APPENDIX A DGML Users Manual
APPENDIX A
DGML Users Manual
Appendix A
DESIGN GROUND MOTION LIBRARY
DGML
Version 2.0
Users Manual
Gang Wang, Ph.D., P.E.
Maurice Power, P.E.
Robert Youngs, Ph.D., P.E.
Zhihua (Lillian) Li, Ph.D., P.E.
AMEC Geomatrix, Inc.
August 2009
TABLE OF CONTENTS
Page
0.
DISCLAIMER.............................................................................................................. A-1
1.
INTRODUCTION........................................................................................................ A-1
2.
GETTING STARTED.................................................................................................. A-3
2.1
DGML 2.0 DVD PACKAGE ............................................................................... A-3
2.2
START DGML ................................................................................................... A-5
2.3
DGML BASIC STEPS ......................................................................................... A-5
3.
CREATE TARGET SPECTRUM................................................................................ A-6
3.1
SELECT SPECTRUM MODEL ............................................................................... A-6
3.2
PEER-NGA SPECTRUM .................................................................................... A-7
3.2.1 Select PEER-NGA Ground Motion Models......................................... A-8
3.2.2 Ground Motion Model Input Parameters ............................................. A-8
3.2.3 Number of Standard Deviations and Conditional Mean Spectrum .... A-11
3.3
USER-DEFINED SPECTRUM .............................................................................. A-14
3.3.1 Creating and Loading a User Defined Spectrum File ........................ A-15
3.3.2 Generate User Defined Spectrum....................................................... A-16
3.4
CODE SPECTRUM ............................................................................................. A-16
3.5
CONTROL PANEL ............................................................................................. A-18
3.6
SAVE TARGET SPECTRUM................................................................................ A-19
3.6.1 Save Target Spectrum Report............................................................. A-19
3.6.2 Save Target Spectrum Plot ................................................................. A-21
4.
SEARCH NGA DATABASE .................................................................................... A-22
4.1
SWITCH BETWEEN MAIN AND SUPPLEMENTARY SEARCH ENGINES ................. A-22
4.2
MAIN SEARCH ENGINE: SEARCH ACCORDING TO ACCEPTANCE CRITERIA ...... A-23
4.3
SUPPLEMENTARY SEARCH ENGINE: SEARCH ACCORDING TO NGA
NUMBERS ETC. .................................................................................... A-25
4.4
SPECIFY SCALING METHOD AND WEIGHT FUNCTION ...................................... A-27
4.4.1 Scaling the Records ............................................................................ A-27
4.4.2 Weight Function (Period Array and Weight Array)........................... A-28
4.5
PERFORM THE SEARCH .................................................................................... A-30
4.5.1 Search for Records and Calculate Average Spectrum........................ A-30
4.5.2 List the Search Result......................................................................... A-31
4.6
SELECTION AND EVALUATION OF RECORDS .................................................... A-33
4.6.1 Highlight an Individual Record .......................................................... A-33
4.6.2 Highlight Response Spectrum of an Individual Record ..................... A-34
4.6.3 Highlight Time History of an Individual Record ............................... A-35
4.6.4 Zoom In Time Function for Examining the Time History of an
Individual Record ............................................................................... A-35
4.6.5 Accept or Reject an Individual Record .............................................. A-37
4.7
GRAPHIC CONTROL ......................................................................................... A-38
4.7.1 Graphic Control Panel ........................................................................ A-38
4.7.2 Change Plot Axes ............................................................................... A-38
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-i
TABLE OF CONTENTS
(Continued)
Page
4.8
SAVE DGML SEARCH RESULT ........................................................................ A-39
4.8.1 DGML Search Report......................................................................... A-39
4.8.2 Save Acceleration Time History Files................................................ A-43
4.8.3 Save the Plot ....................................................................................... A-43
5.
DGML EXAMPLES .................................................................................................. A-45
5.1
RECORD SELECTION AND MODIFICATION ........................................................ A-45
5.2
USE DGML SUPPLEMENTARY SEARCH ENGINE .............................................. A-52
6.
DGML MATLAB SOURCE CODE.......................................................................... A-56
6.1
INTRODUCTION ................................................................................................ A-56
6.2
RUN DGML IN MATLAB ................................................................................. A-57
6.3
VIEW GUI OBJECT CALLBACK FUNCTIONS ..................................................... A-57
REFERENCES ....................................................................................................................... A-61
TABLES
Table 1:
Table 2:
Table 3:
Table 4:
Table 5:
Parameters for PEER-NGA Ground Motion Models ...........................................A-9
Parameters for DGML Search Engine ................................................................A-24
Listed Record Information of DGML Search Result..........................................A-32
Selected Ground Motion Records .......................................................................A-49
Descriptions of DGML Matlab Source Codes ....................................................A-56
FIGURES
Figure 1:
Figure 2:
Figure 3:
Figure 4:
Figure 5:
Figure 6:
Figure 7:
Figure 8:
Figure 9:
Figure 10:
Figure 11:
Figure 12:
Figure 13:
Figure 14:
Figure 15:
Figure 16:
Figure 17:
Figure 18:
Directory Structure on DGML Version 2.0 DVD Package ..................................A-3
DGML Start Window............................................................................................A-5
Create Target Spectrum Window..........................................................................A-6
Selection of Spectrum Model................................................................................A-7
Input Panel for Creating a PEER-NGA Spectrum ................................................A-7
Definition of Fault Geometry and Distance Measures........................................A-11
Constructing a Constant Epsilon Spectrum ........................................................A-12
Example of Constant Epsilon Spectrum .............................................................A-12
Input Needed to Construct a Conditional Mean Spectrum .................................A-13
Example of Conditional Mean Spectrum............................................................A-14
Selecting a User-Defined Spectrum Model ........................................................A-14
Example of User-Defined Spectrum File............................................................A-15
Selecting a User-Defined Spectrum File From the File Menu............................A-15
Direct Input of the File Name and Path for a User-Defined Spectrum...............A-16
Plot of User-Defined Spectrum Using Data Listed in Figure 12 ........................A-16
ASCE/SEI 7-05 Code Spectrum .........................................................................A-17
Selecting the ASCE Code Spectrum Model .......................................................A-17
Input Needed to Construct a Code Spectrum......................................................A-17
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-ii
TABLE OF CONTENTS (continued)
FIGURES (continued)
Figure 19:
Figure 20:
Figure 21:
Figure 22:
Figure 23:
Figure 24:
Figure 25:
Figure 26:
Figure 27:
Figure 28:
Figure 29:
Figure 30:
Figure 31:
Figure 32:
Figure 33:
Figure 34:
Figure 35:
Figure 36:
Figure 37:
Figure 38:
Figure 39:
Figure 40:
Figure 41:
Figure 42:
Figure 43:
Figure 44:
Figure 45:
Figure 46:
Figure 47:
Figure 48:
Figure 49:
Figure 50:
Figure 51:
Figure 52:
Figure 53:
Figure 54:
Figure 55:
Figure 56:
Figure 57:
Figure 58:
Figure 59:
Figure 60:
ASCE/SEI 7-05 Code Spectrum Created Using the Values Listed in Figure 18A-18
The Control Panel ...............................................................................................A-19
Click “Save Target” Button to Save Target Spectrum........................................A-20
Example of DGML Target Spectrum Report......................................................A-20
Save DGML Target Spectrum Plot .....................................................................A-21
DGML Search Engine Window ..........................................................................A-22
Switch Between Main and Supplementary Search Engines ...............................A-23
Main Search Engine (Default) User Interface.....................................................A-24
Supplementary Search Engine User Interface ....................................................A-26
Examples of Specifying Weight Function ..........................................................A-29
Example of Average Spectrum of Selected Records ..........................................A-31
Output List Window............................................................................................A-31
Close-up of Output List Window........................................................................A-32
Highlight an Individual Record on the Output List ............................................A-33
Highlight the Response Spectrum of an Individual Record................................A-34
Plot of the Scaled Acceleration/Velocity/Time History of an Individual Record in
Fault Normal, Fault Parallel and Vertical Directions .........................................A-35
Zoom In Time History Plot.................................................................................A-36
Steps to Accept or Reject an Individual Record .................................................A-37
Graphic Control Panel.........................................................................................A-38
Example of Four Options to Change Plot Axes ..................................................A-39
Save DGML Search Result .................................................................................A-40
An Example of DGML Search Report................................................................A-41
Save Acceleration Time History Files ................................................................A-43
Steps to Save the Plot..........................................................................................A-44
Select Spectrum Model .......................................................................................A-45
Specify Spectrum Parameters .............................................................................A-45
The Target Design Spectrum ..............................................................................A-46
Specify the Search Criteria .................................................................................A-47
DGML Search Result..........................................................................................A-48
Modify the List of Selected Records and Re-average.........................................A-49
Average Spectrum of 7 Selected Records...........................................................A-50
The Acceleration and Velocity Time Histories of Selected Records..................A-52
Search by NGA Sequence Using The Supplementary Search Engine................A-53
Plot Spectra of Selected 7 Records .....................................................................A-54
Search by Event Name Using the Supplementary Search Engine ......................A-54
Update to Include the Northridge Record ...........................................................A-55
Plot Spectra of Updated 7 Records .....................................................................A-55
Use Matlab “Guide” Command to Open DGML GUI .......................................A-58
GUI Layout of “Target Spectrum” Page.............................................................A-58
GUI Layout of “Search Engine” Page ................................................................A-59
View Callback Functions of a GUI Object .........................................................A-59
Source Code of a Callback Function...................................................................A-60
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-iii
0. DISCLAIMER
DGML Version 2.0 (Release Date: August 2009) was developed based on knowledge believed
to be accurate and reliable. Due to the possibility of human or mechanical error as well as other
factors, this software is provided “as is” and the authors make no representation, express or
implied, as to the accuracy, reliability, completeness, or timeliness of this software, and are not
responsible for any loss or damages incurred by parties using this software.
The acceleration time history database used in DGML is the PEER-NGA strong-motion
database, which was rotated to fault-normal and fault-parallel directions by a third party.
Further QA/QC may be needed to ensure the quality of this database. Any comments,
questions, bug reports regarding the DGML software and the Users Manual can be addressed to
the authors of the program (email: [email protected]); [email protected];
[email protected]; [email protected]).
1. INTRODUCTION
DGML Version 2.0 is an interactive software package that allows the user to select sets of
strong ground motion acceleration time histories that are representative of design ground
motions. The user specifies the design ground motions in terms of a target response spectrum
and the desired characteristics of the earthquake ground motions in terms of earthquake
magnitude, source-to-site distance and other general characteristics. The DGML tool then
selects acceleration time histories from the PEER-NGA database for rotated fault-normal and
fault-parallel acceleration time histories that satisfy the user-specified selection criteria and
provide good fits to the target response spectrum.
DGML Version 2.0 has been developed based on the capability of Matlab’s Graphic User
Interface (GUI) (Matlab ® version 7.2). The software was compiled using Matlab Compiler ®
into a stand-alone program so that it can be executed in any Windows PCs independent of
Matlab environment.
Several features of DGML Version 2.0 are highlighted as follows:
Graphic User Interface: GUI provides a user-friendly interface for data input and processing.
The user’s operation involves checking boxes, selecting pop-up menus, and pushing buttons.
There are numerous features that are designed to facilitate easy usage. One example is the
software can automatically retrieve previously specified input data.
Interactive Plotting: Results in each step can be visualized in real time, and results from
different sets of input parameters can be easily compared. The software provides a list of the
selected records with important information. By clicking the list, users can visually inspect the
response spectrum, acceleration/velocity/displacement time-history of each individual record
for each component.
Flexibility: The DGML Version 2.0 provides users flexibility to exercise different criteria to
select the design records. Users have the options to scale or not to scale the records; to select
the record according to the geometric mean of fault normal and fault parallel components or
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-1
according to a single component. The user can choose to select records with pulses or without
pulses. The number of output records can be user-specified, and users have the capacity to
manually select and adjust the results to meet their specific requirement. In addition to the
features provided in the previous release, DGML version 2 provides a supplementary search
engine to allow users to search the database by NGA numbers, earthquake names, and station
names. The new search engine gives users flexibility to inspect any record in the database.
Easy Output: DGML version 2 provides easy ways to output search results, plots and tables.
The software can automatically generate a “Target Spectrum Report” and a “Search Result
Report” and save them into text files or csv (Comma Separated Values) files. The DGML
reports summarize search criteria, scale factors, scaled spectra of selected records, and other
record information, and they can be opened by Microsoft Excel. The response spectra plots and
time history plots can be exported into figure files of different formats. Finally, files containing
acceleration time histories of selected records can be saved for each project.
Efficiency: The algorithm of the DGML Version 2.0 package is robust and efficient. The
search engine can scan and sort the NGA database within a few seconds.
Extendibility: The DGML Version 2.0 package is directly connected to the NGA flatfile and
strong motion database, so it can be easily upgraded to accommodate future development of the
NGA database.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-2
2. GETTING STARTED
2.1 DGML 2.0 DVD PACKAGE
The DGML Version 2.0 DVD contains the compiled executable DGML program, PEER-NGA
rotated fault normal/fault parallel dataset, and electronic version of DGML report and User’s
Manual. The directory structure of the released DVD is shown on Figure 1.
Figure 1: Directory Structure on DGML Version 2.0 DVD Package
The batch file “DGML.bat” is used for users to launch the DGML graphic interface.
Although DGML source codes are written in Matlab® (v7.2), they are compiled using Matlab
Compiler® (v7.10) to create a stand-alone executable program that does not require Matlab.
The compiled DGML programs are located in “DVD Drive:\DGML\exe” directory. “DVD
Drive:\DGML\mcr” contains expanded files from “Matlab Component Runtime (v7.10)”,
which are required to run the compiled program.
DGML source codes contain all *.m files, *.fig files in “DVD Drive:\DGML\” directory and
all files in “DVD Drive:\DGML\Utility\” subdirectory. DGML can also be executed directly
in Matlab environment from DVD Drive:\DGML\DGMLStart.m
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-3
The NGA rotated fault normal/fault parallel acceleration record dataset is also contained in the
DVD. It is stored in the “DVD Drive:\FNFPDataset\” directory. The NGA original (unrotated)
database is stored in the “DVD Drive:\NGADatabase\” directory. These records can be viewed
by a text editor.
IMPORTANT NOTES FOR THE FIRST TIME USER
DGML Version 2.0 is designed to run directly on the DVD drive. User should have
administrator privilege to run DGML, since it involves operation to write files (very minimum
but necessary) on the local disk. Once the program starts, DGML automatically creates a
directory on the user’s local disk to store intermediate files and output files. This directory is
named “C:\DGML\”. It is safe to delete “C:\DGML\” directory and files after running the
program.
For the first-time user to run DGML, user needs to install Visual C++ 2005 Redistributable
Package first if the package is not already installed in the local machine. Otherwise,
DGML.bat cannot be properly executed. DGML is written in Matlab language and was
compiled through Matlab Compiler v7.6 to create a stand-alone program, which requires this
new installation procedure.
The installer is located in “DVD:\installer\” directory. For 32-bit CPU (applies to most
Pentium, Athlon, AMD CPUs etc. in most personal computers), run vcredist_x86.exe, For 64bit CPU (eg. super-computers, stations, servers etc..), run vcredist_x64.exe. The Microsoft
Visual C++ 2005 Redistributable Package installs runtime components of Visual C++
Libraries required to run applications developed with Visual C++ on a computer that does not
have Visual C++ 2005 installed. The file does not do any harm to the user's computer. User
only needs to install Visual C++ 2005 Redistributable Package once.
The program is best viewed on a 15- inch or bigger screen with resolution 1280 by 1024 pixels.
Graphic quality may deteriorate under other lower resolution settings.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-4
2.2 START DGML
To start DGML Version 2.0, double click the batch file “DVD Drive:\ DGML.bat” to launch
the program from the DVD. The start window will show up as Figure 2. Click the “START”
button to start the program.
Click START
Figure 2: DGML Start Window
2.3 DGML BASIC STEPS
There are two basic steps in the use of the DGML software. Step one is the creation of the
target response spectrum. The process of creating the target spectrum is described in
Section 3.0 of the Users Manual. Step two is to search the PEER-NGA database for recordings
that satisfy user-specified selection criteria and whose response spectra are similar to the target
spectrum. Section 4.0 of the Users Manual describes the process of specifying selection criteria
and developing sets of acceleration time histories. A user can inspect each ground motion
record to finalize the selection.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-5
3. CREATE TARGET SPECTRUM
The TARGET SPECTRUM window is shown in Figure 3. The window contains the following
main parts: (1) Select Spectrum Model; (2) PEER-NGA Spectrum; (3) User Defined Spectrum;
(4) Code Specification; (5) Plot Control panel; (6) Spectrum plot; (7) Explanation of notations;
(8) Save Target Spectrum button; (9) Go to Next Step to Perform DGML Search.
2
6
1
3
7
4
5
Figure 3: Create Target Spectrum Window
8
9
3.1 SELECT SPECTRUM MODEL
The first step is to select the spectrum model to generate the target spectrum. Three options are
available by mouse-clicking the drop menu at location (1) in Figure 3:
(1) PEER-NGA spectrum;
(2) User defined spectrum;
(3) ASCE/SEI Standard 7-05 code specified spectrum.
Once a spectrum model is selected, the corresponding panel for parameter input will be
highlighted in yellow color.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-6
Click drop menu to
select spectrum
models
Figure 4: Selection of Spectrum Model
3.2 PEER-NGA SPECTRUM
The “PEER-NGA Spectrum” model creates a target response spectrum using the PEER-NGA
ground motion models published in the February, 2008 issue of Earthquake Spectra. Five NGA
empirical models are employed in DGML Version 2: Abrahamson-Silva (A&S, 2008), BooreAtkinson (B&A, 2008), Campbell- Bozorgnia (C&B, 2008), Chiou-Youngs (C&Y, 2008a), and
Idriss (2008). The spectrum is defined for a specific scenario earthquake defined in terms of
magnitude, distance, style of faulting, and site conditions as specified in the PEER-NGA
ground motion models. This option is selected by clicking the drop menu and selecting “PEERNGA spectrum” option. The panel shown at location (2) in Figure 3 allows the user to input the
necessary parameters to generate a PEER-NGA spectrum. An expanded view of this input
panel is shown in Figure 5. The required input parameters are described below.
Check box to select
NGA model(s), see
Sec 3.2.1.
Input boxes to
enter NGA model
parameters, see
Sec. 3.2.2
Input boxes to specify
number of standard
deviations, see
Sec .2.3
Click button to specify
method of averaging
selected NGA models,
see Sec. 3.2.1
Click button to enable or Input box to specify the period
disable use of
of conditional mean method,
conditional mean
see Sec. 3.2.3
method, see Sec. 3.2.3
Figure 5: Input Panel for Creating a PEER-NGA Spectrum
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-7
3.2.1 Select PEER-NGA Ground Motion Models
Use check boxes to select the PEER-NGA ground motion models to be used. The user has the
option to use any single, or any combination, of these NGA models, except that the Idriss
(2008) model is not applicable to cases where VS30 is less than 450m/s, or fault type is normal
or normal oblique faulting. In such cases, DGML will automatically not select the Idriss (2008)
model. If the Idriss (2008) model is selected by the user and is applicable for the conditions
specified by the user (i.e. Vs30>=450m/s and strike slip faulting or reverse or reverse oblique
faulting), then the Idriss model will be used.
If more than one model is selected, the user can further specify the resulting average target
spectrum to be the arithmetic or the geometric mean of the spectra produced by the selected
models.
The arithmetic mean of n spectra is defined as follows
SA(T ) =
1⎛ n
⎞
⎜ ∑ SAi (T ) ⎟
n ⎝ i =1
⎠
where SAi (T ) is the spectrum produced by model i; The geometric mean of spectra can be
viewed as the arithmetic mean of the logarithm-transformed values (i.e., the arithmetic mean of
the logarithms), and then using exponentiation to return the logarithm values to the natural
scale actual values. The geometric mean of n spectra is defined as follows
ln SA(T ) =
1⎛ n
⎞
⎛1 n
⎞
SA(T ) = exp⎜ ∑ ln SAi (T ) ⎟
⎜ ∑ ln SAi (T ) ⎟
n ⎝ i =1
⎠ or
⎝ n i =1
⎠
3.2.2 Ground Motion Model Input Parameters
Depending on the PEER-NGA ground motion models selected, the user must enter up to twelve
parameters in the data input boxes to construct the response spectrum. Parameters required by
each ground motion model are listed in Table 1. DGML will only display the necessary
parameters according to the model selection specified by the user.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-8
Table 1: Parameters for PEER-NGA Ground Motion Models
A&S
B&A
C&B
C&Y
Magnitude
9
9
9
9
9
Fault Type
9
9
9
9
9
Dip (deg)
9
9
9
Dip angle of rupture plane
Width (km)
9
9
9
Down-dip width of rupture plane
ZTOR (km)
9
9
9
Depth to top of rupture plane
RJB (km)
9
9
9
Joyner-Boore distance to rupture plane
RRup (km)
9
9
9
RX (km)
9
VS30 (m/s)
9
estimated
9
9
Specifying VS30 is estimated or measured
Z1.0 (km)
9
9
Depth to VS=1.0 km/s horizon
9
Idriss
9
9
9
9
9
Moment magnitude of the earthquake
Types of fault mechanism. Options are:
(1) Strike Slip; (2) Normal or Normal
Oblique; (3) Reverse or Reverse Oblique
Closest distance to rupture plane
Site coordinate w.r.t. top of rupture
9
9
Z2.5 (km)
Explanations
Epsilon
9
9
9
9
9
T_eps
9
9
9
9
9
Average shear wave velocity of top 30 m
Depth to VS=2.5 km/s horizon
Number of standard deviations away from
the median spectrum
The period upon which conditional mean
spectrum is conditioned
Remarks:
1. Magnitude, Fault Type, Dip, Width and ZTOR are parameters that define the seismic
source. RJB, RRup, RX define different types of distance measures from the site to the
fault rupture. VS30, Z1.0 and Z2.5 are used to describe the site condition.
2. Definition of RJB, RRup, and RX are shown in Figure 6 for strike-slip faulting and for
reverse or normal faulting with the site on the hanging-wall and the foot-wall side. Rrup
is the closest distance to the rupture plane; RJB is Joyner-Boore distance defined as the
closest horizontal distance to the trace of the rupture that is vertically projected to the
ground surface. RRup and RJB are always non-negative numbers. Please note that RX is
the horizontal coordinate of the site with respect to the top of the rupture. For a reverse
or normal fault, if the site is on the hanging wall side, RX is a positive value; if the site
is on the footwall side, RX is a negative value. Therefore, DGML does not provide a
separate hanging wall flag, instead, the sign of RX is used to indicate a hanging wall
condition for a reverse or normal fault. For a strike-slip fault, it does not matter whether
RX is positive or negative.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-9
3. Parameters Z1.0 and Z2.5 specify the depths at which the shear wave velocity reaches 1.0
km/s and 2.5km/s at the site, respectively. A user can specify the values of Z1.0 and Z2.5
directly as shown in the following check boxes. The “default” checkboxes will be
automatically unchecked once a user keys in numbers.
user-specified Z1.0 and Z2.5 values
4. Default values of Z1.0 and Z2.5 can also be used by checking the “default” checkboxes.
Empirical relationships are used to estimate Z1.0 value based on VS30, and the default
values are automatically displayed in the text boxes thereafter.
use default values of Z1.0 and Z2.5
A&S model uses the following empirical equations to estimate Z1.0, where the unit
of Z1.0 is in km, and VS30 is in m/s.
⎧ exp (6.745) / 1000
⎪
Z1.0 = ⎨ exp (5.394 - 4.48 *1n(VS30 /500))/1000
⎪ exp (6.745 - 1.35 *1n(V /180))/1000
S30
⎩
if
if
if
VS30 < 180 m/s
VS30 > 500 m/s
500 m/s ≥ VS30 ≥ 180 m/s
C&Y model uses the following relationship to estimate Z1.0,
Z1.0 = exp (28.5 – 3.82/8 *1n(VS308 + 378.88))/1000
Please note the difference of empirical correlations used by A&S and C&Y, where the
estimate of Z1.0 from C&Y is always smaller than that from A&S. If both A&S and
C&Y models are specified in DGML, the above relation is used accordingly for each
model. For simplicity, DGML only displays Z1.0 values estimated by A&S model, but
different Z1.0 values are used as the default for C&Y model. Parameter Z2.5 is used only
by C&B model. Default value of Z2.5 is determined based on the value of Z1.0. If Z1.0 is
specified by the user, the following relation is used to estimate Z2.5 based on Z1.0 (both
in units of km)
Z2.5 = 0.519 + 3.595 * Z1.0
Otherwise, DGML estimates Z1.0 first using relationship proposed by A&S, and then
Z2.5 is estimated using above equation.
5. It is the user’s responsibility to ensure the input parameters are correct. The DGML
does NOT check the consistency of the input data.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-10
RX=RJB (positive)
RX (positive)
Surface
RJB
Site
ZTOR
Surface
Site
RRup
ZTOR
RRup
DIP
DIP
Fault
Width
Width
Fault
(a) Strike slip faulting
(b) Reverse or normal faulting, hanging-wall site
RX (negative)
RJB
Surface
Site
RRup
ZTOR
DIP
Fault
Width
(c) Reverse or normal faulting, foot-wall site
Figure 6: Definition of Fault Geometry and Distance Measures
3.2.3 Number of Standard Deviations and Conditional Mean Spectrum
The PEER-NGA ground motion models provide relationships for the median ground motion
amplitude (mean value in log space) and for the aleatory variability about the median value.
The aleatory variability is defined as the standard deviation of the natural log of the spectral
acceleration. Parameter epsilon is used to define the number of standard deviations away
from the median represented by a particular ground motion level. For example, the 84thpercentile spectrum is a spectrum where the ground motion levels are one standard deviation
above the median at all spectral periods (epsilon = 1.0).
The user has two options for incorporating aleatory variability in the target spectrum developed
from the PEER-NGA models. The first option is to use a constant value of epsilon at all
spectral periods (e.g. epsilon=1.0 for an 84th-percentile spectrum). The second option is to
develop a conditional mean spectrum (Baker and Cornell, 2006) in which the user specifies the
value of epsilon at a specific spectral period and the correlation model developed by Baker and
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-11
Jayaram (2008) is used to compute the expected value of epsilon at other spectral periods.
(1) Constant Epsilon Spectrum
The constant epsilon spectrum is constructed by specifying a single value of epsilon for all
periods (e.g. an 84th-percentile spectrum). To construct the constant epsilon spectrum the user
provides the desired epsilon value and disables the method of conditional mean spectrum, as
shown in Figure 7.
Click the button to “NO”
to disable conditional
mean method
Enter the Epsilon value
Click to choose between “geometric mean” or “arithmetic
mean” of selected NGA model spectrum
Figure 7: Constructing a Constant Epsilon Spectrum.
Figure 8 shows examples of constant epsilon spectra generated using the average of five NGA
models and epsilon = 0, 1, and 2. Model parameters are specified as shown in Figure 5. Epsilon
is changed to be 0, 1 and 2 for each case as shown in Figure 8. The three cases are plotted
together in the same graph for easy comparison, by using “Hold on” button and “change color
of lines menu” provided in the Control Panel, see Sec. 3.5.
1
10
Spectral Accleration, Sa (g)
0
10
Epsilon=2
−1
10
Epsilon=1
Epsilon=0
−2
10
−3
10
−2
10
−1
0
10
10
1
10
Period, T (sec)
Figure 8: Example of Constant Epsilon Spectrum
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-12
(2) Conditional Mean Spectrum (CMS)
A conditional mean spectrum can be constructed using NGA models and the correlation
coefficient model by Baker and Jayaram (2008) (see text of the report for a complete
description). The inputs needed to construct a conditional mean spectrum are shown in Figure
9.
Figure 9: Input Needed to Construct a Conditional Mean Spectrum
Figure 10 shows an example of a conditional mean spectrum (CMS) created for an epsilon
value of 1.0 at a spectral period of 0.5 seconds. Model parameters are specified as shown in
Figure 5 using input parameters for the CMS are the same as shown in Figure 9. The solid
black line shows the generated CMS for epsilon=1 and T_eps=0.5 sec. Constant epsilon spectra
for epsilon=0, 1, as shown previously in Figure 8, are also plotted in Figure 10 for comparison
with the CMS.
The value of epsilon may be selected in a variety of ways. One approach would be to use the
results of epsilon deaggregation from a PSHA calculation. Alternatively, the user may specify a
target spectral acceleration at T_eps and then enter trial values of epsilon until the resulting
average conditional mean spectrum matches the target value.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-13
Epsilon=0
Conditional Mean
Spectrum, Epsilon=1,
T_eps=0.5 sec
Epsilon=1
Figure 10: Example of Conditional Mean Spectrum
3.3 USER-DEFINED SPECTRUM
A user-defined spectrum is any target response spectrum the user wishes to use. It may
represent a uniform hazard spectrum from a PSHA or a scenario earthquake spectrum created
using other ground motion models. There is no limit for the number of spectral periods and
spectral acceleration values that may be entered.
Select a “user-defined spectrum” model by clicking the drop menu (Location ① in Figure 3)
and select a “user-defined spectrum” option.
Figure 11: Selecting a User-Defined Spectrum Model
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-14
3.3.1 Creating and Loading a User Defined Spectrum File
A file containing a user defined spectrum is created using a text editor. An example file is
shown in Figure 12.
# User Specified Spectrum
# T(sec) Sa(g)
# -----------------------------------0.01
0.54
0.02
0.54186
0.075 0.8481
0.1
1.01966
0.2
1.31426
0.3
1.3162
0.5
1.10297
1
0.74228
1.5
0.5379
2.0
0.40813
3
0.24773
4
0.16592
10
0.05
Comment lines are the lines that contain any
characters (as long as it is not pure numbers)
Data field, each line contains spectral
period (T) and spectral acceleration (SA)
data in pairs, separated by blanks or tabs.
There is no limit for the number of data
that maybe entered.
Figure 12: Example of User-Defined Spectrum File
DGML provides two methods to load a user-defined spectrum. (1) Click the “Open File” button
to select from directory. The following figure shows an example:
Select the directory that
contains the user-defined
spectrum file
Select the file and click “Ope
Click “Open File” button
Figure 13: Selecting a User-Defined Spectrum File From the File Menu
or (2) key in complete path and name of the file in the window box. The following is an
example:
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-15
Type the complete path and
name of the file, and hit
“Enter” on keyboard
Figure 14: Direct Input of the File Name and Path for a User-Defined Spectrum
3.3.2 Generate User Defined Spectrum
Click Create button in Control Panel, and the user defined spectrum will be plotted in the
plotting area, as shown in Figure 15.
User Defined Spectrum
1
Spectral Accleration, Sa (g)
10
0
10
−1
10
−2
10
−2
10
−1
0
10
10
1
10
Period, T (sec)
Figure 15: Plot of User-Defined Spectrum Using Data Listed in Figure 12.
3.4 CODE SPECTRUM
The code specified design response spectrum is in accordance with ASCE Standard ASCE/SEI
7-05 specified in the "Minimum Design Loads for Buildings and Other Structures", published
by the American Society of Civil Engineers, 2006. The code specification requires three points
to construct the spectrum, as follows: site-adjusted short period (0.2 sec) spectral acceleration
(SDS), site-adjusted one-second period (1.0 sec) spectral acceleration (SD1), and the transition
period (TL) between constant spectral velocity and constant spectral displacement regions of the
spectrum. These parameters are illustrated in Figure 16.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-16
Figure 16: ASCE/SEI 7-05 Code Spectrum
The procedures, equations, and parameters for constructing each branch of the spectrum are
given in ASCE Standard ASCE/SEI 7-05.
Example:
Step 1: To activate the function to generate a target spectrum according to the code
specification, users need to select the spectrum model by selecting “ASCE code specification”
from the menu at location ① in Figure 3.
Figure 17: Selecting the ASCE Code Spectrum Model
An illustration will appear in the plotting area to explain the symbols of the spectrum as
specified by the code.
Step 2: Key in Sds (g), Sd1 (g), TL (sec) in the Input Windows for the Code Spectrum
Figure 18: Input Needed to Construct a Code Spectrum.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-17
Step 3: Press the Create Button at Control panel.
Step 4: A target spectrum is then generated in the plotting area as illustrated in Figure 19.
Figure 19: ASCE/SEI 7-05 Code Spectrum Created Using the Values Listed in
Figure 18.
3.5 CONTROL PANEL
The control panel is used to provide options for automatic data input and plot adjustment.
Figure 20 displays the layout of the control panel, and each of the functions will be described in
the following:
(1) “Create” button is used to create and plot the target spectrum using current spectrum model
and parameters.
(2) “Clear” button and “Mem” button are designed for automatic data input. Clicking “Clear”
button will fill in “0” in all input boxes of Figure 5. Clicking “Mem” button will
automatically fill in all input boxes of Figure 5 with data from the previous run.
(3) Drop menus in the right side of the panel are used to control the plot of the target spectrum:
drop menu A to change the plot axis, e.g. from log-log plot to linear plot; drop menus B and
C to select the color and style of the lines to be plotted.
(4) “Hold On” checkbox allows a new graph to be plotted on top of the previous plot for easy
comparison. If Hold On box is checked, the plot will be displayed in the plot window
without erasing the previous plot. If Hold On box is unchecked, each time the new plot is
created, the previous one is erased. “Grid On” checkbox is to display the grid lines on the
plot window. “Only Average” checkbox, if checked, displays only the average spectrum on
the plot window if more than one PEER-NGA model is used. “Normalized” checkbox, if
checked, will plot the normalized target spectrum (normalized with respect to PGA).
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-18
Clear: Click to clear all input data boxes
Mem: Click to automatically refill all input data boxes
A
B
C
Click to change
style of lines
Click to Create
target spectrum
Click to change
plot axis
Click to change
color of lines
Hold On: Check to allow the next plot to superimpose on the previous one
Grid On: Check to show grid lines on the plot
Only Average: Check to only plot the average spectrum if more than one PEER-NGA
models is selected.
Normalize: Create target spectrum normalized with respect to PGA
Figure 20: The Control Panel
3.6 SAVE TARGET SPECTRUM
3.6.1 Save Target Spectrum Report
DGML Version 2 allows users to output the generated target spectrum in .csv format (Comma
Separated Variables file) or .txt (ASCII file) by clicking “Save Target” button in the lower right
side of the window, as shown in Figure 21. The CSV format file can be directly opened by
Microsoft Excel or any text editor, providing the most convenient way for data export.
Figure 22 demonstrates a DGML automatically generated target spectrum report using
parameters prescribed in Figure 5. In this file, the parameters that are used in generating the
target spectrum are listed. In this example, the VS30 value is annotated as “estimated”, and
default values of Z1.0 and Z2.5 for A&S, C&Y and C&B model are reported separately.
Spectrum values by individual NGA model together with the mean spectrum are listed for
periods of 0.01, 0.02, 0.03, 0.04, 0.05, 0.075, 0.1, 0.15, 0.2, 0.25, 0.3, 0.4, 0.5, 0.75, 1, 1.5, 2, 3,
4, 5, 7.5, and 10 seconds.
For a code-specified target spectrum, only the periods and spectrum values are reported. The
exported values are interpolated at period intervals that are evenly spaced in log space, with
100 points per decade.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-19
Figure 21: Click “Save Target” Button to Save Target Spectrum
Figure 22: Example of DGML Target Spectrum Report
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-20
3.6.2 Save Target Spectrum Plot
The spectrum plot can be saved to the disk as figure files. Right click mouse on the plot area,
and a “Save Plot As” box will show up. Left click the box to pop up a window, select the
proper directory, and enter the name of the graphic file to save. The plot can be saved as .bmp,
.jpg, .tiff, .eps, or pdf format by selecting the proper “Save as type”. Figure 23 illustrates the
steps to save the spectrum plot.
Figure 23: Save DGML Target Spectrum Plot
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-21
4. SEARCH NGA DATABASE
The DGML SEARCH ENGINE window contains eight main parts, as labeled in Figure 24: (1)
Search Engine to specify the record acceptance criteria and perform search over NGA database;
(2) Specification of Weight Function used for scaling records; (3) Spectra plotting window; (4)
Weight Function plot; (5) Acceleration/ Velocity/ Displacement time history plotting of a
selected record; (6) Ground motion record information output list; (7) Graphic control panel for
line styles and display of ground motion components; (8) Buttons to accept, reject individual
record, save the search results and selected acceleration time history files.
7
1
3
2
5
4
6
8
Figure 24: DGML Search Engine Window
4.1 SWITCH BETWEEN MAIN AND SUPPLEMENTARY SEARCH ENGINES
DGML Version 2 provides two types of search engines. Figure 25 illustrates the switch
between two search engine interfaces by clicking “Switch” button. The main search engine
searches the database according to the specified acceptable ranges for the characteristics of the
recordings (e.g. the appropriate magnitude, distance range etc, termed as “acceptance criteria”).
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-22
The supplementary search engine searches the database according to specified NGA sequence
numbers, event names, and station names. The two search engines are described in the
following sections. By default, DGML displays the main search engine.
Figure 25: Switch between Main and Supplementary Search Engines
4.2 MAIN SEARCH ENGINE: SEARCH ACCORDING TO ACCEPTANCE CRITERIA
The acceptance criteria for recordings are entered in the search engine data boxes shown in
Figure 26 (location 1 in Figure 24). As described below, acceptance criteria are specified by
indicating the allowable range or restriction for the parameters listed in Table 2. Any recording
satisfying these criteria will be passed on to the second stage of comparison with the target
spectrum.
Data field input format
•
The range of acceptable values for a recording parameter is entered into the data box
with a yellow background color. The acceptance range is specified by the minimum and
the maximum values. The minimum and maximum values should be delimited by a
comma or space. For example:
•
An input box in yellow colors can be left BLANK if no search restriction is imposed on
that data field. For example:
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-23
Switch the search engine
Drop menu or text box to
specify the minimum and
maximum values of the
search restriction, see
Table 2
Press button to perform the
search
Text box to
specify the
number of
output, see
Table 2
Specify scaling method and
weight function, see sec. 4.4
Figure 26: Main Search Engine (Default) User Interface.
Table 2: Parameters for DGML Search Engine
Data Field
Explanations
Magnitude
Restrict range of moment magnitude, input in the format of [min, max] or
leave as blank for no restriction.
Fault Type
D5-95(sec)
R_JB (km)
R_rup (km)
Vs30 (m/s)
Components
Types of fault mechanism. Options are: (1) All types of fault; (2) Strike Slip;
(3) Normal or Normal Oblique; (4) Reverse or Reverse Oblique; (5)
Combination of (2, 3); (6) Combination of (2,4); (7) Combination of (3,4).
Restrict range of the significant duration of the records, input in the format
of [min, max], or leave as blank for no restriction. The duration is defined as
the time needed to build up between 5 and 95 percent of the total Arias
intensity.
Restrict range of Joyner-Boore distance, input in the format of [min, max],
or leave as blank for no restriction.
Restrict range of closest distance to rupture plane, input in the format of
[min, max], or leave as blank for no restriction.
Average shear wave velocity of top 30 meters of the site.
Restrict the directional component of the records to search. Options are: (1)
Use fault normal and fault parallel in pair, and use geometric mean of fault
normal and fault parallel components for spectral matching; (2) Use only
fault normal component; (3) Use only fault parallel component; (4) Use
arbitrary component, either fault normal or fault parallel, regardless of the
direction.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-24
Data Field
Pulse
Factor Limit
Ts (sec)
Period Array
Weight Array
Total Num.
Output
Total Num.
Averaged
Explanations
Restrict the pulse characteristics of the searched record. Options are: (1) Any
record; (2) Only pulse-like record; (3) No pulse-like record.
Restrict range of scale factors, input in the format of [min, max], or leave as
blank for no restriction. The parameter is applicable only if “scaling” button
is “YES” and “single period” button is “NO” (i.e. the records are scaled, but
are not scaled to a single period).
The period whose target spectral value all records are scaled to. The
parameter is applicable only if “scaling” button is “YES” and “single period”
button is “YES” (i.e. the records are scaled, but are scaled only to a single
period).
A real number sequence of periods used to specify the weight function. The
number sequence is in ascending order between [0.01, 10].
A real number sequence of weights used to specify the weight function. The
number sequence is in one-to-one correspondence to that of the Period
Array.
Specify the total number of records to display in the “ground motion record
information display window”. If the number of qualified records is less than
the specified number, the Total Num. Output will be reset to the maximum
possible. Please note that the current version of DGML only displays at most
up to 200 records. The restriction is necessary to avoid abuse of the software.
Specify the total number of record spectra to compute the average spectra.
The number will be automatically reset to the maximum possible if the total
output record number is less.
4.3 SUPPLEMENTARY SEARCH ENGINE: SEARCH ACCORDING TO NGA NUMBERS ETC.
Supplementary Search Engine is a new feature in DGML Version 2, which allows users to
search, scale, and rank the records according to specified NGA sequence numbers, event name,
or station name. This feature is particularly useful for users to inspect any particular record,
especially convenient to fine-tune the search results from the Main Search Engine.
NGA sequence number is a unique number assigned to each pair of FN/FP records in the NGA
flatfile. The sequence number can combine with the “Components” drop menu to specify which
components are specified, and “Pulse” drop menu for specification of pulse characteristics. The
NGA sequence number should be positive numbers, except that if the “Components” is chosen
to be “Any Component, Either FN or FP”, one needs to distinguish between the FN and FP
components of the record which have the same NGA sequence number. In this case, positive
numbers represent fault normal component, and negative numbers represent fault parallel
component. For example, in the case that the “Components” menu is chosen to be “Any
Component, Either FN or FP”, the NGA sequence input of “-179, 719” specifies two singlecomponent records: FP component from NGA #179 record, and FN component from NGA
#719 record. Please note that, as explained in the DGML report, some records have been
removed from the DGML database so those records are not searchable.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-25
NGA sequence numbers can be input using the following formats: (1) number array separated
by comma, eg. “1, 2, 3, 4, 5”. (2) number range delimited by colon, eg. “1:300” specifies NGA
number range from #1 to #300. (3) combination of format (1) and (2), eg. “1:100, 200, 300”
specifies NGA numbers from #1 to #100, and #200, #300. DGML will automatically eliminate
any duplicate numbers in the NGA sequence number input.
Event Name and Station Name require string input. The input string should be contained in the
list of event names and station names as specified in NGA flatfile. For example, if user wants
to search all “Imperial Valley” records, he can input “Imperial Valley” (input the string inside
quotation marks), or “imperial valley”, or “Imperial”, or simply “Imp”. The DGML uses exact
character matching to search the records, but it won’t distinguish upper- or lower-case letters.
The input boxes can be left as blank, which imposes no restriction in that field. If more than
one input field is filled in, the search results are the logical “AND” of these multiple conditions.
Figure 27 illustrates an example using the supplementary search engine to search and scale all
“Imperial Valley” event from NGA #1-#3000, using geometric mean spectra of FN FP
components. The search results in 49 records, and the number is automatically displayed in
“Total Num Output” box. In this case, “Total Num Output” is not controlled by the user. Please
note that DGML limits the maximum number of output to be less than 200 to avoid abuse of
the program. “Total Num. Averaged” is set by the user.
Switch the search engine
Text boxes to specify the
range of NGA sequence
numbers, Event Name and
Station Name
Text box to
display the
number of
output
Figure 27: Supplementary Search Engine User Interface.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-26
4.4 SPECIFY SCALING METHOD AND WEIGHT FUNCTION
The degree of similarity between the target and recording spectra is measured by the mean
squared error between the two spectra. The user has several options for scaling the recordings
before computation of the mean squared error. These are described in the following section.
4.4.1 Scaling the Records
Record scaling in the DGML is accomplished by applying a linear scale factor that does not
alter the relative frequency content of the acceleration time history, and thus does not change
the shape of the response spectrum of the time history. Two options are provided for scaling the
records to match the target response spectrum. The user also has the option to use unscaled
records.
(1) Scale the record to match the target spectrum over a period range
In this approach the record (or pair of records) is scaled by a factor that minimizes the mean
squared error between the spectrum (spectra) of the scaled record(s) and the target spectrum.
Calculation of the mean squared error is described in the main text of the DGML report.
Specify the max and min of scale
factors. Leave the box blank if scale
factor is not restricted
Click to YES
Click to NO
Click the Scaling button to “YES”, the Single Period button to “NO”, and specify the limits on
scale factors, if desired. In this scheme, the scale factor is computed to minimize the weighted
squared residuals between the scaled record and the target spectrum (see Section 2.3.2 of the
report). Specification of the weight function is described in Section 4.4.2 of this manual. If the
range (minimum, maximum) of scale factor is specified by the user, and the computed scale
factor is greater (or less) than the specified maximum (or minimum), then the scale factor takes
the maximum (minimum) value. If the factor limit is left blank, no restriction is imposed on the
scale factor. The above example limits the minimum scale factor to 0.5, and maximum scale
factor to 2.0.
(2) Scale the records to match the target spectrum to a single period
In this approach, the records are scaled to match the target spectrum at a specific spectral
period, called Ts (sec). In this scheme, the scale factor (f ) is computed such that the record
spectrum matches the target spectrum at the single period Ts, i.e.,
f =
SA t arg et (Ts )
SA record (Ts )
The mean squared error is computed for the scaled record as described in Section 2.3.2 of the
report.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-27
To scale the records to match the target spectrum to a single period of Ts, click Scaling button
to “YES”, Single Period button to “YES”, and specify the Ts value as follows,
Specify the period to which all the
records are scaled.
Click to YES
Click to YES
The above example shows all records are to be scaled to match the target spectrum at period 1.0
sec.
(3) No scaling
The third option is to consider only unscaled records. The mean squared error between the
spectrum (spectra) of the recordings and the target is computed as described in Section 2.3.2 of
the main report with the scale factor set to 1.0.
Click to NO, and all other fields
become invisible
Click Scaling button to turn off scaling method. Original records are used in this case.
4.4.2 Weight Function (Period Array and Weight Array)
A weight function is used to compute the scale factor for scaling option 1 in Section 4.4.1 and
in the computation of the mean squared error for all scaling options. We emphasize that the
user needs to specify the weight function even if the records are to be scaled to match the target
spectrum at a single period, or if there is no scaling at all. In the last two cases, although the
weight function is not involved in computing the scale factors, it is used to compute the mean
squared error to order the results with respect to degree of similarity of target spectrum and
spectra of recordings.
The user needs to specify Period Array and Weight Array to construct the weight function. The
Period Array or the Weight Array each has at least two data points. The Period Array is a
period sequence (T1, T2, … Tn) in ascending order between [0.01,10], with each element
delimitated by space or comma. The Weight Array data set is a sequence of non-negative
numbers that have a one-to-one correspondence to the Period Array data set, and thus shall
have the same number of elements (W1, W2, … Wn). Weights beyond the range of the Period
Array are automatically set to zero.
In DGML, the weight function specified by Period Array and Weight Array is discretized by
linearly interpolating over evenly-spaced discrete period points in log scale. To maintain
sufficient accuracy, one log-cycle is discretized by 100 points. Therefore, there are 301 discrete
period points (end-points included) in total to range from 0.01-10 sec. The weight function only
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-28
represents relative weights assigned to various discrete periods and are normalized in the
program such that the summation of the weight function over discrete period points equals
unity. Therefore, the absolute value of the weight function is immaterial.
The following three examples illustrate input of the weight function and the visualization
produced by the DGML software.
Example 1
Equal weight between period 0.03
sec. to 5.0 sec.
The weight function is visualized in
Area 4 of Figure 24 when the Search
button in Figure 26 is pressed.
Example 2
More weight is put on the short
period range (0.1-1 sec) than long
period range (1-3 sec)
Example 3
Discrete weight on the short period
range (0.03-0.3 sec) and long period
range (1-5 sec). Because the Period
Array needs to be a monotonically
increasing sequence, step function
should be approximated by a very steep
ramp, as illustrated.
Figure 28: Examples of Specifying Weight Function
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-29
4.5 PERFORM THE SEARCH
4.5.1 Search for Records and Calculate Average Spectrum
Once the search restriction, scaling method and weight function are specified, press “Search”
button to perform the search.
Press button to perform the
search
Text box to specify the
number of output
A progress bar will appear and shows the size of the data bin that satisfies all the specified
search restrictions, and specified number of records for output.
The progress bar shows there are total of 33 available records that satisfy the acceptance criteria
specified in the Search Engine Panel. The best 30 (specified in Total Num. Output box) records
will be displayed for inspection, and the best 7 records (specified in Total Num. Averaged box)
are used to compute the average spectrum. If the total number of records that satisfy the
acceptance criteria is less than the value of Total Num Output specified by the user, the value of
Total Num Output is automatically reset to the maximum number of acceptable recordings in
the database. DGML restricts the display of outputs up to 200 records to avoid abuse of the
program.
The spectra for all 30 output recordings will be displayed in the spectrum plot window together
with the average spectrum. Both the geometric mean and algebraic mean of the selected
number of spectra are plotted against the target spectrum, see Figure 29 for an example.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-30
Lines denoting the period
range of the weight function
Distribution of total
num output (30)
records
Averaged spectrum of
the selected (7)
records
Figure 29: Example of Average Spectrum of Selected Records
4.5.2 List the Search Result
DGML lists the search results in the output list window for the total number of output records
specified by users. The record list is ranked in order of increasing MSE.
Individual Record Information
Scroll bar horizontally to
see all output information
field
Scroll down to see all
the output record
Figure 30: Output List Window
Since user has specified Total Num Averaged (=7), the first 7 records (NGA #s 184, 165, 1116,
549, 719, 725, 159) in the list are selected to compute the average spectrum, with an asterisk
(*) shown in front of each selected record. The user can manually add or reject an individual
record from the list of output records. See sec. 4.6.4.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-31
Records
with * in
front are
selected
records to
compute
average
spectrum
Figure 31: Close-up of Output List Window
The listed information for each record is tabulated in Table 3.
Table 3. Listed Record Information of DGML Search Result
*
The asterisk (*) is used to designate the record is selected to compute the
averaged design spectrum.
Comp.
Component indicator:
FN: Record of fault normal (FN) direction
FP: Record of fault parallel (FP) direction
GM: Record of both FN and FP directions; use geometric mean of two
components.
NGA#
NGA number. A unique number assigned to each NGA record for
identification purposes.
MSE
Computed Mean Squared Error (MSE) of the selected record with respect to
the target spectrum.
ScaleF
Scale factor of the record computed by DGML
Pulse
Pulse Indicator: binary code to indicate if the unscaled record has velocity
pulse.
0 for non-pulse-like record
1 for pulse-like record
If the search is for two components in pair, the pulse indicator is shown for
both components. The first binary is for fault normal record, and the second
binary is for fault parallel record.
Tp (sec)
The period of the velocity pulse. No number is assigned for a non-pulse
record. If the search is for two components in pair, pulse periods (if any) for
both FN and FP components are given in order.
D5-95 (sec)
Significant duration, the time needed to build up between 5 and 95 percent of
the total Arias intensity. If the search is for two components in pair, durations
for both FN and FP components are given.
Event
Name of the earthquake event
Year
Year of earthquake
Station
The unique name of strong-motion station
Mag.
Moment magnitude of earthquake
Mechanism
Type of Fault Mechanism. Available mechanisms are: Strike-Slip, Normal,
Normal-Oblique, Reverse, Reverse-Oblique.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-32
Rjb (km)
Joyner-Boore distance to rupture plane
Rrup (km)
Closest distance to rupture plane
VS30 (m/s)
Average shear velocity of top 30 m
Low. Freq. (Hz)
The recommended lowest usable frequency for the record. (see Footnote (1)
below)
Horiz. Acc. File Name
or
FN Acc. File Name
FP Acc. File Name
The horizontal acceleration time history record file name in \FNFPDataset\
directory. If the search is for FN/FP components in pair, the file names for
both FN FP components are given (labeled as “FN Acc. File Name” and “FP
Acc. File Name”, respectively).
Vertical Acc. File
Name
The vertical acceleration time history record file name in \NGADatabase\
directory. If the vertical file is missing, the file name is displayed as
“nonexistent”.
4.6 SELECTION AND EVALUATION OF RECORDS
4.6.1 Highlight an Individual Record
The response spectrum and acceleration, velocity, and displacement time histories for scaled
individual records can be easily visualized in DGML, providing users the capacity to inspect
the records in greater detail. The user can simply click on the record line in the list window, and
the selected record line will be highlighted. The following example shows a particular record
(NGA#165) selected. Record #165 is ranked 2nd in terms of closeness of the spectrum shape to
the target spectrum, where geometric mean of the FN and FP components (Comp.=GM) are
used to calculate the scale factor and MSE. The scale factor (ScaleF) is computed to be 0.7674.
The pulse indicator shows that both components of this record is not pulse-like (Pulse=0 for
both FN FP component).
Click to highlight an individual record
Figure 32: Highlight an Individual Record on the Output List
Footnote (1): The recommended lowest usable frequency is related to filtering of a record by the record processing
organization to remove low-frequency (long-period) noise. Filtering results in suppression of ground motion
amplitudes and energy at frequencies lower than the lowest usable frequency such that the motion is not
representative of the real ground motion at those frequencies. It is a user’s choice on whether to select or reject a
record on the basis of the lowest usable frequency. Because of the suppression of ground motion at frequencies
lower than the lowest usable frequency, it is recommended that selected records have lowest usable frequencies
equal to or lower than the lowest frequency of interest.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-33
4.6.2 Highlight Response Spectrum of an Individual Record
By specifying suitable options in the “Control” panel, the scaled response spectrum of an
individual record can be highlighted in the Spectrum Plot Window. Figure 33 shows the graphic
control to highlight the scaled fault-normal (in blue color), fault-parallel (in red color) and
vertical (in green color) response spectra for record NGA#165. Please note that the vertical
component is scaled using the same scale factor that was obtained for the horizontal
components.
(a) Graphic control to highlight fault-normal and fault-parallel spectrum of selected record
(b) Graphic control to superimpose vertical response spectrum
Figure 33: Highlight the Response Spectrum of an Individual Record
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-34
4.6.3 Highlight Time History of an Individual Record
Once an individual record is highlighted, the scaled acceleration/velocity/displacement time
history is automatically plotted in Time History Plot Window. One can select the drop menu to
plot acceleration or velocity or displacement time history. Check boxes in the Control panel
give users the option to display fault-normal (in blue color), fault-parallel (in red color) and
vertical (in green color) components. Figure 34 shows the scaled time histories for a
highlighted record (NGA#165 in this case).
Click drop menu
to select to plot
acceleration or
velocity or
displacement
time history
Check box
to select
which
component
to display
Figure 34: Plot of the Scaled Acceleration/Velocity/Time History of an Individual Record in
Fault Normal, Fault Parallel and Vertical Directions
4.6.4 Zoom In Time Function for Examining the Time History of an Individual Record
The “Zoom In Time” function is provided to help the user inspect the details of the time history
plot. The function can be activated by selecting the option “Zoom In Time”. As shown in
Figure 35, an input box appears to prompt input in the form of “min, max”. The example shows
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-35
that the acceleration time history is zoomed in between 5-7 seconds. To zoom out, one can
input blank in the input box, or click the “Zoom In Time” to deactivate the function.
(a) Click on radio button “Zoom In Time” for the input box to appear.
(b) It is prompted to input data pair in the format of [min, max]
(c) Key in the time range to zoom in. The acceleration time history zooms in
between 5-7 seconds in this example.
Figure 35. Zoom In Time History Plot
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-36
4.6.5
Accept or Reject an Individual Record
The user has the option to remove or add an individual record from the selected list by using
“Accept” and “Reject” button. First, mouse-clicking the record in the list window highlights the
individual record. If the asterisk (*) shows the record is currently in the selected list, then
clicking Reject button removes the record from the list, and the asterisk (*) is also removed in
the front to signify the record was rejected. Similarly, clicking the Accept button will add the
highlighted record into the selected list, and the asterisk (*) symbol appears in the front after
the operation. The average spectrum of the selected list is automatically updated after each
“Accept” or “Reject”, however, one can always check the updated spectrum using the
“Refresh” button. The steps of operation are illustrated in Figure 36.
Click to highlight the
individual record
Star is removed from the
front to show the record is
rejected
Click Reject to remove the
record from selected list
Star appears in the front
to show the record is
added
Highlight record and click Accept to add the
record from selected list
Click Refresh to see the
average spectrum of the
updated selected list.
Figure 36: Steps to Accept or Reject an Individual Record
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-37
4.7 GRAPHIC CONTROL
4.7.1 Graphic Control Panel
The graphic control panel provides the functionality to manipulate the spectrum plot and the
time history plot. Basic functions of the control panel are summarized in Figure 37.
Change axes of spectrum plot, see sec 4.5.2.
Change line style of averaged and target spectrum
Change line color of total output in spectrum plot
Check to show grid on the spectrum and time history plot
Check box to select which component to display in the
spectrum plot and time history plot. See Sec. 4.6.3. The
“FaultNormal” and “FaultParallel” check boxes are
available only if searching for FN and FP record in pair.
“Vertical” component box is unchecked by default.
Figure 37: Graphic Control Panel
4.7.2 Change Plot Axes
The axes of the spectrum plot can be changed from log scale to linear scale by selecting options
in the drop menu. The following plots in Figure 38 illustrate four options available to change
the axes of a same plot.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-38
(a) log-log axes
(b) semilogx axes
(c) linear axes
(d) semilogy axes
Figure 38: Example of Four Options to Change Plot Axes
4.8 SAVE DGML SEARCH RESULT
4.8.1 DGML Search Report
The DGML search criteria and summary of the selected records can be exported by clicking
“Save Search Result” button (cf. Figure 39). If “Save Vertical Records” button is selected, the
corresponding vertical component information will be exported together with the horizontal
components. A pop-up window allows the user to specify the file name and file type (“csv’ or
‘txt’ format). “cvs” (Comma Separated Variables) file format is preferred since it can be
opened using Microsoft Excel, or any text editor.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-39
Figure 39: Save DGML Search Result
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-40
Figure 40: An Example of DGML Search Report
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-41
Figure 40 illustrates an example of the DGML search report (named as
SaveSearchResultExample.csv in this case) opened by Microsoft Excel. The search report
features the following data blocks:
(1) Summary of DGML search criteria. All user-defined search criteria are listed in this data
field, including the magnitude range, fault type, specified D9-95 range, specified Rrup and Rjb
ranges, specified VS30 range, component specification, pulse characteristics, scale factor limit,
period array and weight array.
(2) Summary of Properties of Selected Horizontal Records. The same information in the
DGML output list window is reported only for selected records (these marked with stars). The
reported properties of each selected record are: Component Indicator, NGA number, Mean
Squared Error (MSE), Scale Factor, Pulse Indicator, Pulse Periods, D5-95 duration, Event
Name, Year, Station Name, Magnitude, Fault Mechanism, RJB, RRup, VS30, recommended
lowest usable frequency, and the acceleration record file names. Please note that if RJB or RRup
is in a squared bracket, it indicates that the value is absent in NGA Flatfile. The reported value
was estimated by Chiou and Youngs (2008b).
(3) Scaled Spectral Acceleration of Selected Horizontal Records. For each selected record
listed in data field (2), their component indicator, NGA number, scale factor, scaled PGA,
PGV, PGD values and scaled spectra acceleration values are reported in this data field. The
scaled spectra acceleration values are reported for the following periods (in seconds):
0.01
0.042
0.07
0.133
0.25
0.4
0.667
1.3
2.5
4.2
8
0.02
0.044
0.075
0.14
0.26
0.42
0.7
1.4
2.6
4.4
8.5
0.022
0.045
0.08
0.15
0.28
0.44
0.75
1.5
2.8
4.6
9
0.025
0.046
0.085
0.16
0.29
0.45
0.8
1.6
3
4.8
9.5
0.029
0.048
0.09
0.17
0.3
0.46
0.85
1.7
3.2
5
10
0.03
0.05
0.095
0.18
0.32
0.48
0.9
1.8
3.4
5.5
0.032
0.055
0.1
0.19
0.34
0.5
0.95
1.9
3.5
6
0.035
0.06
0.11
0.2
0.35
0.55
1
2.0
3.6
6.5
0.036
0.065
0.12
0.22
0.36
0.6
1.1
2.2
3.8
7
0.04
0.067
0.13
0.24
0.38
0.65
1.2
2.4
4
7.5
If the search component is “FN/FP in Pair”, the spectra values of the Geometric Mean (GM)
spectra and Fault Normal (FN), Fault Parallel (FP) are all listed separately. If the search
component is “FN” or “FP”, only the spectra values of FN or FP component are listed.
(4) Scaled Spectral Acceleration of Selected Vertical Records. The data block is available only
if “Save Vertical Records” button is selected. The data block lists the same information for
vertical component as is in data field (3).
(5) Scaled Average Spectral Acceleration of Selected Horizontal Records. The data field
reports the target spectrum (horizontal) values, and both geometric mean spectrum and
arithmetic mean spectrum of selected horizontal records (see definitions of geometric mean and
arithmetic mean of spectrum, Section 3.2.1 of Users Manual). Please note that the reported
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-42
target spectrum is re-interpolated to the same period sequence as used in data field (3),
therefore, the values may be slightly different from user’s original input.
4.8.2
Save Acceleration Time History Files
DGML (version 2) allows users to save the selected acceleration time history file. Please note
that the saved time history file contains unscaled original data as in the PEER NGA
database. The steps are illustrated in Figure 41.
Step (1): user clicks to highlight the record to be saved. By default, only horizontal time history
files will be saved. One must check “Save Vertical Records” button if the corresponding
vertical time history file is to be exported.
Step (2): click “Save Unscaled .acc” button. A pop-up window allows the user to specify the
save directory and file name. The pop-up window shows a default file name assigned by
DGML using the convention of “FilePath.Filename” as was used in the PEER NGA database to
distinguish records that might have the same file name but different file paths in the database. If
the search component is set as “FN/FP in Pair”, the pop-up window will show up sequentially
for each of the FN and FP acceleration time history file, and for the vertical file in the last if
“Save Vertical Records” button is checked.
Figure 41: Save Acceleration Time History Files
4.8.3 Save the Plot
Both spectrum and time history plots can be saved to the disk as figure files. The steps are
illustrated in Figure 42. Step (1): Right click mouse on the plot area, and a “Save Plot As” box
will show up. Step (2): Left click the box to pop up a window, select the proper directory, and
enter the name of the graphic file to save. The plot can be save as .bmp, .jpg, .tiff, or .eps
format.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-43
Right click mouse in the plot field
Select the directory to save the plot
Enter the name of the file to save
Figure 42: Steps to Save the Plot
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-44
5.
DGML EXAMPLES
Examples are provided below to illustrate the major steps in selecting suitable ground motion
records using the DGML. The examples are served mainly for the purpose of demonstration
and not for the purpose of any practical application.
5.1 RECORD SELECTION AND MODIFICATION
Step 1. Developing the Target Spectrum
The site selected for the example is Site Class D at a location in San Francisco approximately
10 km from the San Andreas fault. The structure was assumed to be a medium- to high-rise
building having a significant period range of 0.3 to 3 seconds. The design response spectrum is
a Code spectrum, and the following steps are needed to construct the spectrum.
(1) Select Spectrum Model: Choose “ASCE code specification” from the drop menu.
Figure 43: Select Spectrum Model
(2) Specify Spectrum Parameters: By referring to the design ground motion maps and
provisions of ASCE standard ASCE/SEI 7-05, parameters Sds, Sd1, and TL needed to
construct the code design spectrum are obtained as follows, where equations, figures, maps,
and tables refer to ASCE (2006):
Maximum Considered Earthquake parameters:
Ss=1.5g (Map, p. 214), Fa= 1.0 (Table 11.4-1), Sms=1.5g (Eq.11.4-1)
S1=0.75g (Map, p.216), Fv=1.5 (Table 11.4-2), Sm1=1.125g (Eq. 11.4-2)
Design earthquake parameters:
Sds=2/3*Sms=1.0g (Eq. 11.4-3)
Sd1=2/3*Sm1=0.75g (Eq. 11.4-4)
TL=12 sec
(Fig. 22-16)
Enter Parameters for the Code
Spectrum
Figure 44: Specify Spectrum Parameters
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-45
(3) Create the Target Design Spectrum: Press “Create” button in the Control Panel to create
the target spectrum, which is shown in the spectrum plot window as follows:
Figure 45: The Target Design Spectrum
Press NEXT button to proceed to the next step.
Step 2. Defining the Search Criteria
The USGS web site, http://earthquake.usgs.gov/research/hazmaps, was used to deaggregate the
site ground motion hazard. At 1-second period, the dominant contributor to the hazard was an
earthquake of approximately magnitude 7.8 occurring on the San Andreas fault. For purposes
of the example, the search was restricted to recordings from strike-slip (SS) earthquakes of
magnitude 7 and higher occurring within 30 km of the site. A wide range of VS30 limits and no
restrictions on significant duration (D5-95) or on scale factor were utilized in order to capture a
sufficiently large set of potential records for examination and scaling and matching of recorded
spectra to the Code spectrum within the significant period range. It is noted that although
specifying the site VS30 and site class is essential in developing the design response spectrum, it
is not so important in selecting records because the search will rank the records with respect to
their match to the design spectrum. It was assumed that matching would be to the Code
spectrum rather than to a conditional mean spectrum. Equal weight was given to matching at all
periods within the significant period range (0.3-3 sec). It was also assumed that for the
structural analyses, only individual horizontal components were required rather than FN/FP
pairs. No restrictions were placed for searching on whether the record should or should not
have pulses. The display of search criteria is shown in Figure 46.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-46
Search within records of magnitude
7-9, strike- slip type of faulting, any
significant duration, distance (R_JB,
R_rup) from 0-30km, and Vs30 from
200-1000 m/s
Search any individual component
and no restriction on pulse
Scaling the record with no
restriction on scaling factors
Give equal weight to all periods
within 0.3-3.0 sec
Output 30 records, and compute the
average spectrum of 7 records best
matching the design spectrum
Figure 46: Specify the Search Criteria
Step 3. Searching for Records
Press “Search” button in the Search Engine. The progress bar shows there are 46 records that
satisfy the search criteria.
Step 4. Inspect, Evaluate, and Finalize the Search Result
As shown in Figure 46, the search was set up to output 30 records and computes the average
spectrum for the 7 records that best match the Code spectrum. Scroll down the list bar in the list
window to see information on all 30 output records, as shown in Figure 47. Clicking on an
individual record allows one to examine the spectrum and acceleration/ velocity/ displacement
time history of that record.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-47
The best-matching 7 records are called “Selected Records”, with a star in the front for
identification. Figure 47 shows the top 7 records (with * symbols) that are selected by DGML.
However, the user has the capability to reject records from or add records to the list of selected
records. The next step is to adjust the list of selected records and finalize the search result.
After inspecting the records, it was decided to reject records #FP1148, #FP 900, # FN 1165,
and add records #FN882, #FN 1176, #FN 850 into the list of selected records. Figure 48 shows
the operation to include or exclude an individual record. Note that the selected records
always have * symbol in the front. The finalized 7 selected records are summarized in
Table 4.
Figure 47: DGML Search Result
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-48
Click to highlight record, and
click Reject button to exclude it
from selected list.
Press Refresh to check averaged
spectrum for the updated selected
records
Click to highlight record and
click Accept button to include it
in the selected list
Press Refresh to check average
spectrum for the updated
selected records
Figure 48: Modify the List of Selected Records and Re-average
Table 4. Selected Ground Motion Records
Rrup
(km)
Vs30 Low.Freq
(m/s)
(Hz)
0
6.6
276
0.10
Strike-Slip
13.6
15.4
276
0.24
7.3
Strike-Slip
26.8
26.8
345
0.28
1992
Mission
7.3
Creek Fault
Strike-Slip
26.8
26.8
345
0.11
Kocaeli,
Turkey
1999
Yarimca
7.5
Strike-Slip
1.4
4.8
297
0.09
37
Landers
1992
North Palm
Springs
7.3
Strike-Slip
26.8
26.8
345
0.28
32
Landers
1992
Desert Hot
Springs
7.3
Strike-Slip
21.8
21.8
345
0.07
Ma
g
Mechanism
Duzce
7.1
Strike-Slip
1999
Duzce
7.5
Landers
1992
North Palm
Springs
40.2
Landers
4.5
15.4
0
---
0
---
Tp
(sec)
D5-95
(sec)
1
5.6
1.638
0
882
4.182
FN
880
FN
NGA#
Scale
Factor
FP
1605
0.966
FN
1158
FP
Event
Year
10.7
Duzce,
Turkey
1999
---
11.7
Kocaeli,
Turkey
0
---
36.3
4.955
0
---
1176
1.445
1
FN
882
3.597
FN
850
3.147
Comp.
Pulse
Station
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
Rjb
(km)
A-49
For the seven selected records, significant duration (5 to 95% of Arias Intensity) ranged from
11 to 40 seconds with an average of about 26 seconds. The average is very close to the median
value of about 27 seconds for a magnitude 7.8 earthquake at 10km distance based on the
correlation of Kempton and Stewart (2006) in Table 1 of the main report. Two pulse records
are included in the data set, with estimated pulse periods in the range of 4.5 to 5.6 seconds.
Although pulse periods of these amplitudes are reasonable for a large-magnitude earthquake
(see Figure 9 of the report), it could be desirable to have a record set with a wider range of
pulse periods, or a larger number of pulse records, depending on the structural characteristics
and response. The average spectrum for the seven selected records is shown in Figure 49, and
the background shows the spectra of all 30 listed records. The averaged spectrum has a close
match to the target spectrum in the significant period range of 0.3 to 3 sec. The acceleration,
velocity, and displacement time histories of the selected records are shown in Figure 50.
1
10
Spectral Accleration, Sa (g)
0
10
-1
10
-2
10
Target Spectrum
Record Geom.Mean
Record Algb.Mean
-3
10
-2
10
-1
0
10
10
1
10
Period, T (sec)
Figure 49: Average Spectrum of 7 Selected Records
Record #FP 1605
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-50
Record # FN 1158
Record # FP 882
Record # FN 880
Record # FN 1176
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-51
Record # FN 882
Record # FN 850
Figure 50: The Acceleration, Velocity and Displacement Time Histories of Selected Records
5.2 USE DGML SUPPLEMENTARY SEARCH ENGINE
The previous example illustrates the steps to use DGML’s default search engine to select and
modify ground motion records. In DGML Version 2, a supplementary search engine is
provided to enhance the search capability, which is documented in Section 4.3 of this manual.
The supplementary search engine is useful to inspect the properties of a set of records
according to specified sequence of NGA numbers, and/or event name, station name.
Step 1. Switch to Supplementary Search Engine
Press “Switch” button to activate the Supplementary Search Engine. The user interface is
shown in Figure 51.
Step 2. Search According to NGA Numbers
The final result from Example 5.1 can be easily reproduced using the Supplementary Search
Engine. First, specify the NGA sequence numbers with the rule set in Section 4.3. The
“Components”, “Pulse”, and other options also need to be set properly as shown in the
Figure 51.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-52
After pressing the “Search” button, only the specified seven records are displayed. The
spectrum plots are shown in Figure 52, where the target spectrum, the average spectrum and
each of the selected records are illustrated in colors. Please also notice the difference between
Figure 52 with Figure 49, where the background in Figure 49 shows the spectra of all 30 listed
records.
Step 3. Search According to Event Name
Assuming that the user particularly wants to include records from Northridge earthquake. one
can use the Supplementary Search Engine to search records according to the “Event Name”, as
shown in Figure 53. All records with “Event Name” containing the key word “Northridge” will
be searched and displayed.
Step 4. Modify Search to Include a Particular Record
The Supplementary Search Engine is also useful to add particular records into a group. Assume
that after inspecting all “Northridge” records in Step 3, the user wants to include FP component
of NGA #1057 record to replace the last record (FN NGA#850) used in Step 2. The task can be
easily achieved by specifying an updated set of seven records by the “NGA Number Sequence”
as shown in Figure 54, where 850 was replaced by -1057. The spectra for the updated set are
shown in Figure 55.
Figure 51: Search by NGA Sequence Using The Supplementary Search Engine
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-53
Figure 52: Plot Spectra of Selected 7 Records
Figure 53: Search by Event Name Using the Supplementary Search Engine
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-54
Figure 54: Update to Include the Northridge Record
Figure 55: Plot Spectra of Updated 7 Records
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-55
6.
DGML MATLAB SOURCE CODE
DGML was developed using Matlab ® version 7.2 Graphic User Interface (GUI). The DGML
Matlab source codes were compiled into standalone executable so that no Matlab environment
is required for common users. The DGML Matlab source codes are also provided in the DVD
release. Advanced users can execute or modify the source codes in Matlab environment. The
section provides brief descriptions of the DGML source codes for advanced programmers, and
it can be skipped by common users. Relevant references on Matlab GUI programming should
be consulted if necessary.
6.1 INTRODUCTION
The Matlab source codes are stored in “DVD:\DGML\” directory, with Matlab utility files
stored in “DVD:\DGML\Utility\” directory. The utility files are Matlab .mat files that store preprocessed NGA database information, such as spectrum values, earthquake magnitudes and
distances, etc. The utility files are called by DGML in real time, serving as input data for
various subroutines. DGML can be updated by simply updating the utility files without recompiling the whole system.
There are two types of source files under “DVD:\DGML\” directory. The *.fig files are Matlab
figure files used to design the GUI layout, and *.m files are Matlab source files that provide the
callback functions of the GUI objects and perform other operations. Brief descriptions of the
source codes are given in Table 5.
Table 5. Descriptions of DGML Matlab Source Codes
Matlab Source Codes
Description
DGMLStart.fig
GUI layout of DGML Start page
DGMLStart.m
Main program of DGML Start GUI
NGATargetSpectrum.fig
GUI layout of Target Spectrum page
NGATargetSpectrum.m
Main program of Target Spectrum GUI
AS2008.m
Subroutine to compute Abrahamson and Silva
(2008) NGA spectrum
BA2008.m
Subroutine to compute Boore-Atkinson(2008) NGA
spectrum
CB2008.m
Subroutine to compute Campbell- Bozorgnia (2008)
NGA spectrum
CY2008.m
Subroutine to compute Chiou-Youngs (2008) NGA
spectrum
Idriss2008.m
Subroutine to compute Idriss (2008) NGA spectrum
ComputeEpsilon.m
Subroutine to compute correlation coefficient based
on Baker & Jayaram (2008)
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-56
DGMLibPostProcessor.fig
GUI layout of DGML Search Engine page
DGMLibPostProcessor.m
Main program of DGML Search Engine
DGMLibSearchEngine.m
Core code segment of DGML search engine.
ComputeAccVelDispTimeHistory.m
Subroutine to read in acceleration time history data
from a file in NGA database, and compute the
velocity and displacement time history.
fixPSlinestyle.m
Subroutine to fix the dashed line style in post script
figure output
6.2 RUN DGML IN MATLAB
Beside running the pre-compiled program on DVD, DGML can be executed directly using
Matlab ® version 7.2 or higher. By setting “DVD:\DGML\” as the working directory, DGML
can be started by calling DGMLStart.m function as follows,
>> DGMLStart
6.3 VIEW GUI OBJECT CALLBACK FUNCTIONS
Matlab command “guide” can be used to open each GUI *.fig file. Type the following
command in Matlab:
>> guide
The command will show existing GUI files that can be opened, as in Figure 56. For example, if
one selects to click open “NGATargetSpectrum.fig”, the GUI layout of “Target Spectrum”
page will be shown in Figure 57. The GUI Layout of “Search Engine” page is shown in Figure
58. Developer can resize each GUI object (buttons, menus, etc.) by clicking and dragging. GUI
programming is “event-driven”, i.e., a user-defined call-back function will be invoked when a
button is clicked, or a drop menu is selected and so on. To view the call-back function of a GUI
object, one can right-click a GUI object, and the related object call-back functions can be
selected from the pop-up menu. Figure 59 illustrates the procedure to view call back function of
the “Create” button on NGATargetSpectrum.fig. Figure 60 shows the source code of the
callback function when “Create” button is pressed.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-57
Figure 56: Use Matlab “Guide” Command to Open DGML GUI
Figure 57: GUI Layout of “Target Spectrum” Page
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-58
Figure 58: GUI Layout of “Search Engine” Page
Figure 59: View Callback Functions of a GUI Object
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-59
Figure 60: Source Code of a Callback Function
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
A-60
REFERENCES
Abrahamson, N.A., and Silva, W.J., 2008, Summary of the Abrahamson & Silva NGA groundmotion relations, Earthquake Spectra, Vol. 24, No. 1, pp. 67 – 97.
Baker, J.W., and Cornell, C.A., 2006, Spectral shape, epsilon and record selection, Earthquake
Engineering & Structural Dynamics, Vol. 35, No. 9, pp. 1077 – 1095.
Baker, J. W. and Jayaram, N., 2008, Correlation of spectral acceleration values from NGA
ground motion models. Earthquake Spectra, Vol. 24, No. 1, pp. 299–317.
Boore, D.M., and Atkinson, G.M., 2008, Ground-motion prediction equations for the average
horizontal component of PGA, PGV, and 5% damped PSA at spectral periods between
0.01s and 10.0s, Earthquake Spectra, Vol. 24, No. 1, pp. 99 – 138.
Campbell, K.W., and Bozorgnia, Y., 2008, NGA ground motion model for the geometric mean
horizontal component of PGA, PGV, PGD and 5% damped linear elastic response
spectra for periods ranging from 0.01 to 10 s, Earthquake Spectra, Vol. 24, No. 1, pp.
139 – 171.
Chiou, B.S.J., and Youngs, R.R., 2008a, Chiou-Youngs NGA ground motion relations for the
geometric mean horizontal component of peak and spectral ground motion parameters,
Earthquake Spectra, Vol. 24, No. 1, pp. 173 – 215.
Chiou, B.S.J., and Youngs, R.R., 2008b, NGA model for average horizontal component of peak
ground motion and response spectra, Report PEER 2008/09, Pacific Engineering
Research Center, University of California, Berkeley.
Idriss, I. M., 2008, An NGA empirical model for estimating the horizontal spectral values
generated by shallow crustal earthquakes, Earthquake Spectra, Vol. 24, No. 1, pp. 217 –
242.
\\Oad-fs1\doc_safe\10000s\10607.000\3000 REPORT\DGML Report\4 Apx A\DGML Users Manual.doc
AMEC Geomatrix, Inc.
61
APPENDIX B
Summary of PEER-NGA Records Included in and Excluded from DGML Database
(tables in electronic form on DVD)
Table B-1
Table B-1a
Table B-1b
Table B-1c
Table B-2
List of PEER-NGA Records Included in DGML Database
List of PEER-NGA Records Included in DGML Database (Taiwan)
List of PEER-NGA Records Included in DGML Database (California)
List of PEER-NGA Records Included in DGML Database (Other Regions)
List of PEER-NGA Records Excluded from DGML Database
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