type of virgin fibers.
•
Produce about 200 lb of recovered carbon fibers for use in an automotive application.
•
Investigate the technical feasibility and advantages of using a hybrid treatment process (a thermal process
followed by a chemical process, and vice versa).
143
FY 2001 Progress Report
Automotive Lightweighting Materials
Process Improvement Study of the Thermal
Treatment Method
An experimental study was conducted to
determine efficient and economical process
operating conditions for recovering fibers from both
urethane and epoxy substrates. Samples as large as
4 × 12 in were treated at temperatures of up to
1750°F in air or in an inert environment. The pilotscale and the bench-scale experiments produced
consistent results.
Test Results Obtained From Treating PMC
Samples Made With Urethane-Based
Substrates
Figure 2. Carbon fibers recovered at 1250° F in air for
5 minutes from a polymer matrix composite
sample with a urethane substrate.
At temperatures greater than 500°F, the carbon
fibers were completely freed from the polymer,
given enough residence time. At 1250°F, the
required residence time was about 5 minutes. About
45% of the weight of the samples was the polymer
substrate that was pyrolized when the PMC samples
were treated in an inert environment at different
temperatures for prolonged periods of time.
Appreciable oxidation of the fibers did not start until
after 8 minutes. To determine the degree of
oxidation, we repeated the experiments at 1250°F in
a nitrogen environment. The results are shown in
Figure 1. The recovered carbon fibers were also
examined under a microscope. Figure 2 is a picture
of recovered fibers at 1250°F in air for 5 minutes.
This picture shows that the fibers are free of
contamination. The individual fibers are also free
from each other, and their surfaces are smooth and
have the same diameter. These results indicate that
no oxidation has occurred at the surfaces of these
fibers.
Test Results Obtained From Treating PMC
Samples Made With Epoxy-Based Substrates
Two types of PMC materials made with epoxy
substrates were tested. The first was supplied by
Oak Ridge National Laboratory (ORNL) and was
about the same thickness (~ 1/16 in.) as the samples
made with urethane substrate; however, it was
cylindrical in shape. The other type was in the form
of plates of about 1/16 in. in thickness, and was
supplied by Hexcel Corporation. The Hexcel
samples had well-specified properties and materials
of construction. The ORNL samples were more
difficult to treat than the samples made with the
urethane substrate, even though the substrate
constituted only about 25% of these pieces by
weight. Further, unlike the urethane-based PMC
samples that produced loose fibers when treated, the
ORNL samples remained somewhat intact and
retained some level of rigidity and stiffness after
treatment. However, when a mechanical force was
applied to the treated samples, such as repeated
manual bending, the fibers became soft and loose;
and when they were aerated using compressed air or
placed in water with agitation, most of the residual
solid flakes fell off. These samples also suffered
some oxidation when treated in air, as shown in
Figure 3.
These data show that the treatment process was
completed in about 10 minutes. When the treatment
was carried out in air, about 8% of the fibers were
lost as a result of oxidation. The reason for the
higher level of oxidation is not clear at this time.
Thermal Treatment of Urethane Based PMC in Air and in Nitrogen
54
Weight Loss %
52
In Air @ 1250 Deg F
50
48
46
In Nitrogen @ 1250 Deg F
44
42
0
5
10
15
Treatm ent Time in Minutes
20
25
Figure 1. Weight loss of polymer matrix composite
samples with urethane-based substrate
treated in air or in nitrogen at 1250°F.
144
Automotive Lightweighting Materials
FY 2001 Progress Report
for about 10–15 minutes will enable the operator to
recover carbon fibers from the PMC scrap with
minimal loss of fibers and without having to use
nitrogen. Changes in these operating conditions may
be necessary to accommodate different types of
scrap.
P M C exp erim en tal D a ta, E po xy S u b strate a t 1 250 D eg F
50
In A ir
PMC Wt. Loss, %
40
30
In A rg o n G as
20
10
Determination of the Disposal and Treatment
Requirements of the Effluent Stream
0
0
10
20
30
40
50
T rea tm ent T im e, M inutes
The PMC waste stream will contain scrap made
with various types of thermoset and thermoplastic
substrates. The thermal treatment method will be
capable of treating all of these types. However,
different substrates will generate different pyrolysis
and oxidation products, including volatile and semivolatile organic compounds. Based on the heating
values of polymers, we estimate that at least 60% of
the energy requirements of the process could be
satisfied by the energy value of this stream. In most
cases, the PMC panels are made with about 50%
polymeric substrate that has a heating value of over
18,000 Btu/lb. The energy released upon combustion
of the polymeric substrate is enough to maintain the
treatment reactor at the desired temperature;
therefore, the system can be self-sufficient in
energy. An afterburner is required to destroy
residual carbon monoxide and organic compounds.
The process does not generate liquid or solid waste
streams.
Figure 3. Weight loss of the ORNL polymer matrix
composite samples with epoxy substrate
treated in air or in argon gas at 1250°F.
The Hexcel PMC samples containing the epoxy
substrate were more readily treatable than the ORNL
samples. Samples of various dimensions were
treated at temperatures in the range of 850 to 1500°F
in air and in a nitrogen environment. These samples
were made with about 30% epoxy substrate by
weight and about 70% fiber. The results of the
treatment are presented in Figure 4. Significant
oxidation of the fibers did take place after prolonged
treatment in air. The treatment was essentially
completed in 5 minutes when conducted in air at
1250°F, and in about 10 minutes when conducted in
nitrogen at the same temperature. We plan to
process about 40 test panels from Hexcel in
FY 2002. The fibers recovered from these panels
will be characterized to determine the physical
properties of recovered fibers relative to those of
“virgin” fibers. The recovered fibers will also be
used to produce 20 “recycled” test panels to
determine the physical properties of the recycled
panels vis-à-vis the properties of panels made of
virgin fibers. The results of the process improvement
study suggest treating PMC scrap at 1250°F in air
60
Development of a Conceptual Design for a
Full-scale Plant
A conceptual design for the process was
developed, including subsystems for loading the
PMC scrap and for collecting the recovered fibers.
We used this conceptual design and the data from
the process improvement study to update the
economic analysis for a full-scale plant processing
1,000,000 lb/year of PMC scrap. The following
assumptions were also made:
Treatment in air at 1250 Deg F
Weight Loss, %
50
40
•
Treatment in Nitrogen at 1500 Deg F
30
Treatment in Nitrogen at 1250 Deg F
20
•
10
0
0
5
10
15
20
25
30
35
Treatment Time, Minutes
•
Figure 4. Weight loss of the Hexcel epoxy sample.
145
The recoverable carbon fiber in the PMC scrap
is 40% of the scrap weight
The scrap contains 50% by weight polymeric
substrate that has a heating value of
18,000 Btu/lb.
The recovered fibers are valued at $1.50 per
pound, based on a value of $3/pound for the
virgin fibers
FY 2001 Progress Report
Automotive Lightweighting Materials
We also estimated the cost of the system to be
$1,000,000. The results, summarized in Table 1,
suggest a payback of less than 2 years.
Milestones
Table 1. Process economics (annual basis) for
carbon fiber recovery from obsolete PMC
materials
Revenues, $1000/year
Recovered carbon fibers @ $1.5/lb
600
Other
0
Total revenues
600
•
Operating costs, $1000/year
Feedstock (credit for avoided disposal)
Waste disposal
Utilities
Labor
Maintenance
Total operating costs
–10.0
0.0
25
50
20
85
Net income, $1000/year
Capital cost, $1000
515
1,000
The following milestones of the proposed effort
are to be conducted in FY 2002.
•
•
Complete testing of the PMC panels made with
recovered carbon fibers.
Complete production of 200 lb of recovered
fibers for testing in an actual application.
Assess the technical feasibility and advantages
of using a hybrid treatment process (thermal
followed by chemical and chemical followed by
thermal).
Conclusions
Work conducted so far has demonstrated that the
recovery of carbon fibers from PMC scrap is
technically feasible and potentially economical.
Establishment of the properties of the recovered
fibers and of products made with recovered fibers is
essential in order to identify applications for the
recovered fibers.
146
Automotive Lightweighting Materials
FY 2001 Progress Report
C. Investigation of the Cost and Properties of Recycled Magnesium Die Casting
for Employment in Automotive Applications
Co-Principal Investigators: John F. Wallace and David Schwam
Case Western Reserve University, Department of Materials Science and Engineering
10900 Euclid Avenue, Cleveland, OH 44106-7204
(216) 368-4222; fax: (216) 368-3209; e-mail: [email protected] and dxs [email protected]
PNNL Contract Manager: Russell H. Jones
(509) 376-4276; fax: (509) 376-0418; e-mail: [email protected]
DOE Program Manager: Joseph A. Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Participants
Qing Ming Chang, Ph.D., Case Western Reserve University
Xiaofeng Su, Ph.D., Case Western Reserve University
Yulong Zhu, Ph.D., Case Western Reserve University
Contractor: Pacific Northwest National Laboratory
Contract No.: DE-AC06-76RL01830
Objective
•
Investigate the recycling of magnesium alloys—AZ91D, AM60, and AM50—from previously cast Class 1 and
Class 2 die-cast material.
OAAT R&D Plan: Task 9; Barriers A, E
Approach
•
Determine the quality of the cast metal from individual melts cast by permanent molding and squeeze die
casting in two recycling shops. The effects of various flux additions and of the refining process used were
determined by the tensile strength and ductility.
•
When the quality of the metal produced in the permanent molds is not adequate, repeat the process using other
fluxes and improved processes.
•
Using this procedure, obtain desirable properties for all three alloys. (Some difficulty was experienced with the
AZ91D composition.)
Accomplishments
•
During the refining process, planned and conducted tests on AZ91D, AM60, and AM50 alloys in the Case
Western Reserve University (CWRU) foundry and at the cooperating business recycling shops at Garfield
Alloys and MagReTech. The overall purpose of the tests was to develop a technical procedure that will produce
good die-casting melt quality in the three alloys from Class 1 and Class 2 scrap.
147
FY 2001 Progress Report
Automotive Lightweighting Materials
Introduction
Table 1.
Effect of alloying elements on magnesium
alloys
Manganese
Does not affect mechanical properties
Promotes corrosion resistance
Magnesium alloys have many advantages for
specific structural applications. Until recently, most
magnesium castings were used in specialized
applications, such as aircraft or hand tools, where
lightness is important to minimize weight, improve
handling, or reduce inertia. Cast products offer
advantages over some wrought products because
work-hardening of the HCP lattice structure in the
latter can cause difficulties. However, the generally
higher premium cost and variable price of
magnesium vis-à-vis aluminum has interfered with
the commercial growth of magnesium alloys. Many
of these economic factors either have been corrected
or are being muted by current developments. The
die-casting process has been an excellent method of
producing magnesium castings because of the rapid
production of parts with close-to-final dimensions.
These improvements have resulted in the
growing use of magnesium casting for more general
usage, particularly in automotive applications
because of their light weight. The present technical
literature1–7 continues to state the advantages of
magnesium alloys, including its increased
availability, improved machineability, and light
weight. These articles do highlight some problems,
such as the need for careful handling or particularly,
with the requested properties, for some purposes.
One of the problems in improving the development
of magnesium alloys has been the reclamation of the
scrap produced.1–7 When the scrap is in the form of
Class I material, such as gates and rejected castings,
it can frequently be processed through normal
procedures recommended for magnesium. However,
lighter Class 2 scrap, such as turnings and thinner
materials, requires considerably improved handling.
The development of improved sludging,
evaporation, and filtering technologies will be of
assistance in processing this Class 2 material so that
it can be used to produce magnesium alloy of the
proper quality.
In addition to these issues, the presence of some
elements (Table 1) has to be controlled. They exert
an effect on the behavior of magnesium alloys with
which they are in close contact.
Silicon
Present in small amounts in most alloys
Improves creep strength by producing Mg2Si phase
Beryllium
Sometimes added at low levels <10 ppm to minimize
oxidation of the molten alloy
Does not affect mechanical properties
Impurities—Iron, Nickel, and Copper
Controlled at ppm levels
Negligible effect on mechanical properties
When in excess of spec cause large loss in corrosion
resistance
arrangement was made with Garfield Alloys and
MagReTech to work with CWRU in preparing and
testing magnesium alloys that are preferable to those
produced with currently used procedures. The
activities have included casting and testing of the
alloys produced with various processing techniques
at these two shops and at CWRU. This work has
included tests on the tensile properties of bars cast
from these alloys. Under the arrangement, CWRU
personnel visit both Garfield Alloys and
MagReTech, and the supervisory personnel of these
plants visit CWRU. Both organizations have
benefited from the arrangement. CWRU has run
several small-scale tests on reclaiming material and
has been aided by the comments of the two
companies. The supervisory personnel from the
plants have made several recommendations for
controlling the processing equipment employed at
CWRU. CWRU has analyzed many of the fluxes
used by these companies and has made specific
recommendations.
In addition, the CWRU Foundry Group has
maintained a close relationship with a local diecasting shop that is using reclaimed scrap from
Garfield Alloys and MagReTech to produce
magnesium castings on a production basis. This diecasting shop, Magnesium Aluminum Corporation,
has been providing magnesium alloys castings for
40 years on a commercial basis. Personnel from
CWRU have visited the die-casting shop and shop
personnel have visited the operation at CWRU.
Background of the Project
The casting group at CWRU has been
proceeding in several directions to satisfy the
requirements in the statement of work. First, an
148
Automotive Lightweighting Materials
FY 2001 Progress Report
Comments from supervisors at the shop have been
used to improve CWRU’s process.
Investigation Processes
The first project task was to develop sludging,
evaporation, and filtering processes to produce
magnesium alloys, using Class 1 scrap from a
commercial magnesium recyclers, with the same
properties as virgin magnesium die-cast AZ91D,
AM50, and AM60. As part of this work, Garfield
Alloys and MagReTech cooperated in a series of
experiments involving Class 1 scrap. The scrap
material was melted in 2300-lb steel crucibles at
Garfield Alloys and 6500-lb steel crucibles at
MagReTech. The Class 1 scrap was melted in the
furnace with a control cover of flux, and a standard
fluxing and sludging treatment were used throughout
the working cycle. Tests were conducted and the
properties of the different alloys measured when the
fluxing started after the Class 1 alloy was charged in
the furnace. A cover and drying flux were used as
the test proceeded. The compositions of these fluxes
were changed somewhat to improve the quality of
the alloys as the treatments proceeded. The fluxes
that were employed followed the general pattern
shown in Table 2.8
The processing of AM60 and AM50 proceeded
without problems. However, adjustments in the
analysis of the treatment slags were found to be
necessary to obtain good properties with AZ91D.
necessary to obtain good properties with AZ91D.
This alloy is very susceptible to difficulties and
receives special treatments when fluxing and
sludging are employed.
The results for AM50, AM60, and AZ91D show
that AZ91D is a more difficult alloy to process than
the others because of the presence of oxide
inclusions that form readily in it. Treating for long
periods of time with the fluxes produces alloys that
are generally clean, without inclusions. Another
problem that occurs is black spots on the fracture
surface of the tensile bars. The composition of these
fluxes and the analysis of these black spots are
shown in Figure 1. A brightmeter has been obtained
Table 2. Fluxes for magnesium-alloy melting
Flux No.
KCl
MgCl2
MgO
230
55.0
34.0
….
232
37.5
42.0
7.5
310
20.0
50.0
15.0
CaF2
2.0
8.5
15.0
Figure 1. Black spots on the fracture surface of PM
AZ91D.
from Daimler Chrysler for use in testing and in
analyzing the results of these tensile tests.
This initial work was conducted with Class 1
scrap at each plant. Subsequent work at Garfield
Alloys was conducted on AM60 alloy using Class 2
scrap to attain the required results. The work with
the Class 2 scrap is considerably more difficult and
time-consuming than the work with the Class 1
material. After some of the difficulties with handling
Class 2 scrap were determined, it was decided to
limit the amount of work being conducted at the two
cooperating shops and conduct the work at CWRU.
The loss of time and production interferes with
production schedules at the commercial shops.
Table 3 shows the properties of the permanent
mold test pieces based on a permanent mold test bar
shown in Figure 2. This mold was used to cast test
bars from the alloy product during the refining
operation that employed the standard fluxing and
sludging at Garfield Alloys and MagReTech. To
obtain good properties, it was necessary to change
the fluxes to those with higher fluoride content and
to slow the operation down. The specialized work to
obtain the AZ91D alloy properties using this special
fluxing and sludging procedures eliminated the
inclusions.
BaCl2
9.0
4.5
….
149
MnCl2
….
….
….
Use
Remelting pots, fluid slag
Remelting pots, fluid slag
Crucible melting, dries during use
FY 2001 Progress Report
Table 3.
Automotive Lightweighting Materials
Mechanical properties of permanent-mold
and squeeze-cast magnesium alloys
Mechanical properties
magnesium alloys
UTS
(ksi)
Elongation
(%)
AM50—Permanent mold
(as cast)
27.7
10.8
AM60—Permanent mold
(as cast)
29.7
11.2
AM60—Squeeze case
(as cast)
31
7.25
AZ91D—Permanent mold
(as cast)
23.4
3
Figure 3. Squeeze-cast mold for test bar.
the structures for these bars show a dendrite arm
spacing (DAS) of about 25 microns in the permanent
mold bar and only 15 to 16 microns in the squeezecast bar. Previous work9 has indicated that the closer
spacing of the DAS squeeze-cast bar should
normally provide better properties in the absence of
defects. However, attaining the appropriate
conditions will require improvements in casting
procedures.
The squeeze-cast bars were produced in the
mold shown in Figure 3, where the general pattern
of feeding from both ends is present. However, the
colder die and rapid solidification did not allow the
test bar to attain the appropriate structure along its
length. It is apparent from these results that it is
necessary to pour the casting into a mold faster at a
higher temperature. The results we have obtained to
date show tensile strengths that approach and even
exceed those of the bars made in permanent molds
in some cases, as shown previously. Because of the
finer DAS of the squeeze-cast bars compared with
the permanent mold bars, a higher tensile strength
was expected. However, the facilities we have
available cannot attain the shot speed or the higher
die temperature on a regular basis. Although these
results will eventually be obtained, it will require
more equipment than is available within the limits of
this agreement. For this reason, the metallographic
problems with the squeeze-cast test parts require that
the permanent mold be widely used. The results
from the permanent mold tests are generally used in
the work on reclaiming magnesium alloys made
from Class 1 and Class 2 types of scrap.
Figure 2. Permanent mold for tensile test bar.
The results for the refined alloys are reported for
the test casting in permanent mold test bars. These
molds produced a constant quality of casting when
the material being cast was properly treated.
Throughout the work being processed, the
reclaimers employed for processing were excellent
at both locations. However, when the squeeze-cast
mold shown in Figure 3 was employed at CWRU,
the strength and ductility of the test bars were
somewhat lower. It became apparent that the bars
did not have a consistent cross-section without the
presence of inclusions or segregations. To improve
the properties of the squeeze-cast bar test, a faster
pour in a hotter mold was needed. A comparison of
150
Automotive Lightweighting Materials
FY 2001 Progress Report
Our experience has indicated that the difficulties
with the Class 2 alloy involve the presence of
inclusions and small particles of scrap material still
remaining in the melt. These types of inclusions
have reduced the properties significantly. Methods
of eliminating these inclusions, so that they do not
lower the tensile properties or appear as defects in
the brightness examination, are to be developed
using our large melting furnace and tapping the alloy
into the ingot molds after further melting and
processing. This equipment is shown in Figure 4.
Figure 5. Ultimate tensile strength and elongation of
fluxless recycled AZ91 at CWRU 032001.
techniques. The tensile test results obtained have
required considerable periods of time for these
operations. The results have included the large
3600-lb and 5400-lb steel melting furnaces at
Garfield Alloys and MagReTech.
Excellent results have also been obtained with
smaller heats processed at CWRU using the same
three alloys. These results are reported based on the
tensile properties and fracture surfaces of the test
bars produced as permanent test bar castings. Both
Class 1 and Class 2 scrap have been tested at
CWRU.
The results obtained with squeeze-cast test bars,
however, have not been consistently up to the
standards attained in the permanent mold test bars.
The difficulty has been that the squeeze-cast bars
have contained areas of eutectic alloy. These areas
have formed during the solidification of the test bars
and have not exhibited the strength expected in this
process.
The overall results of the work indicate that the
reclamation at Garfield Alloys, MagReTech, and
CWRU produced quality products. The primary
work with the magnesium processor was for Class 1
scrap with alloys of AM50, AM60, and AZ91D. The
work conducted with CWRU included both Class 1
and Class 2 alloys. The processing at Garfield
Alloys and MagReTech used the conventional
fluxing procedure. The CWRU process included
treatment of the AZ91D alloy with a non-fluxing
method that bubbled argon through the melt and
removed the impurities by skimming.
Figure 4. Experimental set-up for recycling magnesium
alloys.
The equipment uses a gaseous, non-salt cleaning
procedure in which inclusions in the melt are
removed by skimming with argon gas through the
melt. The skimming causes the inclusions to rise to
the surface so that impurities can be removed via a
non-fluxing skimming method before casting. This
equipment has been used experimentally for nonfluxing procedures to produce the good-quality test
runs shown in Figure 5. This method has shown its
efficiency in producing the better properties of the
clean alloy.
Conclusions
At present, the investigation that we have
undertaken has shown that excellent tensile
properties and quality of metal can be obtained for
AM60, AM50, and AZ91D with proper reclamation.
These investigations have included both Class 1 and
Class 2 scrap from other die casting operations.
Attaining these properties required the use of
fluxing, sludging, and skimming removal
151
FY 2001 Progress Report
Automotive Lightweighting Materials
7. J. C. Grebetz and A. G. Haerle, “Measuring
the Cleanliness of Recycled Magnesium for
Automotive Applications: A Detailed Examination
of the Fracture Brightness Technique,” pp. 60–69 in
Light Metal Age, Fellom Publishing Co., San
Francisco, August 1997.
8. R. W. Heine, C. R. Loper, Jr., P. C.
Rosenthal, Principles of Metal Casting, American
Foundrymen's Society, McGraw-Hill Book Co.,
New York, 1967.
9. H. Sasaki, M. Adachi, T. Sakamoto, and
A. Takimoto, Effect of Solidified Structure on
Mechanical Properties of AZ91 Alloy, pp. 86–92.
10. M. W. Dierks, C. Kuhn, and J. Etling,
Enhanced Mechanical Properties of Die Cast AM
Series Magnesium: Through Part Design, Die
Design and Process Control, pp. 311–38.
11. S. C. Erickson, J. F. King, and T. Mellerud,
“Conserving SF6 in Magnesium Melting Operations:
A Summary of Best Practices in the Industry for
Using SF6 as a Protective Atmosphere and Ideas for
Reducing Consumption and Emissions,” Foundry
Management & Technology, pp. 38–49, June 1998.
References
1. P. M. Pinfold, D. L. Albright, D. O. Karlsen,
“Refined Recycled High-Ductility Magnesium
Alloys,” Proceedings of the 1995 NADCA Congress,
Paper T95-051.
2. T. K. Aane, P. Bakke, H. Westengen,
T. Melherud, D. L. Albright, “Potential for
Improving Recycling Friendliness of High-Priority
Magnesium Die Casting Alloys,” Proceedings of
1997 NADCA Congress, Paper T95-051.
3. J. Grebetz, T. Day, D. Haerle, “Qualification
of Recycled Magnesium Class 1 Alloys,” presented
at the National Magnesium Association, 1998.
4. J. C. Grebetz and D. L. Albright, “Recycled
Magnesium Alloys for High-Ductility Automotive
Applications,” SAE Paper 960413, pp. 1–7, Society
of Automotive Engineers, Detroit, Michigan, 1996.
5. A. G. Haerle, B. A. Mikucki, and W. E.
Mercer II, “A New Technique for Quantifying NonMetallic Inclusion Content in Magnesium,” pp. 22–
29 in Light Metal Age, Fellom Publishing Co., San
Francisco, August 1996.
6. A. G. Haerle, R. W. Murray, W. E. Mercer II,
B. A. Mikucki, and M. H. Miller, “The Effect of
Non-Metallic Inclusions on the Properties of Die
Cast Magnesium,” SAE Paper 9970331, pp. 74–84,
Society of Automotive Engineers, Detroit,
Michigan, 1997.
152
Automotive Lightweighting Materials
FY 2001 Progress Report
D. Recycling Assessments and Planning
Principal Investigator and Program Manager: Edward J. Daniels
Argonne National Laboratory
9700 S. Cass Ave., Argonne, IL 60439
(630) 252-5279; fax: (630) 252-1342; e-mail: [email protected]
Project Manager, Composites: C. David Warren
Oak Ridge National Laboratory
P.O. Box 2009, Oak Ridge, TN 37831-8050
(865) 574-9693; fax: (865) 574-0740; e-mail: [email protected]
DOE Program Manager: Joseph A. Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: Argonne National Laboratory
Contract No.: W-31-109-Eng-38
Objectives
•
Establish priorities for cost-effective recycling of advanced automotive technology (AAT) materials and
components associated with PNGV and beyond.
•
Focus research on near-term recycling on the technologies critical to achieving 85% recyclability when the first
wave of the PNGV fleet reaches end-of-life status.
•
Focus long-term efforts on enhanced recovery/sorting procedures and advanced recycling technologies that are
enabling factors in a vision for 100% recyclability.
OAAT R&D Plan: Task 9; Barrier E
Approach
•
Consult with automotive manufacturers and recycling industries, the U.S. Council on Automotive Research
(USCAR) and its affiliates, national laboratories, universities, and other relevant organizations to assess critical
recycling needs/barriers.
•
Develop a recycling research and development (R&D) program plan that will serve as a working document to
guide DOE/OAAT in establishing priority goals, with an initial emphasis on lightweight body and chassis
materials.
•
Assist DOE in establishing advanced recycling R&D initiatives and provide technical oversight to ensure that
priority objectives/goals are accomplished.
Accomplishments
•
Organized the AAT Recycle Roadmapping Workshop during the fourth quarter of FY 2000. The workshop was
facilitated through a subcontract with Energetics. The roadmap document, A Roadmap for Recycling End-ofLife Vehicles of the Future, was completed in May 2001 and will now provide the basis for structuring specific
projects consistent with the needs of the key stakeholders.
153
FY 2001 Progress Report
Automotive Lightweighting Materials
•
Identified related federally funded research projects, one funded by the U.S. Department of Commerce and
three funded by the National Science Foundation during 1997 through 2000. The principal investigators of
those projects have been contacted and reports from their projects have been requested.
•
Continued to work with the Automotive Aluminum Alliance on its project to develop technology to sort
aluminum alloys by class using color sorting and/or laser-induced breakdown spectroscopy. Both technologies
are being tested on a large scale at Huron Valley Steel Corp.
Future Direction
•
Structure a suite of prioritized projects over the next year in collaboration with the stakeholders.
Continue work on development of the AAT Recycling Program R&D Plan through collaboration with the
USCAR Vehicle Recycling Partnership and associated USCAR affiliates, the Advanced Battery Consortia,
the Automotive Parts Rebuilders Association, remanufacturing industries, materials trade organizations,
and other appropriate industries involved in automotive recycling.
Continue evaluations of recycling technology progress in Europe and Japan, along with assessments of
critical recycling technologies as new needs are identified.
Continue to work with the Automotive Aluminum Alliance, which is working with Huron Valley Steel
Corporation to demonstrate technology developed by Huron Valley Steel for separation of aluminum
alloys.
•
Continue project efforts to assist DOE in preparation of AAT planning documents, priority recycling R&D
needs, proposal reviews, and related tasks throughout FY 2002.
Recycling Partnership and other organizations
(including the Aluminum Association, American
Plastics Council, Institute of Scrap Recycling
Industries, Automotive Recyclers Association,
Automotive Parts Rebuilders Association, and the
federal government) have been working both
collaboratively and independently to address
technical, institutional, and economic issues that
limit the recycling of ELVs. Progress has been made
toward understanding some of these issues, and
technology has been developed that can impact the
level of ELV recycling.
The recyclability of ELVs is currently limited by
a lack of commercially proven technical capabilities
to cost-effectively separate, identify, and sort
materials and components, and by a lack of
profitable post-use markets. While nearly 75% by
weight of ELVs currently are recycled in some form,
the remaining 25% is sent to landfills each year.
Over the next 20 years, both the number and
complexity of ELVs are expected to increase, posing
significant challenges to the existing recycling
infrastructure. The automobile of the future will use
significantly greater amounts of lightweight
materials (e.g., ultralight steels, aluminum, plastics,
composites) and more sophisticated/complex
components.
Summary
The recycling program objective is to establish
priorities and develop cost-effective recycling technologies and strategies in support of long-term DOE
OAAT objectives and goals. Automobile recycling
is the final productive use of end-of-life vehicles
(ELVs). The obsolete car has been a valuable source
of recycled raw materials and useable parts for
repair as long as cars have been mass-produced.
Today, cars that reach the end of their useful service
life in the United States are profitably processed for
materials and parts recovery by an existing recycling
infrastructure. That infrastructure includes
automotive dismantlers who recover useable parts
for repair and reuse, automotive remanufacturers
who remanufacture a full range of components—
including starters, alternators, and engines—to
replace defective parts, and ultimately the scrap
processor who recovers raw materials such as iron,
steel, aluminum, and copper from the remaining
auto “hulk” after components have been recovered
for recycling. Each of these activities contributes to
the recycling of obsolete vehicles.
Today, less than 25% by weight of obsolete cars
is not profitably recoverable for recycling and is
therefore landfilled. Over the past 10 years, the
original equipment manufacturers—Ford, General
Motors, and DaimlerChrysler—through the Vehicle
154
Automotive Lightweighting Materials
FY 2001 Progress Report
•
New technology is and will continue to be
needed to improve vehicle recyclability. Using the
Roadmap as a framework, we will work with the key
stakeholders over the next year to structure a suite of
projects to be undertaken in this program area.
The following strategy was developed as part of
the Roadmap to maximize the value recovered from
ELVs.
•
•
•
•
•
•
•
Come together as a unified recycling community
to share the costs of the development of required
new technologies.
Incorporate reuse, remanufacturing, and
recycling into the design phase for cars
whenever possible.
Recycle as early in the recycling stream as
possible while relying on the market to optimize
the value and amount recycled at each step.
Maintain a flexible recycling process that can
adapt to diverse model lines fabricated with
different techniques and materials from various
suppliers.
•
•
•
155
Develop automated ways to recover bulk
materials.
Emphasize R&D on post-shredding material
identification, sorting, and product recovery
Focus R&D efforts on materials not currently
recycled by sorters (e.g., post-shredding glass,
rubber, fluids, textiles, plastics).
Develop uses for recovered materials (whether
in the same or different applications) and testing
specifications.
Encourage investment in the infrastructure
needed to achieve the recyclability goal. Build
on the existing infrastructure.
Develop a means to prevent the entry of PCBs
and other hazardous materials into the recycling
stream and promote acceptable limits in
shredder residues.
Consider the recycling requirements of new
technologies entering fleets as early as possible.
Automotive Lightweighting Materials
FY 2001 Progress Report
7. ENABLING TECHNOLOGIES
A. Durability of Carbon-Fiber Composites
James M. Corum (Principal Investigator), R. L. Battiste, M. B. Ruggles, and Y. J. Weitsman
Oak Ridge National Laboratory, P.O. Box 2009
Oak Ridge, TN 37831-8051
(865) 574-0718; fax: (865) 574-0740; e-mail: [email protected]
Automotive Composites Consortium Contact: Edward M. Hagerman
General Motors R&D, 30500 Mound Road 1-6
Warren, MI 48090-9055
(810) 986-1208; fax: (810) 986-1207; e-mail: [email protected]
Project Manager, Composites: C. David Warren
Oak Ridge National Laboratory, P.O. Box 2009
Oak Ridge, TN 37831-8050
(865) 574-9693; fax: (865) 574-0740; e-mail: [email protected]
DOE Program Manager: Joseph A. Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: Oak Ridge National Laboratory
Contract No.: DE-AC05-00OR22725
Objective
•
Develop experimentally based, durability-driven design guidelines to ensure the long-term (15-year) integrity of
representative carbon-fiber-based composite systems that can be used to produce large structural automotive
components. Durability issues being considered include the potentially degrading effects of cyclic and sustained
loadings; exposure to automotive fluids; temperature extremes; and the low-energy impacts of such events as
tool drops and kickups of roadway debris on structural strength, stiffness, and dimensional stability.
OAAT R&D Plan: Task 4; Barriers A, C
Approach
•
Characterize and model the durability behavior of a progression of three representative carbon-fiber
composites; each has the same thermoset urethane matrix but with a different reinforcement type and
configuration: (1) reference crossply, (2) quasi-isotropic, and, (3) random chopped fiber.
•
Replicate on-road conditions in laboratory tests of each composite to generate durability data and models.
•
Develop and publish durability-based design criteria for each composite.
157
FY 2001 Progress Report
Automotive Lightweighting Materials
Accomplishments
•
Published a report on durability-based design criteria for the reference crossply carbon-fiber composite
(ORNL/TM-2000/322).
•
Completed the required experimental characterization of the durability of quasi-isotropic carbon-fiber
composite.
•
Developed a suitable chopped-carbon-fiber composite system that met Focal Project 3 design requirements and
that will be the basis for durability characterization and modeling [Automotive Composites Consortium
(ACC)].
Future Direction
•
Develop durability-based design criteria for quasi-isotropic carbon-fiber composite and publish a report.
•
Publish a report on the development of a mechanical model for predicting the long-term, time-dependent,
nonlinear response of directed-carbon-fiber composite from short-time tests.
•
Complete durability characterization and modeling of Focal Project 3 chopped-carbon-fiber composite and
publish a design criteria report.
•
Address the durability of thermoplastic carbon-fiber composites.
Zoltek. Further, the high productivity requirements
(100,000 units per year) have similarly directed
work toward liquid molding and the programmable
powdered performing process (P4) chopped-fiber
technology demonstrated in the ACC Focal Project 2
truck box program. The material requirements
resulting from these developments are for a
chopped-carbon-fiber–reinforced composite
containing 40 vol % fiber and capable of being
molded at a thickness of 1.5 mm. In chopped-fiber
composites, both the 40 vol % fiber content and the
1.5-mm thickness represent substantial technology
challenges.
Several chopping and molding runs were carried
out using high-tow fibers from Hexcel, Toray, and
Zoltek. Also, several sizings were used. Sizings are
materials introduced by the fiber manufacturers to
hold and stabilize fiber tows (bundles) during
handling and shipping. “Hard” sizes keep the bundles tightly together during molding; “soft” sizes
allow the bundles to break down into individual
fibers or smaller fiber bundles. Soft sizes lead to
better properties, but they result in more difficult
preforming and higher molding pressures.
Ultimately, Zoltek Panex 33, seven-split 46K
carbon fiber with an X8 size package, was found to
perform the best, both in terms of ease of
performance using the ACC P4 machine and in
terms of the resulting mechanical properties. Using a
Material Development
The three molded thermoset composites being
addressed are differentiated by their carbon-fiber
preforms, which are depicted in Figure 1. All three
have the same Bayer 420 IMR urethane resin matrix.
The required experimental characterization of the
quasi-isotropic composite is complete. The choppedfiber composite will next be addressed.
Figure 1. Preforms used in progression of carbon-fiber
thermoset composites being characterized and
modeled.
In developing a viable chopped-carbon-fiber
composite, the ACC Materials group has focused its
work on the requirements being developed for Focal
Project 3. A 60% mass reduction and cost parity
with steel have directed work toward the low-cost,
high-tow carbon products from suppliers such as
158
Automotive Lightweighting Materials
FY 2001 Progress Report
1500-ton press, the 1.5-mm-thick plaques have been
produced with fiber volumes ranging from 42 to
46%. An ultimate tensile strength (UTS) of 250
MPa, tensile modulus of 45 GPa, and failure strain
of 0.6% easily satisfy Focal Project 3 design
requirements.
This composite has consequently been identified
as the main stream material for the ACC Focal
Project 3 low-mass body-in-white structure. The
next step is to produce a quantity of plaques of the
chosen material for complete ACC characterization,
for durability characterization and modeling, and for
university programs.
survival probability at the minimum stress is 90% at
a confidence level of 95%. The resulting roomtemperature S0 value is 201 MPa. Values at other
temperatures are obtained by using the
aforementioned multiplication factors.
The effects of two standard bounding fluid
exposures1000 h in distilled water and 100 h in
windshield washer fluid (70% methanol)on
baseline properties were investigated. The distilled
water had the greatest effect; a factor of 0.94 bounds
the effects on strength and stiffness in tension,
compression, and shear.
The effects of temperature cycling, such as that
experienced by automotive structures, were
investigated by cycling groups of tensile,
compressive, and shear specimens between –40°C
and 120°C twenty-five times prior to testing. The
only significantly deleterious effect of the cycling
before testing was on shear stiffness, which was
reduced by 25%.
Some bending is unavoidable in structures.
While the quasi-isotropic composite is isotropic in
the plane, it is anisotropic in bending because of the
directionality of the outer surface fibers. Thus
uniaxial bending tests were performed on specimens
with surface fibers oriented both perpendicular and
parallel to the specimen axis. As expected, the
lowest strength values were for specimens with
transverse surface fibers. The modulus of rupture,
which is the elastically calculated maximum bending
stress (assuming an isotropic, homogeneous
material), was 1.8 times the UTS at room
temperature and 1.2 times the UTS at 120°C. These
values will be utilized in the design criteria. The
flexural strength at 120°C was 0.44 times that at
room temperature. The same factor was obtained in
biaxial flexure tests.
Basic Properties and Environmental Effects
Basic short-time properties used to develop
design criteria include tensile, compressive, shear,
and flexural strength. These have been developed for
the quasi-isotropic composite (40 vol % fiber) over
the automotive design temperature range of –40°C
to 120°C. Also, the effects of fluids on each property
have been determined. Average tensile,
compressive, and shear properties are tabulated in
Table 1 at three key temperatures. Multiplication
factor curves have been derived from these, and
from data at other temperatures, to allow the
determination of a property at any temperature from
the corresponding room-temperature value.
Table 1.
Baseline tensile, compressive, and shear
properties
Temperature
–40°C
–23°C
120°C
Tension
Modulus, GPa
Strength, MPa
33.7
292
32.4
336
29.8
272
Compression
Modulus, GPa
Strength, MPa
32.7
230
32.1
225
27.3
131
Shear
Modulus, GPa
Strength, MPa
12.1
240
12.2
226
10.9
133
Cyclic Fatigue
As was done for the basic property
determinations, cyclic fatigue tests were performed
at temperatures over the –40°C to 120°C range of
interest and in distilled water with a 1000-h presoak
and in windshield washer fluid with a 100-h
presoak. The setup for performing tests in a fluid is
shown in Figure 2. Tensile fatigue curves at room
temperature and at 120°C are shown in Figure 3.
Similar curves exist for –40°C and 70°C and for the
two fluid exposures. Factors relating the failure
The basic short-time allowable stress, S0, for
design is defined as two-thirds of the minimum
UTS. The B-basis minimum UTS used is based on
statistical treatment of 86 strength values so that the
159
FY 2001 Progress Report
Automotive Lightweighting Materials
Table 2.
Fatigue strength multiplication factors
Environment
Cycles to failure
10
2
104
106
108
–40°C air
0.89
1.00
0.96
0.88
70°C air
1.00
0.97
0.89
0.82
120°C air
0.97
0.78
0.62
0.50
Distilled water
0.92
0.91
0.97
1.03
Windshield
washer fluid
0.97
0.92
0.95
0.98
Long-Term Creep-Rupture Strength
Long-term creep-deformation and creep-rupture
tests are performed in deadweight lever-arm test
machines. Most specimens are instrumented with
strain gages to measure long-term time-dependent
creep. Tensile creep test results have been generated
at room temperature and at 120°C and in distilled
water and windshield washer fluid. Creep-rupture
curves at room temperature and at 120°C are shown
in Figure 4. The curve at room temperature is very
flat, but there is a significant increase in slope at
120°C.
Figure 2. Tensile fatigue specimen immersed in fluid
during cyclic loading. Extensometer probes
pass through the container to measure stiffness
loss during cycling.
Figure 3. Cyclic-fatigue curves at room temperature and
at 120°C in air.
stresses from these curves to those from the roomtemperature curve are tabulated in Table 2.
A room-temperature design curve was
developed by placing two reduction factors on the
room-temperature curve shown in Figure 3. First, a
factor of 20 was placed on cycles to failure.
Maximum stresses were then multiplied by 0.90, the
ratio of minimum to average UTS. The resulting
curve and the factors in Table 2 can then be used to
determine design-allowable cyclic stresses for other
temperatures and environments.
Figure 4. Creep-rupture curves at room temperature and
at 120°C in air.
Table 3 is a tabulation of multiplication factors
for relating tensile creep-rupture strength at 5000 h
and 15 years, for 120°C and for the two fluid
environments, to the corresponding strength at room
temperature in air. The 5000-h time corresponds to
the maximum assumed operating time of a vehicle;
15 years is the assumed design life.
160
Automotive Lightweighting Materials
Table 3.
FY 2001 Progress Report
Creep-rupture strength multiplication
factors
Environment
Rupture time
5000 h
15 year
120°C
0.73
0.70a
Distilled water
0.98
0.99
Windshield washer fluid
0.95
0.94
a
This is an unlikely condition because 5000 h
is the maximum time a component would be hot.
A room-temperature creep-rupture design curve
is derived by taking 80% of the stresses from a
minimum curve determined from the roomtemperature data in Figure 4. The factors in Table 3
can then be applied to the design curve. This,
together with reduction factors for compressive
stresses, provides the time-dependent allowable
stresses in the durability-based design criteria.
Figure 5. Typical impact damage area on backside of
impact plate specimen.
To determine the degradation of strength due to
impact damage, special 76.2-mm-wide compressionafter-impact specimens were cut from the impacted
plate specimens and tested. The results are shown in
Figure 6 as the ratio of the degraded strength to that
of undamaged specimens. A “design curve,” shown
in the figure, was developed to conservatively
estimate the degradation.
Damage Tolerance
Damage tolerance is defined as a measure of a
structure’s ability to sustain a level of damage and
still safely perform its function. Types of damage of
concern in automotive composite structures include
low-level impact damage or the presence of a flaw.
Both have been investigated for the quasi-isotropic
composite.
In the case of impact damage, pendulum drop tests
were used to represent such things as tool drops,
while an air gun with a small projectile was used to
represent the other extreme, such as kickups of
roadway debris. Figure 5 is a photograph of the
backside of a typical impacted plate specimen.
Lower-energy impacts can also produce damage that
is not visibly detectable. Ultrasonic C-scans were
used to quantify the damage areas, which in turn
were correlated with impactor mass and velocity. A
single design curve was developed that
conservatively predicts damage, including damage
in specimens impacted at a temperature of –40°C
and in specimens impacted by dropped bricks.
Figure 6. Residual compression-after-impact strength vs
impact damage area.
161
Automotive Lightweighting Materials
FY 2001 Progress Report
B. Creep, Creep Rupture, and Environment-Induced Degradation of Carbon
and Glass-Reinforced Automotive Composites
John M. Henshaw
The University of Tulsa
600 South College, Tulsa, OK 74104
(918) 631-3002; fax: (918) 631-2397; e-mail: [email protected]
Véronique Lombart
The University of Tulsa
e-mail: [email protected]
ACC Project Contact: Dan Houston
Ford Motor Company
20000 Rotunda Dr., MD 3182 SRL, P.O. Box 2053, Dearborn, MI 48121-2053
(313) 323-2879; fax: (313) 323-0514; e-mail: [email protected]
Project Manager, Composites: C. David Warren
Oak Ridge National Laboratory, P.O. Box 2009
Oak Ridge, TN 37831-8050
(865) 574-9693; fax: (865) 574-0740; e-mail: [email protected]
DOE Program Manager: Joseph A. Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: U.S. Automotive Materials Partnership (cooperative agreement)
Contract No.: DE-FC05-97OR22545
Objectives
•
Design and develop low-cost, reliable fixtures and methods for the characterization of creep-rupture behavior of
automotive composites with and without environmental exposure. Confirm results generated by the new fixture
with those from conventional testing systems.
•
Incorporate these fixtures and methods into industry-standard test methods for automotive composites.
•
Use results of short-term tests to develop predictive models for lifetime property degradation.
•
Investigate the fundamental damage mechanisms in polymer-matrix, E-glass, and carbon-fiber composites as a
function of specific varied mechanical loading with concurrent environmental exposure.
OAAT R&D Plan: Task 4; Barriers A, C
Approach
•
Design and develop a compact fixture system for creep-rupture testing and confirm its performance.
•
Use the new fixture system to develop a creep-rupture database.
•
Develop a standard procedure for creep-rupture testing using the new system.
•
Develop and verify damage and creep-rupture models for structural automotive composites.
163
FY 2001 Progress Report
Automotive Lightweighting Materials
Accomplishments
•
Tested the first pre-prototype. This process revealed the necessary concept modifications.
•
Developed a new feasible concept using pulleys and a lever arm.
•
Fabricated the final prototype.
•
Tested the final concept using Bayer programmable powdered performing process (P4) materials.
•
Implemented various improvements to the fixture in the areas of loading procedure and load measurement.
Future Direction
•
Continue prove-out of fixture system.
•
Evaluate varying environments on the fixture system and the test materials.
•
Use the new fixture to perform modeling.
results compared well with those of Tulsa University
and Oak Ridge National Laboratory (ORNL)
models.
This pre-prototype was assembled in order to
test the concept of using two pulleys and a lever arm
for the load multiplier. However, the use of a lever
arm within the load multiplier showed some
disadvantages such as travel limitation. Therefore,
the lever arm was replaced by a pulley.
Introduction
Because of insufficient information concerning
the long-term durability of lightweight composite
materials, reliable methods and models requiring
relatively short-term tests are essential if composites
are to achieve their full potential in the automotive
industry. The purpose of this project is to develop
simple low-cost fixtures and methods for the creep
and creep-rupture characterization of automotive
composites and to confirm the in situ creep-test
fixture results with those obtained using
conventional testing methods.
Final Concept of the Creep-Rupture Fixture
Based on the results obtained from the first
prototypes, some improvements were made. A final
concept was produced. This later prototype consists
of two pulleys, two sprockets, and a lever arm
(Figure 1) and is designed to achieve a ratio of up to
216:1. The actual ratio is currently being measured
to account for friction within the system.
A series of tests were performed on Bayer P4
composite specimens. These tests successfully
verified the ability to creep and to rupture, therefore
proving the feasibility of the concept selected.
Furthermore, test data showed encouraging results
and compared well with ORNL results generated on
standard deadweight machines.
A series of improvements was implemented on
the prototype. A clutch was added to adjust the
weight height during a test without perturbing the
specimen. The loading procedure was also modified
to obtain the desired loading rate. Initial tests after
these modifications showed the need for further
improvement. This is currently being addressed.
Initial Design Concept for the Creep-Rupture
Fixture
The first pre-prototype included a speed
multiplier coupled to a lever arm. Testing of this
design concept for the creep-rupture fixture showed
difficulty maintaining constant load as strain
increased. This prototype emphasized the critical
need to minimize friction in a creep-rupture fixture.
From these observations, a second design concept,
combining two pulleys and one lever arm was
developed.
This second concept showed much improvement
in the ability to maintain a constant load while a
specimen was undergoing creep. This modified
concept consisted of a load multiplier composed of
two pulleys and a vertical lever arm that fed into the
creep fixture. Results from a series of creep tests
compared well with accepted values for the tested
material. These experimental results were
encouraging. Specimens were able to creep, and the
164
Automotive Lightweighting Materials
FY 2001 Progress Report
•
•
•
•
ability to test specimens in various environments
(water, etc.)
maintenance of ≥ 95% of the initial load
inexpensive in comparison with a conventional
testing machine: fixture cost less than
approximately $4000 without the data
acquisition system
easy conversion to the spring-loaded in situ
creep fixture previously developed (patent
pending, U.S. Patent Office no. 09/821,280)
A load multiplication of at least 120:1 reduces
the required input load to a maximum of no
more than 50 lb, for a maximum specimen load of
5650 lb.
The final prototype is composed of two main
parts: the fixture and the load multiplier, as shown in
Figure 1. One lever arm, two pulleys, and two
sprockets enable the fixture to achieve the desired
ratio of 180:1. The lever arm is connected to a small
pulley located on the upper axle of the load
multiplier. This axle is coupled to the lower axle
using two sprockets and a chain. A small amount of
tension was applied on the chain to reduce friction
within the system. A weight hangs from a large
pulley next to the small sprocket on the lower axle.
Because of the load multiplication of the fixture, the
weight moves much farther than any of the other
components in the system. For this reason, it is
useful to be able to raise the weight without
disturbing the other components. To accomplish
this, a one-way clutch was installed on the large
pulley, as shown in Figure 3.
Figure 1. Final prototype.
Finally, in addition to strain, temperature and
humidity, load is now integrated in the data
acquisition system and can be measured
simultaneously (Figure 2).
Figure 2. Schematic of the fixture and data acquisition
system.
Creep-Rupture Test Fixture Details
As stated in the previous report, to achieve the
goal of developing a compact, inexpensive creeprupture fixture, several design specifications were
targeted. These specifications are as follows:
•
•
•
•
•
capability of testing materials such as aerospacegrade carbon-fiber composites
maximum load application of 5650 lb on a test
specimen (assuming an ultimate tensile strength
of 90 ksi for a specimen of 0.5 × 0.125 in. cross
section)
load multiplication of at least 120:1
geometric envelope: 500 × 250 × 500 mm
lightweight: less than 45 lb
Figure 3. Pulley modification.
165
FY 2001 Progress Report
Automotive Lightweighting Materials
A load cell was placed between the lever arm
and the cable to record any load variation during a
test, as shown in Figure 1. A Labview program was
developed to record the load simultaneously with
strain, temperature, and humidity during a test in
order to assess any environmental influences on the
overall fixture. Finally, a first attempt was made to
load the specimen at a constant chosen loading rate.
A threaded rod was placed horizontally in the load
multiplier and touched the lever arm (Figure 4) to
stop it from deflecting. To load the specimen, the
threaded rod would turn at a constant rate to slowly
release the lever arm, thereby placing the specimen
in tension. However, high-compression forces on the
threaded rod made loading the specimen difficult. In
fact, the rod could barely turn. Following this
experiment, a second version of that concept was
developed and is currently installed on the fixture
and ready for testing.
Figure 5. Results of five creep tests conducted at various
levels of stress using the new creep-rupture
test fixture. These results are compared with
upper and lower bounds from an ORNL
model.
Figure 6. Test data compared with those of a nonlinear
model.
test on the fixture. These results were encouraging in
comparison with ORNL results.
For each test, the load was recorded to verify the
ability of the fixture to maintain load. Figure 7
shows some results obtained from the load
measurement. The load loss during a test remains
under 5% on average. However, some fluctuations
in the load do occur. These are currently being
studied.
When examining the data recorded during a
typical experiment, it was noted that temperature
and humidity varied significantly throughout the
test. Figure 8 is an example of these variations. The
load fluctuation is being further analyzed to see if
temperature and humidity affect the load variation.
If the load is affected by the environmental
conditions, it will be necessary to better understand
the fixture and account for it.
Figure 4. Loading procedure modifications (first
attempt).
Test Results Using the Creep-Rupture Test
Fixture
Figure 5 shows results of five creep tests on
Bayer P4 composite made using the new fixture.
Also shown in Figure 5 are the upper and lower
bounds of a nonlinear model developed by ORNL
for the same P4 material.
Figure 6 shows the curve fit of test data obtained
with the new fixture at a stress level of 104 MPa.
For this stress level, the ORNL nonlinear model was
applied, showing the upper and lower bound for the
tested material. With one exception, the strain curve
fell within or on the boundary of the model for each
166
Automotive Lightweighting Materials
FY 2001 Progress Report
Ongoing Work
A modification to the loading procedure was
recently implemented in the fixture and is ready to
be tested. The load multiplication ratio of the creeprupture test fixture will be experimentally verified.
Finally, a series of tests accounting for
environmental effects on the creep-rupture fixture
system will be performed to better understand how
the fixture behaves in different environmental
conditions. Because the load varies (i.e., does not
only decrease) during a test, it is possible that an
increase or decrease in temperature and/or humidity
might have an influence on the fixture. It is
important to determine what this influence might be
in order to account for it when a test is performed.
This series of tests will consist of a designed
experiment having two parameters: temperature and
humidity. For each parameter, two values will be
chosen. For example, room and elevated
temperatures could be selected, and the humidity
levels could be 25% (or 50%) and 90%. Creep tests
will be run at each set of conditions, and the results
will be used to determine environmental influences
on the fixture.
Figure 7. Load fluctuation for four different tests.
Figure 8. Possible effect of environmental condition on
load and strain. (No scale is shown for the
strain.)
167
Automotive Lightweighting Materials
FY 2001 Progress Report
C. NDE Tools for Evaluation of Laser-Welded Metals (Steel and Aluminum)
Project Leader: William Charron
Ford Motor Company
24500 Glendale Avenue
Redford, MI 48239
(313) 592-2296; fax: (313) 592-2111; e-mail: [email protected]
Project Administrator: Constance J. S. Philips
National Center for Manufacturing Sciences
3025 Boardwalk
Ann Arbor, MI 48108-3266
(734) 995-7051; fax: (734) 995-1150; e-mail: [email protected]
DOE Program Manager: Joseph Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: U.S. Automotive Materials Partnership
Contract No.: DE-FC05-95OR22363
Objectives
•
Develop fast, accurate, robust, noncontact nondestructive evaluation (NDE) tools and methodologies to replace
current manual destructive testing of laser-welded sheet and structures in zinc-coated steel and aluminum.
•
Demonstrate accuracy and repeatability of the technologies developed or applied.
•
Eliminate the need for a highly trained/experienced NDE evaluator.
OAAT R&D Plan: Task 6, Barriers A, C, D
Approach
•
Phase 1: assess state-of-the-art technologies, down-select, conduct validation testing using fabricated welded
coupons, correlate NDE test results with destructive test results, and select technologies.
•
Phase 2: develop and build of bench prototypes of the selected NDE methodologies, conduct further validation
testing using laser-welded production parts, and correlate NDE test results with destructive test results.
Accomplishments
•
Completed preliminary weld flaw characterization of metals from original equipment manufacturers.
•
Completed selection of target laser weld application.
•
Completed preliminary NDE system functional specifications.
•
Completed initial assessment of NDE state-of-the-art technologies.
169
FY 2001 Progress Report
Automotive Lightweighting Materials
system ready for installation in a factory setting for
in-plant evaluation by the end of FY 2002.
Approved for incorporation into the
USAMP/AMD project portfolio in January 2001 and
launched in May 2001, this project has two primary
investigative missions: (1) evaluate and develop new
NDE tools for laser-welded steel (zinc or organically
coated) and (2) evaluate and develop new NDE tools
for laser-welded aluminum. The investigative
strategy is to
• Conduct a comprehensive assessment of existing
and emerging NDE tools for steel and
aluminum.
• Down-select technologies to the most promising
methodologies.
• Conduct validation testing of the selected
methodologies against the destructive test
method currently in use first for fabricated
welded coupons and then production laserwelded parts.
• Demonstrate the selected methodologies through
bench prototype systems.
Introduction
Laser welding has been widely accepted by the
automotive industry as an industrial process. Uses
range from welding of tailor-welded blanks to
transmission components to air-bag inflation
modules. The use of lasers continues to increase,
with some manufacturers considering the use of
lasers for welding of the “body in white.”
While laser welding has been accepted by the
industry, cost factors have always been an issue,
particularly in the area of weld defects. Given the
high speed and high volume of laser welding
coupled with the relatively small critical flaw size,
finding defects can be time-consuming and difficult
and, hence, expensive. If not detected before
subsequent processing and/or use, these defects
could cause failures to occur later during processing
or during use. For example, pinhole porosity in a
laser-welded tailor blank can cause failure during the
stamping operation, which can in turn damage the
dies and thus cause the press to be out of operation
for a period of time. To prevent this, laser welding
needs to be accomplished with no detrimental flaws
and be monitored with a high degree of reliability.
There have been a number of systems developed
to monitor laser welding systems in real time.
Generally, these systems examine the by-products of
the laser-to-metal interaction to determine the
quality of the weld. These detection methods may
include examining the frequency and intensity of the
light that is given off and comparing it with a known
“acceptable” weld. Most of these monitoring
systems use this “training” method as a basis for
determining acceptable welds versus welds to be
rejected. Generally, these systems interpolate
between known parameter variations. Some systems
use software that is neural network based rather than
function based. As of this writing, these systems
have had limited success in the production
environment.
The U.S. Automotive Materials Partnership
(USAMP) is exploring NDE tools for use upon
completion of the weld. Of specific interest is the
identification of progressive and emerging
technologies in NDE. As part of this effort, the
AMD 303 project performed a technology
assessment of NDE. From this assessment, specific
technology recommendations are being reviewed,
with the goal being to have an NDE prototype
Details of Phase 1
Phase 1 is the assessment of state-of-the-art
technologies, down-selection, validation testing
using fabricated welded coupons, correlation of
NDE test results with destructive test results, and
technology selection. As of this writing, the
following Phase 1 activities had been accomplished.
The NDE system sought was conceptually
defined in terms of its functionality, and the laser
weld flaws or defects that it must be able to detect
were characterized. Some examples of weld defects
can be seen in Figures 1–6.
An initial assessment of NDE technologies was
completed. Available commercial off-the-shelf
systems were identified in addition to emerging
NDE technologies. The goal of this project is to
identify techniques that can be applied to steel and
aluminum structures. The techniques identified by
our initial assessment are applicable to both
materials. However, one should bear in mind that
procedures or equipment developed for one material
would likely require extensive redesign before they
could be applied to the other material. For example,
ultrasonic transducers developed for steel would
need to be completely redesigned for aluminum due
170
Automotive Lightweighting Materials
FY 2001 Progress Report
Figure 1. Pinholes in a laser weld.
Figure 4. Incomplete penetration in laser weld.
Figure 2. Porosity in cross-sectional sample of a laser
weld.
Figure 5. Laser weld with solidification defect at weld
centerline.
Figure 3. Centerline cracking of laser weld.
Figure 6. Contaminated laser weld.
171
FY 2001 Progress Report
Automotive Lightweighting Materials
•
to the difference in sound characteristics between the
two materials. Another example would be for
radiographic testing. Aluminum requires much less
radioactive energy than steel to achieve an image of
sufficient density for interpretation. Consequently,
an X-ray tube suitable for aluminum inspection may
not be adequate to penetrate steel to achieve
sufficient density in a timely manner. Factors such
as these will be given serious consideration as the
project progresses toward down-selection of the
technologies.
Preliminary findings arising out of our initial
assessment have provided further insight into the
functionalities required from the eventual NDE
prototype system built by this project. The prototype
system should be
•
•
•
Capable of autonomously evaluating and
assessing welds in the production system with
minimal operator intervention and a high degree
of reliability.
Adaptable to a variety of weld joint geometries
and materials.
Capable of being mounted on a robot arm.
Capable of detecting and evaluating
discontinuities in accordance with a uniform
standard.
The AMD 303 project team continues to
evaluate the findings arising from the initial
technology assessment study, and down-selection of
the methodologies to those that merit validation
testing is imminent.
172
Automotive Lightweighting Materials
FY 2001 Progress Report
D. Modeling of Composite Materials for Energy Absorption
Edward Zywicz and Steve DeTeresa
Lawrence Livermore National Laboratory
Livermore, CA 94550
(925) 423-5632; fax: (925) 424-2135; e-mail: [email protected]
Srdan Simunovic, Haeng-Ki Lee, J. Michael Starbuck, and Raymond G. Boeman
Oak Ridge National Laboratory
Oak Ridge, TN 37831-6359
(865) 241-3863; fax: (865) 574-7463; e-mail: [email protected]
Project Manager, Composites: C. David Warren
Oak Ridge National Laboratory
P.O. Box 2009, Oak Ridge, TN 37831-8050
(865) 574-9693; fax: (865) 574-0740; e-mail: [email protected]
DOE Program Manager: Joseph A. Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: Oak Ridge National Laboratory, Lawrence Livermore National Laboratory
Contract No.: DE-AC05-OR22725, W-7405-ENG
Objective
•
Develop analytical and numerical tools that efficiently predict the behavior of carbon-fiber–based composites in
vehicular crashworthiness simulations. These predictive tools are intended to decrease the automotive design
process time and cost by reducing component testing and to increase the simulation accuracy of carbon-fiberreinforced structures. The developed tools are used in conjunction with existing crash simulation software.
OAAT R&D Plan: Task 7; Barrier C
Accomplishments
•
Developed a nonlinear undulation model, incorporated it into the previously developed triaxially braided
constitutive model, and validated it against detailed finite element simulations of the braid’s unit cell.
•
Developed a constitutive-level enhanced-strain element formulation for the braid model to simulate tensile
failure.
•
Formulated and implemented a compressive damage surfaces and evolution equations for the braid model.
Future Direction
•
Validate the braid material model by simulating previously conducted strip and tube crush experiments.
173
FY 2001 Progress Report
Automotive Lightweighting Materials
depicts the complex fiber geometry using a
simplified unit-cell approach. The material within
the braid’s unit cell is partitioned into three distant
layers, and then each layer is individually
homogenized. The behavior within each layer is
simulated using a tow-level constitutive model,
which represents the basic elastic, plastic, and
damage behavior of the straight fibers; and an
undulation model, which replicates how the curved
fiber regions respond.
In FY 2001, the basic tow-level constitutive
model was enhanced in several ways. First, a
nonlinear undulation model was formulated and
integrated into the tow model. Second, the matrix
behavior was expanded to include rate-dependent
plastic deformation. Third, compressive and tensile
damage mechanisms were derived and incorporated
into the constitutive relationship.
A micro-mechanics–based nonlinear undulation
model was formulated, implemented, and
numerically validated. Detailed elasto-plastic finite
element (FE) simulations of the curved tow region
(shown in Figure 1) were used to distill the macro
undulation response under uni-axial tensile and
compressive loadings. The newly developed
nonlinear undulation model idealizes this complex
behavior. It accounts for limited plastic deformation
and stiffening in tension (due to tow straightening).
Figure 2 shows the excellent agreement achieved
between the macro stress-strain response of the
undulation region obtained by the FE simulations
Introduction
Automotive structures manufactured from
carbon-fiber composites (CFCs) offer the potential
for significant advantages in weight, durability,
design flexibility, and investment cost. Although
there is substantial experience with graphite-fiber
laminated composites in the aerospace community,
there is little knowledge of how CFCs would
respond in automotive applications during impactinduced “crash” loading conditions (i.e., “crush”).
Furthermore, predictive analytical and numerical
tools required to accurately evaluate and design
carbon-fiber automotive structures for crush do not
exist. This project aims to understand and quantify
the basic deformation and failure mechanisms active
in carbon-fiber materials during vehicular crush
conditions. The project entails modeling, numerical,
and experimental components and deals exclusively
with automotive-type materials: braided, textile, and
chopped random-fiber architectures.
Experimental Investigation of Braided
Carbon-Fiber Composites
Basic material tests were performed on the
braided CFC at quasi-static and elevated rates to
quantify and characterize the material behavior.
While the resin demonstrated a marked dependence
upon strain rate, little rate dependence was observed
in the failure behavior of the 0o/±30o braided CFC
tested. In addition to the physical experiments,
detailed microscopic examinations were performed
on the failed specimens tested at Lawrence
Livermore National Laboratory (LLNL), strip
specimens crushed by Oak Ridge National
Laboratory (ORNL), and automotive tubes
dynamically crushed by the Automotive Composites
Consortium (ACC). The controlling failure
mechanisms appear to be inter-tow rather than intratow ones. Results from all of LLNL’s traditional and
non-traditional material testing and post-mortem
examinations of the present CFCs are summarized in
ref. 1.
Pure
Elasto-Plastic
Tow Model
Damage Model for Braided Carbon-Fiber
Composites
A framework to numerically simulate the
behavior of large-tow triaxially braided CFC was
developed in FY 2000.2 It represents the layered
composite material on a layer-by-layer basis and
Figure 1. Model of an undulation region in a 0o/±45o
triaxially braided carbon fiber composite.
174
Automotive Lightweighting Materials
FY 2001 Progress Report
and transverse directions, as well as for different
mechanisms in tension and compression.
While both tensile and compressive loadings
generate inter- and intra-tow micro and macro
cracking, commingling of the debris and the
undamaged material causes the compressive stressstrain response to reach a non-zero saturation level
after the peak load. In the fiber direction, this is
simulated by abruptly lowering the maximum stress
the fiber can carry to a prescribed level after the (inplane) axial fiber strain reaches are at a critical level.
Similarly, in the transverse direction, the matrix
flow strength is reduced to a saturation value after
the onset of damage. Motivated by recent triaxial
experimental work on uni-directional CFC material,
the matrix stresses are used in a J-2 plasticity-like
criterion to determine when transverse damage
commences.
The post-peak tensile response in each direction
is replicated using a pseudo enhanced-strain element
formulation.4 In this approach, the total
displacement field is additively decomposed into a
material displacement field and a crack displacement
field. The material strains, calculated from the
material displacements, are used by the tow-level
constitutive model to predict the bulk stresses, while
the crack displacements and a rigid-softening
traction-displacement law replicate how the “crack”
evolves with deformation. Auxiliary constraints,
which ensure that the bulk stresses and crack
tractions are consistent, complete the formulation.
The present tensile damage formulation ensures
that a fully damaged material supports no load
across the postulated crack and is fully consistent
with fracture mechanics concepts. Figure 4 shows
Figure 2. Simulated (finite element) and predicted
(model) macro response of an undulation
region in a 0o/±30o tri-axially braided carbon
fiber composite.
and from the new undulation model for a 0o/±30o
braid. The undulation model produced similar
behavior and agreement for the other architectures
examined.
The nonlinear undulation model is necessary to
accurately represent braided CFC under shear and
transverse loadings. Figure 3 shows the transverse
extensional response obtained from detailed FE
simulations of the braid’s representative volume
element (RVE)3 and the newly enhanced three-layer
model. The agreement between the two results is
greatly improved by the nonlinear undulation model,
as is the shear response (not shown in figure).
Figure 3. Three-layer model and detailed finite element
representative volume element predictions of
the transverse extensional response of various
triaxially braided carbon fiber composites.
Guided by experimental observations, two
compressive and two tensile damage mechanisms
were derived and incorporated into the elasto-plastic
tow-level constitutive model. They allowed separate
formulations to be developed for damage in the fiber
Figure 4. In-plane stress vs applied (x-direction) strain
for a uni-directional carbon fiber composite
lamina. The fibers are oriented 40( from the
x-axis.
175
FY 2001 Progress Report
Automotive Lightweighting Materials
crush tests of 0o/±30o braid architectures. The model
will then be used to predict drop-tower tube crush
tests of 0o/±45o and 0o/±60o braided CFC
architectures.
the predicted in-plane stresses, as a function of the
applied strain, for a uni-directional CFC lamina
subjected to in-plane uni-axial straining. A square
shell element, aligned with the x-y coordinate
system, is used, and the fibers are oriented at 40(
from the x-axis. Notice that the resultant stress-strain
curves reach a peak and then decay to a zero stress
level as desired. Similar behavior was demonstrated
for fiber orientations between –180 and 180( (i.e.,
any angle).
Budget
FY
2001
Conclusions
A three-dimensional micro-mechanical–based
finite-deformation constitutive model has been
development and implemented into the nonlinear,
FE code DYNA3D. The model attempts to replicate
progressive damage in CFCs and is intended for use
in automotive crashworthiness applications. The
new undulation component added in FY 2001 allows
the model to accurately predict the elasto-plastic predamage response of braided CFC, while the tensile
and compressive damage mechanisms that were
added facilitate modeling of the post-peak material
response.
Carryover
New
money
Current
spent
YTD
spent
íN
325K
169.5k
306k
References
1. DeTeresa, Allison, Cunningham, Freeman,
Saculla, Sanchez, and Winchester, Experimental
Results in Support of Simulating Progressive Crush
in Carbon-Fiber Textile Composites, UCRL-ID,
2001.
2. Zywicz, A Tow-Level Progressive Damage
Model for Simulating Carbon-Fiber Textile
Composites: Interm Report, UCRL-ID-139828,
2000.
3. Zywicz, O’Brien, and Ngyuen, “On the
Elasto-Plastic Response of a Large-Tow Triaxial
Braided Composite,” in Proceedings of the Fifteenth
Annual Meeting of the American Society of
Composites, 2000.
4. Zywicz, “A Constitutive-Level EnhancedStrain Element Formulation for Simulating Brittle
Damage,” presented at the workshop Multiscale
Modeling of Materials: Strength and Failure,
Bodega Bay, Calif., UCRL-145514, 2001.
Future Work
With the basic formulations of the braid model
complete, the future emphasis will be upon
numerically “tuning” the model as necessary and
validating it against experimental results, as well as
preparing documentation on its theoretical
formulation. Using the existing experimental data
for a 0o/±30o braid CFC and uni-directional CFC
tows, the necessary material properties will be
determined and used to validate the model against
previously conducted strip tests and drop-tower tube
176
Automotive Lightweighting Materials
FY 2001 Progress Report
E. Composite Crash Energy Management
Principal Investigator: Richard Jeryan
Ford Research Laboratories
2101 Village Road
MD2115 SRL, Rm 2621D
Dearborn, MI 48124-2053
(313) 594-4903; fax: (313) 845-4724; e-mail: [email protected]
Project Manager, Composites: C. David Warren
Oak Ridge National Laboratory
P.O. Box 2009, Oak Ridge, TN 37831-8050
(865) 574-9693; fax: (865) 574-0740; e-mail: [email protected]
DOE Program Manager: Joseph Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: U.S. Automotive Materials Partnership (cooperative agreement) Automotive Composites
Consortium Energy Management Working Group
Contract No.: DE- FC05-95OR22363
Objectives
•
Experimentally determine the effects of material, design, environment, and loading on macroscopic crash
performance to guide design and development of predictive tools.
•
Determine the key mechanisms responsible for crash energy absorption and examine microstructural behavior
during a crash to direct the development of material models.
•
Develop analytical methods for predicting energy absorption and the crash behavior of components and
structures.
•
Conduct experiments to validate analytical tools and design practices.
•
Develop and demonstrate crash design guidelines and practices.
•
Develop and support design concepts for application in demonstration projects.
•
Experimentally determine the effects of material, design, environment, and loading on carbon fiber–reinforced
tubes to guide automotive design and analysis development.
•
Conduct microscopic characterization of crush mechanisms for a range of material types and loading to identify
the key mechanisms responsible for crash energy absorption and to direct the development of material models.
•
Experimentally determine the sensitivity of tube crush performance to impact velocity.
•
Determine the mechanisms that control the observed impact velocity effect to support the development of
analytical models.
•
Experimentally determine the relative energy absorption due to each failure mode of a progressively crushed
composite tube and isolate that portion of energy absorption due to friction. Then evaluate differences in
friction energy absorption between quasi-static and dynamic crush.
177
FY 2001 Progress Report
Automotive Lightweighting Materials
•
Experimentally determine strain fields around an advancing damage zone emanating from a notch in a braided
carbon fiber composite under biaxial load. Determine the effects of load biaxiality and microstructure on
damage accumulation and assess the potential for developing continuum damage mechanics formulations.
•
Evaluate the performance of bonded structures under crash loads.
•
Examine the influence of bond design concepts, impact velocity, environmental conditions, adhesive type, and
other material issues.
•
Fabricate new molding tools to produce simulated automotive structures.
•
Evaluate the performance of carbon composites and section geometries under angular impact.
•
Study the mode of collapse of carbon composite structures subjected to combined loading.
•
Establish the operating envelope for structures that will maintain progressive axial crush mode of failure for
maximum energy absorption.
•
Investigate the viability and crashworthiness of novel sandwich composite concepts for automotive
applications.
OAAT R&D Plan: Task 7; Barrier C
Approach
•
Conduct experimental projects to enhance understanding of the global and macro influences of major variables
on crash performance.
•
Create crash intuition, guidelines and rules of thumb, and data for validation of analysis developments.
•
Conduct microscopic experimental characterization to define the mechanisms that occur during and as a result
of the crash process.
Accomplishments
•
Completed material property characterization tests of low-cost carbon fiber composites.
•
Completed drop tower tests of 2-in.-square tubes.
•
Fabricated additional tubes (2- × 4-in. and 4-in.-square cross section) for axial and off-axis dynamic impact
testing.
•
Obtained approval of a final design of the intermediate-rate machine. The build phase of the project has been
initiated.
•
Completed off-angle compression tests on thick unidirectional carbon/vinyl ester specimens to examine failure
mechanisms at the fiber/resin interface.
•
Performed preliminary testing of 2-in.-square tubes for relative energy absorption and friction energy
absorption.
•
Conducted off-axis tests to characterize homogenized inelastic stress-strain response of the composites.
•
Completed fabrication of test samples used to evaluate the performance of carbon composite and section
geometries under angular impact.
Future Direction
•
Focus efforts on Partners’ needs in the areas of materials, processes, and applications.
•
Determine bending and friction energy absorption modes.
178
Automotive Lightweighting Materials
FY 2001 Progress Report
high-angle specimens. This result is shown to be due
to the nonlinear stress-strain behavior of the matrix.
Project Accomplishments
Carbon-Reinforced Tube Performance
Experimental determination of the effects of
material, design, environment, and loading on
carbon fiber reinforced tubes to guide automotive
design and analysis development. Conduct
microscopic characterization of crush mechanisms
for a range of material types and loading to identify
the key mechanisms responsible for crash energy
absorption and to direct the development of material
models.
The material property characterization tests of
the low-cost carbon fiber composites have been
completed. Materials tested included composites
with unidirectional, stitched, and woven fabric as
well as biaxial and triaxial braids of Fortafil #556
80k fibers and some Grafil 34-700 12k fibers.
Plaques with 40k fiber are being fabricated for
additional material property characterization tests.
To begin investigating the crush mechanisms
associated with dynamic axial crush, drop tower
tests of 2-in.-square tubes were completed.
Additional tubes of 2 × 4 in. and 4-in.-square cross
section were fabricated during the year for axial and
off-axis dynamic impact testing. Fiber architectures
were selected based on the 2-in.-square tube results.
Dynamic impact tests of these new geometries have
been initiated and will be completed in FY 2002.
Energy dissipation mechanisms include resin
fracture and plastic deformation, fiber fracture and
pullout, and inter-ply delamination.
Friction Effects on Crash Performance
Experimentally determine the relative energy
absorption due to each failure mode of a
progressively crushed composite tube and isolate
that portion of energy absorption due to friction.
Evaluate differences in friction energy absorption
between quasi-static and dynamic crush.
This project was begun in July 2001.
Preliminary testing using 2-in.-square tubes of
(0/900/90/0) fiberglass reinforced vinyl ester resin
composite has been performed. Baseline energy
absorption of the tube in axial crush has been
measured at 4950 J. Tensile test results have been
used to calculate the energy absorption due to tube
corner splitting at 98 J or 2% of the total energy
absorbed. End-notch shear test results were used to
calculate energy absorption due to ply delamination
at 377 J or 7.6% of the total energy absorbed.
Finally, a shallow-angle (5º) trigger was used, with
the total energy absorbed measured at 764 J. The
shallow-angle trigger test effectively eliminated the
energy absorption due to bending and much of the
friction energy absorption. The remaining energy
absorption modes, bending and friction, have yet to
be measured. A standard tube crush trigger and a
shallow-angle tapered trigger coated with a lowfriction diamond-like carbon coating will be used in
future tube crush tests to further discriminate the
energy absorption due to friction. A test method to
measure bending energy absorption will be
developed.
Static vs Dynamic Performance
Experimentally determine the sensitivity of tube
crush performance to impact velocity. Determine the
mechanisms that control the observed impact
velocity effect to support the development of
analytical models. This work is supported by an
additional university research project and a planned
intermediate-rate testing facility at ORNL.
A final design of the intermediate-rate machine
has been approved, and the build phase of the
project has been initiated. The intermediate rate
machine will be delivered to ORNL by mid-2002.
Off-angle compression tests on thick unidirectional
carbon/vinyl ester specimens have been completed
to examine the failure mechanisms at the fiber/resin
interface. These tests showed that there is a
transition in the failure mode between low- and
Biaxial Material Studies
Experimentally determine strain fields around an
advancing damage zone emanating from a notch in a
braided carbon fiber composite under biaxial load.
Determine the effects of load biaxiality and
microstructure on damage accumulation. Assess the
potential for developing continuum damage
mechanics formulations based on biaxial test data
for braided carbon composites.
A detailed study of the braided carbon fiber
strips from the tubes revealed an understanding of
the size effects. It also provided a simple benchmark
for verifying detailed nonlinear analyses. Analysis
of the detailed 3D model was carried out to
comprehend the role played by material nonlinearity
179
FY 2001 Progress Report
Automotive Lightweighting Materials
the mode of collapse of carbon composite structures
subjected to combined loading. Establish the
operating envelop for structures that will maintain
progressive axial crush mode of failure for
maximum energy absorption.
The fabrication of the test samples was
completed. The tubes are constructed of carbon
fibers and Hetron resin. Square and rectangular
section tubes (100 mm × 100 mm, 100 mm ×
50 mm) will be tested statically and dynamically at
impact angles between 0° and 20°. The study will be
completed in the first quarter of FY 2002.
and natural geometric imperfections in the braided
structure. Off-axis tests to characterize homogenized
inelastic stress-strain response of the composites
were carried out. Biaxial tests with plaques having
the same thickness as the tubes revealed a bending
problem. As a result, thicker plates are being used in
biaxial testing. Continuum damage mechanics
models are being studied to relate to biaxial test data
for predictive capability. The project has been
extended by 6 months at no cost. Detailed study of
the biaxial test data and relationships to Continuum
Damage Mechanics will be reported at the end of the
project.
Sandwich Structures
Impact Performance of Bonded Structures
To investigate the viability and crashworthiness
of novel sandwich composite concepts for
automotive applications.
A review of the literature on the subject was
conducted. Various topics were highlighted as being
areas requiring further research. A specialized subteam was formed, and additional independent
external researchers were identified and contacted
for possible collaboration. Various primary and
secondary vehicle components were identified for
possible sandwich composites employment, as well
as potential material candidates for the face sheets
and core. An experimental testing program is
currently being set up in which different specimen
sizes and loading configurations are being
considered.
Evaluate the performance of bonded structures
under crash loads. Examine the influence of bond
design concepts, impact velocity, environmental
conditions, adhesive type and other material issues.
Fabricate new molding tools to produce simulated
automotive structures. This is a joint program with
the Automotive Composites Consortium (ACC)
Joining Work Group and part of Focal Project 3
design studies.
Supplier involvement has been obtained to
conduct this effort jointly with Focal Project 3 and
ACC Joining. The program has been planned and
will begin in FY 2002.
Off-Axis Impact
Evaluate the performance of carbon composites
and section geometries under angular impact. Study
180
Automotive Lightweighting Materials
FY 2001 Progress Report
F. Intermediate-Rate Crush Response of Crash Energy Management Structures
Raymond G. Boeman
Oak Ridge National Laboratory
53-1/2 W. Huron St., Suite 213
Pontiac, MI 48342
(248) 452-0336; fax: (248) 452-8992; e-mail: [email protected]
Project Manager, Composites: C. David Warren
Oak Ridge National Laboratory
P.O. Box 2009, Oak Ridge, TN 37831-8050
(865) 574-9693; fax: (865) 574-0740; e-mail: [email protected]
DOE Program Manager: Joseph A. Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: Oak Ridge National Laboratory
Contract No.: DE-SL05-01OR22866
Objectives
•
Develop a unique characterization facility for controlled, progressive crush testing, at intermediate rates, of
automotive materials (polymer composites, high-strength steels, aluminum) and structures.
•
Study the deformation and failure mechanisms of automotive materials subjected to crush forces as a function of
impact velocity for different velocity profiles, including constant velocity.
•
Obtain specific energy absorption and strain data and correlate them with deformation and failure mechanisms
to describe the unknown transitional effects from quasi-static to high-rate for polymer composites.
•
Characterize the strain rate effects for metallic materials and components.
OAAT R&D Plan: Task 7, 10; Barriers A, B, C
Approach
•
Develop a unique high-force (50,000 lb), high-velocity (160 ips) servo-hydraulic to conduct progressive crush
tests on tubes and cones at intermediate rates.
•
Use high-speed imaging to observe and document deformation and damage mechanism during the crush event.
•
Obtain specific-energy absorption as a function of impact velocity.
•
Take strain measurements at discrete locations and investigate full-field measurements of strains and curvatures.
•
Coordinate polymer composites investigations with the Automotive Composites Consortium (ACC) Energy
Management Group via Ari Caliskan of Ford.
•
Coordinate steel investigations with the Automotive Steel Partnership via Srdan Simunovic of Oak Ridge
National Laboratory.
181
FY 2001 Progress Report
Automotive Lightweighting Materials
Accomplishments
•
Developed specifications for a high-force (60,000 lb), high-rate (160 ips @ 60,000 lb) hydraulic test machine.
•
Developed specifications for the Interactive Dynamic Inverse Model software and its interaction with the
Iterative Drive File Generation software for pre-optimization of the servo drive file prior to testing and final
optimization via learning algorithms to achieve the specified velocity profile.
•
Let a contract to MTS for procurement of the machine in February 2001.
•
Completed the mechanical design and released it to manufacturing in July 2001.
•
Procured a hydraulic pump for the machine (in-kind contribution from the ACC) and installed it at the National
Transportation Research Center.
Future Direction
•
Procure a high-speed imaging system in the first quarter of FY 2002.
•
Procure a high-speed synchronous data acquisition system during the first quarter of FY 2002.
•
Take delivery of the machine early in the second quarter of FY 2002.
•
Test programs for steel and composites to be initiated in the third quarter of FY 2002.
•
Survey full-field measurement techniques for strain (in-plane) and curvatures (out-of-plane) for applicability and
explore them to the extent funds are available.
that will lead to a predictive capability for
composites subjected to crash events. Previous ACC
research has demonstrated a dramatic difference in
the energy absorption of composite tubes when they
are tested quasi-statically (2 in./min) and
dynamically (3–10 m/s). The difference between
quasi-static and dynamic crush is as much as 2 to 1,
respectively. However, it is not clear how the
transition from one failure mode to another occurs at
intermediate rates (0–3 m/s) (Figure 1). Currently,
there is no practical opportunity to study the crush
response of tubes tested at intermediate rates. Under
this project, the lack of test capability at
intermediate rates will be addressed to facilitate the
acquisition and documentation of data describing
the transition from quasi-static to high-rate.
Specifications for a high-force (60,000 lb), high-rate
(160 ips @ 60,000 lb) hydraulic test machine have
been developed (Figure 2).
Background
Polymer matrix composites show great promise
for crash energy management in transportation
applications. Researchers worldwide have reported
higher specific energy absorption values for
composites materials with respect to traditional
metal counterparts. However, the energy absorption
mechanisms are much different than in metals; and
relatively little is known regarding the influence of
material and structural parameters on the damage
mechanisms that develop during a crash event. This
knowledge must be gained before composites can be
most efficiently utilized for crash management
structures.
The ACC’s Energy Management Group, with
the assistance of national laboratories and selected
universities, is undertaking the development of the
knowledge base, both analytical and experimental,
182
Automotive Lightweighting Materials
FY 2001 Progress Report
Figure 1. It is unclear how transition from one failure
mode to another occurs at intermediate crash
rates.
Figure 2. This high-force, highvelocity servo-hydraulic
will be used to conduct
progressive crush tests on
tubes and cones at
intermediate rates.
183
Automotive Lightweighting Materials
FY 2001 Progress Report
G. Long-Life Electrodes for Resistance Spot Welding of Aluminum Sheet Alloys
and Coated High-Strength Steel Sheets
Project Manager: Eric Pakalnins
DaimlerChrysler Corporation
Body Materials Engineering
800 Chrysler Dr.
CIMS 482-00-11
Auburn Hills, MI 48326
(248) 576-7454; fax: (248) 576-7490; e-mail: [email protected]
DOE Program Manager: Joseph Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: U.S. Automotive Materials Partnership
Contract No.: DE-FC05-95OR22363
Objectives
•
Survey the currently available technology for achieving long electrode life.
•
Comparatively test a broad selection of existing and developmental electrode technologies that have technical
merit.
•
Confirm the test results through electrode metallography and computer modeling.
•
Evaluate the best practices through beta-site testing of automobile applications.
OAAT R&D Plan: Task 6; Barriers A, C
Approach
•
Conduct benchmarking (Phase 1) in order to produce a state-of-the-art report on electrode wear. This will be
accomplished through a review of open literature and available corporate literature, along with interviews of
industry experts.
•
Screen candidate electrode technologies, conduct in-depth testing of electrodes, and perform beta testing of
selected electrode technologies in an automotive production environment (Phase 2).
•
Conduct computer modeling of the electrode metallurgical and mechanical changes that occur as a result of
electrode wear (Phase 3). The models will be used to define the mechanism(s) of electrode wear when resistance
spot welding aluminum and high-strength coated steels.
Accomplishments
•
Began the process of forming the project team.
•
Obtained approval of the work plan.
185
FY 2001 Progress Report
Automotive Lightweighting Materials
•
Began to conduct interviews with industry experts and to perform a review of the literature relating to electrode
wear.
•
Obtained approval for Oak Ridge National Laboratory (ORNL) to conduct computer modeling.
team. In January 2001, representatives from Geneal
Motors (GM), Ford, and DaimlerChrysler met to
discuss the program. They selected Edison Welding
Institute (EWI) to manage the program and perform
a benchmark study (Phase 1). Brush-Wellman, a
copper producer, and electrode producers Nippert
Company and Huys Industries are currently part of
the program as well. Additional electrode and
copper producers are being sought. ORNL will be
performing the computer modeling of the electrodes.
Representatives from both steel and aluminum
producers may also be invited to participate.
Introduction
Resistance spot welding has been widely
adopted by the automotive industry because of its
relatively low capital and operating costs and its
potential for high production rates. However,
electrode wear of coated steels and aluminum has
been a continuing and significant problem.
Electrode wear adversely affects the cost and
productivity of automotive assembly welding
because it reduces the reliability and robustness of
the weld product. This requires additional
inspection and mandates more strict control of the
welding parameters to ensure quality. Ultimately,
worn electrodes can result in deteriorated weld
performance. Consequently, the production
manufacturing cost savings of doubling the
electrode life are substantial.
As technology has developed over years of use,
few engineering solutions have been successfully
introduced into the manufacturing process to
manage the issues resulting from electrode wear.
Weld current steppers and electrode cap dressers
have been used for many years to minimize the
effects of electrode wear, but these techniques do
not address the underlying causes of electrode
degradation. More-recent efforts to remedy
electrode wear have resulted in the development of
new electrodes, equipment, and sheet metal
surfaces. Innovations in electrode technology
include new compositions, inserts on the electrode
face, coated electrodes, new manufacturing
processes, and new electrode geometries. The
objective evaluation of these types of existing and
developmental technologies is the subject of the
present initiative.
Work Plan Approval
The team was presented with a work plan,
which has been approved by GM, Ford, and
DaimlerChrysler. This plan calls for three phases,
which are illustrated in Figure 1. Phase 1 is a
comprehensive benchmarking study that will
include a literature review, interviews with industry
experts, a review of the available internal corporate
documents and data, and identification of
commercial and developmental technology for
inclusion in Phase 2. This activity will be completed
by December 2001 and will be the foundation for
the next two phases.
Phases 2 and 3 will run concurrently and will
test those electrode geometries and compositions
that could technically provide significant
improvements in electrode life. These technologies
will be selected via analysis of the information made
available during the benchmarking study. Phases 2
and 3 include activities such as electrode
metallography, computer modeling, and beta-site
testing in an effort to provide both practical and
technical confidence for successful implementation
in automotive production. The anticipated
completion date of the initiative is the 4th quarter of
FY 2003.
Work Team Development
A work team is being assembled, with a
representative from DaimlerChrysler leading the
186
Automotive Lightweighting Materials
Phase 1
Benchmarking
Literature
Review
Industry
Expert
Interviews
Review of Internal
Company Documents
Identification of
Commercial Technology
FY 2001 Progress Report
Phase 2
Core Testing Program
Life Testing of
Selected
Technology
Computer Model for
Galvanized and
Aluminum
Sequential Life Testing
Electrode
Deformation
Modeling
Metallographic
SEM Analysis
and
Preliminary
Analysis of Results
Beta-Site Testing
Evaluation of Candidate
Technologies
Phase 3
Computer Modeling
Final
Analysis and
Report
Microstructural
Modeling
Model Predictions for
Selected Electrodes
Analysis of Results
Figure 1. Long electrode life initiative.
Phase 1—State-of-the-Art Benchmarking
Study
Phase 2—Core Testing Program
Electrode life testing using standard techniques
will be performed to assess the electrode life of
competing technologies. The results will be
evaluated against the minimum standards
established in Phase 1. Those technologies that
appear to substantially improve electrode life will be
evaluated further. The data from this task will also
be used as input to the computer modeling work in
Phase 3.
For selected technologies that perform at or
above the selected minimum expectations,
interrupted electrode life tests will be performed.
Standard life testing will be performed using
identical procedures but will be interrupted at 5, 10,
25, 50, and 100 welds, etc. At each break point, the
electrodes will be permanently removed from the
holders and electrode wear measurements will be
assessed. These electrodes will be
metallographically sectioned and compared using
optical and scanning electron microscopy (SEM)
techniques to assess the stages of electrode wear.
The data will be analyzed to determine the nature of
electrode life improvement. These results will be
compared with those from the computer modeling in
Phase 3.
The objectives of Phase 1 are to collect all
pertinent data and technologies for consideration in
an evaluation of technologies that improve electrode
wear. The open literature will be reviewed to
identify the current theories of electrode wear
mechanisms and phenomena. Experts in the
automobile industry will be interviewed to
determine the issues of electrode wear in practice,
the solutions, and the production constraints within
which potential solutions must work.
Data on electrode wear phenomena and
potential solutions to electrode wear will be
collected during these interviews. If made available,
internal corporate documents will be reviewed. Both
corporate test results and nonproprietary production
welding practices will also be summarized.
The existing commercial technologies
suggesting improvement in electrode wear will be
surveyed in an attempt to obtain test results,
determine the mechanism of electrode wear
improvement, and assess implementation
requirements.
The information from the above tasks will be
reviewed, and specific candidate technologies will
be recommended for inclusion in Phase 2. The
minimum requirements for improved electrode wear
will be established.
187
FY 2001 Progress Report
Automotive Lightweighting Materials
modeling results from Phase 3, will be analyzed.
The technologies offering the best opportunity for
improved electrode life when welding aluminum
and galvanized steel will be selected for beta-site
testing. The tests of selected technologies will be
performed in an automotive production
application(s) on aluminum and coated highstrength steels. The existing operating conditions
and electrode life will be assessed before
implementation of the new technology.
Phase 3—Computer Modeling of Electrode
Wear
An existing computer model of electrode
performance will be adapted for use with galvanized
steel and aluminum. This model will be enhanced to
provide for both the global and electrode-sheet
deformation events central to electrode wear
phenomena. The model will be further enhanced to
include the microstructural changes that occur
between zinc and copper during the electrode wear
phenomena and will be used to predict the
deformations and alloy layer development for the
selected electrodes. These predictions will then be
compared with the metallographic results. After
assessing the validity of the model, it will be used to
better understand the causes of electrode wear and
interpret the mechanism of electrode life
improvement as evidenced by the results. These
results will be used to help select one or more betatesting candidate technologies.
Summary of Current Ongoing Activities
Interviews with industry experts are presently
under way to better define electrode wear and how it
occurs in practice, to obtain input on the factors
contributing to electrode wear, and to identify
potential existing and developmental technologies
that may help to avoid the effects of electrode wear.
At this time, 25 interviews have taken place and
more than 15 are scheduled.
The level of activity with regard to the literature
review is accelerating. Papers are being received,
and a few have been reviewed.
Project Summary and Beta-Site Testing
The data from the electrode life and
metallographic testing, as well as the computer
188
Automotive Lightweighting Materials
FY 2001 Progress Report
H. Plasma Arc Welding of Lightweight Materials
Project Manager: William Marttila
DaimlerChrysler Corporation
Materials Engineering
800 Chrysler Dr.
CIMS 482-00-11
Auburn Hills, MI 48326
(248) 576-7446; fax: (248) 576-7490; e-mail: [email protected]
DOE Program Manager: Joseph Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: U.S. Automotive Materials Partnership
Contract No.: DE-FC05-95OR22363
Objectives
•
Develop and verify the joining technology required for the joining of lightweight materials (aluminum and
magnesium) utilizing plasma arc spotwelding technology.
•
Develop the necessary weld parameters and techniques required for a robust joining process.
•
Develop guidelines for testing mechanical properties for this new technology that are appropriate for the various
anticipated applications.
OAAT R&D Plan: Task 6, 13; Barriers C, D
Approach
This project has been divided into three key phases: testing, analysis and summary.
•
Phase I: produce approximately 5000 material coupons that will be used for tensile, shear, and metallurgical
analysis and testing.
•
Phase II: modify the material alloys based on Phase I results.
•
Phase III: produce guidelines for welding parameters and quality control.
Accomplishments
•
Formed project team and conducted official meetings.
•
Finalized test matrix (material selections and combinations).
•
Obtained material from ALCAN to produce coupons for testing.
•
Secured agreement from ALCAN and ALCOA to supply required material for the program.
•
Secured ABB robots for the program [accomplished by General Motors (GM)].
•
Designed and fabricated fixture to ensure uniform coupon production.
•
Redesigned plasma arc torch (reduced tip size) to allow for reduction in flange size.
189
FY 2001 Progress Report
Automotive Lightweighting Materials
Future Direction
•
Proceed with Phase I testing.
•
Conduct Phase II testing based on Phase I results.
•
Document and report results of the research.
Additional team members were solicited from
the automotive industry. Rite-on Industries, ABB
Robots, ALCAN, and ALCOA joined the team and
are currently a part of the program. Additional
suppliers may be also be invited to participate in the
program.
Introduction
The current joining technology for aluminum
relies heavily on resistance spotwelding, which at
best is a marginal process. High maintenance and tip
wear continues to be a major concern. Rivets and or
mechanical clinching are costly alternatives that
require high capital investment. The viability of
plasma arc spotwelds has entered the picture
through the efforts of Arc Kinetics, Ltd. Arc
Kinetics is the leading developer of the plasma arc
welding process in both steel and lightweight
materials. The company developed its original
process for single-sided plasma arc spotwelding for
Jaguar Motor Cars for Jaguar’s sheet metal floor
pan assembly, which was not accessible with
conventional resistance spotwelding equipment. In a
separate effort, the company developed a process
called aluminum plasma arc welding (APAW). By
combining single-sided spotwelds and APAW, Arc
Kinetics developed a process that demonstrates
excellent potential for the joining of lightweight
materials, aluminum and magnesium.
Work Matrix Approval
A work matrix (material) presented to the team
has been approved by GM, Ford, and
DaimlerChrysler. This plan calls for initially two
phases, which require producing material coupons
for various tests: tensile, shear, and metallurgical
analysis/testing.
Evaluation of these test results will determine
the next phases of the program. Phase II will be a
repetition of Phase I with modification to the
material alloys based on Phase I results. Phase III
will be the reporting phase of the program.
Completion of the initiative is expected in the fourth
quarter of 2002.
Phases I and II—Base Testing Programs
Team Development
Coupon testing will consist of tensile, shear, and
metallurgical analysis. Coupons will be evaluated to
ensure weld integrity and establish equipment
parameters to be used in future applications.
The project team was formed with a
representative from DaimlerChrysler as the leader.
Representatives from GM, Ford, DaimlerChrysler,
and Arc Kinetics met to discuss the program in
February 2001. They selected a previous team
member, F. G. LaManna, (LaManna Enterprises,
L.L.C.) to manage the program.
Project Summary
The data from the testing in Phases I and II will
be compiled and used to produce guidelines for
welding parameters and quality.
190
Automotive Lightweighting Materials
FY 2001 Progress Report
I. Performance Evaluation and Durability Prediction of Dissimilar Material
Hybrid Joints
D. L. Erdman
Oak Ridge National Laboratory
P.O. Box 2009, Oak Ridge, TN 37831-8048
(865) 574-0743; fax: (865) 574-8257; e-mail: [email protected]
Project Manager, Composites: C. David Warren
Oak Ridge National Laboratory
P.O. Box 2009, Oak Ridge, TN 37831-8050
(865) 574-9693; fax: (865) 574-0740; e-mail: [email protected]
DOE Program Manager: Joseph Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: Oak Ridge National Laboratory
Contract No.: DE-AC05-00OR22725
Objective
•
Develop new experimental methods and analysis techniques to enable the use of hybrid joining as a viable
attachment technology in automotive structures. The work includes evaluating the mechanical behavior of
composite/metal joints assembled using a variety of hybrid joining methods and quantifying the resultant
damage mechanisms under environmental exposures, including temperature extremes and automotive fluids,
for the ultimate development of practical modeling techniques that offer global predictions for joint durability.
OAAT R&D Plan: Task 6; Barriers C, D
Approach
•
Characterize the structural hybrid joint under quasi-static load conditions.
•
Characterize the response to fatigue, creep, and environmental exposures.
•
Conduct predictive analysis.
Accomplishments
•
Identified a candidate hybrid joint with the Partnership for a New Generation of Vehicles (PNGV) Joining Task
Force.
•
Procured mechanical fasteners (rivets), steel and composite substrates, and adhesive system for manufacture of
the candidate hybrid joint.
•
Completed the fabrication of test specimens and mechanical characterization (tensile and shear properties) of
the joint adhesive in conjunction with the University of Michigan and the Automotive Composites Consortium
(ACC).
191
FY 2001 Progress Report
Automotive Lightweighting Materials
•
Developed basic mechanics and finite element models to estimate stress fields in the hat-section geometry.
•
Completed the mechanical test set-up and associated instrumentation for the first series of structural hat-section
tests.
Future Direction
•
Complete mechanical testing to characterize structural behavior under various in-service load conditions and
environmental conditions.
•
Identify failure modes and damage accumulation and progression and correlate with measurable physical
response.
•
Develop realistic, practical models to predict joint durability and/or failure under a variety of load conditions
and environmental exposures.
which serve as locations for crack starters. Hybrid
joining methods can also provide additional joint
continuity to allow increased spacing between
fasteners or welds.
Although numerous benefits are derived from
using hybrid joining techniques, and the joining of
dissimilar materials is becoming a reality, little or
no practical information is available concerning
the performance and durability of hybrid joints.
Therefore, this project has taken on the task of
developing new technologies to quantify joint
toughness and predict long-term durability. This
will necessitate identifying and developing an
understanding of key issues associated with hybrid
joint performance, such as creep, fatigue, and the
effects of environmental exposure.
Introduction
Automobile weight can be reduced and fuel
efficiency increased without compromising
structural integrity or utility by incorporating
innovative designs that strategically utilize modern
lightweight materials—such as polymeric
composites—in conjunction with traditional
structural materials such as aluminum,
magnesium, and steel. Despite the advantages
associated with dissimilar or hybrid material
systems, there is a reluctance to adopt them for
primary structural applications. In part, this
reluctance can be attributed to the limited
knowledge of joining techniques for such
disparate materials. Traditional fastening methods
such as welding, riveting, screw type fasteners,
and bolted joints may not be appropriate.
One solution to this problem is the use of
hybrid joining techniques in which a combination
of two or more fastening methods is employed to
attach similar or dissimilar materials. One example
is a mechanically fastened joint (i.e., bolted or
riveted) that is also bonded with adhesive. These
types of joints could provide a compromise
between using a familiar mechanical attachment
that has proven reliability, and reducing
problematic issues such as stress concentrations
and crack nucleation sites introduced by using
mechanical fasteners with polymeric composites.
The use of hybrid joining could also lead to
other benefits such as increased joint rigidity,
which contributes to overall stiffness gains and
reduced vehicle mass. Additionally, adhesives
used in conjunction with mechanical fasteners
could significantly reduce stress concentrations,
Candidate Hybrid Joint Selection
To initiate this study, it was necessary to
choose a candidate hybrid joint representative of
those typically encountered in automobiles.
Because of their wide applicability in automotive
structures, several combinations of hat-section
geometries were considered. Hat sections can be
incorporated into a variety of generic automotive
structural components, such as crush-tubes or
frame rails, when they are bonded and
mechanically fastened to other geometries. For the
current study, the Joining Task Force selected a
composite hat-section bonded and riveted to a
steel base as shown in Figure 1. This selection was
made on the basis of general applicability to a
192
Automotive Lightweighting Materials
FY 2001 Progress Report
Figure 1. Candidate hybrid joint consisting of a
composite hat-section bonded and riveted to
a steel plate.
variety of automobile structural components.
Members of the Joining Task Force identified
industry partners for sources for the steel, rivets,
composite hat section, and adhesive.
Figure 3. Torsion test setup in the AT machine (note:
the setup used a 600 in.-lb torque cell).
Adhesive Characterization
Characterization of the adhesive for the hybrid
joint was carried out at Oak Ridge National
Laboratory (ORNL) and at the University of
Michigan in a joint project sponsored by the ACC.
The motivation for this collaboration was a mutual
need of the two programs for material properties
and the desire of the Hybrid Joining Project to
obtain independent data from a second source.
Testing consisted of flat coupon tensile tests
for axial stiffness, strength, and stress-strain
behavior and torsional testing of cylindrical
specimens for shear properties. Failed tensile
specimens are shown in Figure 2, where typical
Figure 4. Exotherm resulting in obvious porosity in
the first exploratory specimen yielded much
lower torsional strength.
Figure 2. Cast-panel tensile dog-bone specimens
revealed the presence of trapped gas
bubbles.
bubbles are observed on the fracture surface as a
result of gases trapped during casting of the flat
specimens. A torsional specimen prior to testing is
shown in Figure 3. Note that all cylindrical
specimens were manufactured using curing
procedures newly developed at ORNL to prevent
exothermic reactions (see Figures 4 and 5), which
result in large material porosity from combustion
of the adhesive. Additionally, it was necessary to
Figure 5. Improved cure with “pre-gel” of next six
specimens resulted in no noticeable porosity
and significant improvement in specimen
torsional strength.
193
FY 2001 Progress Report
Automotive Lightweighting Materials
establish new test procedures and apparati to
accommodate the cylindrical specimens.
Typical results for the tensile and shear tests
are plotted in Figures 6 and 7. Reduction of the
data indicated that consistent stiffness and strength
results were obtained between the tests carried out
at ORNL and the University of Michigan; the
consistent results indicated that specimen
preparation and testing were comparable. In
addition, calculation of Poisson’s ratio from the
tensile and shear data produced results typical of
the adhesive system.
Modeling Efforts
A primary long-term goal of this project is to
predict the durability of the hybrid joint structure
under load, using detailed models that account for
interactions between the dissimilar materials,
adhesive bonds, and mechanical fasteners.
However, we decided to initiate the analysis for
this project by constructing simpler models to
estimate stresses prior to running the first tests.
The results from this work can be used to
determine the locations of stress concentrations
and likely damage initiation sites. Additionally,
this analysis will provide some indications of the
types of failure modes (e.g., crack growth,
material yielding) that can be anticipated during
the initial testing.
The first of these models is a simple
composite beam analysis based on a twodimensional mechanics-of-materials approach,
which accounts for the dissimilar material
properties in the response of the structure but
ignores the adhesive and the rivets. The interface
between the disparate materials is considered a
rigid joint. The second model is a threedimensional finite element analysis that includes
the adhesive layer but excludes the rivets. This
model provides details of the stress field
throughout the structure, especially variations of
the shear stresses, which are not addressed in the
first model.
These models will serve as an important
starting point for the long-term modeling goals
projected to be initiated in the final years of the
research program. A typical plot of shearing
stresses obtained from the finite element model is
shown in Figure 8.
654 ETG Stress-Strain Curves
60
654-1
654-3
50
654-5
654-12
654-13
Stress [MPa]
40
654-14
30
20
10
0
0.0
0.5
1.0
1.5
2.0
2.5
Strain [%]
Figure 6. Stress–strain curve for the flat adhesive tensile
specimens.
E654 ETG Shear-Stress vs. Shear Strain Curves
ENGINEERING SHEAR STRESS (MPa)
70
60
50
40
Mechanical Test Set-Up
30
SPECIMEN #1
SPECIMEN #2
SPECIMEN #3
SPECIMEN #4
SPECIMEN #5
SPECIMEN #6
20
10
To carry out the hybrid joint structural tests, it
was necessary to modify the load train fixtures to
accommodate the newly developed rail geometry
of the stiffened hat-joint specimen. Additionally,
the data acquisition system (computer and
associated instrumentation) was updated to the
current version of LabView (a data acquisition and
control graphical programming language). All data
acquisition and control software is being recompiled to be consistent with the latest versions
of the programming software.
0
0
2
4
6
8
10
12
14
16
ENGINEERING SHEAR STRAIN (%)
Figure 7. Shear stress vs shear strain test results for the
torsional adhesive specimens.
194
Automotive Lightweighting Materials
FY 2001 Progress Report
Upon the arrival of the steel being supplied,
expected in the near future, all aspects of the
mechanical test system will be ready and will be
ebugged to begin the first exploratory tests with
the new specimen.
Conclusions
Although this project was started relatively
late in the fiscal year, a proficient start has yielded
excellent progress and all planned tasks were
accomplished on schedule. Specifically, the
identification of a candidate joint meaningful to
automobile industry partners has been selected,
preliminary testing of the adhesive system has
been completed, and preliminary modeling of the
hybrid joint has provided the necessary
information to proceed with the static tests slated
to start during the first quarter of FY 2002.
Additionally, all necessary modifications to the
test apparatus identified at this time have been
completed. With the groundwork now completed,
it is estimated that there will be little or no delay in
accomplishing the next year’s objectives.
Figure 8. Preliminary finite element model results: shear
stress for the hybrid joint hat-section
specimen.
Preliminary testing within the linear region of
a hat section without the steel base was conducted
to estimate the required load capacity. Based on
the mechanical response of the specimen, this
exploratory test warranted installation of a larger
load cell (10 kip, 2.2 kN) to achieve the desired
levels of displacement for the upcoming tests.
195
Automotive Lightweighting Materials
FY 2001 Progress Report
J. Joining of Dissimilar Metals for Automotive Applications: From Process to
Performance
Principal Investigator: Moe Khaleel
(509) 375-2438; fax: (509) 375-6605; e-mail: [email protected]
USCAR Joining Team Point of Contact: Jim Quinn
(248) 680-4732; fax: (248) 680-2874; e-mail: [email protected]
DOE Program Manager: Joseph A. Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: Pacific Northwest National Laboratory
Contract No.: DE-AC06-76RL01830
Objectives
•
Develop and evaluate different joining technologies for dissimilar aluminum alloys and for joining aluminum to
steel.
•
Characterize the performance of these joints.
•
Develop a unified modeling procedure to represent these joints in vehicle structural simulation. The steel
materials include mild, high-strength low-alloy (HSLA), and dual-phase steels.
OAAT R&D Plan: Task 6; Barriers C, D
Approach
•
Further develop and/or enhance self-piercing rivets and resistance spot welding, with and without adhesives, for
joining dissimilar metals.
•
Develop a database for the static, dynamic, fatigue, and corrosion behavior for dissimilar material joints
consisting of different material selections and different joining techniques.
•
Incorporate and represent the joint performance data into current computer-aided engineering (CAE) codes for
evaluation of the impact and fatigue performance of joint components.
•
Develop design guidelines in the form of tables and charts for use in joint structural and crash designs.
Accomplishments
•
Selected two promising joining techniques for joining dissimilar metals.
•
Experimentally and numerically investigated the effects of specimen design on reported joint performance.
Future Direction
•
Join dissimilar aluminum alloys (AA5182 and AA6111) and steel (SAE 1010, HSLA, dual-phase) using selfpiercing rivets and resistance spot welding with and without adhesives.
•
Develop interlayer material for the use in spot welding of dissimilar metals.
197
FY 2001 Progress Report
Automotive Lightweighting Materials
•
Conduct mechanical and microstructural evaluation of dissimilar metal joints subjected to static, dynamic, and
fatigue loads. In addition, investigate the effect of structural adhesives, temperature, and loading rates.
•
Develop joint failure criteria for CAE safety and fatigue simulations based on the performance database
achieved in the experiments.
•
Develop design guidelines in the form of a design space that contains strength, energy absorption, normalized
dimensional parameters, material combinations, and so on.
Introduction
Joint Selection and Development
This project, which started in April 2001, is a
collaborative effort among DOE, Pacific Northwest
National Laboratory (PNNL), and the metals joining
team of the U.S. Council on Automotive Research.
The automotive industry envisions that the
optimized vehicle, in terms of performance and cost,
can be achieved only by using different materials at
different vehicle locations to utilize the materials’
functionalities to the fullest extent. Currently,
aluminum and steel are the most important
construction materials for the mass production of
automotive structures. High-volume non-steel
joining is a significant new problem to the industry.
For joining dissimilar aluminum alloys, the leading
candidate joining methods are spot welding and selfpiercing rivets, with or without adhesives. The major
concerns regarding aluminum spot welding are its
high energy consumption, low electrode life, and
structural performance concerns related to weld
porosity. The technology for joining aluminum to
steel with which the industry is currently
comfortable is self-piercing rivets (with and without
adhesives). However, there are a number of barriers
to the widespread exploitation and high-volume
production of the riveting technology, one of which
is limited performance data relative to automotive
applications.
On the other hand, in order to shorten the
vehicle development cycle, more and more CAE
analyses are performed before the actual prototype is
built. The question CAE engineers ask most often is
how to represent the structural joints in crash
simulation and fatigue simulation. Currently, there is
no unified approach to representing structural joints
that works for different material combinations under
multi-axial loading. This is particularly true for
dissimilar material joints, where even the basic
performance information on the joint coupon level
does not exist.
The U.S. Automotive Materials Partnership
(USAMP) joining team and PNNL staff selected the
most relevant material set and material gage based
upon the most relevant vehicle applications. The
material set used to fabricate dissimilar metal joints
includes
• steel (mild steel: SAE 1010; HSLA; ultra-high
strength steel: dual-phase steel)
• aluminum (AA 5182 and 6111)
• combinations of aluminum and steel
The type of adhesives to be used in the joint
fabrication has also been identified by the project
team. During the first phase of the program, Dow’s
Betamate 4601 and Eftec’s WC2309 will be
examined. These adhesives will be applied in
addition to the structural joint (weld or rivet);
therefore, the role of the adhesives on the impact,
fatigue, and corrosion performance of the joint can
be evaluated.
Joint Performance Characterization
It is well known in the automotive welding
community that joint performance can be greatly
influenced by different attributes of the joint sample
design. However, the lack of a unified joint design
standard—including the American Society for
Testing and Materials (ASTM), American Welding
Society, and International Institute of Welding
(IIW)—has hampered the integration of joint
performance data obtained from different sources.
To overcome the joint design issues and ensure the
validity of the performance data obtained during the
later phase of the program, a coupon design
sensitivity study is being performed for tensile shear
specimens of both steel spot welds and aluminum
self-piercing rivets.
Coupon design factors such as coupon width,
sheet overlap, and the effect of forming a C-channel
on the coupon flange have been investigated
198
Automotive Lightweighting Materials
FY 2001 Progress Report
experimentally and numerically. Since some of the
joining team members are also members of the
Auto/Steel Partnership (A/SP) task force on highstrength steel joint performance, and since A/SP has
used the C-channeled tensile shear design
previously, coupons with C-channels are also
included in our initial coupon design table. For
example, Figure 1 shows the measured quasi-static
load versus displacement curves for 1.0-mm
SAE1008 tensile shear spot-welded coupons with
different coupon widths and C-channels. The weld
diameter used is 5.0 mm.
absorption should be evaluated using a coupon
width of larger than 50 mm or a coupon with a Cchannel for this joint.
Similar experiments have been carried out for
tensile shear riveted aluminum coupons. For
example, Figure 2 shows the quasi-static testing
results for 2-mm AA6111 riveted tensile shear
coupons with different widths and C-channels. A
similar conclusion can be derived that a coupon
width of 50 mm should be sufficient to reduce the
sheet bending effect and, therefore, capture the true
rivet strength and displacement at failure for the
energy absorption.
Experimental Results Showing the Effect of Specimen Design on Measured Mechanical Properties
Specimens were Spot Weld Coupons Prepared from Mild Steel - Test Speed was Quasistatic
2000
C-shaped
AA6111-T4
Sheet Thickness = 2mm, Self-Piercing Rivet (5mm)
40mm Wide
1800
2000
35mm Wide
30mm Wide
25mm Wide
1800
20mm Wide
1400
1600
1200
1400
Axial Load (lbs)
Axial Load (lbs)
1600
1000
800
L-36-25MM
50mm Wide
50mm Wide
C-shaped
20mmx105mm-1"
25mmx105mm-1"
30mmx105mm-1"
35mmx105mm-1
40mmx105mm-1
1200
50mmx105mm-1
C-shapedx105mm-1
20mm W ide
1000
800
L-36-25MM
600
L-45-30MM
600
L-14-20MM
L-11 FORMED
400
L-13 FORMED
400
L-16 35MM
200
L-44 40MM
L-40 50MM
200
L-39 50MM
0
0.25
0.35
0.45
0.55
0.65
0.75
0
-1.4
Displacement (in)
-1.3
-1.2
-1.1
-1
-0.9
Displacement (in)
Figure 1. Experimental quasi-static measurement of
tensile-shear spot-welded samples of different
coupon designs.
Figure 2. Experimental quasi-static measurement of
tensile-shear riveted aluminum samples of
different coupon designs.
Static weld tensile shear measurements shown in
Figure 1 validate the conclusion that weld
performance varies with joint coupon design. The
joint strength—that is, the peak load value of the
load-displacement curve—increases with increasing
coupon width and converges to a single curve after a
coupon width of 50 mm is reached. The
displacement at peak load increases with decreasing
coupon width because the overall cross-section
moment of inertia is reduced when coupon width is
reduced. Therefore, coupon rotation is more
prominent during the deformation process. This
large displacement at failure does not indicate the
weld ductility; rather, it is caused by the rotation of
the two sheets during the deformation process. On
the other hand, when the coupon width is increased
or a C-channel flange is introduced, the overall
cross-section moment of inertia is increased, and
there is little or no coupon rotation during weld
failure. Therefore, the true weld strength and energy
Based on these experimental results, the joining
team and PNNL staff decided to use the flat tensileshear samples with a 50-mm coupon width to
achieve the convergent joint strength.
Similar work on joint design issues of crosstension samples is being conducted at PNNL and
will be reported during the next quarterly review
meeting.
Joint CAE
To allow flexibility in investigating the
influence of different weld attributes to determine
the critical coupon width for tensile shear tests, a
generic finite element procedure was developed. It
allows the coupon width, weld diameter, sheet
thickness, and sheet metal overlap to be changed
parametrically so that the effect of different design
attributes can be studied. This work parallels the
experimental work, and it enables the study of more
199
FY 2001 Progress Report
Automotive Lightweighting Materials
parameter combinations than the experiment could
easily achieve. The model is validated by comparing
the predicted load versus displacement curves. The
experimental results are shown in Figure 1.
Figure 3 shows the model geometry and the
detailed failure mode of the weld region. This
generic analysis tool is used to study the coupon
width and overlap effect. Figure 4 shows the
predicted coupon width effect on the overall coupon
rotation at failure, and Figure 5 shows the predicted
overlap effect on static load-displacement behavior.
The results show that the overlap width is not a
critical factor influencing the static strength of the
joint. Above 20 mm, the predicted static behavior of
the samples with different overlap widths is almost
the same. However, it should be mentioned that the
minimal coupon width needs to be observed to avoid
the bearing deformation of the parent near the rivet
(see Figure 6).
Figure 3. Model geometry and detailed failure mode of weld region.
Width = 20mm
Width = 70mm
Figure 4. Predicted coupon-width effect on tensile shear samples.
200
Automotive Lightweighting Materials
FY 2001 Progress Report
P redicted effect of the value of coupon overlap
2000
1800
1600
Loa d (lbs.)
1400
1200
w= 50m m , overlap=35m m
1000
800
w= 50m m , overlap=20m m
600
400
w= 50m m , overlap=50m m
200
0
0
0.5
1
1.5
2
2.5
Displa ce m e nt (m m )
Figure 5. Predicted effect of coupon overlap.
Figure 6. Bearing deformation
occurs in the base
metal since the overlap
width is too small.
201
Automotive Lightweighting Materials
FY 2001 Progress Report
K. Technical Cost Modeling
Sujit Das
Oak Ridge National Laboratory
Oak Ridge, TN, 37831-6205
(865) 574-5182; fax: (865) 574-8884; e-mail: [email protected]
DOE Program Manager: Joseph A. Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: Oak Ridge National Laboratory
Contract No.: DE-AC05-00OR22725
Objectives
•
Address the economic viability of new and existing lightweight materials technologies.
•
Develop technical cost models to estimate the cost of lightweight materials technologies.
OATT R&D Plan: Task 8, Barrier A
Approach
•
Address the economic viability of lightweight materials technologies supported by the Automotive Lightweight
Materials (ALM) program.
•
Use cost modeling to estimate specific technology improvements and major cost drivers that are detrimental to
the economic viability of these new technologies.
•
Derive cost estimates based on a fair representation of the technical and economic parameters of each process
step.
•
Provide technical cost models and/or evaluations of the “realism” of cost projections of lightweight materials
projects under consideration for ALM Program funding.
•
Examine technical cost models of lightweight materials technologies including (but are not limited to) aluminum sheet; carbon fiber precursor and precursor processing methods; fiber-reinforced polymer composites; and
methods of producing primary aluminum, magnesium, and titanium, and magnesium alloys with adequate hightemperature properties for powertrain applications.
Accomplishments
•
Conducted a limited cost analysis of alternative precursor materials for the low-cost carbon fiber development
program.
•
Examined the economic viability of nondestructive testing (NDT) methods for testing of joints.
•
Examined the economic viability of polymer composites and how the R&D supported by ALM is responding
to the needs of the industry from an economic viability perspective.
•
Identified and tested methods appropriate for estimating the benefits attributable to research and development
projects funded by ALM.
203
FY 2001 Progress Report
Automotive Lightweighting Materials
Future Direction
•
Focus on the fabrication technologies for composites.
•
Explore the economic viability of a carbon fiber composite-intensive body-in-white.
Continue individual project-level cost modeling to identify specific technology improvements and major
cost drivers that are detrimental to the economic viability of these technologies.
Continue the benefit evaluations of the R&D projects funded by the ALM Program.
material and preprocessing either by chemical
modification or radiation can reduce the carbon fiber
cost by $1.50/lb and $1.30/lb, respectively. The cost
reduction potential is lower in the case of radiation
preprocessing, mainly because of the higher capital
cost of electron beam processing. Sulfonation of
linear low-density polyethylene has the most costreduction potential—about $1.60/lb, mainly due to
higher overall process yield (70% vs. 48% for the
conventional process). The preliminary cost analysis
further confirmed that a combination of low-cost
precursor material and enhancements in the
processing speed (particularly at the stabilization
step) would be necessary to achieve the target
carbon-fiber cost for automotive applications.
Major Accomplishments
The major accomplishments of this effort
include cost assessments of alternative precursor
materials for low-cost carbon fiber and NDT
methods for testing of joints. Another highlight was
benefit evaluations of the R&D projects supported
by the ALM Program. In addition, the composites
R&D portfolio of the ALM program was examined
in the context of satisfying the current needs of the
industry from an economic perspective.
Analysis of Alternative Precursor Materials
To produce a low-cost carbon fiber for
automotive applications in the range of $3 to $5 per
lb, several alternative precursor materials are under
consideration. To achieve the target cost, a
significant reduction is needed in precursor and
capital costs, which, combined, currently contribute
to more than 60% of the total carbon fiber
production cost. Melt-processable acrylonitrile
copolymers have the potential to reduce costs by
about $0.60/lb, compared with the conventional
solution-spinning process, by eliminating safety and
recovery costs associated with the use of solvents.
Melt-spun lignin, relatively less expensive than
polyacrylonitrile and a major by-product of the
paper and pulp industry, can further reduce the
carbon fiber cost by about $0.25/lb. The use of
commodity textile acrylic tow as the alternative
precursor material lowers the precursor cost
significantly, but the overall cost reduction potential
is only about $0.90/lb because of reduced overall
process yield and higher capital costs resulting from
a slower stabilization process and higher processing
temperature.
Other approaches considered for low-cost
carbon fiber include the preprocessing of precursors
by chemical modification or radiation in order to
enhance the stabilization step of the carbon-fiber
production process. The combination of using
commodity textile acrylic tow as the precursor
Viability of NDT Joint-Testing Methods
The cost-effectiveness of current (i.e.,
conventional pry) and potential (i.e., ultrasonic and
global resonance) NDT methods was based on a
potential demonstration of these methods on a
particular application (i.e., a compact automobile
door), assuming spot welds are being tested. The
part-test cost is estimated to be $0.60 for the global
resonance method, the least out of three NDT
processes considered in the study. The part-test costs
of the other two offline methods are considerably
higher, mostly because of the labor-intensive nature,
in the case of the conventional pry method. Labor
dominated the total part-test cost (accounting for
more than 85%); therefore, the automated online
global resonance method allowing computer-based
determination of joint integrity showed the most
favorable economics. There is no doubt that NDT
methods are more cost-effective, as more ultrasonic
testing is being used for the evaluation of spot welds
in the automobile industry today. In the future, the
automobile industry’s use of advanced lightweight
materials that require adhesive bonding will
necessitate the development of on-line
nondestructive evaluation methods to provide a
204
Automotive Lightweighting Materials
FY 2001 Progress Report
Automotive Research, is sponsoring research under
the ALM Program that seeks to overcome the
barriers to more widespread use of composites in
automotive applications. DOE is attempting to take a
comprehensive look at the research needs of the
composites industry and has prioritized certain
areas, such as low-cost carbon fiber production,
thermoplastic structural composites, and the
development of new reinforcement technologies
such as nanocomposite technology. Its research
portfolio is appropriately focused both on its
ongoing research and its 5-year research plan that
covers five major barrier areas: cost,
manufacturability, design data and test
methodologies, joining and inspection, and recycling
and repair. Although cost reduction is a pervasive
factor in all the composites R&D activities, it is
appropriate to focus the cost area on materials,
primarily carbon fiber. To improve the
manufacturability of polymer composites,
development of high-volume production
manufacturing processes should remain one of
DOE’s research priorities. More research is needed
in the areas of carbon-fiber–reinforced polymer
composites for these to be economically viable
automotive materials. It is clear that for polymer
composites to become the material of choice for
automakers, an aggressive R&D portfolio should be
followed to achieve major breakthroughs that are
necessary for several orders of magnitude of cost
reduction.
means of controlling the production process and
ensuring the integrity of the vehicles. The costeffectiveness of global resonance for spot-welds
estimated in this study can be extrapolated to weldbonded and adhesively bonded joints because the
testing methodology would be the same for these
types of joints.
Economic Viability of Polymer Composites
A literature review of the cost studies of
composites shows some general qualitative trends in
the economic viability of composites. To date, most
of the cost analyses of polymer composites are for
body-in-white (BIW) applications because of the
significant weight-reduction potential these offer.
The viability of composites is still seen
predominantly in non-structural elements, such as
the bolt-on exterior panels of today’s vehicles; and
most composites are glass-fiber–reinforced
thermoset polymers, such as sheet-molded
composites, used in niche-market vehicles with
annual production volumes of less than 80,000. At a
higher annual production volume of 250,000, for
example, an evaluation of the most efficient
composite monocoque design indicates that the cost
of glass-fiber–reinforced thermosets and carbonfiber–reinforced thermoplastics are 62% and 76%
higher, respectively, than the cost of a conventional
steel unibody.
Even on a life-cycle basis, the cost of polymer
composites is estimated to be higher than that of
steel unibodies. For composites to be costcompetitive in a part-by-part substitution,
improvements are necessary in cycle times and
material utilization, which in some cases have been
estimated to contribute 60% and 21%, respectively,
of the total cost of carbon-fiber–reinforced
thermoplastics. The material cost plays a key role in
the economic viability of polymer composites,
particularly at higher production volumes and for
carbon-fiber–reinforced thermoplastic composites.
The carbon-fiber cost needs to drop by 50% (i.e., to
the $3–$5/lb range), and smaller cost reductions in
other thermoplastic materials are needed for these
composites to be economically viable.
Estimation of Benefits
Through a thorough literature review, three
methods were identified to estimate benefits
attributable to ALM R&D projects: qualitative
assessment, National Research Council indicators,
and benefit-cost analysis. This combination
addresses all-important aspects of the benefits of
R&D projects. The qualitative assessment addresses
short-term project outcomes at the project level. The
indicators address standard measures associated with
the quality of research projects. The benefit-cost
analysis addresses both short-term and long-term
benefits associated with commercializing new
technologies. Since ALM projects encompass both
the creation of new knowledge and the
commercialization of new technologies, both the
indicator and benefit-cost approaches were found to
Responding to the Needs of the Composites
Industry
DOE, in partnership with the Automotive
Composites Consortium of the U.S. Council for
205
FY 2001 Progress Report
Automotive Lightweighting Materials
Quantity of Composites (million
lbs)
be appropriate for a comprehensive benefits
assessment.
Three ALM R&D projects evaluated for the
application of these benefit assessment methods
were low-cost, continuous-cast aluminum sheet;
advanced forming technologies for aluminum; and
manufacturing of composite automotive structures.
Figure 1 shows the market forecasts for the
manufacturing of composite automotive structures,
which were used to estimate its benefits and costs.
The three projects assessed all appear to have
yielded high levels of benefits. Based on the
qualitative assessment, all met their technical goals,
increased knowledge, and led to increased
collaboration. It is not likely that these projects
would have been undertaken without federal
support, or at least undertaken as soon as they were.
The only open question pertains to
commercialization. With respect to the National
Research Council indicators, there is a mixed rate of
publication, but each project is shown to improve
U.S. competitiveness in its respective area. The
benefit-cost analysis yields impressive benefit-tocost ratios, ranging from 168:1 to 699:1, when
environmental benefits are taken into consideration.
In order to improve the assessment of long-term
benefits, publications and market penetration rates
should be periodically revisited, and case studies
need to be undertaken to identify any valuable spinoff technologies.
600
500
400
Semi-structural Passenger
Cars
Open cargo truck boxes
300
200
100
0
2001
2005
2009
2013
2017
2021
2025
Year
Figure 1. Forecasts of manufacturing of automotive
composite structures.
206
Automotive Lightweighting Materials
FY 2001 Progress Report
L. Nondestructive Evaluation Techniques for On-Line Inspection of Automotive
Structures
Principal Investigator: Deborah Hopkins, Ph.D.
Lawrence Berkeley National Laboratory
1 Cyclotron Road, MS 46A-1123B, Berkeley, CA 94720
(510) 486-4922; fax: (510) 486-4711; e-mail: [email protected]
Project Manager, Composites: C. David Warren
Oak Ridge National Laboratory
P.O. Box 2009, Oak Ridge, TN 37831-8050
(865) 574-9693; fax: (865) 574-0740; e-mail: [email protected]
DOE Program Manager: Joseph A. Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Participants
NDE Steering Committee composed of representatives from DaimlerChrysler, Ford Motor Co.,
and General Motors
James Triplett, Director of Engineering, Lawrence Berkeley National Laboratory
Frédéric Reverdy, Ph.D., Lawrence Berkeley National Laboratory
Daniel Türler, M.S., Lawrence Berkeley National Laboratory
Murat Karaca, Ph.D., Lawrence Berkeley National Laboratory
Contractor: Lawrence Berkeley National Laboratory
Contract No.: DE-AC03-765F0095
Objective
•
Evaluate and develop nondestructive evaluation (NDE) and testing techniques that are sufficiently fast, robust
in manufacturing environments, accurate, and cost-effective to be suitable for on-line inspection of automotive
structures.
OAAT R&D Plan: Task 6, Barriers A, C, D
Approach
•
Evaluate the state of the art of existing NDE technologies in conjunction with efforts to develop improved
systems and new techniques that are necessary either to meet industry requirements or to meet the challenges
posed by the introduction of lightweight materials.
•
To develop the necessary technology, perform laboratory experiments in parallel with the development of
sensors, models, data-acquisition systems, and data post-processing software toward the goal of producing
prototype inspection systems for specific automotive-manufacturing applications.
Accomplishments
•
Carried out site visits to commercial vendors of spot-weld-inspection systems and phased-array technologies.
207
FY 2000 Progress Report
Automotive Lightweighting Materials
•
Conducted laboratory experiments using acoustic and thermographic techniques to evaluate conventional and
phased-array spot-weld inspection systems.
•
Developed a finite-difference code that allows modeling of wave propagation in anisotropic and viscoelastic
media to study propagation of acoustic waves in spot-welded samples.
•
Evaluated the state of the art of phased-array technologies and their suitability for spot-weld inspection.
Future Direction
•
Evaluate commercially available spot-weld inspection systems; summarize in a report the results of round-robin
evaluations completed by vendors and researchers.
•
Identify and describe technologies viable in a manufacturing environment with the short- and long-term
potential to improve the accuracy, robustness, speed, cost-effectiveness, and ease of use and implementation of
spot-weld inspection systems.
•
In conjunction with automobile manufacturers and suppliers, and government and university researchers,
develop a roadmap to identify barriers to increased used of NDE in automobile manufacturing and determine
strategies for implementing NDE inspection and process-monitoring techniques for specific applications.
Technologies, Inc.; Edison Welding Institute; and
Kumamoto University (Japan).
Introduction
The NDE Industry Steering Committee is
composed of representatives from DaimlerChrysler,
Ford, and General Motors with expertise in
manufacturing, materials, and NDE. The committee
asked Lawrence Berkeley National Laboratory
(LBNL) to evaluate the state of the art in spot-weld
inspection and to determine the potential of acoustic
phased arrays for inspecting welds. To accomplish
these tasks, LBNL was asked to attend the annual
Review of Progress in Quantitative Nondestructive
Evaluation (QNDE) to learn more about recent
advances in the development of phased-array
techniques; evaluate commercially available spotweld inspection systems; visit those
vendors/institutions with the most promising
conventional and phased-array systems; and
evaluate the results of a round-robin experiment in
which test specimens prepared by Ford are being
inspected by vendors and institutions with applicable
technologies.
To date, LBNL has visited IRT Scanmaster
Systems; Krautkramer, Inc.; Sonoscan, Inc.;
Imasonic SA; and Imperium, Inc. Meetings have
been held with representatives from R/D Tech;
Panametrics, Inc.; Commission de l’Energie
Atomique (CEA, France); FORCE Institute
(Denmark); Laboratoire de Mécanique Physique
(University of Bordeaux, France); and Ohio State
University. LBNL is in contact with Innerspec
Project Deliverables
•
•
•
Report summarizing the results of round-robin
sample evaluations performed by commercial
vendors.
Paper summarizing the laboratory experiments
and modeling studies, suitable for publication.
Roadmap of strategies to increase the use of
NDE in automotive applications.
Planned Approach
Assessment of the state of the art in
conventional and phased-array spot-weld inspection
is being accomplished by attending professional
technical meetings; consulting with industry,
government, and academic experts; visiting vendors
and institutions with promising technologies; and
analyzing the results of measurements performed by
commercial vendors on test specimens provided by
industry partners. In addition, LBNL is conducting
laboratory experiments and modeling studies as
necessary to evaluate phased-array technology and
commercially available spot-weld inspection
systems.
The round-robin testing being performed by
commercial vendors is focused on determination of
the size of the weld nugget, which is an industry
criterion for distinguishing between good and bad
welds.1 Work being performed by LBNL is more
208
Automotive Lightweighting Materials
FY 2001 Progress Report
broadly focused; NDE techniques are being
evaluated based on
•
•
•
•
•
their ability to identify the many different kinds
of defects that occur in spot welds, including
inclusions and voids
their sensitivity to surface conditions, including
zones of zinc expulsion and surface roughness
their ability to inspect welds used in a variety of
configurations, including triple-thickness joints
and joints formed between sheets with different
material properties and thicknesses
their potential for miniaturization as necessary to
inspect welds in locations where access is
severely restricted
the level of operator training required.
Figure 1. Five spot welds joining stainless-steel plates.
The quality of the welds was varied by varying
the welding current between 5,800 (first weld)
and 10,500 A (fifth weld).
more complicated to evaluate because the zinc
coating has a lower melting temperature than the
base material; the zinc tends to be expelled from the
fused zone, forming a weak solder joint around the
weld nugget.
The acoustic microscopy results for the spotwelded sample shown in Figure 1 are displayed in
Figure 2. Blue corresponds to areas where
reflections have relatively low amplitude compared
with areas with high-amplitude reflections, shown in
red. The scans were performed with the transducer
focused at the interface between the two steel sheets.
In the area of the weld nugget, where the steel sheets
are fused together, only small-amplitude reflections
are observed. Outside of the fused zone, the
interface between the sheets generates highamplitude reflections. Defects in the weld nugget are
also visible; they can be observed in the first three
images, where defects are evident as small bumps,
indicating reflections with relatively high
amplitudes. Work under way includes determining
the relationship between the low-amplitude fused
zone visible in the images, and the size of the weld
nugget measured when the joint is peeled open.
A finite-difference model was developed to
study acoustic-wave propagation across complicated
boundaries such as those around the weld nugget
and inclusions inside the weld. The code allows
modeling of wave propagation in anisotropic and
viscoelastic media. A staggered grid was
implemented to increase the accuracy of the model
without increasing the computation time. The
method achieves second-order accuracy in time and
fourth-order accuracy in space. Viscoelastic
behavior is modeled using a tensor that allows
different attenuation coefficients to be specified for
different directions of propagation. Figure 3 shows a
micrograph of a spot weld containing zinc
inclusions. The finite-difference code described
earlier was used to model the weld pictured in
Figure 3. Figure 4 shows two snapshots of acoustic
Viable systems must also be fast, robust, and
able to issue a warning if they are not properly
deployed.
Emerging phased-array technologies and
commercial spot-weld-inspection systems are
ultrasonic techniques. Evaluating these systems
requires a thorough understanding of acoustic-wave
propagation in spot-welded joints, including how
acoustic waves are affected by defects in the welds,
changes in material properties in the heat-affected
zone, surface roughness, coupling between the
sensor and test specimen, and orientation of the
sensor. Laboratory and modeling studies were
performed to help us understand these issues and
interpret the results obtained with commercial
systems.
Acoustic microscopy is a high-resolution
technique used to create images of surfaces and bulk
microstructures.2,3 A focused transducer generating
acoustic waves is used to scan the test specimen;
images are generated by measuring the amplitude of
reflected waves, which is directly proportional to
changes in impedance in the sample. Acoustic
microscopy measurements were made at the
Laboratoire de Mécanique Physique in France. A
stainless-steel test specimen with five spot welds is
pictured in Figure 1. The quality of the welds was
varied by varying the welding current between 5,800
and 10,500 A. The surface of the test specimen was
polished because of the sensitivity of the method to
surface conditions. Experiments are being conducted
in several types of steel; stainless was used here to
obtain baseline data. Welds in galvanized steel are
209
FY 2000 Progress Report
Automotive Lightweighting Materials
Figure 2. Acoustic microscopy images of the five spot welds pictured in Figure 1. Blue and red correspond to low- and
high-amplitude reflected waves, respectively. The elements in the grid are 0.25-mm square.
Conclusions
Commercially available ultrasonic spot-weld
inspection systems are based on time of flight and
wave attenuation. These systems are relatively
inexpensive and are widely used in Europe.
Measurements are very sensitive to positioning of
the probe, and operator training is critical. Work by
LBNL and others suggests that the measurements
are also affected by surface roughness and
indentation of the surface caused by the welding
electrodes. Existing commercial systems may not be
viable for inspecting welds in aluminum or highstrength steel. However, modifications to these
systems as required to improve performance may be
a more cost-effective approach for certain
applications than development of phased-array
systems.
Industry’s NDE Steering Committee asked
LBNL to evaluate ultrasonic phased arrays against
several criteria, including sensor technology,
maturity of technology, industrial applications,
robustness, and cost. Although phased arrays
provide the ability to scan, focus, and steer with a
fixed probe, they do not change what can be
measured; that is, things that cannot be measured
with conventional ultrasonic techniques cannot be
measured with phased arrays. Based on site visits,
discussions with experts, and a review of the
literature:
• The number of elements in an array is not an
issue except for cost, which is primarily in
electronics.
• Although different materials are available and
used for transducers, piezocomposites appear to
be the most appropriate material for arrays
because it is easier to minimize acoustic
interaction between elements in them than with
other materials.
Figure 3. Cross section of a spot weld joining steel
plates. Two inclusions in the weld nugget are
clearly visible.
waves propagating in the sample; the dark lines
represent the propagating waves, before (Figure 4a)
and after (Figure 4b) the waves interact with the
interfaces and defects in the sample that cause
reflection and diffraction of the waves.
Work performed previously demonstrated that
thermographic images of spot welds can be analyzed
to provide information on weld quality4–7. For work
in progress, thermographic images are being used to
provide independent data to help interpret acoustic
measurements. Infrared (IR) imaging radiometers
are used to measure thermal-radiation energy; any
change in thermal resistance in the direction of the
heat flux results in a change in the temperature
gradient. The spot-welded sample that produced the
acoustic images displayed in Figure 2 was also
evaluated using steady-state IR thermography. The
resulting thermogram is shown in Figure 5; red
corresponds to relatively warm areas compared with
surrounding cooler areas, indicated in blue. The
results show close agreement with acoustic
microscopy measurements. Three spot welds appear
to be satisfactory, while the two others appear to be
undersized. These results will be confirmed when
the welds are peeled open to reveal the size of the
weld nugget.
210
Automotive Lightweighting Materials
FY 2001 Progress Report
Figure 4a. Snapshot in time of acoustic waves propagating in a spot-welded sample
(Figure 3) calculated using a finite-difference model developed to help interpret
results obtained from spot-weld inspection systems. The image shows
longitudinal, shear, conical, and surface waves generated from a source located
at the center of the bottom surface. The image shows the waves propagating
from the source before they interact with the weld nugget (red box) and
interfaces between the two steel sheets.
Figure 4b. Snapshot of propagating acoustic waves at a later time than that captured in
Figure 4a, showing interaction of the waves with the weld nugget, inclusions
inside the nugget, and the interface between the steel sheets. Reflections, mode
conversion, and diffraction can be observed.
Figure 5. Thermograms of the five spot welds pictured
in Figure 1. Red in the welds corresponds to
relatively warm areas compared with
surrounding cooler areas, shown in blue.
211
FY 2000 Progress Report
•
•
•
•
•
•
Automotive Lightweighting Materials
3. R. A. Lemons and C. F. Quate, “Acoustic
Microscope—Scanning Version,” Applied Physics
Letters, 24 (4), pp. 163–165 (February 1974).
4. S. Satonaka, H. Ohba and K. Shinozaki,
“Nondestructive Evaluation of Weld Defects by
Infrared Thermography,” Materials Engineering,
vol. 3, pp. 305–312, American Society of
Mechanical Engineers, 1995.
5. S. Satonaka, K. Shinozaki, S. Arima,
T. Nishiwaki and Y. Kohno, “Nondestructive
Evaluation of Spot Welds by Infrared
Thermography,” International Institute of Welding
Doc.3, pp. 1064–1096.
6. Daniel Türler, “Predicting the Geometry and
Location of Defects in Adhesive and Spot-Welded
Lap Joints Using Steady-State Thermographic
Techniques,” in Proc. SPIE, the International
Society for Optical Engineering, Vol. 3700, Dennis
H. LeMieux and John R. Snell, Jr., eds., 1999.
7. Daniel Türler, Deborah Hopkins, Seiji
Nakagawa, António Valente, and Kurt Nihei,
“Thermographic and Acoustic Imaging of SpotWelded and Weld-Bonded Joints,” Review of
Progress in Quantitative Nondestructive Evaluation,
vol. 18, D. O. Thompson and D. E. Chimenti, eds.,
1999.
Commercially available arrays all allow
scanning, as well as steering and focusing of the
acoustic beam with a single probe.
At present, there are limitations on element size
and frequency. The smallest element available is
about 0.5 mm; smaller elements can be
manufactured, but as the size decreases, the
stress acting on the element increases, making it
more difficult to control the frequency. The
highest frequency that can be obtained at present
is on the order of 18 MHz.
Electronics are the most expensive components
in a phased-array system, typically costing more
than ten times as much as the probe.
Costs are likely to drop in the future with
increased competition.
Surface roughness and coupling are still issues.
Problems in interpretation are similar to those
with conventional systems; for example, for
galvanized steel, it may be difficult to
differentiate between the heat-affected zone,
zinc solder joint, and weld nugget.
At present, no 2-dimensional matrix array is
commercially available; systems have to be
custom-ordered and built for a particular
application.
Publications
Because the measurements made with phased
arrays are comparable to those made with
conventional ultrasonics, existing inspection systems
and laboratory measurements can help determine the
viability of phased arrays. Issues including cost and
manufacturability may limit the viability of phased
arrays for spot-weld inspection in the near future. If
so, modifications to existing systems to improve
their performance, and ease of use may be a costeffective solution for the near term. In addition,
recent advances in computer science, statistics, and
signal processing provide new opportunities for inprocess monitoring and for reducing the reliance on
trained operators.
1. Deborah Hopkins, Daniel Türler, Seiji
Nakagawa, Kurt Nihei, and Guillaume Neau, “OnLine Nondestructive Inspection Techniques for
Lightweight Automotive Structures,” Society of
Automotive Engineers, Paper no. 00FCC-124, 2000.
2. Daniel Türler, Deborah Hopkins, Seiji
Nakagawa, António Valente, and Kurt Nihei,
“Nondestructive Evaluation of Spot-Welded and
Weld-Bonded Joints,” submitted to the Society of
Automotive Engineers, 2000.
3. Guillaume Neau, Deborah Hopkins, Seiji
Nakagawa, and Kurt Nihei, “Complications of Using
Resonance–Frequency Shifts to Detect Defective
Joints,” Review of Progress in Quantitative
Nondestructive Evaluation, vol. 19, D. O. Thompson
and D. E. Chimenti, eds.
4. Daniel Türler, “Predicting the Geometry and
Location of Defects in Adhesive and Spot-Welded
Lap Joints Using Steady-State Thermographic
Techniques,” in Proc. SPIE, the International
Society for Optical Engineering, Vol. 3700, Dennis
H. LeMieux and John R. Snell, Jr., eds., 1999.
References
1. T. M. Mansour, “Ultrasonic Inspection of
Spot Welds in Thin-Gauge Steel,” Materials
Evaluation, pp. 650–658, April 1988.
2. L. W. Kessler, P. R. Palermo, and A. Korpel,
“Practical High Resolution Acoustic Microscopy,”
Acoustical Holography, vol. 4, pp. 51–71, G. Wade,
ed., New York: Plenum Press, 1972.
212
Automotive Lightweighting Materials
FY 2001 Progress Report
Wavefield,” Review of Progress in Quantitative
Nondestructive Evaluation, vol. 18, D. O. Thompson
and D. E. Chimenti, eds., 1999.
7. Deborah Hopkins, Seiji Nakagawa, Kurt
Nihei, and Daniel Türler, “Imaging Flaws in
Adhesive Joints Using Acoustic Techniques and
Infrared Thermography,” in Review of Progress in
Quantitative Nondestructive Evaluation, vol. 17,
D. O. Thompson and D. E. Chimenti, eds., 1998.
5. Daniel Türler, Deborah Hopkins, Seiji
Nakagawa, António Valente, and Kurt Nihei,
“Thermographic and Acoustic Imaging of SpotWelded and Weld-Bonded Joints,” Review of
Progress in Quantitative Nondestructive Evaluation,
vol. 18, D. O. Thompson and D. E. Chimenti, eds.,
1999.
6. Kurt Nihei, Seiji Nakagawa, and Deborah
Hopkins, “Defect Detection Using the Reverberant
213
Automotive Lightweighting Materials
FY 2001 Progress Report
8. HIGH-STRENGTH STEELS
A. Enhanced Forming Limit Diagrams
Project Manager: Pat J. Villano
Auto/Steel Partnership
2000 Town Center Drive, Suite 320
Southfield, Michigan 48075-1123
(248) 945-4780; fax: (248) 356-8511; e-mail: [email protected]
Project Co-chairman: John Siekirk
DaimlerChrysler Corporation
CIMS 484-34-01, 800 Chrysler Drive
Auburn Hills, Michigan 48326-2757
(248) 576-2567; fax: (248) 576-2155; e-mail: [email protected]
Project Co-chairman: Bernard S. Levy
B. S. Levy Consultants, Ltd.
1700 E. 56th St., Apt. 3705
Chicago, Illinois, 60637
(773) 752-4306
DOE Program Manager: Joseph Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: U.S. Automotive Materials Partnership
Contract No.: DE-SL05-01OR22866
Objective
•
Document previously unavailable relationships between stamping die variables and increased forming limits that
will be necessary to facilitate the manufacture of lightweight steel vehicles. This objective will be accomplished
by using unique capabilities developed for this project, which continue to be refined.
OAAT R&D Plan: Task 12, 13; Barriers B, C
Approach
•
Phase 1: Extend the predictive relation for aluminum-killed draw quality (AKDQ) steel to the high-strength
steels critical to the lightweighting of steel passenger car bodies.
•
Phase 2: Develop an application methodology for use on the original equipment manufacturer’s vehicle
programs.
•
Phase 3: Develop training modules for effective transfer of the technology to the manufacturing floor.
215
FY 2001 Progress Report
Automotive Lightweighting Materials
Accomplishments
•
Designed experimental equipment and developed procedures to quantify improvements in the breakage limits.
•
Used a channel draw die to produce open-ended channel sections using varying draw bead geometries.
•
Completed a predictive model for AKDQ steel.
•
Measured and analyzed data from the channel draw tests.
•
Performed longitudinal and transverse tensile tests from the wall of the channel draw specimens. These included
19 conditions and 2 directions, in triplicate, for a total of 114 tensile tests.
•
Analyzed the data and extended the predictive model to the high-strength steels needed for the A/SP
lightweighting initiatives.
This technology reduces die tryout costs by avoiding
the delays associated with solving forming
problems, and it can reduce manufacturing costs
through improved blank utilization. In addition, this
effort provides a framework for future studies of
sheet metal stamping that would broaden the range
of parts that can be successfully produced with
advanced high-strength steels.
Background
The Enhanced Forming Limits project team was
formed to quantify increases in breakage limits to
allow greater use of the intrinsic formability of sheet
steel. This goal is particularly important for mass
reduction initiatives because the required highstrength steels exhibit reduced formability. It thus is
critical to use all of the formability that is available.
216
Automotive Lightweighting Materials
FY 2001 Progress Report
B. High-Strength Steel Stamping Project
Program Manager: Jack Noel
Auto/Steel Partnership
2000 Town Center, Suite 320, Southfield, MI 48075-1123
(248) 945-4778; fax: (248) 356-8511; e-mail: [email protected]
Co-chairman: James Fekete
General Motors Corporation— Metal Fabricating Division
100 Kirts Blvd., Troy, MI 48007-5001
(248) 696-1176; fax: (248) 696-1101; e-mail: [email protected]
Co-chairman: Changqing Du
DaimlerChrysler Corporation
800 Chrysler Drive, Auburn Hills, MI 48326
(248) 576-5168; fax: (248) 576-7910; e-mail: [email protected]
DOE Program Manager: Joseph Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: U.S. Automotive Materials Partnership
Contract No.: DE-FC05-95OR22363
Objectives
•
Determine how to accurately predict the degree of springback in a variety of flanging conditions on parts made
from high-strength steel (HSS) sheets prior to the construction of the production tooling.
•
Identify part designs and manufacturing processes that will reduce springback and other part distortions and
recommend them to design and manufacturing engineers.
OAAT R&D Plan: Task 12, 13; Barrier B
Approach
•
Enhance HSS stamping springback predictability through finite element analysis (FEA).
•
Improve HSS stamping springback control by developing the required knowledge of part design geometries that
affect flange springback and die processes that help to control springback.
•
Constructed, or obtained on loan, dies offering different panel conditions for the Auto/Steel Partnership for
process development.
Accomplishments
•
Studied FEAs of part designs and computer simulations of manufacturing processes to determine factors causing
springback of HSS sheet metal stampings.
•
Established optimum stamping die processes and die metal clearance values through comparison of simulation
output to actual stamping results. Although the most accurate springback prediction was noted using a die
217
FY 2001 Progress Report
Automotive Lightweighting Materials
clearance of 1.4 times the metal thickness (1.4t), the least amount of actual part springback occurred when the
die clearance was reduced to 1.1 times metal thickness (1.1t). The simulation group is working to improve the
FEA predictive accuracy with varying metal clearances to produce the least amount of springback and variation.
•
Conducted springback control experiments using tooling designed to replicate actual production stamping die
processes. These experiments confirmed the importance of optimizing part geometry and the stamping process
in order to reduce residual stress, flange springback, and part-to-part variation.
•
Measured parts and recorded and analyzed the data. Results will be published in hard copy and on the
Auto/Steel Partnership web site.
60 KSI—were used to stamp rails, and the finished
panels were measured. A correlation between actual
and simulation data quantified the differences. The
simulation task team proposed that predictions be
developed for the die segment, stampings be
manufactured with varying material clearances in
the die, and simulation predictions be compared
with the measured stamping data. The segment
findings will then be extrapolated over the full-size
panel (Figure 1). The purpose of this proposal is to
allow the continuation of stamping trials and
computer simulation comparisons, but with the
additional element of working with known die gaps,
rather than assumed clearances between the punch
and the die.
Introduction
The use of HSS in automotive applications has
been increasing over the last 10 years. Owing to the
mechanical properties of HSS, the springback after
forming and the geometric dimensional control of
the stamped parts have been critical issues in
stamping tool construction and in the stamping
process. Because the actual dimensions of the HSS
stamping off the tool (e.g., for a structural rail-type
stamping) are unpredictable with current tools and
technology, the average die face re-machining
process may be four to six times that required for
mild steel applications and result in 2 to 3 months of
added tryout time.
Computer simulation technology has been
widely applied in the stamping industry and has
been recognized as a common virtual stamping tool
to identify the formability issues and evaluate the
solutions before the actual stamping dies are made.
Actual experience has demonstrated that computer
simulation data have not been reliable in predicting
the degree and modes of the springback for rail-type
stampings, even with small form radii of two to
three times the metal thickness.
Details
To assess the accuracy of the current
forming/springback simulation technology, a fullsize structural rail stamping die and a smaller die
representing a segment of the full-size die were
built. Three different HSS materials—40, 50 and
Figure 1. Full-scale rail stamping die (lower half).
218
Automotive Lightweighting Materials
FY 2001 Progress Report
C. Hydroform Materials and Lubricants Project
Program Manager: Jack Noel
Auto/Steel Partnership
2000 Town Center, Suite 320, Southfield, MI 48075-1123
(248) 945-4778; fax: (248) 356-8511; e-mail:[email protected]
Chairman: Mike Thorpe
Stelco, Inc., Hilton Works, P.O. Box 2030
Hamilton, Ontario, Canada L8N 3T1
(905) 527-8335; fax: (905) 308-7020; e-mail: [email protected]
DOE Program Manager: Joseph Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: U.S. Automotive Materials Partnership
Contract No.: DE-FC05-95OR22363
Objectives
•
Develop mechanical test procedures for tubes.
•
Improve the accuracy of and confidence in finite element modeling of tubular hydroforming.
•
Develop an understanding of steel and lubricant requirements for hydroforming, using a combination of
experiments and finite element modeling.
OAAT R&D Plan: Task 8, 13; Barriers B, C
Approach
•
Conduct a series of experiments to determine the free expansion and corner fill capabilities of various mild and
high strength steel tubes using internal pressurization.
•
Conduct other experiments for tube expansion to determine the effects of axial compression in combination with
internal pressurization and the effects of pre-bending and pre-forming on subsequent formability.
•
Use the collected data to develop forming limit diagrams for tubular hydroforming. These are required for the
computer simulation of hydroforming.
Accomplishments
•
Worked with the Industrial Research and Development Institute (IRDI) of Midland, Ontario, Canada, to perform
free expansion and corner fill experiments with mild and high-strength steel. Existing mechanical property
testing methods for tubes have been evaluated, and new testing methods are being developed as needed.
•
Started evaluating the validity of the traditional forming limit curve for sheet metal stamping and developing
correction factors for tubular hydroforming.
•
Initiated a designed experiment using both mild and high-strength steels to gain an empirical understanding of
material and lubricant requirements in order to develop test cases for validation with finite element modeling.
219
FY 2001 Progress Report
Automotive Lightweighting Materials
The experiment covers free expansion and corner filling of tubes, as well as internal pressurization and axial
compression with straight and bent tubes.
•
Completed the free expansion and corner fill portions of the experiment. As of October 30, 2001, tubes are in
the process of being made and bent for the ongoing internal pressurization and axial compression
experimentation that will continue into 2002.
Furthermore, additional information is needed
on the fundamental material attributes that control
the hydroformability limits of steel. This knowledge
will allow steelmakers to develop new steel grades
or apply existing steel grades to hydroforming
applications (Figure 1). Experiments will be
conducted using the facilities of the IRDI of
Midland, Ontario, Canada.
Background
The formability limits for steel in tubular
hydroforming applications are poorly understood.
Practitioners are using the sheet steel forming limits
developed for conventional stamping processes to
assess the formability of hydroformed components.
As an example, the effect of axial compression on
the forming limits of steel is unknown. Likewise,
the effect of prior strains induced in the tubeconversion and tube-bending processes that precede
the hydroforming operation have been ignored. The
accuracy of material characteristics needs to be
assessed and improved to allow the optimum
application of tubular hydroforming in vehicle
lightweighting activities.
In addition, the tests that tube producers are
using to evaluate tube quality do not properly
address metal forming issues relevant to tubular
hydroforming. The current tests focus on the weld
quality and dimensional accuracy of the tubes, but
they do not adequately describe the amount of
ductility left in the tube for post-forming operations
such as hydroforming. There is a need to develop
and evaluate tube-testing methods that are suitable
for evaluating and comparing material formability
and are also suitable for input into finite element
models.
Figure 1. Section of die and part for hydroforming corner-fill experiments.
220
Automotive Lightweighting Materials
FY 2001 Progress Report
D. Sheet Steel Joining Technologies
Program Manager: Thomas D. Mackie
Auto/Steel Partnership
2000 Town Center, Suite 320
Southfield, MI 48075-1123
(248) 945-4781; fax: (248) 352-1740; e-mail: [email protected]
Co-Chairman: James Dolfi
Ford Motor Company
17000 Oakwood Blvd. P.O. Box 1586, Mail code #E1425
Dearborn, MI 48121
(313) 323-0033; fax: (313) 845-6117; e-mail: [email protected]
Co-Chairman: Philip Coduti
Ispat Inland Inc.
3001 East Columbus Drive, Mail code MC9-000
East Chicago, IN 46312
(219) 399-6111; fax: (219) 399-6562; e-mail: [email protected]
DOE Program Manager: Joseph Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: U.S. Automotive Materials Partnership
Contract No.: DE-FC-59OR22363
Objective
•
Facilitate the increased application of advanced high-strength steels (AHSSs) in Auto/Steel Partnership (A/SP)
lightweighting projects through the development of weld parameters and schedules that will produce quality
welds in 12 different grades of these advanced steels. The project will be pursued from several perspectives,
including weld lobe evaluation under peel test conditions to determine the optimum welding for weld time and
current. Weld-bonded joints will be evaluated, as well as straight-welded joints. Impact fracture and fatigue
characteristics of these welds will be documented and compared with the weld strengths (cross-tension and
tensile shear) in the development of a standard for partial thickness fractures.
OAAT R&D Plan: Task 6, 8; Barrier D
Approach
•
Establish a weld window for the 12 different AHSSs using two resistance spot-welding systems, scissor gun, and
“C” gun, and evaluate the welds through peel test, tensile shear, cross-tension, microhardness, and
metallographic examination.
•
Develop a fatigue test coupon and evaluate weld fatigue properties for the various AHSS materials.
•
Develop an impact test model and evaluate impact characteristics for the various AHSS materials.
221
FY 2001 Progress Report
Automotive Lightweighting Materials
•
Characterize the scissor gun and C gun machines as to rigidity and weldability.
•
Characterize the effects of electrode design, metallurgy, and coatings on the weldability of the particular steels.
Accomplishments
•
Established the procedure for weld lobe study, testing, and machine characterization.
•
Acquired 200 ft2 of material for 10 of the 12 materials and delivered the material to the test facility, where it has
been sheared into coupons.
•
Completed the weld lobe study to the 75% level, including weld peel, shear coupons, and micro-examination.
•
Prepared metallographic samples and took photographs as peel tests were completed.
for the new steels qualifies these for various
automotive components. Extensive work has been
attempted on spot welding of mild steels during the
Intelligent Resistant Welding Project of the Auto
Body Consortium, and various A/SP resistance
welding and adhesive bonding projects.
Background
The use of AHSSs is integral to the steel
industry effort to make automotive components
more lightweight while not sacrificing strength or
escalating cost, an issue with the application of
other lightweight materials. The development of
welding parameters and physical testing parameters
222
Automotive Lightweighting Materials
FY 2001 Progress Report
E. Sheet Steel Fatigue Characteristics
Program Manager: Gene Cowie
Auto/Steel Partnership
2000 Town Center, Suite 320, Southfield, MI 48075-1123
(248) 945-47798; fax: (248) 356-8511; e-mail: [email protected]
Benda Yan, Chairman
Staff Research Engineer, Product Applications
Ispat Inland Inc.
3001 E. Columbus Drive, East Chicago, IN 4312
(219) 399-6922; fax: (219) 399-6562; email: [email protected]
DOE Program Manager: Joseph Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: U.S. Automotive Materials Partnership
Contract No.: DE-FC05-95OR22363
Objectives
•
Investigate the work-hardening effects of stamping and the strengthening effects of paint baking on the grades
exhibiting such effects on constant-amplitude strain-controlled fatigue properties, and compare them with
similarly strained and baked conventional high-strength steels.
•
Investigate the variable-amplitude performance of low-carbon steel compared with that of interstitial-free steel.
•
Investigate other new advanced high-strength steels, possibly with various joining methods.
OAAT R&D Plan: Task 8, 13; Barriers B, C
Approach
•
Characterize the fatigue properties of selected grades of steel by identifying three fatigue coefficients and three
fatigue exponents.
•
Compile the results into a database that can be used by designers and design engineers to retrieve and compare
the fatigue lives of sheet steel structural components in the design phase.
Accomplishments
•
Completed phase 2 and phase 3 testing.
•
Developed a sheet steel fatigue database and populated it with fatigue data on 54 coils of steel. Team members
are evaluating the database as a resource for fatigue information.
223
FY 2001 Progress Report
Automotive Lightweighting Materials
Introduction
In the era when vehicle mass was not a
drawback, and in fact was sometimes considered an
advantage, the primary concern in body design was
rigidity. More recently, the need to reduce body
mass in order to comply with mandated corporate
average fuel economy (CAFE) standards has caused
design engineers to reexamine design procedures
and materials. High-strength steels, judiciously
selected and applied, emerged as potential low-cost,
reliable materials for reducing vehicle mass. As
structural components are optimized and thinnergauge, higher-strength materials are assessed,
component fatigue life becomes a consideration. To
assess the performance of a material in the design
phase, its fatigue characteristics must be known.
This project is addressing steel grades identified
by the lightweighting initiatives (closures and front
end structures) of the Auto/Steel Partnership. The
objective of this 3-year program is to continue the
investigation of particular formable and highstrength steels used in automotive components.
Advanced high-strength steels are being investigated
by the Lightweight Front End Structure Project
Team for their potential to reduce weight and
manage crash energy. The work will also investigate
the performance of steels currently being used under
variable-amplitude conditions. The results are
expected to improve durability modeling and
simulation.
Figure 1. Design of the machined test sample.
Selection of qualified test laboratories capable
of performing the work is even more critical. Since
participating project team members had extensive
experience with a number of qualified machine
shops and laboratories, phase 1 was completed on
March 1, 1997. To maximize the consistency among
testing services, a standard test procedure was
written for use by all test laboratories.
Phase 2 was the testing, based on procedures
established in phase 1, to determine the fatigue
properties of the following steel grades:
• Interstitial-free hot-dipped galvanized
• Drawing-quality, special-killed, cold-rolled
• High-strength low-alloy (HSLA) 50 ksi
(345 MPa) hot-dipped galvanized
• High-strength low-alloy 50 ksi (345 MPa)
galvannealed
• Interstitial-free hot-dipped galvannealed
• Structural-grade 40 ksi (275 MPa) galvannealed
Details
The long-term strategy for the project to
research sheet steel fatigue characteristics consists
of four phases of work. Phase 1 objectives were to
develop and establish confidence in the testing
procedure. This phase included selecting machine
shops and laboratories to prepare the test samples
and perform the tests. Test samples must be
prepared in a manner that does not alter the as-rolled
properties of the sheet steel. For example, stamping
the samples will induce work hardening in the
immediate area of the cut edge. The work hardening
is essentially eliminated by machining the samples
according to carefully determined procedures. The
design of the machined test sample is shown in
Figure 1.
The tests for each grade of steel were normally
based on samples taken from material supplied by
three different steel suppliers. Samples from each
supplier were taken from three different locations on
each of three coils of steel from the supplier.
Therefore, each data set for each grade of steel was
based on samples taken from 3 × 3 × 3 = 27
locations. A sufficient number of test coupons were
prepared from each location that each fatigue test
was run at multiple strain amplitudes. The results
were plotted on a log-log scale of cycles versus
strain amplitude, and a theoretical curve was
constructed.
224
Automotive Lightweighting Materials
FY 2001 Progress Report
•
This work was reported as completed at the
January 10, 2001, team meeting. An Access
database was developed, and all the material
properties, including strain based fatigue, were
entered.
Phase 3 is the verification testing to determine
and demonstrate the benefits of the work accomplished in phases 1 and 2. Verification was performed
on DDQ+ steel. Two SAE spectra were tested:
transmission time history and suspension time
history. The test matrix consisted of three replicates
at five strain/load levels for each spectrum for a
total of 30 tests. The range of strain/load for each
spectrum was approximately 3 to 1.
Phase 4 was initiated upon agreement of the
team members. This phase of work comprises two
projects:
Project 1 is an analysis of the strain + bake
properties of dual-phase (DP) 600 vs.
transformation-induced plasticity 600 vs. HSLA 340
(50 ksi). Two prestrain conditions were specified.
•
Balanced biaxial stretch with an equivalent
strain of approximately 2% for the DP 600+ and
HSLA 340 steels
Balanced biaxial stretch with an equivalent
strain of approximately 8% (or ~6% if 8%
cannot be achieved) for all three materials
This work is scheduled for completion in March
2002.
Project 2 is a comparison of low-carbon
drawing steel to the interstitial-free steel tested in
Phase 3 under variable amplitude loading
conditions. This work is scheduled for completion
on October 31, 2001.
The sheet steel fatigue database, noted above
under Phase 2, will be an essential product of this
program. That database is intended to provide a
user-friendly compilation of the fatigue data
generated in Phase 2, which was reported on a series
of spreadsheets. Compilation into the database will
give the user access to data in one location. It also
provides features such as data sorting, plotting and
data averaging.
225
Automotive Lightweighting Materials
FY 2001 Progress Report
F. Strain Rate Characterization
Project Manager: Pat J. Villano
Auto/Steel Partnership
2000 Town Center Drive, Suite #320
Southfield, Michigan 48075-1123
(248) 945-4780; fax: (248) 356-8511; e-mail: [email protected]
James R. Fekete, Chairman
General Motors Corporation—Metal Fabricating Division
100 Kirts Blvd., Troy, Michigan 48007-5001
(248) 696-1176; fax: (248) 696-1101; e-mail: [email protected]
DOE Program Manager: Joseph Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: U.S. Automotive Materials Partnership
Contract No.: DE-FC05-95OR22363
Objective
•
Embark on a 4-year program that will result in the development of test methods, constitutive materials models,
and finite element analysis (FEA) guidelines for industry-wide use.
OAAT R&D Plan: Task 8; Barrier C
Approach
This project addresses two aspects of the simulation accuracy problem:
•
Accurate representation of material properties:
Develop a standard test for sheet metal at high strain rates to characterize materials with minimum testing
efforts.
Evaluate Hopkinson bar and servo-hydraulic load frames to capture the behavior of sheet metal at typical
crash strain rates, to understand the relationship between tension and compression, and to understand and
effectively model the hardening behavior of sheet steel under typical crash conditions.
Develop a database of properties for existing and new materials being developed for use in automotive
construction, including advanced high-strength steels.
•
Optimization of material constitutive models and finite element models:
Develop the necessary guidelines for structuring finite element problems to ensure that all inputs, including
material properties, are accurately reflected in the results.
Develop validation experiment(s) for evaluating the different materials and material models that they
represent.
Optimize the use of new element technology, time-step control and integration, modeling of contact
conditions, and modeling of joints.
Evaluate stochastic approaches to crash modeling.
•
Evaluate the influence of part variability on the stability of collapse modes in components and in assemblies.
227
FY 2001 Progress Report
Automotive Lightweighting Materials
•
Recommend robust, accurate computationally efficient materials models for FEAs.
•
Evaluate the influence of work hardening, bake hardening, and strain-rate sensitivity effects on constitutive
models.
•
Develop models for non-ferrite materials such as dual-phase and transformation-induced plasticity steels, as well
as models for multi-axial loading.
Accomplishments
•
Completed a project through the University of Dayton Research Institute for compression and tensile “split
Hopkinson bar” tests at strain rates of 300/s and 1000/s on drawing-quality, special-killed and DF140T steel
grades. The final report was delivered.
•
Completed review of the Callidine and Avalle papers to arrive at recommendations on the test procedures for
the cylindrical tube crush test.
•
Continued to support the Oak Ridge National Laboratory project “Strain Rate Sensitivity Study for Automotive
Structures.”
•
Completed specification development work for quoting the axial hinge plate and the cylindrical tube crush
experiments.
•
Submitted purchase order requests for axial hinge plate crush and cylindrical tube crush experiments.
•
Issued a purchase order to a supplier for the manufacture of steel tubes that will be used in the tube crush
experiments. The high-strength steel materials to be used in experiments—drawing-quality mild steel, highstrength, low-alloy 350 Mpa steel, and dual-phase 600 Mpa steel—will be supplied by partnership member
companies.
The accurate representation of the mechanical
properties of the sheet steels used in current and
future vehicles is a key element in the development
of accurate math-based tools. One key aspect of
these properties is the strain rate dependence of the
stress/strain behavior. Previous Auto/Steel
Partnership research has experimentally
demonstrated the positive influence of strain rate
sensitivity on dynamic performance of sheet steel
structures. Efforts to accurately simulate this
behavior using math-based tools have not, however,
been completely successful.
Background
Over the past several years, evaluating
crashworthiness in new vehicle designs has become
more and more dependent upon math-based
engineering tools, including FEA. The use of these
tools enables rapid optimization of product designs,
thereby reducing development time and cost while
minimizing product mass. The growing dependence
on these techniques and the desire to accelerate the
replacement of physical testing with analytical
validation of design solutions is driving the need to
improve the predictive capability of the math-based
tools.
228
Automotive Lightweighting Materials
FY 2001 Progress Report
G. High-Strength Steel Tailor-Welded Blanks
Project Manager: Jack Noel
Auto/Steel Partnership
2000 Town Center, Suite 320, Southfield, MI 48075-1123
(248) 945-4778; fax: (248) 356-8511; e-mail: [email protected]
Co-Chair: Aleksy Konieczny
U.S. Steel Group, a unit of USX Corp.
5850 New King Court, Troy, MI 48098-2692
(248) 267-2541; fax: (248) 267-2581; e-mail: [email protected]
Co-Chair: Mariana Forrest
DaimlerChrysler Corporation
2730 Research Drive, Rochester Hills, MI 48309
(248) 838-5256; fax: (248) 838-5338; e-mail: [email protected]
DOE Program Manager: Joseph Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: U.S. Automotive Materials Partnership
Contract No.: DE-SL05-01OR22866
Objectives
•
Investigate the formability and weld properties of high-strength steel (HSS) and advanced HSS (AHSS) tailor
welded blanks (TWBs).
•
Determine the fatigue properties of laser and mash seam welds in AHSS.
•
Investigate the weight and cost savings potential for patch-type vs laser-welded TWBs.
OAAT R&D Plan: Task 6, 8, 13; Barriers B, C
Approach
•
Prove the concept of using patch-type TWBs on door inner panels.
•
Make stampings from automotive door inner dies to produce doors with patch-type TWBs.
•
Conduct fatigue tests on AHSS blanks with laser and mash seam welds.
Accomplishments
•
Worked with Oxford Automotive to make stampings with patch-type blanks from automotive door inner panel
dies.
•
Determined testing procedures and began identifying prospective testing laboratories for fatigue tests.
229
FY 2001 Progress Report
Automotive Lightweighting Materials
Introduction
The tailor welding of sheet steel blanks—that is,
the use of varying gauges and grades of steel within
a specific panel “tailored” to provide added strength
in select areas—continues to gain favor in the
product design community. Industry data show that
production of TWBs has grown from just over 1
million in 1993 to nearly 24 million in 1999, with
the market potential estimated at 50 million blanks
by the year 2003. TWBs have gained acceptance in
the automotive industry as a means of
simultaneously achieving cost and weight reduction.
The use of conventional high-strength low-alloy
steels, HSSs, and AHSSs—such as dual-phase and
transformation-induced-plasticity steels—offers the
potential for further weight reduction. However,
uncertainty exists regarding the weldability of these
products and the resultant ductility of the weld and
heat-affected zone. Manufacturing techniques must
be developed and validated to produce reliable,
high-volume HSS and AHSS TWBs.
The purpose of this focused 3-year project is to
investigate the manufacturability and formability of
HSS and AHSS TWBs and demonstrate that these
products can produce a front-end structure of
significantly lower mass.
In addition to supporting the work of the
Auto/Steel Partnership Lightweight Front End
Structure Project, the Tailor Welded Blank Project
Team is investigating the wider application of patchtype blanks to reduce the weight and cost of
automotive door assemblies for the Auto/Steel
Partnership Light Weight Closures Project.
Figure 1. Door inner hinge reinforcement made
from a patch-type blank.
be made to the smallest required irregular shape,
whereas the laser weld must be a straight line,
making the heavy metal member of the blank much
larger and heavier than necessary. The cost of spot
welding the patch-type blank is also much less than
the cost of laser welding.
The TWB Project Team worked with Oxford
Automotive to make stampings from automotive
door inner dies to produce this type of door. Initial
reviews by representatives from General Motors and
DaimlerChrysler were generally positive, with
requests to repeat the stamping trials on dies for
doors with more complex configurations.
Fatigue test planning has been done in
conjunction with the needs of the Lightweight Front
End Structures Project. As they defined the
materials they would use for tailor welded blanks in
the motor compartment rails (Figure 2), the Tailor
Welded Blank Project Team has determined the
required test procedures and is determining
prospective laboratories to perform this work.
Approach
The TWB project team has worked with the
Lightweight Closures Project Team to prove the
concept of using patch-type TWBs on door inner
panels (Figure 1). This type of tailored blank
reduces weight because the spot-welded patch can
230
Automotive Lightweighting Materials
FY 2001 Progress Report
Results
Most of the work undertaken in the first 6
months of 2001 involved the patch-type tailored
blank for the Lightweight Closures Project. Patchtype blanks have been used in some less complex
production applications to integrate parts and reduce
tooling cost. The stamping die tryout work with
Oxford Automotive indicates the potential for this
process to reduce the cost and weight of select door
assemblies. Doors with a straight hinge pillar
surface, such as rear doors, appear to have the most
potential at this time. Minor distortions at the
sealing surface are a problem, but they can be
corrected in subsequent die operations.
Figure 2. Example of a patch blank on an
underbody rail part.
231
Automotive Lightweighting Materials
FY 2001 Progress Report
H. Tribology
Project Manager: Pat J. Villano
Auto/Steel Partnership
2000 Town Center Drive, Suite #320
Southfield, Michigan 48075-1123
(248) 945-4780; fax: (248) 356-8511; e-mail: [email protected]
Project Chairman: Dominick Zaccone
Ford Motor Company
Vehicle Operations General Office
P.O. Box 1586—Room C290
Dearborn, Michigan 48121
(313) 594-0569; fax: (313) 390-6425; e-mail: [email protected]
DOE Program Manager: Joseph Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: U.S. Automotive Materials Partnership
Contract No.: DE-FC05-95OR22363
Objective
•
Conduct stamping simulation tests to study the effects of tribology—the interrelationship of lubrication, friction
and wear—on the stamping performance of advanced high-strength steels. (Stamping performance in this
project is defined as minimizing die wear and maximizing dimensional stability. Understanding of the
contribution of lubricants to errors in springback prediction when using finite element analysis will also improve
dimensional performance.
OAAT R&D Plan: Task 8, 13; Barrier B
Approach
Phase one:
• Select the sheet high-strength steel material grades for the tests.
•
Secure the material and ship it to TribSys Inc. to do the testing. Using its twist compression tester and draw bead
simulator, TribSys will perform friction testing.
Phase two:
• Compare the wear rate of draw beads using different lubricants and high-strength steel materials. Restraining
force will be compared with temperature, contact area, and wear on the draw beads. These data will be
correlated with the results of the twist comparison test and draw bead simulator test conducted in phase one.
233
FY 2001 Progress Report
Automotive Lightweighting Materials
Accomplishments
•
Changed the name of the project team from sheet steel lubricants to tribology.
•
Invited lubricant suppliers to address the project team with their product recommendations along with
supporting performance data. Participating companies included Quaker Chemical, Diversified Chemical
Technology, Inc., Fuchs Chemical, and Midstate Chemical.
•
Member company Dofasco Steel did a presentation titled “ Lubricant/Steel Characterization Study Using the
Draw Bead Simulator”
•
Installed Alan Pearson of General Motors as the new project team chairman on July 1, 2001
process parameters associated with die wear, such as
heat build-up, die scoring, and dimensional variation
caused by springback variation. The contribution of
interfacial friction to springback variation is not
known. Advanced high-strength steels may require
different lubricant and/or die materials to minimize
the friction and die wear.
Introduction
Lubricants have historically been used in the
forming of sheet metal stampings to reduce friction,
improve formability, and minimize die wear. They
also provide corrosion protection. The anticipated
increase in the use of advanced high-strength steels
places a greater emphasis on understanding the
234
Automotive Lightweighting Materials
FY 2001 Progress Report
I. Modeling of High-Strain-Rate Deformation of Steel Structures
Srdan Simunovic
Oak Ridge National Laboratory
Oak Ridge TN 37831-6359
(865) 241-3863; fax: (865) 574-7463; e-mail: [email protected]
DOE Program Manager: Joseph A. Carpenter
(202) 586-1022; fax: (202) 586-6109; e-mail: [email protected]
ORNL Technical Program Manager: Philip S. Sklad
(865) 574-5069; fax: (865) 576-4963; e-mail: [email protected]
Contractor: Oak Ridge National Laboratory
Contract No.: DE-AC05-00OR22725
Objective
•
Develop numerical modeling guidelines to realistically assess the influence that the properties of strain-ratedependent materials exert in crashworthiness computations. The dynamic loading problems are modeled using
diverse combinations of modeling approaches (sub-models) that are essential in describing strain-rate
sensitivity in computational simulations.
OAAT R&D Plan: Task 7, 8, 13; Barrier C
Accomplishments
•
Performed a literature review for progressive crushing and characterization and modeling of strain-rate effects
in automotive impact.
•
Developed computational models for steels under various strain rates based on experimental data.
•
Developed computer programs for calculating instantaneous strain rates that are used for evaluation of
constitutive models in computer programs.
•
Evaluated finite element model (FEM) element performance for modeling of axisymmetric crushing of circular
tubes.
•
Determined the effects of various material parameters on the crush behavior of axisymmetric tubes.
•
Developed experiments for characterizing structural and material aspects in component crashworthiness.
Future Direction
•
Analyze transitions from axisymmetric to antisymmetric mode in progressive crushing and evaluate the
performance of FEM formulations for modeling of transition problems.
•
Determine the effects of FEM shell element discretization and formulation on model accuracy.
•
Determine optimal FEM formulations for modeling of crushing of rectangular tubes.
•
Develop an experimental program for crushing of rectangular tubes.
•
Develop modeling guidelines for rectangular tubes.
235
FY 2001 Progress Report
Automotive Lightweighting Materials
development of a mathematical model of the
stress-strain behavior, using empirical or
fundamental equations; or by a combination of the
two. The models must allow the computer code to
efficiently and accurately select the appropriate
material properties for the stress-strain situation at
any given time step in the crash simulation. The
models should also comprehend the changes in
properties due to processing during vehicle
manufacture, including work hardening (under
various strain paths) and bake hardening. These
models are evaluated by way of a validation test,
where a structure is deformed in a known manner,
and FEM modeling is used to predict the
deformation behavior. Accurate validation tests
have been developed and documented for bulk
materials, but they need to be developed for sheet
materials.
Finite element codes. The FEMs must
accurately predict the stress, strain and strain rate
for each element in the model under sometimes
violently changing loading conditions and under
conditions of rapid buckling. This requires robust
finite element codes, meshing practices, and
elements. The introduction of strain-rate-sensitive
properties adds new complexity to these
calculations, which must be understood. In the
end, guidelines are needed for setting up and
running FEM crash analyses, which maximize the
accuracy of all aspects of the model, including the
material properties.
The objective of the project is to develop
numerical modeling guidelines for strain-ratedependent materials in crashworthiness
computations. The scope of the project is to study
specific, experimentally well-defined structural
problems in automotive impact. The dynamic
loading problems will be modeled using diverse
combinations of modeling approaches (submodels) that are essential in describing strain-rate
sensitivity in computational simulations. Submodels to be examined include finite element
formulations, constitutive materials models,
contact conditions, and so on. The trends,
influences, and direct effects of the modeling
techniques employed will be identified and
documented. The relative significance of the submodels employed will be established, particularly
in relation to the strain-rate effect resulting from
the material constitutive models.
Introduction
In the past several years, the development of
the crashworthiness of new vehicles has become
more and more dependent upon math-based
engineering tools, including finite element
analysis. The use of these tools enables the rapid
optimization of product designs, reducing
development time and cost, as well as minimizing
product mass. The growing dependence on these
techniques, and the desire to accelerate the
replacement of physical testing with analytical
validation of design solutions, is driving the need
to improve the predictive capability of the mathbased tools.
The accurate representation of the mechanical
properties of the sheet metals used in
contemporary vehicles is a key element in the
development of accurate math-based tools. An
important aspect of these properties is the strainrate dependence of the stress-strain behavior.
Previous research has experimentally
demonstrated the positive influence of strain-rate
sensitivity on the dynamic performance of sheet
steel structures.1, 2 However, efforts to accurately
simulate this behavior using math-based tools
have not been completely successful. Three
aspects to this problem must be addressed:
Accurate material properties. Several test
methods are available to measure material
properties under stress-strain conditions
representative of crash events, such as the
Hopkinson pressure bar and high-capacity servo
hydraulic load frames coupled with advanced
extensometry for strain measurement. The
relatively low thickness of sheet metals typically
used in vehicle structures presents a special
problem in accurate mechanical testing of these
materials. The characteristics of load transfer from
grips to sample in high-strain-rate tensile tests are
not well known. Compression tests must be done
through the thickness. Nonetheless, accurate
properties are the cornerstone of accurate
simulations. Recommendations are needed for
robust, accurate testing methods for sheet
materials.
Constitutive material models. The
mechanical properties of the materials must be
accurately represented in the computational
simulation programs. This is done either by direct
representation of the stress-strain curves; by
236
Automotive Lightweighting Materials
FY 2001 Progress Report
In particular, the project will
•
•
•
•
•
•
•
•
framework within which the constraints of
analytical approaches can be largely eliminated,
and radically more complex deformation and
material response can be modeled.
Body-centered cubic (BCC) metals, such as
steel, exhibit high strain-rate sensitivity that can
increase flow stress several times over the quasistatic values for deformation rates of interest in
automotive impact.4 Therefore, incorporation of
strain-rate sensitivity into material models for
crashworthiness analysis is the first logical step in
improving the accuracy of analysis.
Currently, the most widely used approach to
modeling strain-rate-sensitive material in tube
crushing is to use isotropic plasticity models with
rate-sensitivity components that have moderate
requirements on the experimental program. The
isotropic plasticity models that incorporate strainrate sensitivity and that were recently reported for
tube crush simulations are the Johnson-Cook
model,5 the Zerilli-Armstrong model,6 and the
piecewise linear strain-rate-sensitive material
model. The models are appealing because they
have been implemented in commercial codes used
for crash simulations and have a limited (less
than 7) number of material parameters that must
be determined by experiments. In the case of the
piecewise linear plasticity model, effective strainstress curves are directly fed into the material
models and require the least amount of effort for
model development.
The standard practice for extracting material
parameters is to perform experiments with
uniaxial loading configurations. The ductility of
uniaxially tensile-loaded specimens is limited by
the geometric instability; and for automotive steel
sheets, it limits the magnitude of uniform plastic
strains to about 25%. However, the strains that are
measured and modeled during tube crushing far
exceed these values and can reach 100%. The
physical reality of such large deformations is not
in question. The biaxial loading and bending
provides additional stability that allows the
utilization of a material’s strain hardening,
ductility, and corresponding large energy
dissipation.
Modeling of tube folds is the area where
material models and finite element formulations
are intrinsically linked. In the case when we want
the kinematics of the deformation to be accurate,
the resulting maximum plastic strains are
Evaluate the behavior of the available strainrate material models on the automotive
component level
Examine the constraints of FEM technology
used in explicit time integration codes with
respect to strain-rate modeling
Use data from existing experimental programs
to verify modeling methodology
Implement, modify and utilize strain-ratedependent material models into a crash
simulation code if the existing model
implementation is not available
Help define new experimental programs to
support the modeling of material strain-rate
effect
Utilize supercomputing resources to provide
sufficient model resolution and fidelity
Develop an Internet-based, passwordprotected project site for project progress
reporting, verification and dissemination of
results
Define the requirements for transfer of project
developments to different commercial FEM
codes
Modeling of Strain Rates in Steels
The response of metals under a multiaxial
state of stress and varying strain rates is still far
from being described by a unified theory. During
multi-axial large plastic deformations, material
undergoes significant changes of microstructure
and texture that lead to changes in material
properties on the macro level. From a practical
standpoint, the best that we should hope for is that
the selected material model is applicable to the
ranges of loading and deformation for the problem
at hand. In tubular crush devices that are used as
energy absorbers in vehicles, metallic sheets are
subjected to large, localized deformations that
organize into global collapse mechanisms. In the
past, analytical and semi-analytical methods were
the only feasible routes for analyzing progressive
structural collapse. These methods start from
kinematics assumptions and idealization of
material response, and derive the expressions for
the structural response based on energy
minimization principles. The advances in FEM 3
and in computational power have provided a
237
FY 2001 Progress Report
Automotive Lightweighting Materials
as opposed to the length scale imposed by the
element mesh spacing that governs the
localization response when a rate-independent
material model is used.
approximately four times the experimental strain
range. On the other end of the spectrum, when the
finite element resolution is coarse, the material
model extrapolation is not a concern because the
plastic strains get smeared over larger volumes
and are within the range of uniaxial experiments.
The problem is then the ability to correctly
describe the kinematics of the problem. In models
based on shell finite elements, which are currently
the standard way of modeling tubular crush, the
issue of element resolution boils down to the
number of bi-linear shell elements that are used
for modeling of plastic folds. The curvature of
plastic folds is of the order of the material
thickness; and for the kinematic representation of
the fold to be accurate, element length should
approach shell thickness. It is at this point that
standard shell theory tends to break down, and
material models extrapolate far outside the
experimentally verified range.
In both of these scenarios, in order to match
the experimental results with models, material
parameters are commonly modified. While the
practice may be appalling to material scientists,
the reason for modification in one case is to
compensate for the inability to represent local
deformations; in the other case, the reason is to
account for the significance of large strain regions
for which experimental data are not available. The
flexibility of computer programs now even
provides methods for definition of optimization
problems where the material parameters are
determined so that they result in the optimal
match between the structural simulations and
experiments. These approaches are inadvertently
linked to specific structural problems, FEM mesh
configurations, and loading situations. The best
that we can hope for is that the deformation modes
and characteristic deformation mechanisms
remain essentially the same, in which case we are
able to cover a certain family of problems.
In addition to better representation of the
material response, strain-rate sensitivity has an
added benefit of promoting the computational
stability of simulations. The pathological mesh
sensitivity that is characteristic of modeling of
localized deformations is alleviated when strainrate-sensitive materials are used. Material rate
dependence introduces a length scale that is
proportional to the length of propagation of elastic
waves. This length scale has a physical character
Impact Simulations using Johnson-Cook
and Zerilli-Armstrong Models
Strain rates in automotive impact problems
are reported to be less than 1000/s. Most of the
published studies on automotive-related impacts
with strain-rate-sensitive material models employ
Johnson-Cook and Zerilli-Armstrong material
models. These two models have been developed
for modeling metals subjected to large strains,
high strain rates, and high temperatures occurring
during projectile impact. The popularity of the
Johnson-Cook model is promoted by its simplicity
and ease of implementation. The ZerilliArmstrong model, on the other hand, is appealing
because of its roots in the dislocation mechanics
and the possibility for physical interpretation of
the model parameters.
The strain-rate sensitivity of BCC materials is
attributed to the rate-controlling mechanism of the
thermal component of the flow stress. Because the
activation volume for BCC metals is considered to
be constant, the increase in strain rates should
theoretically be described by an upward
translation of the quasi-static strain-stress curve.
Experimental results show there is also a
corresponding reduction of the strain-hardening
rate so that the parallel strain-stress curves can be
viewed as a limit case.
The Johnson-Cook model is expressed in a
multiplicative form of strain, strain-rate, and
temperature terms:
.
σ = ( A + Bε n )(1 + C ln ε )(1 − T m )
(1)
The consequence of the multiplicative form is
that the strain hardening rate at a certain strain
will increase when the strain rate increases. The
strain-stress curves for increasing strain rates will
tend to “fan out,” which is in contrast to the
experimental results for steel sheets. The material
parameters can be selected so that this differential
hardening is insignificant and curves become
parallel in the limit case. However, the overall
feature still cannot accommodate the global trend
of reduction of the strain hardening rate with
238
Automotive Lightweighting Materials
FY 2001 Progress Report
increasing strain rate that is observed in
experiments.
The Zerilli-Armstrong model for BCC metals
is written in an additive form akin to the standard
materials science expression for flow stress as an
aggregate of various strengthening mechanisms:
Single Finite Element Under Constant
Strain-Rate Loading
Figure 1 shows a numerical test configuration
for evaluation of a material constitutive model.
.
σ = c0 + B0 e − ( β o − β1 ln ε )T + K ε n
(2)
The strain hardening for various strain rates is
constant; therefore, the corresponding strain-stress
curves are parallel. The correlation of yield with
the Zerilli-Armstrong model has been shown to be
very good. However, the inherent inability to
model the dependence of strain hardening on
strain rate will unavoidably lead to adjustment of
the element discretization and selection of specific
strain intervals for material parameter derivation,
as it does for the Johnson-Cook model. Because of
the inability of the models to match the
experimental results across all ranges of strain and
strain rate, these models were not selected for the
detailed tube crush simulations that follow.
It was shown that the strain hardening has a
crucial role in development of a collapse pattern
and bifurcation between the modes in tube
crushing. Material models such as Bodner-Partom
and Khan-Liang can offer added flexibility for
modeling of strain hardening, but they have not
yet been implemented in commercial crash codes.
In the future work, these models will be
implemented, and their performance for the
modeling of progressive crush problems will be
evaluated.
Figure 1. One-element strain-rate test.
The imposed velocity that will result in a
constant strain rate is
.
.
v = ε L0 exp(ε t )
(3)
where Lo denotes element length and t denotes
time. The experimental true plastic strain-stress
data under various strain rates for drawing-quality,
special-killed steel were provided by the
Auto/Steel Partnership and are shown in Figure 2.
900
750
8
Stress [MPa]
600
Impact Simulations using Piecewise Linear
Plasticity Model
Another popular approach for modeling of
strain-rate sensitivity is to use a tabulated form of
strain-stress curves for different strain rates. The
resulting stress in simulations is interpolated
between the known values of strains and strain
rates. The blessing and the curse of this approach
is that the experimental data fit the model exactly,
and that any testing artifacts or errors contained in
the experimental data will be carried over to the
simulations. The highest strain rate in
experimental data acts as a saturation plateau for
strain-rate effects.
7
6
5
4
3
2
1
450
0.001
0.1
5
12
70
85
90
3500
300
150
0
0
0.03
0.06
0.09
0.12
0.15
0.18
0.21
0.24
0.27
0.3
Plastic Strain
Figure 2. Drawing-quality, special-killed steel material
data.
Simulation of the constant strain-rate test
shows the exact overlap of experimental data and
simulation in Figure 3.
239
FY 2001 Progress Report
Automotive Lightweighting Materials
1 microsecond. The material constitutive relations
are evaluated at each computational time
increment. The strain rates are also calculated at
this smallest time scale where wave propagation
effects are important. The experimental data are
determined on the higher time scale where the
wave effects are not measured except for the
highest-rate experiments. Therefore, it is
important to determine the values of the strain
rates in the explicit time integration calculations in
order to define the extent of the material
characterization experimental program. It is also
important to investigate the effects of FEM
formulations and element discretization on the
magnitudes and distribution of plastic strains and
strain rates in the areas of large plastic
deformations that determine the crush mode. So
far, no studies have been reported on the analysis
of strain-rate calculations in explicit FEM
programs.
We have developed a set of programs that
take the data from the FEM computations at each
time increment, process them on the fly, and store
results for further analysis. The development of
new programs was necessary because storing
information for each time increment is not
feasible.
The results show a clear dependence of the
strain-rate calculations on the element
discretization. Strain-rate calculations are shown
to be much more sensitive to FEM formulations
than are plastic strains. The magnitudes of strain
rates are of the order 103/s, so that the
experimental maximum strain rate of that order is
required for the material model development.
Currently, we are analyzing distributions of strains
and strain rates in the area of tube folds for
different discretizations. The simulations will be
compared with tube crush experiments on the
ORNL Intermediate Strain Rate Machine.
Figure 3. Plastic strain evolution in the element.
The only variation comes from the effects of
contact conditions on the base; but, as can be seen,
they can be neglected so long as the time
increment and contact penalty parameters meet
the stability criteria.
The important question in implementation of
the piecewise linear plasticity model is the
maximum strain rate for which the stress-strain
data should be tested. Another is the number and
arrangement of curves across the strain-rate range.
In Figure 2, it is clear that one curve per order of
magnitude would suffice. The material model uses
linear interpolation between the curves, whereas it
has been shown that the stress increment relation
is logarithmic. However, when sufficient
experimental data are available, the effects of
linear interpolation should not be significant. One
possible improvement over the existing model is
to develop a logarithmic fit for the available
curves with respect to strain rate and then
extrapolate it outside the plateau value.
Strain-Rate Calculation in FEM
Simulations
Crashworthiness simulations are almost
exclusively done using FEM programs with
explicit time integration.7 The explicit time
integration algorithms are only conditionally
stable, meaning that the integration time step must
be smaller than the time for a disturbance to travel
across the smallest finite element in the model.
For example, for steel and a finite element
characteristic length of 5mm, the condition
requires that the time step be smaller than
Axisymmetric Crushing of a Steel Tube
The axisymmetric crush mode energy has
larger energy dissipation than antisymmetric
modes and would be preferred for energyabsorbing devices. The mode is difficult to
achieve in practical automotive impact problems,
but it is useful for studying interplay between
materials and structures in the crush problems.
The axisymmetric folding is assisted by the
240
Automotive Lightweighting Materials
FY 2001 Progress Report
localized deformations driven by low strain
hardening. The new high-strength steels for
automotive applications, such as dual-phase steel,
exhibit very high strain hardening, which tends to
precipitate bifurcation into antisymmetric collapse
modes. This makes the modeling of high-strength
steel more sensitive to details of material models
and element formulations.
The objective of our study on axisymmetric
crushing was to investigate the trends and
magnitudes in strain rates as calculated by the
finite element program. As stated earlier, this
information is important in deciding the extent of
the experimental program for derivation of
material models. Contrary to the plastic strains
that tend to monotonically increase, strain rate
exhibits intermittent bursts of activity in the area
of evolving folds. The effects of different cutoff
strain rates on the evolving deformation are
currently under investigation and will be reported
in a journal publication.
A typical setup for the tube crush experiment
is shown in Figure 4. The tube is clamped on the
bottom and is impacted on the top with a moving
mass. The experimental configuration from ref. 8
was used. The tube is made of mild steel that is
modeled using the piecewise linear plasticity
model with strain rate tabulated curves as shown
in Figure 2. The tube thickness is 1.2 mm, the
radius is 28 mm, and the length is 180 mm.
The axisymmetrically crushed tube is shown
in Figure 5.
Figure 4. Tube crush setup.
Effects of Strain Rate on Tube Crush
Figure 5. Axisymmetric tube crush.
Because of the axisymmetry of the problem,
we can simplify the problem by analyzing a
segment of the tube. The tube is modeled by a 2°wide cylinder segment with symmetry boundary
conditions imposed on the radial sides. Figure 6
shows the deformed tube segment for simulations
using material with and without strain-rate
sensitivity. It is clear that the material strain-rate
sensitivity is an important effect and must be used
for crush simulations.
deformation of the tube for different element
discretizations. The case when 23 elements are
used across the tube length is apparently too
coarse to describe the resulting crush mode. As a
consequence, when the entire circumference of the
tube is modeled, the crush mode will not be
axisymmetric or regular. For the case when
46 elements are used along the tube length, the
general features of the crush are acceptable but are
not in agreement with the experimental results.
For 92 and 160 elements, the results are in
Effects of Element Discretization
In order to investigate the effect of element
discretization on the accuracy of the FEM
simulations, we used 23, 46, 92, and 160 elements
along the segment length. Figure 7 shows final
241
FY 2001 Progress Report
Automotive Lightweighting Materials
1,600
0.035
Element Discretization
46
92
160
1,200
0.0325
800
Curvature [1/m]
X [m]
0.03
1
0.0275
0.025
Material Models
Quasi-Static
Strain-Rate Dependent
400
2
3
4
0
-400
-800
0.0225
-1,200
-1,600
0.02
0
0.008
0.016
0.024
0.032
0.04
0.048
0.056
0.064
0
0.072
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
Figure 8. Curvature of the tube for different element
discretizations.
Figure 6. Deformed configurations for different material
models.
The calculated curvature for the finest element
mesh is larger than the expected value. However,
experimental measurements were not available for
the curvature measures. The experiments in FY
2002 will provide information on deformation
curvatures, and the data will be used for FEM
validation.
7
5
6
1
5
0.02
Tube Length [m]
Z [m]
5
2
5
Number of Elements
23
46
92
160
4
5
Effects of FEM Element Formulations on
Tube-Simulated Tube Deformation
3
Simulations have shown that the FEM shell
formulation has significant influence on the
predicted mode of deformation. Shell
formulations in the FEM program LS-DYNA3D 7
were used for different tube configurations, and it
was shown that the fully integrated shell element
based on Bathe-Dvorkin formulation has superior
performance compared with other element
formulations. This element formulation gives a
good approximation of warping and transverse
shear fields and predicts the correct mode of
deformation with increasing mesh resolution. The
performance was analyzed on existing and new
benchmark problems and on the full-tube crush
simulations.
Figure 9 shows the antisymmetric mode of
deformation that was predicted by an FEM shell
element formulation commonly used in FEM
crush simulations. This formulation predicts the
axisymmetric mode for coarse element
discretizations; but for fine discretizations, the
mode bifurcates.
3
4
5
0
0.008
0.016
0.024
0.032
0.04
0.048
0.056
0.064
0.072
0.08
Tube Length [m]
Figure 7. Deformed configurations for different FEM
element discretizations.
agreement with experiments where the finest
mesh shows the smooth deformation characteristic
of the tube crush.
However, when 160 elements along the length
are employed, the element length is of the order of
element thickness; and the shell element theory
shows certain deficiencies for modeling of large
bending, as shown in the following figures. The
curvature of the tube fold is of the order of its
thickness, and the standard shell element theory
with small strain assumptions is not able to model
the strain fields correctly. The curvature along the
undeformed length of the tube for different
element discretizations is shown in Figure 8.
242
Automotive Lightweighting Materials
FY 2001 Progress Report
Figure 9. Crush simulation with Hughes-Liu shell
element formulation.
Figure 10. Crush simulation with Hughes-Liu shell
element formulation.
When the Bathe-Dvorkin shell element
formulation is used, the mode remains the same
with increasing mesh density, as shown in
Figure 10. The element is computationally more
expensive than the standard reduced integration
shell finite elements; for the detailed models of
the crush, the increase in computational cost is
offset by the increase in accuracy of the crush
mode prediction.
Effects of Strain Hardening and StrainRate Sensitivity on Tube Crush
Figure 11. Crushed tube for strain-rate sensitive
material with strain hardening.
One of the objectives of the project is to
evaluate the influence of various material
parameters on tube crush. The effect of strain rate
and strain hardening on tube crush has been
analyzed. It has been shown that material
hardening has a dominant effect on fold length,
while strain-rate sensitivity influences the number
of folds that is created. Strain hardening and
strain-rate sensitivity both have relatively equal
importance for onset of bifurcation. These
findings are illustrated in Figures 11–16.
Figures 11 and 12 show the tube crush for strainrate sensitive material with strain hardening.
When material does not strain-harden, the fold
lengths are reduced, as shown in Figures 13 and
14. The antisymmetric mode after mode
bifurcation has more periods than the previous
case. The onset of mode bifurcation is similar in
both cases.
Figure 12. Fold profile for strain-rate sensitive material
with strain hardening.
243
FY 2001 Progress Report
Automotive Lightweighting Materials
Figure 13. Crushed tube for strain-rate sensitive
material without strain hardening.
Figure 16. Fold profile for strain-rate insensitive
material without strain hardening.
Future Work
The work on the project in FY 2002 will focus
on two aspects: development of modeling
guidelines for modeling of steel tube crushing and
the support of the new intermediate-strain-rate
experiments. The experiments will provide the
crush data for high-strength, low-alloy steels and
dual-phase steels. These two high-strength steels
have comparable yield stresses but drastically
different strain hardening characters.
On the modeling front, we will analyze
transitions from axisymmetric to antisymmetric
mode in progressive crushing and evaluate the
performance of the FEM shell formulations for
modeling of transition problems. We will analyze
the effects of the FEM shell element
discretizations and formulations on model
accuracy through computational simulations and
validations using experimental data.
Figure 14. Fold profile for strain-rate sensitive material
without strain hardening.
Figures 15 and 16 show tube deformation for
rate-insensitive material without strain hardening.
The model shows no bifurcation from
axisymmetric into antisymmetric crush mode.
References
1. R. G. Davies and C. L. Magee, “The Effect
of Strain-Rate Upon the Tensile Deformation of
Materials,” Journal of Engineering Materials and
Technology, 97(2), pp. 151–155 (1975).
2. R. G. Davies and C. L. Magee, “The Effect
of Strain-Rate Upon the Bending Behavior of
Materials,” Journal of Engineering Materials and
Technology, 99(1), pp. 47–51 (1977).
3. T. Belytschko, W. K. Liu, B. Moran,
Nonlinear Finite Elements for Continua and
Structures, John Wiley and Sons, New York,
2000.
4. J. D. Campbell, Dynamic Plasticity of
Metals, Springer, New York, 1972.
5. G. R. Johnson, and W. H. Cook, “A
Constitutive Model and Data for Metals Subjected
Figure 15. Crushed tube for strain-rate insensitive
material without strain hardening.
244
Automotive Lightweighting Materials
FY 2001 Progress Report
7. J. O. Hallquist, LS-DYNA3D, An Explicit
Finite Element Nonlinear Analysis Code for
Structures in Three Dimensions, LSTC Manual,
1995.
8. W. Abramowicz, and N. Jones, “Dynamic
Axial Crushing of Circular Tubes,” Int. J. of
Impact Eng., 2(3), pp. 263–281 (1984).
to Large Strains, High Strain-rates and High
Temperatures,” pp. 541–547 in Proceedings of the
Seventh International Symposium on Ballistic,
Hague, Netherlands, 1983.
6. F. J. Zerilli and R. W. Armstrong,
“Dislocation-Mechanics-Based Constitutive
Relations for Material Dynamics Calculations,”
Journal of Applied Physics, 61(5), pp. 1816–1825
(1987).
245
Automotive Lightweighting Materials
FY 2001 Progress Report
Appendix A. ACRONYMS AND ABBREVIATIONS
A/SP
AAT
ACC
AHSS
ALCOA
ALM
ALTC
AMD
ANL
APAW
ASTM
Auto/Steel Partnership
Advanced Automotive Technology
Automotive Composites Consortium
advanced high-strength steel
Aluminum Company of America
Automotive Lightweighting Materials
Automotive and Light Truck Committee
Automotive Metals Division
Argonne National Laboratory
aluminum plasma arc welding
American Society for Testing and Materials
BHF
BIW
blank holder force
body-in-white
CAE
CAFE
CFC
CRADA
CWRU
computer-aided engineering
Corporate Average Fuel Economy
carbon-fiber composite
cooperative research and development agreement
Case Western Reserve University
DOE
DPF
U.S. Department of Energy
direct powder forging
E
EGN
ELV
EMF
experimental
exfoliated graphite nanostructure
end-of-life vehicle
electromagnetic forming
FE
FEA
FLD
FY
finite element
finite element analysis
forming-limit diagram
fiscal year
GASP
GM
grazing angle surface polarimeter
General Motors
HDPE
HSLA
HSS
HTML
HVSC
high-density polyethylene
high-strength low alloy
high-strength steel
High Temperature Materials Laboratory
Huron Valley Steel Corporation
ICS
IIW
IRDI
ITP
LANL
Industrial Ceramic Solutions
International Institute of Welding
Industrial Research and Development Institute
International Titanium Powders
Los Alamos National Laboratory
247
FY 2001 Progress Report
Automotive Lightweighting Materials
LBNL
LCCF
LIBS
LLDPE
LLNL
Lawrence Berkeley National Laboratory
low-cost carbon fiber
laser-induced breakdown spectroscopy
linear low-density polyethylene
Lawrence Livermore National Laboratory
MAP
MMC
MMCC
MPa
microwave-assisted plasma
metal matrix composite
Metal Matrix Cast Composites
megaPascal (a unit of pressure)
NDE
NDT
NRCAN
nondestructive evaluation
nondestructive test
Natural Resources of Canada
OAAT
OEM
OHVT
OIT
ORNL
OTT
Office of Advanced Automotive Technologies
original equipment manufacturer
Office of Heavy Vehicle Technologies
Office of Industrial Technologies
Oak Ridge National Laboratory
Office of Transportation Technologies
P4
P&S
PAN
PC
PE
PID
PM
PMC
PMPRA
PNGV
PNNL
pp
PQTI
PRA
PVB
PVC
programmable powder preform process
press and sintering
polyacrylonitrile
personal computer
polyethylene
proportional integral derivative
powder metallurgy
polymer matrix composite
powder metallurgy particle-reinforced aluminum
Partnership for a New Generation of Vehicles
Pacific Northwest National Laboratory
polypropylene
Plasma Quench Titanium, Inc.
particle-reinforced aluminum
polyvinyl butyral
polyvinyl chloride
R&D
ROI
RVE
research and development
return on investment
representative volume element
SA/RD
SAMPE
SCMD
SEM
SNL
sales and administrative/research and development
Society for the Advancement of Material and Process Engineering
structural cast magnesium development
scanning electron microscope
Sandia National Laboratories
248
Automotive Lightweighting Materials
FY 2001 Progress Report
SOM
SRIM
solid-oxygen-conducting membrane
structural reaction injection molding
TCM
TDM
TGA
technical cost model
Troy Design and Manufacturing
thermogravimetric
USAMP
USCAR
UTS
U.S. Automotive Materials Partnership
U.S. Council for Automotive Research
ultimate tensile strength
YSZ
yttria-stabilized zirconia
249
• Office of Advanced Automotive
Technologies FY 2001 Program Highlights
• Vehicle Propulsion and Ancillary Subsystems
• Automotive Lightweighting Materials
• Automotive Propulsion Materials
• Fuels for Advanced CIDI Engines and Fuel Cells
• Combustion and Emission Control for
Advanced CIDI Engines
• Fuel Cells for Transportation
• Advanced Technology Development
(High-Power Battery)
• Batteries for Advanced Transportation
Technologies (High-Energy Battery)
• Vehicle Power Electronics and Electric Machines
• Vehicle High-Power Energy Storage
AUTOMOTIVE L IGHTWEIGHTING M ATERIALS
• Spark Ignition, Direct Injection Engine R&D
2001 A NNUAL P ROGRESS R EPOR T
Office of Transportation Technologies
Series of 2001 Annual Progress Reports
AUTOMOTIVE
LIGHTWEIGHTING
MATERIALS
2001
ANNUAL
PROGRESS
REPORT
• Electric Vehicle Batteries R&D
U.S. Department of Energy
Energy Efficiency and Renewable Energy
Office of Transportation Technologies
www.cartech.doe.gov
DOE/EERE/OTT/OAAT - 2001/003
ACKNOWLEDGEMENT
We would like to express our sincere appreciation to Argonne
National Laboratory and Computer Systems Management, Inc.,
for their artistic contributions in preparing the cover of this
report and to Oak Ridge National Laboratory for its technical
contributions in preparing and publishing this report.
In addition, we would like to thank all our program participants
for their contributions to the programs and all the authors who
prepared the project abstracts that comprise this report.
This document highlights work sponsored by agencies of the U.S. Government. Neither the U.S.
Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied,
or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately
owned rights. Reference herein to any specific commercial product, process, or service by trade name,
trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the U.S. Government or any agency thereof. The views and opinions of authors
expressed herein do not necessarily state or reflect those of the U.S. Government or any agency thereof.
Printed on recycled paper
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement