Graduation_Report_Harikumaran_4118855.

Graduation_Report_Harikumaran_4118855.
Master of Science Thesis
Sizing and Charge Control Strategies for a Grid-Connected Micro-grid
with Electric Vehicles
Jayakrishnan Harikumaran
4118855
Supervisors:
Dr.Ir.P.Bauer (TU Delft)
Ir. Arne Kaas (TU Delft)
June 2012
Sizing and Charge Control Strategies for a Grid-Connected Micro-grid
with Electric Vehicles
by
Jayakrishnan Harikumaran
Thesis submitted to the faculty of Electrical Engineering, Mathematics and Computer Science (EEMCS)
group of Electrical Power Processing (EPP) in partial fulfillment of the requirements for the degree of
Master of Science
in
Sustainable Energy Technology
June 25, 2012
Delft, The Netherlands
Thesis committee:
Prof. dr. J.A. Ferreira (TU Delft)
Prof. dr.ir. P. Bauer (TU Delft)
Dr. ir. M. Popov (TU Delft)
Faculty of Electrical Engineering, Mathematics and Computer Science (EEMCS)
Electrical Power Engineering
Electrical Power Processing (EPP)
Visiting address
Mekelweg 4
2628 CD Delft
Postal address
P.O. Box 5031
2600 GA Delft
The Netherlands
http://www.ewi.tudelft.nl
Author:
Jayakrishnan Harikumaran
Student number
E-mail
4118855
[email protected]
Acknowledgements
This master’s thesis work was completed in the department of Electrical Engineering at Delft University
of Technology. The thesis is supported by the DIEMIEGO2 Project.
I want to thank my supervisor Professor Pavol Bauer for the interesting subject and support throughout
the project. I wish to thank my daily supervisor Arne Kaas on whose work my model was built upon. His
guidance and overview was crucial in the successful completion of my work.
I wish to thank my friend Venugopal Prasanth for the many discussions I had with him for my work and
also for proof reading my thesis. I want to thank my colleague, Gyorgy Vereczki with whom a paper was
written jointly describing our works. I would also like to thank my colleagues in the Electrical Sustainable
energy group for the collaboration and good times during my thesis work. Finally I wish to thank my
mom, my brother and sister-in-law for the moral support and financial assistance during my studies.
Delft
Jayakrishnan Harikumaran
25.05.2012
i
TABLE OF CONTENTS
ACKNOWLEDGEMENTS ....................................................................................................................... I
1
INTRODUCTION............................................................................................................................... 1
1.1
OPPORTUNITIES FOR SYNERGY BETWEEN ELECTRIC VEHICLES AND RENEWABLE POWER SOURCES............2
1.1.1 Interaction between electric vehicles and existing renewable energy policies ........................................2
1.1.2 Electric Vehicles as energy buffer for renewable electricity ...................................................................2
1.1.3 Controlling EV charging patterns to the benefit of Utilities ....................................................................2
2 STATE OF THE ART OF BATTERY ELECTRIC VEHICLES AND CHARGING
INFRASTRUCTURE ................................................................................................................................. 4
2.1
ENERGY STORAGE FOR TRANSPORTATION....................................................................................................4
2.1.1 Cost of Li-Ion Batteries............................................................................................................................4
2.2
EV CHARGING INFRASTRUCTURE .................................................................................................................5
2.2.1 Conductive Charging ...............................................................................................................................5
2.2.2 Inductive Charging ..................................................................................................................................7
3
MOBILITY ANALYSIS................................................................................................................... 10
3.1
3.2
3.3
3.4
4
TRIP CLASSIFICATION ACCORDING TO TRIP DURATION .............................................................................. 11
TRIP CLASSIFICATION ACCORDING TO TRIP LENGTH .................................................................................. 13
TRIP CLASSIFICATION BY TRIP PURPOSE .................................................................................................... 15
EV PARAMETER ASSUMPTIONS ................................................................................................................... 18
EV FLEET MODELING METHODOLOGY ............................................................................... 20
4.1
REFERENCE TRIP DATA GENERATION ......................................................................................................... 20
4.2
TRIP SCHEDULING SUB-ROUTINE ............................................................................................................... 21
4.2.1 Trip Distance and Trip Speed Assignment ............................................................................................. 21
4.2.2 Location Tracking .................................................................................................................................. 22
4.3
EV CHARGING POWER REQUIREMENT ESTIMATION SUB-ROUTINE ............................................................ 22
4.3.1 Dumb Charging and Green Charging ................................................................................................... 23
4.4
SOC TRACKING AND UPDATE SUB-ROUTINE ............................................................................................. 26
5
LOW VOLTAGE GRID IMPACTS AND MOBILITY RESULTS ............................................ 28
5.1
TRIP DISTRIBUTION PROFILES OBTAINED FROM REFERENCE DATA GENERATION SUB-ROUTINE............... 28
5.2
VEHICLE LOCATION DISTRIBUTION ............................................................................................................. 30
5.3
EV CHARGING POWER PROFILE ON LOW VOLTAGE (LV) GRID.................................................................. 31
5.3.1 EV Simulation with Dumb Charging ..................................................................................................... 31
5.3.2 EV Simulation with Smart Charging...................................................................................................... 33
6
QUICK CHARGE REPLENISHMENT......................................................................................... 37
6.1
DC FAST CHARGING ................................................................................................................................... 37
6.1.1 Fast Charge Power Profile and Share of Energy from Fast Charge for Uncontrolled charging with a
range of 130 km and charging enabled only at home ......................................................................................... 39
6.1.2 Fast Charge Power Profile and Share of Energy from Fast Charge for Uncontrolled charging with a
range of 130 km and charging enabled at home and work ................................................................................. 40
6.1.3 Fast Charge Power Profile and Share of Energy from Fast Charge for Uncontrolled charging with a
range of 170 km and charging enabled at home only ......................................................................................... 41
6.1.4 Fast Charge Power Profile and Share of Energy from Fast Charge for Uncontrolled charging with a
range of 170 km and charging enabled at home and work ................................................................................. 42
ii
6.1.5 Variation of Energy Share from Fast Charging .................................................................................... 43
6.2
BATTERY SWAP .......................................................................................................................................... 43
6.2.1 Switching Process Algorithm ................................................................................................................. 45
6.2.2 Switching Station Power Management .................................................................................................. 46
6.2.3 Swap Station Power Profile and Share of Energy from Fast Charge for Uncontrolled charging with a
range of 130 km and charging enabled at home only ......................................................................................... 47
6.2.4 Swap Station Power Profile and Share of Energy from Fast Charge for Uncontrolled charging with a
range of 130 km and charging enabled at home and work ................................................................................. 48
6.2.5 Swap Station Power Profile and Share of Energy from Fast Charge for Uncontrolled charging with a
range of 170 km and charging enabled at home only ......................................................................................... 49
6.2.6 Swap Station Power Profile and Share of Energy from Fast Charge for Uncontrolled charging with a
range of 170 km and charging enabled at home and work ................................................................................. 50
6.2.7 Extra Batteries Required For Battery Switching ................................................................................... 51
6.3
COMPARISON OF THE QUICK CHARGE REPLENISHMENT STRATEGIES ......................................................... 53
6.3.1 Energy Contribution from Quick Charge Replenishment ...................................................................... 55
6.4
QUICK CHARGE REPLENISHMENT ANALYSIS WITH HIGHER RESOLUTION TRAVEL PROFILE ...................... 55
6.4.1 Trip Profile Resolution Increase in ‘50km or more’ Category .............................................................. 55
6.4.2 Comparison of Quick Charge Replenishment Requirements with new Trip Profile .............................. 57
6.5
QCR PEAK POWER AND SERVING UNITS COMPARISON ................................................................................ 60
7
MICRO-GRID WITH ELECTRIC VEHICLE CHARGING LOADS ....................................... 62
7.1
TOPOLOGY OF THE MICRO-GRID WITH EV LOADS ..................................................................................... 62
7.2
LOAD .......................................................................................................................................................... 64
7.3
INVERTERS .................................................................................................................................................. 64
7.4
SOLAR PANELS ........................................................................................................................................... 64
7.5
WIND TURBINES ......................................................................................................................................... 64
7.6
BATTERIES .................................................................................................................................................. 64
7.7
EV CHARGING LOAD .................................................................................................................................. 65
7.8
EXPLANATION OF RENEWABLE ENERGY TERMINOLOGIES USED ................................................................ 65
7.8.1 Potential Energy Mix (PEM) ................................................................................................................. 65
7.8.2 Renewable Energy Ratio (RER) ............................................................................................................. 65
8
SIMULATION OF MICRO-GRID SYSTEM AND SENSITIVITY ANALYSIS ...................... 66
8.1
8.2
8.3
8.4
8.5
9
NORMALIZATION OF COMBINED LOAD ....................................................................................................... 66
SIMULATION RESULT WITH 80 PERCENTAGE EV PENETRATION AND BATTERY SWITCHING ........................ 67
SIMULATION RESULT WITH 80 PERCENTAGE EV PENETRATION AND FAST CHARGING ................................ 70
ENERGY MANAGEMENT AT SWITCHING STATIONS ..................................................................................... 73
EFFECT OF FLEET RANGE ON MICRO-GRID ENERGY USAGE ....................................................................... 78
CONCLUSION AND DISCUSSION .............................................................................................. 81
9.1
SUGGESTED FUTURE WORK ........................................................................................................................ 82
10 BIBLIOGRAPHY ............................................................................................................................. 83
11 APPENDIX ........................................................................................................................................ 85
11.1
APPENDIX A EV LOAD MODELLING FLOW CHARTS ................................................................................... 85
11.1.1
Trip Scheduling and Tracking Flow Chart ........................................................................................ 85
11.1.2
EV Power Estimator Flowchart ........................................................................................................ 86
11.1.3
LV Power Estimator Flow chart........................................................................................................ 87
11.1.4
SOC Update Sub-Routine .................................................................................................................. 88
iii
11.1.5
SOC LV Sub-Routine Flow Chart...................................................................................................... 89
11.2
APPENDIX B QUEUING THEORY AND QCR INFRASTRUCTURE REQUIREMENTS .......................................... 89
Table of Figures
Figure 1-1 Estimated Increase in demand for non-conventional Oil ............................................................ 1
Figure 2-1 Electric Vehicle Charging Infrastructure .................................................................................... 5
Figure 2-2 DC and AC Charging Stations Differences................................................................................. 6
Figure 2-3 Inductive Contactless EV Charging System ............................................................................... 8
Figure 2-4 Renewable Energy Powered Contactless Power Transfer System .............................................. 8
Figure 3-1 Time of Day Variation in Trips ................................................................................................. 11
Figure 3-2 Percentage of Trips of Varying Trip Duration .......................................................................... 11
Figure 3-3 Cumulative Distribution of Trips vs Trip Duration................................................................... 12
Figure 3-4 Estimated Speed of Trips vs Trip Duration ............................................................................... 12
Figure 3-5 Trips Classified according to Distance Driven.......................................................................... 13
Figure 3-6 Derived Average Trip Length in Different Trip Ranges ........................................................... 14
Figure 3-7 Percentage Distribution of Different Trip Lengths ................................................................... 14
Figure 3-8 Cumulative Percentage Distribution of Trips vs Trip Length ................................................... 15
Figure 3-9 Trips Classified According to Trip Purpose .............................................................................. 16
Figure 3-10 Average Trip Distance for Different Purposes ........................................................................ 16
Figure 3-11 Seasonal Variation in Mobility ............................................................................................... 17
Figure 3-12 Weekly Variations in Mobility................................................................................................ 18
Figure 3-13 Energy Flow in EV.................................................................................................................. 19
Figure 4-1 Reference Trip Generation for the Year .................................................................................... 20
Figure 4-2 Trip Scheduling and Location Tracking.................................................................................... 21
Figure 4-3 Dumb Charging of EVs ............................................................................................................. 23
Figure 4-4 EVs Drawing Power vs Connection at Charging Point ............................................................. 24
Figure 4-5 Smart Charging Behaviour for Delayed Charging .................................................................... 25
Figure 4-6 EV Charge Power Requirement Estimator................................................................................ 26
Figure 4-8 EV Charging and State of Charge Update................................................................................. 27
Figure 5-1 Trip Profile by Purpose of Netherlands..................................................................................... 28
Figure 5-2 Trip Profile by Purpose of USA ................................................................................................ 29
Figure 5-3 Weekly trip Profile of Netherlands ........................................................................................... 29
Figure 5-4 Spatial Distribution of Vehicles by time of day ........................................................................ 30
Figure 5-5 Spatial Distribution of Vehicles over a week ............................................................................ 31
Figure 5-6 Dumb Charging SOC Profile with/without work charging ....................................................... 32
Figure 5-7 SOC Profile of vehicles with dumb charging ............................................................................ 32
Figure 5-8 Dumb Charging power Profile with/without work charging..................................................... 33
Figure 5-9 Delayed Smart Charging SOC Profile with/without work charging ......................................... 33
Figure 5-10 SOC Profile of vehicles with delayed smart charging............................................................. 34
Figure 5-11 Delayed Smart Charging power Profile with/without work charging ..................................... 34
Figure 5-12 EV charging power profile from Ecotality .............................................................................. 35
Figure 5-13 Effect of Green Range on Charging Power ............................................................................. 35
Figure 5-14 SOC Profile Over a week with Delayed Smart Charging ....................................................... 36
iv
Figure 5-15 Charging Power Profile for a week with Delayed Smart Charging ........................................ 36
Figure 6-1 Fast Charging Process ............................................................................................................... 38
Figure 6-2 .................................................................................................................................................... 39
Figure 6-3 .................................................................................................................................................... 39
Figure 6-4 .................................................................................................................................................... 40
Figure 6-5 .................................................................................................................................................... 40
Figure 6-6 .................................................................................................................................................... 41
Figure 6-7 .................................................................................................................................................... 41
Figure 6-8 .................................................................................................................................................... 42
Figure 6-9 .................................................................................................................................................... 42
Figure 6-10 Effect of Providing Work Charging on Fast Charge Requirements ........................................ 43
Figure 6-11 Battery Switching Station........................................................................................................ 43
Figure 6-12 Battery Switching Process....................................................................................................... 45
Figure 6-13 Battery Switching Station Power Management....................................................................... 46
Figure 6-14 .................................................................................................................................................. 47
Figure 6-15 .................................................................................................................................................. 47
Figure 6-16 .................................................................................................................................................. 48
Figure 6-17 .................................................................................................................................................. 48
Figure 6-18 .................................................................................................................................................. 49
Figure 6-19 .................................................................................................................................................. 49
Figure 6-20 .................................................................................................................................................. 50
Figure 6-21 .................................................................................................................................................. 50
Figure 6-22 Estimated battery requirements with 3.6 kW charging ........................................................... 51
Figure 6-23 Estimated battery requirements with 10.8 kW charging ......................................................... 52
Figure 6-24 Estimated battery requirements with 50 kW fast charging ..................................................... 52
Figure 6-25 Number of visits and Time per visit at fast charge stations .................................................... 53
Figure 6-26 Number of visits at battery swap stations................................................................................ 54
Figure 6-27 Energy Share of EV usage from Quick Charge Replenishment.............................................. 55
Figure 6-28 Higher resolution data for trips with distance more than 50 km ............................................. 56
Figure 6-29 Distance modeled for trips of distance greater than 50 km ..................................................... 57
Figure 6-30 Number of visits and Time per visit at fast charge stations with higher resolution data......... 57
Figure 6-31 Number of visits at Battery Swap stations with higher resolution data .................................. 58
Figure 6-32 Estimated battery requirements with 10.8 kW charging with higher resolution data ............. 58
Figure 6-33 Battery Capacity Required to support battery switching......................................................... 59
Figure 6-34 Energy Share of EV usage from Quick Charge Replenishment with higher resolution data .. 60
Figure 6-35 Serving units required per QCR station .................................................................................. 60
Figure 6-36 Peak Power Requirement of QCR stations.............................................................................. 61
Figure 6-37 Effect of higher charging power on number of fast charging units ......................................... 61
Figure 7-1 Micro-grid Model with DC-Coupled Battery Storage .............................................................. 63
Figure 7-2 Micro Grid Model with AC-Coupled Battery Storage .............................................................. 63
Figure 7-3 PEM and RER visualization...................................................................................................... 65
Figure 8-1 Green Charging Behaviour........................................................................................................ 66
Figure 8-2 Renewable Energy Direct Consumption with Battery Switching ............................................. 67
Figure 8-3 Energy share of Battery Switching............................................................................................ 68
v
Figure 8-4 EV Charging by Grid Control with Battery Switching ............................................................. 69
Figure 8-5 EV Charging Requested with Battery Switching ...................................................................... 69
Figure 8-6 SOC Profile of the fleet with micro grid simulation and battery switching .............................. 70
Figure 8-7 Renewable Energy Direct Consumption with Fast Charging ................................................... 70
Figure 8-8 Energy share of Fast Charging .................................................................................................. 71
Figure 8-9 EV Charging Requested with Fast Charging ............................................................................ 71
Figure 8-10 EV Charging by Grid Control with Fast Charging.................................................................. 72
Figure 8-11 Peak EV charging power on Low Voltage Network ............................................................... 72
Figure 8-12 Effect on direct renewable energy usage with green charge behaviour .................................. 73
Figure 8-13 Optimized Battery Charge Management at Switching Stations .............................................. 74
Figure 8-14 Renewable Energy Direct Usage with Optimized Battery Switching ..................................... 75
Figure 8-15 Energy share of Battery Switching with optimized battery charging...................................... 76
Figure 8-16 EV Charging by Grid Control with Optimized Battery Switching ......................................... 76
Figure 8-17 EV Charging Requested with Optimized Battery Switching .................................................. 77
Figure 8-18 Effect of EV penetration on direct Renewable energy usage in wind majority system .......... 77
Figure 8-19 Effect of EV penetration on direct Renewable energy usage in solar majority system .......... 78
Figure 8-20 Renewable Energy Direct Usage in relation to range of the fleet ........................................... 79
Figure 8-21 Energy Share of Battery Switching Stations in relation to range of fleet................................ 79
Figure 8-22 EV charging by grid control in relation to range of fleet ........................................................ 80
Figure 8-23 Effect of PEM on EV Charging by Grid Control .................................................................... 80
Table of Tables
Table 3-1 Dutch Mobility Statistics ............................................................................................................ 10
Table 3-2 Derived Quantities from MON Survey....................................................................................... 10
Table 3-3 Assumptions about Electric Vehicles ......................................................................................... 18
Table 6-1 Comparison of Fast Charging vs Battery Switching .................................................................. 37
Table 8-1 Demand Increase with EV Penetration ....................................................................................... 67
Table 8-2 Renewable Energy Direct Usage with optimized charging at swap stations .............................. 75
Table 8-3 Renewable Energy Direct Usage without optimized charging at swap stations......................... 75
vi
1
Introduction
Fossil fuels are the most important primary energy source in most countries of the world. The
transportation sector and especially the individual transport are highly dependent on them. Increased CO2
emission, the proceeding scarcity of crude oil leading to rising prices and political discussions about
future energy security initiated a green economy. Road based transport currently accounts for 22 percent
of UK CO2 emissions and therefore it is a major action area to reduce the overall CO2 emissions
(Department for Business Enterprise and Regulatory Reform & Department of Transport UK, 2008). For
Europe the European Environmental Agency has estimated that, if the present growth of greenhouse gas
emissions from the transport sector is extrapolated to 2050, by that time the transport sector greenhouse
gas emissions will exceed the total emission target for Europe.
The increasing energy use in the transport sector is adding also to the dependency of countries on oil
imports. Despite continuing improvements of the average vehicle fuel efficiency, the sheer increase in
vehicles numbers and kilometers driven, in particular in upcoming economies, is expected to keep
pushing up demand for oil in the transport sector. From an economic point of view as well as from the
perspective of resource depletion the transport sector's dependence on (imported) fossil fuels is a growing
problem. Unconventional oil sources are still abundant but can only be exploited at high costs. This
directly leads to the issue of energy security and supply and as per IEA report of 2009 (IEA, 2009) the
required oil demand with the current growth rate would demand production of oil from sources that are
not yet explored.
Figure 1-1 Estimated Increase in demand for non-conventional Oil
The main candidates for a transition towards renewable energy supply for road transport are identified as
the following by (TNO - Science and Industry,RWTH - Institute for High Voltage Technology,ECN Policy Studies, 2010).
1
•
•
•
•
Vehicles with advanced combustion engines running on (blends of conventional and) biofuels
Battery-electric vehicles using renewable electricity
Plug-in hybrid vehicles running on renewable electricity fuelled by (blends of conventional and)
biofuels in combination with renewable electricity
Fuel cell vehicles using hydrogen produced from renewable sources.
Out of this environment, the electrification of vehicle fleet has been proposed as a mitigating solution for
the energy problem. It is estimated, that electrification of the whole transportation sector could shrink
energy consumption to one fifth of current consumptions (MacKay, 2009). The price of driving on
electricity is factors lower than driving on gasoline. Furthermore adoption of Electric Vehicles would help
in promoting sustainable ways of electric energy generation, such as solar and wind energy.
In this study mainly the Battery-electric vehicles(BEV) are analyzed in detail as it is the technology which
has already reached market maturity and unlike the Combustion engine and Plug in Hybrid vehicles,
BEVs can operate completely without fossil fuels if the grid can supply green power.
1.1
Opportunities for Synergy between Electric Vehicles and Renewable Power
Sources
The Retrans report (TNO - Science and Industry,RWTH - Institute for High Voltage Technology,ECN Policy Studies, 2010) identifies the possible ways in which the renewable electric power production and
electric vehicles can have a synergetic effect.
1.1.1
Interaction between electric vehicles and existing renewable energy policies
In Europe electricity production is part of the EU Emissions Trading Scheme (EU-ETS) for CO2.
Furthermore the EU and various other countries have renewable energy target expressed as percentage
shares in the total energy supply. When energy consumption for transport is shifted from oil-based fuels
to electricity, the demand for electricity increases. In that case both types of policies lead to requirements
for increased capacity of renewable electricity production.
1.1.2
Electric Vehicles as energy buffer for renewable electricity
This is a major area where the renewable energy deployment and the electric vehicles can have a very
beneficial role for each other. The renewable power sources like solar, wind, ocean wave energy are all
intermittent in nature. Since the power output cannot be regulated it would increase the demand for
buffering capacity. With an increasing share of renewable power in electricity production this may create
problems in matching the supply pattern to the demand pattern for electricity. It can be solved by
maintaining fast transitioning intermediate and peak power plants to maintain power balance. But in the
long run with increased renewable penetration, it will be a very costly option.
Another option is to use large scale energy storage. Since electric vehicles can act as decentralized energy
buffers, future smart grids can utilize them to accelerate the transition to a sustainable electricity grid.
1.1.3 Controlling EV charging patterns to the benefit of Utilities
Electric Utilities rely on the ability to predict load patterns to schedule generation assets as well as
maintaining the stability of the electricity network. This ability directly affects the security of the supply
and revenue for utility. A shift to electric mobility will increase the overall demand for electricity. The
Electric Power Research Institute (EPRI) analyzed a scenario with 20% of ‘Vehicle Miles Travelled’ in
the United States powered by electricity in 2030; the modeled electric generating capacity was just 1.7%
higher in this scenario compared to the base case scenario that assumed no EVs or PHEVs (Duvall &
Knipping, 2007). One of the major concerns of electric utilities about the electrification of mobility is the
dynamic nature of EV loads with respect to location and time.
2
Many studies have estimated the potential impacts of uncontrolled EV charging on utility (Ashtari,
Shahidinejad, & Molinski). Uncontrolled dumb charging methods where EVs charge whenever they are
plugged in might lead to increased Electric Peak loads for which electricity producers would need to
utilize more expensive generation units. Furthermore even if the power generators are producing enough
power, the network on the medium and low voltage level might be unable to deliver the demand in some
network areas (Farmer, 2010).
The idea of a smart grid which can control the power flow is a crucial component for integrating dynamic
components like renewable power sources and EV charging loads into the electrical distribution system.
In this respect estimation of electric vehicle usage pattern is the first step in predicting potential EV load
on the grid. Many studies are currently on-going to establish the usage of BEVs as energy storage for
renewable power. The EDISON project of Denmark is profiling the vehicle usage pattern to predict the
availability of EVs as energy buffers as the Danish power mix is heading towards 50% renewable energy
mix.
Once the vehicle usage pattern is established, there are two major ways in which the EVs can be used by
the future smart grids.
1. Preferential Charging
In this option a smart grid controls the charging profile of electric vehicles, and switches chargers on and
off in function of the supply of renewable electricity. In this way renewable electricity is preferentially
guided towards electric vehicles every time peaks in production exceed the overall demand. This
renewable electricity is then used for driving the vehicles
2. Vehicle to Grid Services
An advanced possibility is that electric vehicles, which are charged with excess renewable electricity in
the way as sketched above, also deliver (part of) this energy back to the grid. This option is often referred
to as vehicle-to-grid or V2G. This would be possible once the cycle life of batteries is improved so that
cost of battery cycling comes down. Electric vehicles may also serve other functions in the electricity
system besides storage of electricity. These include frequency stabilization and providing spinning
reserve.
As mentioned before it is important to obtain the possible EV usage profile to realize the full potential of
future smart grids. It can also be argued that the success of electric vehicles will drive and depend on the
smart grid technologies. Smart grids are able to communicate with energy consuming as well as energy
supplying equipment connected to the grid. On the basis of real-time demand and supply information
together with pricing signals the smart grid influences / controls the moments at which this equipment
consumes or produces electricity. Distributed energy generation and storage systems can provide peak
power at premium prices or absorb excess power at off-peak moments as well as provide smoothing of
short term variations in the supply-demand balance.
In this thesis study the potential usage pattern of electric vehicles in the Dutch scenario is modeled and it
is used along with the smart charging method mentioned in chapter 3 to achieve preferential charging of
the electric vehicles during off-peak hours as well as to take up excess renewable power. The impacts of
quick charge replenishment technologies on the grid are also analyzed. The mobility analysis results
could be used by future Vehicle to Grid service providers to predict potential buffering capability of
BEVs and also by network planners in utilities to plan the deployment of the charging infrastructure.
3
2
State of the art of Battery Electric Vehicles and Charging Infrastructure
A brief overview of the state of the art in battery technology for electric vehicles and the different
charging infrastructure required to facilitate the EV usage is described in this chapter. The assumptions
listed here will determine the power usage demand on the utility. The other components of EVs are
standard components like the power converters and motors for propulsion etc.
2.1 Energy Storage for Transportation
The Fuel cell powered electric vehicles are not considered in this study and hence technology review is
focused only on energy storage in batteries.
Energy storage for transportation applications can be loosely divided into two primary categories: high
power/rapid discharge and high energy/extended discharge. High power devices provide short, rapid
discharges for vehicle starting and acceleration. While they cannot provide continuous discharge for
electrified transport, they can dramatically improve fuel efficiency, as demonstrated by the current
generation of hybrid electric vehicles. Currently deployed technologies for these applications include
lithium-ion and nickel-based aqueous batteries. The other technologies being explored include super
capacitors, flywheels etc. Some of these technologies, such as capacitors, may also be used as a fastresponding “buffer” between the electric drive system and the battery in an electric vehicle (EV).
The high energy capacity batteries, where stored electricity is actually used to provide a significant
fraction of the driving energy is the most crucial and expensive component of the electric vehicle. It
constitutes up to 40 percent of the weight of the EV. Hence it is important to select a battery technology
which has the highest specific energy capacity and appropriate specific power capacity (for rapid
charging/discharging) as well as with sufficient number of battery cycles so as to last at least as long as
the car.
Most commercially available and proposed EVs and PHEVs (such as the Chevrolet Volt and Nissan Leaf)
use lithium-ion batteries. In addition to the ability to store more energy per unit weight, the car makers
also benefits from this chemistry’s high power, efficiency, and long cycle life capability (Wagner,
Lakshmanan, & Mathias, 2010). Research and development efforts are focused primarily on reducing cost
and increasing energy density as well as safety of lithium-ion technology. Earlier deployed technologies,
such as lead-acid used in older EVs and nickel metal hydride used in current HEVs, are not considered
likely candidates in future EVs due to fundamental limits of energy density. Concerns have been
expressed about the large-scale availability of several metals used in lithium-ion batteries, as well as its
concentration in a few geographic regions. In the longer term, lithium-metal and metal-air batteries are in
the research and development phase, with the potential of much higher energy density than currently
available battery types.
2.1.1 Cost of Li-Ion Batteries
Costs for the current generation of Li-ion battery packs for electric vehicles are between $1000 and
$1200/kWh, including the battery cell, integration, thermal management, and other system costs (Boston
Consulting Group, 2010). The same report predicts based on future expected EV penetration for the price
to drop to $360-$440 per kWh. But recent cost estimates puts it at the $700/kWh (Brooker, Thornton, &
Rugh, 2010) .The cost of the pack relative to typical vehicle costs is a significant issue. For example, the
necessary pack size for a small- to mid-size electric sedan is on the order of 30 kWh, implying a battery
4
cost on the order of $20,000, which exceeds the total cost of many vehicles in this class. The battery
packs are estimated to have a cycle life of 1000-3000 (Department for Business Enterprise and
Regulatory Reform & Department of Transport UK, 2008).
2.2 EV Charging Infrastructure
The EVs require a charging infrastructure to tap the energy from the electricity network. The development
of charging infrastructure will need to keep pace with the developing market to ensure consumer
confidence in the ability to recharge their vehicles with minimal inconvenience. There should be
standardization of recharging systems to maximize commonality and minimize development of
manufacturer specific systems.
In general the charging technologies fall into the categories: slow charge; fast charge; and conductive,
inductive. The essential difference between conductive and inductive charging to the consumer is that
conductive charging requires a connection via a plug, whereas inductive charging requires no direct
plugged connection, only proximity. The charging infrastructure of the future is envisaged to include
three categories – Slow charging at extended parking duration, Quick Charge Replenishment to complete
trips and On-road Contactless Power transfer. A small discussion on each of the charging methods
follows.
2.2.1 Conductive Charging
This section describes the different charging methods in conductive charging. They fall under this
category when the power transfer happens to the battery via a physical connection to the charging point.
An overview is provided by (Senan, 2009).
Figure 2-1 Electric Vehicle Charging Infrastructure
5
The difference between slow charging and quick charging is described in the diagram shown below
(Dorresteijn, 2011)
The major difference between AC and DC fast charging is that it is better from a cost perspective to have
DC as the power link for fast charging over AC. The capital cost can be shared among all the EV users
saving cost and weight for the EVs. With an AC coupled fast charge system, the on-board charger needs
to be rated for the higher power of fast charging, while fast charging is envisaged to be used for a very
short time.
Figure 2-2 DC and AC Charging Stations Differences
2.2.1.1 Slow Charging
This charging level corresponds to charging using single phase connection at home or at public locations
where the vehicle is about to be parked for a long time. Generally the household fuse is rated at 16A.
Hence the power level at this level of charging is 3.6 kW in Europe (at 230V) and at 1.8 kW in USA (at
110V). At this power level depending on the battery capacity a typical EV could take 6-8 hours to charge
to full capacity. Since the vehicles are stationary for the majority of the time, if the charging time is
planned carefully, the EV charging can be done in a beneficial way for the utility, by utilizing off-peak
hour electricity or excess renewable power production. If the charging is done using a three phase
connection, the charging power can be 10.8 kW in the Europe.
An intermediate standard known as opportunity charging is also mentioned in literature (Senan, 2009). At
this charging level a three phase connection is made and a current of 32A is drawn (21.6kW power). This
can be used at places like restaurants or movie theatres where the parking time is generally between 30
minutes to 2 hours.
2.2.1.2 Quick Charging
Quick charging is required when the range of the car is not sufficient to complete a trip or the vehicle’s
current Battery State of Charge does not permit to complete the trip. In this case the vehicles would need
a quick way to increase the energy stored in the battery. Inevitably it means a sudden huge transfer of
6
energy (implying a high power transfer of energy). As an illustration let us estimate the power level at
petrol station if we assume that filling 50 liters of gasoline takes 5 minutes.
𝐸𝑛𝑒𝑟𝑔𝑦 𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑒𝑑 𝑖𝑛 𝑙𝑖𝑡𝑒𝑟 𝑜𝑓 𝑔𝑎𝑠𝑜𝑙𝑖𝑛𝑒 = 35 𝑀𝐽/𝑙
So the power transfer at the petrol station equals
𝑃𝑜𝑤𝑒𝑟 𝑙𝑒𝑣𝑒𝑙 𝑜𝑓 𝑟𝑒𝑓𝑢𝑒𝑙𝑖𝑛𝑔 =
������ ������� �����������
���� �� ��������
=
��∗��
�∗��
= 5.83 𝑀𝑊
This section gives an indication of the challenge involving Quick Charging. Even if battery technology
improves, the grid impact at the same level of power transfer will need massive upgrading of the
electricity network. For example, the medium voltage transformers can handle a power level of 10 100MVA (Siemens) while low voltage transformers can handle up to 2.5 MW (The Ural Company,
2006). So if the fast charging is required to support the same power level of a gasoline station, a
connection to the high voltage network or via multiple medium voltage transformers would be required.
The power consumption is for a single vehicle recharging. Hence the power transfer level at the Quick
charge stations would be much lower than the power transfer level of gasoline station.
There are two major methods under consideration for quick charging. One is fast charging using an onboard fast charger (not cost effective) or an off-board DC fast charger. Another method is to replace the
depleted battery and insert a fully charged battery in its place as proposed by (BetterPlace, 2012). Both
methods have their own advantage and disadvantage and they are described in detail in chapter 6.
2.2.2 Inductive Charging
The EV users will not prefer to stop for charging, rather to charge while they stop. This is a new charging
technology that has been proposed to counter the disadvantages of conductive charging and also to
counter the effect of high Li-ion battery cost as has been described in section 2.1. The following main
disadvantages are identified by (Wu, Gilchrist, Sealy, Israelsen, & Muhs, 2011).
•
•
•
The cable and connector typically delivers 2-3 times more power than standard plugs at home,
and this increases risk of electrocution especially in wet and hostile environments.
The long wire poses a trip hazard and gives to poor aesthetics for such systems.
In harsh climates that commonly have snow and ice, the plug-in charge point may become frozen
onto the vehicle.
The EV user experience can be improved if on-road charging of EVs is possible. It can extend the range
of the vehicle or provide all the propulsion power required, while driving by either providing charging
spots at stop lights or by building charging infrastructure on the road as proposed by (Chopra & Bauer,
2011). This completely eliminates the problem of range anxiety with EV’s as all the required power by
the vehicle traveling on freeways can be supplied by the grid directly through the roadway.
7
Figure 2-3 Inductive Contactless EV Charging System
An example of an Inductive Charging System as described by (Wu, Gilchrist, Sealy, Israelsen, & Muhs,
2011)
An example for built environment integrated contactless power transfer system is proposed by (Chopra &
Bauer, 2011). A graphical illustration is given below.
Figure 2-4 Renewable Energy Powered Contactless Power Transfer System
The technique has been tested at various power levels from 2.5 kW to 60 kW and it has efficiency from
grid to battery in the range 80-85 percent at a distance of 100-200 mm separation between the primary of
8
the coil and the secondary. The safety limit of Magnetic field exposure to human body has been met with
sufficient margin for the studies by KAIST and University of Auckland (Wu, Gilchrist, Sealy, Israelsen,
& Muhs, 2011).
This method is being explored as a comfort scenario for the EV integration and it would also enable to
circumvent the battery limitation by using in-motion power transfer systems which require a much
smaller battery size. The infrastructure requirement for in-motion charging system could be relatively
small since the U.S. interstate highways make up only 1% of roadway miles, yet they carry 22% of all
miles traveled.
9
3
Mobility Analysis
This section describes the mobility patterns analyzed for estimating the grid impacts of Electric Vehicles.
The Mobility Research Survey of Netherlands 2009 is the basis for the EV mobility model developed
during this work. The key figures extracted from this survey are as follows.
The survey has two major result forms. The ministry has published a report with the key figures like
average distance driven, average no of trips per day, average no of trips according to time of day etc. The
key values used in the model are plotted below. The MON survey provides information for researchers
and policy makers in the field of transport and mobility of Dutch population. The major results from
MON 2009 is summarized as follows
Data
Total Dutch Population
Total no of Personal Cars
Total no of households
Average no of Car trips per person per day
Average distance travelled in car per person per
day
Value
16,319,000
7,588,000
7,300,000
0.97
16.59 km
Table 3-1 Dutch Mobility Statistics
From the above data set the following important parameters can be derived
Parameter
𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛
𝑃𝑒𝑟𝑠𝑜𝑛𝑠 𝑝𝑒𝑟 𝑉𝑒ℎ𝑖𝑐𝑙𝑒 =
𝑇𝑜𝑡𝑎𝑙 𝑛𝑜 𝑜𝑓 𝑐𝑎𝑟𝑠
𝐴𝑣𝑔 𝑛𝑜 𝑜𝑓 𝑡𝑟𝑖𝑝𝑠 𝑝𝑒𝑟 𝑣𝑒ℎ𝑖𝑐𝑙𝑒 𝑝𝑒𝑟 𝑑𝑎𝑦
= 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑛𝑜 𝑜𝑓 𝑐𝑎𝑟 𝑡𝑟𝑖𝑝𝑠 𝑝𝑒𝑟 𝑝𝑒𝑟𝑠𝑜𝑛 𝑝𝑒𝑟 𝑑𝑎𝑦
∗ 𝑃𝑒𝑟𝑠𝑜𝑛𝑠 𝑝𝑒𝑟 𝑉𝑒ℎ𝑖𝑐𝑙𝑒
𝐴𝑣𝑔 𝑑𝑖𝑠𝑡𝑎𝑛𝑐𝑒 𝑝𝑒𝑟 𝑣𝑒ℎ𝑖𝑐𝑙𝑒 𝑝𝑒𝑟 𝑑𝑎𝑦
= 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑑𝑖𝑠𝑡𝑎𝑛𝑐𝑒 𝑡𝑟𝑎𝑣𝑒𝑙𝑒𝑙𝑑 𝑖𝑛 𝑐𝑎𝑟 𝑝𝑒𝑟 𝑝𝑒𝑟𝑠𝑜𝑛 𝑝𝑒𝑟 𝑑𝑎𝑦
∗ 𝑃𝑒𝑟𝑠𝑜𝑛𝑠 𝑝𝑒𝑟 𝑉𝑒ℎ𝑖𝑐𝑙𝑒
𝑇𝑜𝑡𝑎𝑙 𝑛𝑜 𝑜𝑓 𝑐𝑎𝑟𝑠
𝑉𝑒ℎ𝑖𝑐𝑙𝑒𝑠 𝑝𝑒𝑟 𝐻𝑜𝑢𝑠𝑒ℎ𝑜𝑙𝑑 =
𝑇𝑜𝑡𝑎𝑙 𝑛𝑜 𝑜𝑓 ℎ𝑜𝑢𝑠𝑒ℎ𝑜𝑙𝑑𝑠
Value
2.15
2.09
35.67 km
1
Table 3-2 Derived Quantities from MON Survey
The MON survey results are provided in per person per day format. This format is used since it is
assumed that the mobility pattern of a population is affected as a whole and not just by the part of the
populace doing that activity. The average number of trips per vehicle and the trip distance per vehicle per
day is the most crucial data as unlike any other modeling, the vehicle fleet needs to be modeled using
individual vehicles and not as a whole fleet. Without individual vehicle modeling the requirement of
charging will follow the patterns we impose on it and will not bring out the many side effects of random
trip occurrences.
The MON survey also characterizes vehicle trips based on a lot of parameters. Some of the most
interesting statistics are shown below. These statistics are used in the model to generate a distance
distribution of the trips and a temporal distribution of trips separated by trip purpose.
10
Percentage of distance covered/ Trips
12.00%
Distribution of distance driven and trips
based on time of day
Percentage Distance Driven
10.00%
Percentage no of Trips
8.00%
6.00%
4.00%
2.00%
0.00%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time of Day
Figure 3-1 Time of Day Variation in Trips
The distribution of trips over time of day data shows clearly that there is a lot of trips during day time and
the trips are mostly peaking during 7-9 am in the morning and between in 4-6 pm in the evening. This
data correlates well with the peak traffic hours.
3.1 Trip Classification According to Trip Duration
The no of trips and distance driven have also been classified according to the trip duration. This data
gives an indication as to the average speed of a trip if the duration of the trip is known. This data is used
in the EV model to get the trip timing.
Percentage dist covered/no of trips
Trips classified according to trip duration
25.0%
20.0%
Percentage no of trips
Percentage distance Covered
15.0%
10.0%
5.0%
0.0%
1-5
5 - 10 10 - 15 15 - 20 20 - 25 25 - 30 30 - 45 45 - 60 60 - 90
90 120
120 or
more
Trip Duration in Mins
Figure 3-2 Percentage of Trips of Varying Trip Duration
The cumulative percentage distribution of the above data set provides information as to the number of
vehicles whose trip duration falls under a certain duration.
11
Cumulative distribution of trips according to trip
duration
Cumulative percentage no of
trips/distance driven
120%
Cumulative Percentage no of trips
100%
Cumulative Percentage of distance driven
80%
60%
40%
20%
0%
1-5
5 - 10
10 - 15 15 - 20 20 - 25 25 - 30 30 - 45 45 - 60 60 - 90 90 - 120 120 or
more
Trip duration in mins
Figure 3-3 Cumulative Distribution of Trips vs Trip Duration
The cumulative distribution shows that 93% of the trips are of duration one hour or less and 63.5% of the
distance is covered by those trips. From this data we can extract the average speed of the vehicles for
different trip durations. The distance per trip in each category is calculated by dividing the distance
covered by the number of trips and the speed is obtained by dividing the average trip distance by the
mean of the time duration. The following distribution of speed is obtained. The speed estimation is a
rough estimation method, but it shows that the average speed shows an increasing profile as expected
when the trip duration increases.
Speed in km/hour
Speed variation with trip duration
90
80
70
60
50
40
30
20
10
0
Speed in km per hour
1 - 5 5 - 10 10 15
15 20
20 25
25 30
30 45
45 60
60 90
90 - 120 or
120 more
Trip duration
Figure 3-4 Estimated Speed of Trips vs Trip Duration
12
3.2 Trip Classification According to Trip Length
The trips are classified according to the trip length and this data provides an insight into the daily usage of
vehicles and the corresponding power requirements as well as fast charging needs. This distribution of
distances and trips are used to assign a distribution of distances to the trips.
Percentage of trips/distance covered
Trip/Distance Covered classified according to
trip range
45%
40%
35%
Percentage no of trips
Percentage distance covered
30%
25%
20%
15%
10%
5%
0%
0.1 - 0.5 -1 1- 2.5 2.5 - 3.7 - 5 5 - 7.5 7.5 - 10 0.5
3.7
10
15
Trip Range in km
15 20
20 30
30 40
40 - 50 or
50 more
Figure 3-5 Trips Classified according to Distance Driven
One of the major data points to be looked at is the distance covered in the ‘50km or more’ category. Even
though the no of trips covered in quite small in the that category, the distance covered in that category is
quite high and consequently the need for fast charging can be quite high with this data set. In most of the
traffic analysis the information on distance covered in each trip range is generally not specified and this
leads to miscalculation on the requirement of fast charging. The average trip length in different distance
ranges can be extracted by using the above dataset. It is calculated as the following equation
𝐴𝑣𝑒𝑟𝑎𝑔𝑒 𝐷𝑖𝑠𝑡𝑎𝑛𝑐𝑒 𝑜𝑓 𝑎 𝑇𝑟𝑖𝑝 𝑤𝑖𝑡ℎ 𝑟𝑎𝑛𝑔𝑒 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 𝑟𝑎𝑛𝑔𝑒(𝑖)𝑎𝑛𝑑 𝑟𝑎𝑛𝑔𝑒(𝑖 + 1) =
(𝑝𝑒𝑟𝑐𝑒𝑛𝑡𝑎𝑔𝑒 𝑑𝑖𝑠𝑡𝑎𝑛𝑐𝑒 𝑐𝑜𝑣𝑒𝑟𝑒𝑑 (𝑖)⁄𝑝𝑒𝑟𝑐𝑒𝑛𝑡𝑎𝑔𝑒 𝑛𝑜 𝑜𝑓 𝑡𝑟𝑖𝑝𝑠(𝑖)) × 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑡𝑟𝑖𝑝 𝑙𝑒𝑛𝑔𝑡ℎ
This information is used in assigning the trip distance for different trips in the vehicle distribution model.
The distances are assigned in such a way that the overall percentage distribution of the different distances
occurring in a year remains the same as in the survey results. To avoid complicated modeling, the
distances are assigned the exact values and tolerance is not introduced. The plot of average trip length
calculated with the above equation is plotted in the bar graph below.
13
Average trip length of different trip ranges
100.00
Trip Length in km
Average Trip length in km
90.00
80.00
70.00
60.00
50.00
40.00
30.00
20.00
10.00
0.00
0.1 - 0.5 -1 1- 2.5 2.5 - 3.7 - 5 5 - 7.5 7.5 - 10 - 15 15 - 20 20 - 30 30 - 40 40 - 50 50 or
0.5
3.7
10
more
Trip Range in km
Figure 3-6 Derived Average Trip Length in Different Trip Ranges
This data is visualized as shown below to get a percentage distribution of trip length as follows.
18%
16%
Percentage no of trips
Percentage no of trips
Percentage no of trips
14%
12%
10%
8%
6%
4%
2%
0%
0.67
1.60
3.00
4.17
6.00
8.14 11.60 16.43 22.78 33.60 47.33 86.38
Trip Distance in km
Figure 3-7 Percentage Distribution of Different Trip Lengths
14
Cumulative percentage no of trips/distance
covered
The cumulative distribution of the trip and distance data is shown below
Trip/Distance Covered classified according to trip
Range
120%
100%
80%
Cumulative no of trips percentage
Cumulative dist percentage
60%
40%
20%
0%
0.1 - 0.5 -1 1- 2.5 2.5 - 3.7 - 5 5 - 7.5 7.5 - 10 - 15 15 - 20 20 - 30 30 - 40 40 - 50 50 or
0.5
3.7
10
more
Trip Range in km
Figure 3-8 Cumulative Percentage Distribution of Trips vs Trip Length
The cumulative distribution shows that 89% of the trips are less than 40 km. This correlates well with the
assertion that Electric Vehicles can cater to the needs of a vehicle user majority of the time even with a
limited range.
3.3 Trip Classification by Trip Purpose
The MON survey also classifies the trips into different purposes and that data is quite useful in
formulating policy for charging infrastructure development. Policymakers will be able to deploy chargers
at the points where the EV users would require them the most resulting in better utilization and more
insight about the load distribution for the utilities
15
Percentage distance covered/trips
Trips and distance covered according to
trip purpose
40.0%
35.0%
30.0%
25.0%
20.0%
15.0%
10.0%
5.0%
0.0%
Percentage of total trips
Percentage of total distance
Trip Purpose
Figure 3-9 Trips Classified According to Trip Purpose
The average trip distance for different purposes is estimated using the number of trips and distance
covered in each category. The equation used is quite straight forward and given as
𝐴𝑣𝑒𝑟𝑎𝑔𝑒 𝑇𝑟𝑖𝑝 𝐷𝑖𝑠𝑡𝑎𝑛𝑐𝑒 (𝑝𝑢𝑟𝑝𝑜𝑠𝑒)
= 𝑝𝑒𝑟𝑐𝑒𝑛𝑡𝑎𝑔𝑒 𝑜𝑓 𝑡𝑜𝑡𝑎𝑙 𝑑𝑖𝑠𝑡𝑎𝑛𝑐𝑒(𝑝𝑢𝑟𝑝𝑜𝑠𝑒)⁄𝑝𝑒𝑟𝑐𝑒𝑛𝑡𝑎𝑔𝑒 𝑜𝑓 𝑡𝑜𝑡𝑎𝑙 𝑡𝑟𝑖𝑝𝑠 (𝑝𝑢𝑟𝑝𝑜𝑠𝑒) × 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑡𝑟𝑖𝑝 𝑙𝑒𝑛𝑔𝑡ℎ
Average Distance per Trip
Average distance per trip in km
30.00
25.00
20.00
15.00
10.00
5.00
0.00
Trip Purpose
Figure 3-10 Average Trip Distance for Different Purposes
16
Percentage trips/Distance Driven per month
Seasonal variations and weekly variations in travel profile are very important in determining the charging
strategy from a renewable energy powered grid. The distribution of trips over different months is
provided below and it is uniform through the year with slightly higher number of trips during the colder
months. This variation is shown for illustrating the changing travel patterns depending on the time of
year. This has been modeled implicitly in the model as the model uses the percentage distribution of trips
for a whole year as provided by the online database of the MON survey.
12.00%
10.00%
No of trips and distance covered
distributed over months
Percentage trips per month
Percentage total distance per month
8.00%
6.00%
4.00%
2.00%
0.00%
Month
Figure 3-11 Seasonal Variation in Mobility
There is strong variation in the number of trips and distance covered on different days of the week. The
distribution shows that the number of trips is substantially lower on Sundays and hence the availability of
cars are higher but the average charge of the fleet would be much higher than on other days.
17
Weekly Variation in Trips
Percentage trips/Distance Driven
20%
18%
Percentage trips per weekdays
Percentage dist per weekdays
16%
14%
12%
10%
8%
6%
4%
2%
0%
Sunday
Monday
Tuesday Wednesday Thursday
Day of Week
Friday
Saturday
Figure 3-12 Weekly Variations in Mobility
3.4 EV parameter assumptions
The following assumptions have been made about the properties of EV in the model (Tuffner & KintnerMeyer, 2011).
Parameter
Energy required to cover one km(for drive train)
Average home charger power consumption
Average home charger efficiency
Average battery charging efficiency
Average battery discharging efficiency
Battery Cycling (Usable portion of battery
capacity)
Maximum State of Charge
Minimum State of Charge
State of charge possible with fast charging
Fast charger power consumption
Average fast charger efficiency
Minimum driving range reserve
Value
0.15 kWh
3.6 kW
92%
85%
86.7%
65%
90%
25%
80%
54.35 kW
92%
10 km
Table 3-3 Assumptions about Electric Vehicles
Energy required to cover one kilometer is the energy consumed by the mechanical system of the vehicle
from the battery. This does not cover the discharge efficiency of the battery and the electrical drive train.
The charging efficiency of the battery is the amount of energy dissipated while charging the battery. This
value does not cover the charging efficiency of the charging point. All these values increase the actual
demand on the grid. Fast charging is assumed to raise the state of charge to 80 percent of the useable
capacity. This is the recommended value by Nissan leaf. The actual capacity of the battery would be
higher than what is disclosed to the user to avoid deep charge-discharge cycles. The parameters used in
the model can be changed in the final result to visualize the effect of lower/higher efficiencies on the total
18
grid usage. These parameters have been derived from the sources listed here. From the above values we
can easily calculate the total EV charging requirement per vehicle per day.
𝐸𝑛𝑒𝑟𝑔𝑦 𝑅𝑒𝑞𝑢𝑖𝑟𝑒𝑑 𝑡𝑜 𝐶𝑜𝑣𝑒𝑟 𝑜𝑛𝑒 𝑘𝑚 𝑓𝑟𝑜𝑚 𝑡ℎ𝑒 𝐵𝑎𝑡𝑡𝑒𝑟𝑦 =
𝐸𝑛𝑒𝑟𝑔𝑦 𝑟𝑒𝑞𝑢𝑖𝑟𝑒𝑑 𝑡𝑜 𝑐𝑜𝑣𝑒𝑟 𝑜𝑛𝑒 𝑘𝑚 𝑓𝑜𝑟 𝑑𝑟𝑖𝑣𝑒 𝑡𝑟𝑎𝑖𝑛 ⁄ 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑏𝑎𝑡𝑡𝑒𝑟𝑦 𝑑𝑖𝑠𝑐ℎ𝑎𝑟𝑔𝑖𝑛𝑔 𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑐𝑦 =
0.173 𝑘𝑤ℎ/𝑘𝑚
𝐸𝑛𝑒𝑟𝑔𝑦 𝑅𝑒𝑞𝑢𝑖𝑟𝑒𝑑 𝑡𝑜 𝐶𝑜𝑣𝑒𝑟 𝑜𝑛𝑒 𝑘𝑚 𝑓𝑟𝑜𝑚 𝑡ℎ𝑒 𝑐ℎ𝑎𝑟𝑔𝑒𝑟 =
𝐸𝑛𝑒𝑟𝑔𝑦 𝑟𝑒𝑞𝑢𝑖𝑟𝑒𝑑 𝑡𝑜 𝑐𝑜𝑣𝑒𝑟 𝑜𝑛𝑒 𝑘𝑚 𝑓𝑟𝑜𝑚 𝑏𝑎𝑡𝑡𝑒𝑟𝑦
�𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑏𝑎𝑡𝑡𝑒𝑟𝑦 𝑐ℎ𝑎𝑟𝑔𝑖𝑛𝑔 𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑐𝑦 =
0.204 𝑘𝑤ℎ/𝑘𝑚
𝐸𝑛𝑒𝑟𝑔𝑦 𝑅𝑒𝑞𝑢𝑖𝑟𝑒𝑑 𝑡𝑜 𝐶𝑜𝑣𝑒𝑟 𝑜𝑛𝑒 𝑘𝑚 𝑓𝑟𝑜𝑚 𝑡ℎ𝑒 𝑔𝑟𝑖𝑑 =
𝐸𝑛𝑒𝑟𝑔𝑦 𝑟𝑒𝑞𝑢𝑖𝑟𝑒𝑑 𝑡𝑜 𝑐𝑜𝑣𝑒𝑟 𝑜𝑛𝑒 𝑘𝑚 𝑓𝑟𝑜𝑚𝑡ℎ𝑒 𝑐ℎ𝑎𝑟𝑔𝑒𝑟
�𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑐ℎ𝑎𝑟𝑔𝑖𝑛𝑔 𝑝𝑜𝑖𝑛𝑡 𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑐𝑦 =
0.222 𝑘𝑤ℎ/𝑘𝑚
𝑇𝑜𝑡𝑎𝑙 𝐸𝑛𝑒𝑟𝑔𝑦 𝑝𝑒𝑟 𝑉𝑒ℎ𝑖𝑐𝑙𝑒
= 𝐸𝑛𝑒𝑟𝑔𝑦 𝑟𝑒𝑞𝑢𝑖𝑟𝑒𝑑 𝑡𝑜 𝑐𝑜𝑣𝑒𝑟 𝑜𝑛𝑒 𝑘𝑚 𝑓𝑟𝑜𝑚 𝑡ℎ𝑒 𝑔𝑟𝑖𝑑 ∗ 𝑇𝑜𝑡𝑎𝑙 𝑛𝑜 𝑜𝑓 𝐾𝑚𝑠 𝑝𝑒𝑟 𝑑𝑎𝑦
= 7.7 𝑘𝑤ℎ/𝑑𝑎𝑦
The following graphical diagram shows the efficiency and energy losses in the different modules of an
electric vehicle.
Figure 3-13 Energy Flow in EV
19
4
EV fleet modeling Methodology
Based on the yearly trip distribution for the year 2009 from the MON survey the vehicles fleet behavior is
modeled. The algorithm to generate the trip profile tracks each vehicle in the fleet as an average SOC
analysis will not bring out the finer effects due to individual SOC levels of the vehicles.
The model is split into four parts. The first part initializes the fleet and calculates the no of trips to be
scheduled over a year. This data is computed for each hour of the day for each purpose. This module is
based on the results described in chapter 1 and the results provided in chapter 1 is utilized along with a
year profile of probability of trips to build the trip profile.
4.1 Reference Trip Data Generation
This sub-section describes how the data from the mobility survey is used to generate a trip profile with
differentiation between trip purposes and maintaining the average trip profile according to the data
provided by MON for a year. The MON survey gives the summary results described in chapter 1. The
online version of the same has another data set which has more resolution. The online database provides
for each of the seven motives described before the percentage of a trip occurring compared to the total
number of trips for each hour of the day, each day of the week for each month in a year. For example the
database gives the percentage for a trip occurring for commute between 5am-6am on a Monday in the
month of January. This information is used to generate a trip profile based on the number of the particular
weekday in the month and splitting the percentages equally among the various occurrence of that day in
the given month. The algorithm can be summarized as follows.
Figure 4-1 Reference Trip Generation for the Year
20
4.2 Trip Scheduling Sub-Routine
The trip scheduling subroutine calculates the distance of the current trip and based on the availability of
the EV selects one vehicle and deducts the energy required from the vehicle and locks it till the return trip
is scheduled. Return trip is scheduled along with the original trip based on the estimated activity duration.
Figure 4-2 Trip Scheduling and Location Tracking
The trip schedule sub-routine assigns the forward trip with trip time including the time required for quick
charge replenishment and the return trip time without quick charge replenishment time. The detailed flow
chart is shown in appendix A. The main parts of the trip scheduling sub routine are as follows.
4.2.1 Trip Distance and Trip Speed Assignment
This function is called upon by the trip scheduling sub-routine to get a distance and speed of the current
trip. This sub-routine randomly assigns one of the distance calculated using the mobility survey result
shown in Figure 3-7 for that particular purpose and based on the distance assigned; the speed is looked up
from the look up table Figure 3-4. This approach makes sure that the randomness in the trip pattern is
maintained for the fleet. One of the reasons, the model does not assign fixed patterns to each vehicle is
that such an assignment will only bring out an ideal scenario and user parameters are modeled as fixed
throughout the year.
21
Once the distance and time is estimated the vehicles are checked whether they are already executing a trip
or whether they have at least minimum distance so that they can quick charge and execute the trip. If such
a vehicle is obtained then the quick charge and location tracking sub-routine makes the changes to
vehicles SOC and vehicle location to reflect the effects of the trip. Quick Charge Replenishment is
explained in detail in chapter 6.
4.2.2 Location Tracking
The location tracking sub-routine assigns the return trip to the vehicle and also marks its location from the
time of forward trip origin to the time it starts the return trip. The time spent at each activity is assigned
from minimum activity duration to maximum activity duration. If a return trip is not found in the trip
distribution database for the year even after maximum activity duration a return trip is force scheduled.
After location tracking sub-routine is finished, the current number of vehicles at different locations are
tabulated and stored to get the special distribution of vehicles over an average period. The following two
sub-routines act on the EV fleet to calculate and schedule the charging of the EV fleet.
4.3 EV Charging Power Requirement Estimation Sub-Routine
This sub-routine estimates three key figures to be used by the Grid control to modulate the EV charging
behavior. It will calculate the minimum charging power required by the connected EVs, the maximum
charging power that can be accommodated by the connected EVs and the quick charge replenishment
charging requirement (fast charging / battery swap station power). The minimum required power is
determined by the charging strategy followed by the EVs. In this model, two behaviors are modeled. One
is the dumb charging scenario where the connected EVs demand power at full level at the instant of
connection and the other is the green charging scenario where charging is modulated taking into account
the current level of SOC and the amount of estimated parking time remaining. In the dumb charging
scenario the required power would be equal to possible EV charging power as all the connected cars are
charging at full power.
The output of this sub-routine is used by the micro-grid control algorithm to modulate the charging power
for the fleet. Based on the remaining power in the grid after servicing normal consumption (which
includes regular electricity demand, the minimum requested EV power and the quick charge
replenishment power) the grid control assigns more power or no extra power to the EV fleet based on the
renewable generation.
The dumb charging scenario gives us the least amount of quick charge replenishment power that can
occur with the introduction of EVs on the grid. One of the key differences with green charging behavior
can be observed when there is no grid control in the green charging scenario. When that is the case the
amount of fast charging demanded would increase as well as over time the amount of minimum power
required will also increase due to the un-served EV demand during low demand periods for the grid or
during over production periods of renewable sources. These results would be shown in Chapter 6 where
quick charge replenishment strategies are discussed.
The goal with the grid control and green charging scenario is to attain the same level of power
contribution from slow charging as is the case with dumb charging. This would maximize the uptake of
renewable power and shift the power usage to off-peak periods for the grid which in turn reduces grid
exchange and/or need for local storage.
22
4.3.1 Dumb Charging and Green Charging
The two charging behaviors are shown below graphically. The dumb charging scenario as explained
above demands full charging power from the low voltage charger if the connected location is charging
enabled.
Figure 4-3 Dumb Charging of EVs
The need for a green charging scenario arises from the fact that EVs can be recharged in a shorter
duration compared to the total time it is plugged into the charger. This is shown in many EV studies as
one such example is shown here.
23
Figure 4-4 EVs Drawing Power vs Connection at Charging Point
The figure is taken from the Q4 report of 2011 of the EV project USA (Ecotality, 2012). We can see that
EVs are drawing power only 17% of the time it is connected to the charger.
The green charging strategy is split into three sections. This division is utilized by the LV estimator subroutine of the EV charging power requirement module. The three sections are as follows.
1. The current range of the car is less than the anxiety range of EV drivers (it is assumed to be
10 km in this model)
• In this scenario the vehicles will charge aggressively (at full rated power) to avoid range
anxiety
2. The current range of the car is greater than range anxiety range but less than a defined green
range. This range can be varied and it signifies the SOC to which the cars will charge
aggressively even in green charging mode.
• In this scenario if the estimated remaining parking time is less than or equal to 4 hours
the cars will charge at full rated power.
• If the estimated remaining parking time is greater than 4 hours then the cars would scale
the charging power to the ratio of remaining energy needed to achieve green range to the
estimated remaining parking time. This charging power can be increased by grid control
3. The current range of the car is greater than the defined green range
• In this scenario if the estimated remaining parking time is less than zero (as an estimate
provided by a Canadian study is used to define expected parking duration), then cars will
charge it full power. This is to maximize the range of the car as well as follows from
Long Lam’s thesis on Li ion battery models.
• If the estimated remaining parking time is between 1-4 hours then the charging power is
scaled to the ratio of remaining energy needed to achieve full range to the estimated
remaining parking time. This charging power can be increased by grid control.
• If the estimated remaining parking time is greater than 4 hours then the vehicles will not
charge unless extra power is provided by the grid control
24
Figure 4-5 Smart Charging Behaviour for Delayed Charging
The EV charging power estimation sub-routine is shown in the below flowchart. The algorithm can be
explained as follows in words.
1. A loop is run over all the vehicles.
2. It is checked if the vehicle selected is not starting a trip and its battery is not full.
3. If the above two conditions are satisfied it checks if the current location is charging enabled
(home and work are the only locations where charging is allowed in this model)
4. If the 3rd condition is also valid then the LV power estimator sub-routine is invoked.
25
5. Otherwise the car is checked whether it has a scheduled return trip and if it the range is not
sufficient the quick charge replenishment required to complete the trip is estimated.
6. All the steps are repeated until all the vehicles under study are covered.
Figure 4-6 EV Charge Power Requirement Estimator
The detailed flow chart is provided in the appendix A.
4.4 SOC Tracking and Update Sub-Routine
The SOC tracking subroutine takes the power provided by the grid (it will at least equal the minimum
required power calculated by the previous sub-routine) and update the SOC of the vehicles connected to
charging enabled locations. This subroutine makes sure the EVs are charged according to the inputs from
the grid control. This sub-routine also updates the SOC changes due to return trips as well as by quick
charge replenishment on return trips. The algorithm followed is very similar to that of EV Power
Requirement Estimation Sub-routine but it gets an additional input from the grid and the SOC information
of all the vehicles are updated and also the location information of vehicles performing return trips are
also updated. The algorithm is graphically described below.
This module is very similar to the previous module and the difference is that the grid control makes
changes to charging power and that the actual charging of EVs and corresponding changes to SOC are
implemented in this sub-routine.
26
Figure 4-7 EV Charging and State of Charge Update
One note on the fast charging mode of quick charge replenishment is that, the power required for fast
charging is estimated with per minute resolution instead of per hour unlike all the other variables in this
model. Since the data on which this model is built only has a per hour resolution it is not straight forward
why the fast charging has to be tracked per minute. The reason is that fast charging rarely lasts for more
than 30 minutes, since it only charges the vehicle to a maximum SOC of 80 percent. It is not
recommended to fast charge beyond this level to avoid thermal and chemical breakdown of the battery. If
the fast charging were to be tracked per hour, then the value we obtain would have been similar to what
we obtain by averaging the per minute data over an hour and that will bring down the peak demand put on
the grid. Even though the exact timing within the hour cannot be obtained, this way of estimation
provides the fast charging requirement with a precision of one hour. This is sufficient to analyze peak
time demand accurately.
The detailed flowchart is provided in appendix A.
27
5
5.1
Low Voltage Grid Impacts and Mobility Results
Trip Distribution Profiles Obtained from Reference Data Generation SubRoutine
The mobility data is utilized to generate a trip profile and the following figure shows the trip patterns on
an average day for 100 vehicles simulated.
Figure 5-1 Trip Profile by Purpose of Netherlands
As we can see the no of trips peaks in the morning and the evening. Commute trips peak at both the times.
The peak for commute occurs in the morning, where as the peak for total no of trips occur in the evening.
This data provides us an insight into the time of possible peak power occurrence. In the case of
uncontrolled charging the charging power would peak in the evening as lot of vehicles would reach home
after a trip and since every trip reduces the range, all such cars can charge if plugged in. Hence with
uncontrolled charging the peak demand which also occurs after the evening rush hour traffic would
increase even further. We can observe this later on from the simulation results corresponding to
uncontrolled charging.
The trip profile obtained compares quite well with the result published by the National Household Travel
Survey of United States (Source: NHTS 2009).
28
Figure 5-2 Trip Profile by Purpose of USA
The trip profile changes depending on the day of the week as well as there are seasonal variations. An
average week profile is shown in the figure below. We can observe that the number of commute trips and
work related trips as expected is negligible on weekends. The number of shopping trips is higher on
Fridays and Saturdays and the total number of trips is quite low on weekends compared to weekdays,
particularly on Sundays. This would mean there are far more stationary cars on weekends and that the
charging strategy can change based on the day of the week. In a future scenario it is assumed that EVs
could be used for regulation and/or storage. Hence this information combined with customer input can be
quite valuable.
Figure 5-3 Weekly trip Profile of Netherlands
29
5.2 Vehicle Location distribution
The trip scheduling and tracking module also keeps track of the number of vehicles connected at different
locations according to time. Knowledge of this information gives an idea in planning charging stations.
The following figure shows the number of vehicles at different locations and its variation with respect to
time. The cars are mostly parked at home and there is a reduction in the cars at home during the mid day
hours. Towards the end of the day the cars return home. This is an expected result, but the important fact
is the information that if the charging is provided at home and at work majority of the fleet can be covered
and a sizeable portion of the EV fleet will be available connected to the grid.
Figure 5-4 Spatial Distribution of Vehicles by time of day
The weekly profile of cars at different location shows that the travel destinations vary according to the
day of week. This would influence the ideal positioning of charging infrastructure. This could also lead to
a scenario where chargers are not always deployed at the same spot and could be shifted spatially to cover
demands according to the day of the week. This could prove to be a costly option, but if the charging
infrastructure needs to be rolled out quickly to meet the increasing demand for EVs this could be a
measure to start the large scale deployment of EVs. This distribution figure matches well with the results
obtained by (Zhao, Prousch, Hubner, & Moser, 2010).
30
Figure 5-5 Spatial Distribution of Vehicles over a week
5.3 EV Charging Power Profile on Low Voltage (LV) Grid
In this section we would analyze the impact of the EV charging load on the Low Voltage Grid. It should
be mentioned here that the EV charging load is represented in separate entities for charging at home and
work. It is represented in a single power profile. As an example some of the key figures of grid simulation
without grid control for both dumb charging strategy and green charging strategy are presented here. As
described before the advantages of green charging scenarios would be described in chapter 8 when grid
controls are available.
As an example the power and SOC profile for an average day for dumb charging and green charging with
fast charging as the mode of quick charge replenishment for a fleet with a range of 130 km is shown here.
Each section would also show the effects of enabling home charging only and enabling home and work
charging. A generalized plot on the effect of both charging scenario along with enabling charging at
different locations as well as range of the fleet will be shown later.
Mean charging power required for EV charging is calculated as follows. The total energy required per
vehicle from the grid was estimated in chapter 3.4. If that power is assumed to be distributed uniformly
throughout the day we will get the mean EV power consumption from all sources (including fast charge
replenishment).
𝑀𝑒𝑎𝑛 𝐸𝑉 𝑃𝑜𝑤𝑒𝑟 =
𝑇𝑜𝑡𝑎𝑙 𝐸𝑛𝑒𝑟𝑔𝑦 𝑅𝑒𝑞𝑢𝑖𝑟𝑒𝑑 𝑓𝑟𝑜𝑚 𝐺𝑟𝑖𝑑 𝑝𝑒𝑟 𝑑𝑎𝑦�
24 = 0.3291 𝑘𝑤
All the power values are normalized with respect to this value for plotting. The average EV requirement is
only 1/10th of the full rated capacity of the LV charger.
5.3.1 EV Simulation with Dumb Charging
The entire fleet is charging aggressively and the mode of quick charge replenishment is fast charging.
The SOC profile of the entire fleet for an average day is shown below. The SOC peaks around 5 am and it
steadily drops during the day and hits a minimum around the evening rush hour at 6pm. There after the
31
SOC steadily increases. This profile holds true for both the charging enabled scenarios even though
charging enabled at work reduces the fall in SOC value through the day. This is a direct consequence of
the fact that 37 % of all the trips are for commuting to work.
Figure 5-6 Dumb Charging SOC Profile with/without work charging
For better understanding of the fleet characteristics it would be interesting to see how the SOC of vehicles
are distributed.
Figure 5-7 SOC Profile of vehicles with dumb charging
Charging power required follows the SOC profile of the fleet and the location of vehicles. The figure
below shows the effect of enabling charging at different locations as well as typical power profile for
dumb charging scenario. The charge power peaks in the evening as all the vehicles get home and
immediately plugs in. The advantage of enabling charging at work is already obvious from the fact that
32
the evening peak reduces from 1.8 times the mean value to 1.4 times the mean value which makes a big
impact on the grid capacity requirement.
Figure 5-8 Dumb Charging power Profile with/without work charging
This result shows the disadvantage of uncontrolled charging. The EV demand peak coincides with the
peak of the conventional demand. This would aggravate the existing situation and highlights the need for
smarter charging strategies to minimize peak demand and to match the demand with generation.
EV Simulation with Smart Charging
The cars charge according to the green charging algorithm described in chapter 2. The power demand is
throttled based on the last trip timing and the current SOC value of the battery. The results are obtained
for a fleet of range 130 km with fast charging as the mode of quick charge replenishment and no fast
charging enabled. Also the green range is defined as 50 percent of the fleet range. The result of the
variation of this parameter is shown later.
5.3.2
Figure 5-9 Delayed Smart Charging SOC Profile with/without work charging
33
The SOC distribution is different with smart charging and it is easily understandable with the SOC class
of different vehicles over a standard day
Figure 5-10 SOC Profile of vehicles with delayed smart charging
The SOC rises during night and peaks around 6 am and keeps falling throughout the day and there is no
sharp rise after the evening rush hour period which signifies that the charging period has been shifted to
night time or off-peak period in the absence of grid controls. One interesting aspect of the charging
enabled at work profile is that the SOC peaks just when the evening rush hour starts. This is in accordance
with the algorithm which makes all the cars charge aggressively around the time of departure.
The charging power profile can be deduced from the SOC pattern as peaking at night followed by
reducing throughout the day for home only charging. With charging enabled at work as well, the overall
power profile (EV demand along with normal demand) will have a flat shape.
Figure 5-11 Delayed Smart Charging power Profile with/without work charging
34
The profile follows the pattern established by the algorithm. The peak of charging has shifted to late night
as well as peak has increased to 2.3 times the mean value compared to 1.8 for home only charging. Since
the peak demand occurs at night, this is not considered as a drawback of this algorithm. The minor peak
in morning occurs when a lot of cars are just about to leave. Since all such cars irrespective of their SOC
would charge, the peak occurs. This charging power profile is actually without grid control and this
profile would change with grid control.
This profile compares well with the results shown by the Ecotality study (Ecotality, 2012) done based on actual EV
charging profile.
Figure 5-12 EV charging power profile from Ecotality
In the figure we can already see that people are scheduling the charging for late night to take advantage of
the evening peak hours. This confirms our assumption to keep the full charge time at 4 hours before
departure.
The effect of changing green range has the effect of increasing the evening EV load slightly higher, while
the night charging is shifted towards night-time (12 am - 5 am) as shown. The figure below shows the LV
EV power profile when the green range is changed from 60 to 100 percent of the range of the fleet with
charging enabled only at home. There is an increase in the night peak, and the charging power in evening
increases as the green percentage is increased, which is to be expected since a higher green percentage
means the cars are going to be charged at small rate always. The average charging energy share is higher at
night (12 am – 5 am) when the green range is lower.
Figure 5-13 Effect of Green Range on Charging Power
35
We can conclude this section by showing the weekly profile for SOC distribution and charging power
profile when charging is enabled at home and work for the smart charging scenario. This can be used to
compare the effectiveness of grid control along with delayed charging.
Figure 5-14 SOC Profile Over a week with Delayed Smart Charging
Figure 5-15 Charging Power Profile for a week with Delayed Smart Charging
36
6
Quick Charge Replenishment
When the vehicles in the fleet do not have enough energy left in the battery or if the range is not sufficient
to cover a journey, the vehicles are quick charged to complete the trip. There are two scenarios considered
in the model. One option is to utilize DC fast chargers to charge the vehicle to 80 percent of its range
using 50kW DC outlet. The second option is to swap out the depleted battery at battery swap stations and
replace that with a fully charged battery. But both the options require significant capital costs to realize –
High Power Chargers and consequent grid reinforcement for Fast Charging and extra batteries in the
network for battery switching. There are advantages and disadvantages to both the approaches. The slow
charging power profiles will not give a complete picture of the potential impact of EVs on the electric
utility. In this section some comparative results are shown where in some of the important outcomes of
the charging strategy has on the share of energy coming from fast charge replenishment. Since the Quick
charging methods has significant additional requirements Quick Charging methods should be regarded as
a rare method of charging for users and therefore can be priced accordingly to regulate its use.
DC Fast Charging
Single Ownership of Battery
High wear and tear of battery due to high charge
rates
Only 80 percent capacity achieved
Charging time of 20 minutes or more depending on
range
No requirement to track individual batteries as
owned by independent users
Peak demand on utility is exacerbated as the peak
arrival rate at fast charger coincide with peak
activity
Battery Switching
Battery is owned by a third party
Battery can be charged in a controlled manner
100 percent range is replenished
Replacement time of 5-10 minutes
Need accurate battery models to predict battery
state of health.
Grid impacts can be greatly reduced and the
charging time can be adjusted to provide grid
services. Eg: regulation, utilizing more renewable
energy etc.
Table 6-1 Comparison of Fast Charging vs Battery Switching
Since both approaches are expected to be present in the future electric grid, both the methods are analyzed
to estimate their grid impacts as well as the economics of both options. The fast charge replenishment
stations (either battery swap or fast chargers) are assumed to be available within a range of 10 km (the
range anxiety limit).
6.1 DC Fast Charging
Simultaneous fast charging of a significant number of EVs, directly from the grid, will impact on the grid
and local distribution particularly at the peak generation period. Fast charging stations used in this manner
would need to be planned to reduce any grid impacts, and located in areas where distribution networks
can cope or are able to be reinforced.
An alternative is to provide local energy storage at the charging station. These could be trickle charged
from the grid at times of low grid utilization, and provide high energy transfer rates direct from the local
storage. The capital cost of the charge stations is likely to be higher using this technique, although this
could be balanced by the reduced need for grid reinforcement.
The DC fast charging is executed as shown in the algorithm below. The vehicle range is checked if it has
enough energy left to complete the trip and an additional range to avoid range anxiety. If the range of the
37
vehicle is insufficient, then the vehicle is scheduled for fast charging. In the fast charging sub routine the
number of times fast charging is required to cover the complete trip is calculated. The sub routine can
calculate the number of vehicles requiring fast charging during the year with per minute resolution.
Figure 6-1 Fast Charging Process
One of the main results from the fast charging is that with charging enabled at home and work, the
percent energy to come from fast chargers decrease. This is expected, but the amount of decrease shows
that from a grid operators perspective the drop in fast charge demand substantiates the investment
required. The somewhat different values for fast charge value from what is listed in the literature can be
explained as follows. One of the reasons for this observation is that the simulation is run based on the
actual trip distribution. Hence all the vehicles do not follow a set trip pattern, in number and distance of
trips. As a result there will be vehicles which do more than one trips quite close to each other. This
scenario is quite realistic as there will be vehicles doing multiple trips while there are vehicles that are
away for vacation or long drives. Since there is only a possibility to enable work charging and home
charging without significantly increasing the cost of EV infrastructure deployment, this share from fast
charging would be necessary to cover the fleet travel requirements. This result also points to the fact that
fast charge replenishment method has to be chosen carefully as only 1 single car doing fast charging can
cause a power demand of 54.34 kW on the grid, even though for only a few minutes. The peak number of
cars at fast charging station as well as potential loading on the medium voltage grid is shown here.
38
6.1.1
Fast Charge Power Profile and Share of Energy from Fast Charge for Uncontrolled
charging with a range of 130 km and charging enabled only at home
Figure 6-2
Figure 6-3
39
6.1.2
Fast Charge Power Profile and Share of Energy from Fast Charge for Uncontrolled
charging with a range of 130 km and charging enabled at home and work
Figure 6-4
Figure 6-5
40
There is a 7 percent reduction in the amount of power transferred via fast charging as well the peak of fast
charging power decreased to less than 50 percent of the peak when charging is enabled at home and work.
Both have significant advantages to the amount of fast chargers required at the charging stations as well
as the peak demand on the power system.
6.1.3
Fast Charge Power Profile and Share of Energy from Fast Charge for Uncontrolled
charging with a range of 170 km and charging enabled at home only
Figure 6-6
Figure 6-7
41
Increasing the range of the EV fleet has a significant impact on the fast charge requirements as well as on
the peak of fast charge demand. For comparison the impact of increasing range and enabling charging at
work is also provided.
6.1.4
Fast Charge Power Profile and Share of Energy from Fast Charge for Uncontrolled
charging with a range of 170 km and charging enabled at home and work
Figure 6-8
Figure 6-9
The peak power demand has reduced substantially and so has the amount of energy required from fast
charging. This is quite advantageous for the user experience as well as for the state of health of the
battery.
42
6.1.5
Variation of Energy Share from Fast Charging
Figure 6-10 Effect of Providing Work Charging on Fast Charge Requirements
This figure shows the dependence of energy supplied by fast charging on the charging possibility at work
as well as the range of the EV fleet for uncontrolled charging scenario. The higher EV range and
possibility to charge at work drastically reduces the demand of fast charging.
6.2 Battery Swap
The battery swap mode of fast charge replenishment works on the basis of switching out the depleted
battery and replacing the same with a full battery. The process involves driving into a battery switching
bay and an automated process will position the vehicle, switch out the current battery and replace it with a
fully charged battery. The depleted batteries are charged in the station for later deployment.
Figure 6-11 Battery Switching Station
43
Switching Bay shown at Yokohama is shown in the picture above (Schwartz, 2010). It shows the battery
swapping lane and the new battery to be switched into the car.
This approach has seemingly quite some advantages over fast charge technology but to get market
acceptance is a challenge. The main reasons are as follows:
1. Requirement of Standardized Battery Interface across multiple car manufacturers.
• This would mean that car manufacturers who would like to differentiate based on
innovative battery technology will have to share that advantage.
• The battery pack for an average passenger car will weigh 250 to 300kg. To provide good
weight distribution and thus safe handling of the car, the battery pack could be
specifically designed for that vehicle and therefore integrated into the structure. This will
make standardization difficult.
2. Consumer acceptance of not owning a battery and having to change the vehicle battery.
• The consumer would be wary of switching to a battery which he/she is unsure of with
respect to its state of health.
3. The battery state of health estimation
• To monetize the battery usage, there should be a fool proof way to estimate the batteries
state of health to check for its usage pattern.
4. Safety Perspective
• The electrical connection between the battery and the vehicle carries a very high current,
and it is this connection that would need to be made and broken each time the battery is
exchanged. At best, it will cause wear and degradation at the key link between the two
components, at worst; it has the potential to cause a massive discharge, with all the
consequences that might ensue.
Despite all of the above stated concerns there are very attractive options for the EV users as well as for
the future utility grids with the deployment of battery swap stations. By analyzing the EV usage pattern it
is estimated that with 35 percent extra batteries compared to number of cars the fast charge replenishment
requirement of a hypothetical EV fleet in Netherlands can be satisfied. The algorithms for battery
switching as well as swap station power management are shown below.
44
6.2.1
Switching Process Algorithm
Figure 6-12 Battery Switching Process
The algorithm can be summarized as follows. If the vehicle needs battery switching to complete the
current trip, it is calculated the number of times it would be required, based on the trip distance and the
current range of the vehicle. The speed of the car will give the time at which the vehicle would reach
switching stations.
45
6.2.2
Switching Station Power Management
Figure 6-13 Battery Switching Station Power Management
In the model it is assumed that the charging process in the swap station is scheduled to occur at off-peak
hours of the normal demand curve. This also highlights one of the strong points of Swap Stations- being
able to provide ancillary grid services. Battery swap on the other hand increases the percentage of energy
coming from fast charge replenishment to increase to 21 percent. But this is not as bad for waiting times
as it would have been for the fast charging case since the battery switching still requires the same amount
of time. Since more energy is being supplied in a centrally controlled manner, it would mean more usage
of renewable energy and shifting of the EV load pattern to the benefit of the grid.
46
6.2.3
Swap Station Power Profile and Share of Energy from Fast Charge for Uncontrolled
charging with a range of 130 km and charging enabled at home only
Figure 6-14
Figure 6-15
47
6.2.4
Swap Station Power Profile and Share of Energy from Fast Charge for Uncontrolled
charging with a range of 130 km and charging enabled at home and work
Figure 6-16
Figure 6-17
The share of energy coming from fast charge replenishment has increased, but that is not very detrimental
to the waiting time at swap stations. The main reason for it is that every time you go for battery swap you
get a fully charged battery, since the major cost at swap stations would be for the number of times you are
48
switching and not really the cost of energy (assuming charging the batteries at 3.3kW Level 1 Charging).
The other major result from the battery swap power profile is the lack of swap station power between
6am-10am and 6pm-10pm. This has been explicitly disabled in the model with no ill effects on the battery
availability. Availability of charging at work reduces the share of energy from swap stations and also
reduces the peak demand from the swap stations. In that aspect the effect of introducing charging at work
is very similar to that on Fast Charging as the mode of fast charge replenishment.
6.2.5
Swap Station Power Profile and Share of Energy from Fast Charge for Uncontrolled
charging with a range of 170 km and charging enabled at home only
Figure 6-18
Figure 6-19
49
6.2.6
Swap Station Power Profile and Share of Energy from Fast Charge for Uncontrolled
charging with a range of 170 km and charging enabled at home and work
Figure 6-20
Figure 6-21
The effect of range increase on the share of energy from swap stations only kicks in once charging is also
enabled at work place. The power profile follows the pattern of the fast charge power profile as the range
of the EV fleet and the charging locations are altered.
50
6.2.7 Extra Batteries Required For Battery Switching
In the next figure an analysis on the amount of extra batteries required to support an EV fleet is estimated.
For the simulation a fleet of 100 vehicles are assumed and 35 percent extra batteries are assumed to be
available at swap stations when the simulation starts. The required batteries are plotted below.
Figure 6-22 Estimated battery requirements with 3.6 kW charging
The number of batteries would be increased if at any point in simulation the QCR demand cannot be
satisfied. In the simulation, the total number of batteries after simulation is finished remains constant at
35. Which means 35 percent extra batteries is an over design. The required number of batteries shows a
decreasing trend as the range of fleet increase. This is expected as the requirement for QCR decreases.
The trend implies that the capital cost for battery swap stations reduce with increasing range of the fleet
and the number of extra batteries required would also reduce as the fleet range increase. The amount of
batteries required would be less than 35 percent in all cases. For example in the case of fleet with a range
of 140 km, the exact required number of extra batteries is only 14%. So we can safely limit the number of
batteries to 20% (with a fifty percent safety margin).
6.2.7.1 Extra Batteries Required with Increased Swap Station power level
The number of batteries required and their availability will change with the charging power at the
switching station. We can see that the minimum number of batteries required also depends strongly on the
travel patterns. This will be shown with higher resolution data for the ‘50 km or more category’ in section
6.4.
51
Figure 6-23 Estimated battery requirements with 10.8 kW charging
Extra batteries required reduces with 10.8 kW (3 phase charging) at switching station for lower vehicle
ranges. But at higher ranges there is not any difference as the peak travel patterns determines the extra
number of batteries required.
Figure 6-24 Estimated battery requirements with 50 kW fast charging
Extra batteries required reduce with 50 kW output fast charging units at switching station for all vehicle
ranges.
The fluctuations are again due to the random trip selection criteria, but the overall picture should be seen
from the results as the analysis is built on a predicted trip profile and not an actual one. Hence the focus is
placed on the overall trends and not the exact values. One of the assumptions in this model is that
batteries are only switched in once they are fully charged. This might not be true since the current better
place pricing structure provides unlimited battery switching and better place pays for the electricity usage
52
for charging at home (Thomas, 2012). In such a scenario the users will be satisfied with not full batteries
as long as the pricing comes down for the package.
6.3 Comparison of the Quick Charge Replenishment Strategies
In the following section some key figures about the fast charge replenishment options – DC Fast
Charging and Battery Switching are considered. The main figures disclosed are about the number of visits
to a either of the facility, the average time connected to the system and also for the battery switch stations
the amount of extra batteries to be put into the system per electric vehicle. All the results are for
uncontrolled charging with charging enabled at home and at work.
The following figure shows the number of visits per year to a fast charging station and the average time
spent at the fast charge station. These numbers affect the decision to plan number of charging points per
station as well as estimation of maximal load due to fast charging. There is a very strong relation with the
number of visits to fast charge station to the range of the EV fleet. If the range is too small, the number of
charging points required would be too high to reduce the waiting times at charging stations. The average
connected to the system does not have a huge impact based on the range.
Figure 6-25 Number of visits and Time per visit at fast charge stations
The same figure for swap stations is provided below. The connection time at a swap station is 5 minutes
irrespective of the energy transferred. The number of visits decreases as the range of the fleet increases.
The sudden fall in number of visits at 125 km fleet range can be explained by the distance profile used in
the model. The distance of the majority of trips modeled in the traffic model in the category of commute
falls below 125 km. Since this increases the number of trips that can be completed on a single charge the
53
number of visits decreases and also the number of fuller batteries increases as the graph above shows the
case where charging is enabled at home and at work.
Figure 6-26 Number of visits at battery swap stations
The sudden drop at 125 km is applicable for battery swap mode too. This is just due to the discrete
numbers used for distance distribution and since that is the only data available from the MON survey, the
model is kept as it is. The trend is of more interest to the modeling exercise.
54
6.3.1
Energy Contribution from Quick Charge Replenishment
Figure 6-27 Energy Share of EV usage from Quick Charge Replenishment
The percentage energy supplied by the fast charge replenishment to the total energy requirement of the
fleet is shown here. The scenario shown here is when charging is enabled at home and at work. There is a
strong dependency on the range of the fleet. The percentage energy supplied by fast charge is always
lower compared to that by swap station due to the difference in the charging behavior. In the case of fast
charging the cars that have a provision to charge at the destination will only top up their range so that they
can complete the trip. The reason for this behavior is that the price of charging at fast charge station
would be higher as well as it is detrimental to battery health. Another factor would be the extra time
required to be spent waiting for the car to charge up. In the case of battery swap since the charge
replenishment time is irrespective of the amount of energy transferred, the vehicles get a full battery
always.
6.4 Quick Charge Replenishment Analysis with Higher Resolution Travel Profile
The trip profile has a limitation that the 50 km or more category does not provide enough resolution and
hence as the range of the vehicles are increased almost all of the trips modeled can be covered within the
range of the vehicle. This introduces the sudden dip in quick charge facility requirements and it would be
interesting to increase the resolution of this category purely to get a better understanding of the
requirement on quick charge infrastructure.
6.4.1 Trip Profile Resolution Increase in ‘50km or more’ Category
A study on the MON 2008 was done by (Lampropoulos, Vanalme, & Kling, 2010). In the work the raw
database of mobility survey was analyzed and a higher resolution of the above mentioned category is
55
available. The study shows the percentage split of the 41.65 percent of the trips in the “50 km or more”
category for 2008 data. Since the mobility data has shown a saturation tendency since 2001 we can
assume the same holds for 2009 data as well.
Percentage of Distance
Travelled/Number of Trips above 50 km
Percentage Distance/trips
14
12
Distance Travelled Percentage
10
Number of Trips Percentage
8
6
4
2
0
50-75
75-100
100-125
125-150
150-175
175-200
200 or
more
Distance Category
Figure 6-28 Higher resolution data for trips with distance more than 50 km
The percentage number of trips is obtained by making sure the average distance travelled per day data as
well as the distance in each category gives acceptable values. For example if the percentage number of
trip in the category “50-75” is more the average trip distance in that category would be less than 50 km.
So we have a band of very close values to choose from for the percentage number of trips and it has been
finalized after simulating and observing the results.
The average distance for a trip in each distance category is obtained as described in chapter 3.
56
Average Trip Distance in Distance Category
Average Trip Distance in km
250.000
200.000
150.000
100.000
50.000
0.000
50-75
75-100
100-125
125-150
150-175
175-200
200 or more
Distance Category
Figure 6-29 Distance modeled for trips of distance greater than 50 km
This data is combined with mobility data to improve the resolution of the higher distance category and the
following results are obtained for Quick Charge Replenishment requirements.
6.4.2 Comparison of Quick Charge Replenishment Requirements with new Trip Profile
The following figure shows the average number of visits and time per visit for fast charge mode of quick
charge replenishment. There is no sudden jump in the pattern as was observed without the higher
resolution profile.
Figure 6-30 Number of visits and Time per visit at fast charge stations with higher resolution data
The requirement for battery switching infrastructure is shown below. The average time spend at the
switching station is again a constant value of 5 minutes.
57
Figure 6-31 Number of visits at Battery Swap stations with higher resolution data
The extra number of batteries required also increases since the distance of trips at the higher distance
range has increased with the improved resolution data.
Figure 6-32 Estimated battery requirements with 10.8 kW charging with higher resolution data
We can calculate the battery capacity in each car as well as the effective battery capacity per car,
considering the extra batteries in the switching stations. It is shown here in the graph below. The battery
58
capacity in each car is calculated using the range of the car and the energy consumption per kilometer
from battery assuming a battery cycling of 70 percent (only 70 percent of the capacity is used to keep the
lifetime of batteries high).
𝐶𝑎𝑝𝑎𝑐𝑖𝑡𝑦 =
Extra Battery Percentage/Battery Capacity vs
Range of Fleet
70
Extra Battery Percentage/ Battery Capacity in kWh
𝑅𝑎𝑛𝑔𝑒 ∗ 𝐸𝑛𝑒𝑟𝑔𝑦 𝑅𝑒𝑞 𝑝𝑒𝑟 𝑘𝑚 𝑓𝑟𝑜𝑚 𝐵𝑎𝑡𝑡𝑒𝑟𝑦
𝐵𝑎𝑡𝑡𝑒𝑟𝑦 𝐶𝑦𝑐𝑙𝑖𝑛𝑔
Extra Batteries Required per Car in Percentage
Bat capacity in each car
60
Bat capacity required per car
50
40
30
20
10
0
80
85
90
95 100 105 110 115 120 125 130 135 140 145 150 155 160 165 170
Range of Fleet in km
Figure 6-33 Battery Capacity Required to support battery switching
We can see that capacity required per car and for the fleet increases as the range of the fleet increases as
expected. There are conflicting scenarios here as higher battery capacity means lower cycling and hence
less ageing effects while higher upfront capital investment. But the capital investment costs are also not
straightforward to calculate as up to 30% of the battery costs are due to packaging (Boston Consulting
Group, 2010).
59
The energy share from the different quick charge technologies also vary with the new trip profile and it
can be seen as follows.
Figure 6-34 Energy Share of EV usage from Quick Charge Replenishment with higher resolution data
6.5 QCR peak power and serving units comparison
The results from the simulation about QCR technologies are used in the work of (Vereczki, 2012) to
calculate the peak power requirements and number of serving units (Fast Chargers/Battery Switching
lanes). The key results are shown here. The detailed description is provided in Appendix B
The following figure shows the number of serving units required assuming 100 QCR stations serving
1million EVs in Netherlands. The Fast Charging Absolute worst case assumes that all EVs once at a Fast
Charge station will charge to 80 percent, where as in the average peak case EVs only charge to cover the
trip as long as there is possibility to charge at the destination.
Figure 6-35 Serving units required per QCR station
We can see that fast charging requires 4 times as much charging points as switching lanes in the case of a
fleet of range 170 km. Fast charging always requires more number of serving units than switching lanes.
This is a direct result of the fact that the waiting time at fast charging station depends on the charge rate
where as in switching stations the time required is always the same.
60
Another result to come out of this analysis is the peak power requirement of such QCR stations.
Figure 6-36 Peak Power Requirement of QCR stations
The peak power requirements at fast charging stations are much higher than that of battery switching
stations and the fact that fast charging demand cannot be shifted in time means battery switching is more
grid friendly. As a result fast charging stations in the mould of conventional gas stations would need a
high power connection to medium voltage grid where as battery switching stations can manage their
charging power, provide grid services and can be connected to the low voltage grid.
One way to improve the waiting time involved for fast charging is to go for higher charging powers. It
directly reduces the charging time. Fast charging at 100 kW is already being considered (NU.nl, 2012).
The figure below shows the effect of increasing the fast charging power. One of the drawbacks of this
model is that, it assumes same charging efficiency for charging at all powers. If we consider that in the
model, then the effect of increasing charging power will be considerably lower.
Figure 6-37 Effect of higher charging power on number of fast charging units
The number of serving units reduces as well as the time spent fast charging. Hence there will be only
marginal increase in the peak load, but this option is only feasible with improvements in battery
technology to handle high power charging.
61
7
Micro-grid with Electric Vehicle Charging Loads
In this study a grid connected micro-grid is simulated along with the EV charging load. The EV load is
added to the demand curve of normal electricity usage curve and based on the renewable production and
grid capacity the EV charging profile is varied to aid the grid in maximizing the renewable energy
utilization. The smart charging scenario includes provision to scale the power usage whenever excess
renewable power is available on the grid. The micro-grid model used here is derived from the master
thesis work of (Kaas, 2011) at TU Delft. The basic model is explained below.
The energy in the hybrid power system is supplied by wind turbines (WG), photovoltaic solar panels
(PV) and the grid. Because the supply of the renewable energy sources (wind and sun) is not always equal
to the load/demand (LD), the excess supply is supplied to the following sinks in the same order – excess
EV charging capacity, Battery (if included), back to the grid (provided it does not exceed grid capacity),
power dump. A multi source hybrid inverter (MHI) stores excessive power in batteries (BAT) and takes
energy from the batteries when required. Batteries are removed from the simulation as EVs can provide
the balancing required making batteries redundant. Hence we can consider a standard grid following
inverter is used.
7.1 Topology of the Micro-Grid with EV Loads
During the thesis study of Leake the optimal topology was identified and it is represented in the figure
below (Leake, 2010). The reason for choosing this topology is that as few power conversion steps as
possible are required. This causes an overall efficiency gain, because each conversion step introduces
extra losses. The solar panels (PV, Photo Voltaics) and batteries (BAT), if included, operate on direct
current (DC) and are therefore connected to a DC-bus. This DC-bus has a voltage that is equal to the
momentary battery voltage, which therefore changes with the state of charge (SOC) of the battery. The
solar panels are connected via a maximum power point tracker (MPPT) to make sure that they operate on
their optimum power point. The MPPT is displayed in the figure as a DCDC converter, because it
converts the DC solar voltage in to the battery voltage. The wind turbines (WG) and loads (LD) and EV
loads (EV-LD) operate on alternating current and are thus connected to the AC-bus. A bi-directional
DCAC inverter connects the AC-bus and DC-bus. The AC bus has connection with the grid and it helps
in maintaining the power balance on the micro-grid.
62
Figure 7-1 Micro-grid Model with DC-Coupled Battery Storage
An Alternative option without MHI inverter which uses standard components would be an AC coupled
system as shown below.
Figure 7-2 Micro Grid Model with AC-Coupled Battery Storage
The system described above does not mean that all the units are combined together and installed as one
entity. For example the solar panels would be separately installed on each household and it is not feasible
to build a DC grid to bring together all the solar panels. Hence the system would be built up as a modular
system so that the micro-grid can be expanded at a later time. Another reason is to operate individual
units’ at the most optimum point. For example if the power demand on 3 separate DC/AC inverters is
63
low, it might be more efficient to use 2 modules running at higher powers instead of 3 running at low
power.
The individual components in the system are explained in detail in the work of Arne Kaas. The
components excluding the EV charging load is explained here shortly. The EV charging load and the
different charging strategies have been explained in chapter 2. The EV charging loads represents the slow
charging loads at home and work as well as the Quick charge replenishment charging loads (Battery
Switching Load/ DC Fast Charging Load).
7.2 Load
The load modelled is derived from the load profile provided by Ecofys. This load is assumed to be
average demand profile for the Netherlands and the EV load is added to it as a controllable entity when
determining power balance. Every home with an EV is assumed to possess a Slow charger of rated power
3.6 kW.
7.3 Inverters
There are two inverters considered in this system. The PV inverter is a grid following inverter. It is sized
to support the maximum power of the connected solar panel. It internally has a DC/DC converter which
does the MPP Tracking for the solar panels. The batteries are connected to the grid using a separate
inverter and that inverter can be a grid forming inverter if there are critical loads in the micro-grid that
should be available even during grid disturbances. The battery inverter stage inserts or removes energy
from the AC-bus as function of the required power on this side provided the battery can handle it. The
inverters in this system are modular and therefore it is possible to have multiple inverters to expand the
system.
7.4 Solar Panels
The solar panels in this system are connected in arrays. If more solar panels are installed, each inverter
has its own solar panel array. An average solar panel efficiency of 11% was used in the thesis. The solar
irradiation data used in this model has been corrected to get the irradiation pattern if the panels are
mounted at an angle (the optimal angle is equal to the latitude of the area).
7.5 Wind Turbines
Wind turbines/generators (WG) are connected to the AC-bus. The WG will supply its energy directly to
the households or it gets stored if the battery can handle it or send back to the grid. The MHI for battery
will take care of energy storage if the power supply of the WG is higher than the load.
7.6 Batteries
The batteries are connected to the DC-bus in between the DC/DC (MPPT) and DC/AC stage if an MHI is
used. In the AC coupled system the batteries would be connected to the battery inverter module. In a
micro-grid in an urban area where the PV panels are going to be geographically distributed an AC
coupled system is more feasible and hence it is assumed that central battery storage systems will be
installed.
64
7.7 EV Charging Load
The EV charging load consists of Slow charging power requirement of EVs connected at charging
enabled locations (in this study at home and at work) and the quick charge replenishment power
requirement (DC fast charging/ Battery Switching). The EV charging power is obtained as described in
chapter 5 and 6. The EV charging load is a variable load as it will have a component that should be met
compulsorily as well as possibility to charge extra. The two way power transfer is not considered in this
model.
7.8
Explanation of Renewable Energy Terminologies Used
7.8.1 Potential Energy Mix (PEM)
This term indicates how much renewable energy is installed in relation to the total demand. For example
if the average demand is 10 kWh per day, a PEM value of 70% indicates that the average power
production from the installed renewable energy systems will be equal to 70 % of the average demand = 7
kWh. Since capacity factors of renewable energy systems are low (typical values – 11% for Solar PV and
20-30 % for wind), the actual installed capacity will be higher.
7.8.2 Renewable Energy Ratio (RER)
RER indicates the split up of the renewable energy systems between Solar PV and Wind Power Systems.
A RER of 0% indicates all the renewable energy installed is wind power while a RER of 100% indicates
all the installed renewable energy is Solar PV based. As an example in the previous case the average
output of renewable energy systems in 7 kWh with a 70% PEM system. Now consider that the RER is
25%. It means 25% of the average renewable energy production is from Solar PV and 75% of the average
renewable energy production is from wind power. The average power production from Solar PV = 1.75
kWh and average power production from wind power = 5.25 kWh.
Figure 7-3 PEM and RER visualization
65
8
Simulation of Micro-Grid System and Sensitivity Analysis
The topology described in chapter 7 is used to model the actual behavior for a year of a micro-grid with
some on-site storage. The sensitivity analysis performed include changing the amount of renewable
energy installed (Potential Energy Mix – PEM), the ratio of renewable technologies (RER), the grid
exchange capacity, on-site storage capacity for the micro-grid. The vehicle fleet is simulated by varying
the range of the fleet from 120km to 170km in steps of 10km, the quick charge replenishment mode, and
the percentage of EV penetration in the area. Charging is enabled at home and work and the green range
is kept at 50 percent of the range. The simulation was run with 40% EV penetration and 80% EV
penetration. The 40% EV system has lesser direct green energy usage compared to 80% system, by 1-2
percentages.
The Green charging strategy is modified to maximize the green energy uptake capability of EVs as
follows.
Figure 8-1 Green Charging Behaviour
8.1 Normalization of Combined Load
The load of the EVs and the normal load is combined together and normalized to have a combined
average value of 1. The following equations describe the normalizing done for each of them.
A quantity named EV load to Normal Load Ratio is defined to be used for normalizing.
𝐸𝑉 𝑡𝑜 𝐿𝑜𝑎𝑑 𝑅𝑎𝑡𝑖𝑜 =
𝐴𝑣𝑔 𝐸𝑉 𝐿𝑜𝑎𝑑 𝑃𝑒𝑟 𝐷𝑎𝑦
7.7
=
= 0.7833
𝐴𝑣𝑔 𝐿𝑜𝑎𝑑 𝑃𝑒𝑟 𝐷𝑎𝑦
9.83
66
𝐿𝑜𝑎𝑑 𝑆𝑐𝑎𝑙𝑒𝑟 = 𝑀𝑒𝑎𝑛(𝐿𝑜𝑎𝑑) ∗ (1 + 𝐸𝑉 𝑃𝑒𝑟𝑐𝑒𝑛𝑡𝑎𝑔𝑒 ∗ 𝐸𝑉 𝑡𝑜 𝐿𝑜𝑎𝑑 𝑅𝑎𝑡𝑖𝑜)
𝐸𝑉 𝐿𝑜𝑎𝑑 𝑆𝑐𝑎𝑙𝑒𝑟 = 𝑀𝑒𝑎𝑛(𝐸𝑉 𝐿𝑜𝑎𝑑) ∗
𝐸𝑉 𝑃𝑒𝑟𝑐𝑒𝑛𝑡𝑎𝑔𝑒 ∗ 𝐸𝑉 𝑡𝑜 𝐿𝑜𝑎𝑑 𝑅𝑎𝑡𝑖𝑜
(1 + 𝐸𝑉 𝑃𝑒𝑟𝑐𝑒𝑛𝑡𝑎𝑔𝑒 ∗ 𝐸𝑉 𝑡𝑜 𝐿𝑜𝑎𝑑 𝑅𝑎𝑡𝑖𝑜)
Once these scalers is applied the combined load (EV load + Normal Load), it has an average value of 1
and is in the per unit system. The load is divided by the respective scalers to normalize them. The
variation of total load as the EV penetration increases is shown here in the table.
EV Penetration in Percentage of total fleet
0
20
40
60
80
100
Average Net Demand per day in kWh
9.83
11.37
12.91
14.45
15.99
17.53
Table 8-1 Demand Increase with EV Penetration
8.2 Simulation result with 80 percentage EV penetration and Battery Switching
Micro-grid was simulated with the above mentioned settings to conclude the actual loop settings to be
kept in the next simulation. The reason to do so is mainly due to the fact that the full simulation takes a
long time and a lot of settings do not have a huge impact on the energy flows. All the EVs are assumed to
have a green charging behaviour.
The results show that a fully solar system will lead to a lot of energy wastage. This will be true if the
entire EV load is considered as one single load as it is done in this model. But solar only systems will
work in perfect synergy with public charging stations and charging stations at locations other than home
since they are used during day time.
Figure 8-2 Renewable Energy Direct Consumption with Battery Switching
67
From the above figure, we can see that the amount of demand directly served by renewable production for
varying amount of installed renewable energy (PEM) and varying share of wind and solar energy in the
installed renewable energy system (RER).
The conclusion we can draw from the figure is that a wind only system matches the demand closely,
while a solar only system would require a good amount of grid support or on-site storage. For a renewable
energy distribution of 25 % solar and 75 % wind the results are very similar to that of a 100 percent wind
system. The above mentioned ratio is recommended since that would ease the installation of renewable
energy. Large scale wind parks and distributed solar generation is easier to achieve than the opposite.
The energy share change of battery swapping can be seen from the following plot. It shows that there is a
valley in the demand from battery swapping as the installed renewable energy is 120 percent of the
demand with a 25 % solar and 75 % wind system.
Figure 8-3 Energy share of Battery Switching
The extra EV charging algorithm is effective when there is sufficient renewable energy that matches the
travel profile. The plot below validates the proposed system further as we can see that the extra EV
charging maximizes with the proposed system.
68
Figure 8-4 EV Charging by Grid Control with Battery Switching
The requested EV charging power (the charging power EVs must have based on parking time and SOC
levels) is also the lowest with the configuration 25% solar 75 % wind and 120 % renewable energy
installed.
Figure 8-5 EV Charging Requested with Battery Switching
The result presented here shows that systems with more wind energy suit EV penetration. This has been
proposed in literature (DeForest, et al., 2009). They argue that wind energy is particularly suitable for
providing electric vehicle charging, as it tends to peak around dawn and dusk when vehicle recharging
will be most convenient and affordable. Hence this EV charging algorithm and the renewable energy mix
can be considered as the basis for EV infrastructure development. The SOC distribution of the fleet is
shown here.
69
Figure 8-6 SOC Profile of the fleet with micro grid simulation and battery switching
We can observe that the fleet SOC profile is maintained with the green charging strategy and what it
implies is that it can maintain the user experience while improving the renewable energy uptake and
shifting the energy usage to off-peak time.
8.3 Simulation result with 80 percentage EV penetration and Fast Charging
The results of renewable energy directly used are very similar for both the modes of quick charging. The
same configuration of 25 percent solar and 75 percent wind maximizes the renewable energy uptake.
Figure 8-7 Renewable Energy Direct Consumption with Fast Charging
70
The share of energy required from fast charging is shown in the figure below. The pattern is same as that
of the battery switching scenario. So the renewable energy configuration suits both the quick charge
replenishment modes.
Figure 8-8 Energy share of Fast Charging
The EV Charging requested is shown below. It is lower than the corresponding value with battery
switching at 25-75 system with 120 percent installed renewable energy.
Figure 8-9 EV Charging Requested with Fast Charging
71
The EV charging power by grid control shows that the amount of power from grid control is more with
fast charging. But, this does not show a key advantage for battery switching. The battery switching
stations have more ability to control the charging behaviour and hence the advantage is shifted from slow
charging to quick charge replenishment stations, while fast charging is an uncontrollable load.
Figure 8-10 EV Charging by Grid Control with Fast Charging
The maximum charge power on the low voltage network determines the network capacity to be planned
by the utility. The simulation shows there is a relation to the renewable energy split up on the peak
charging requirement that occurs on the low voltage side. This has an effect on the capacity of the
network to be built.
Figure 8-11 Peak EV charging power on Low Voltage Network
72
There are variations as the PEM varies while there is a clear increasing trend as the RER changes. The
figure shows that for any amount of installed renewable energy, the 25% solar 75% wind system results in
the lowest peak demand on the low voltage side. The fluctuations among PEMs do not convey any data as
that is due to simulation variations.
For the greenest system – PEM 120 and RER 25, the amount of green energy used directly increases as
the green charging behaviour of the fleet increase. This proves the effectiveness of the green charging
algorithm.
Figure 8-12 Effect on direct renewable energy usage with green charge behaviour
8.4 Energy Management at Switching Stations
Battery switching despite being a disruptive technology has gained a lot of attention and with a smart
energy management; the excess batteries could be used as a valuable resource. The grid services that can
be supplied include load shaping, regulation and energy reserves. Of this we would look at load shaping
as it requires an algorithm that will optimize between the charging power demand and service capability
of switching stations.
A change to the charging algorithm is proposed where the serviceability is retained while maximizing
renewable energy uptake for charging the batteries. This energy management system could improve the
renewable energy usage for quick charge replenishment instead of being a dumb load on the system. The
charging estimation algorithm is described below.
73
Figure 8-13 Optimized Battery Charge Management at Switching Stations
Please note that battery swapping accounts for only a fraction of the EV charging load and EV charging
load is less than 50 percent of the combined electrical load. The results of the optimizer should be viewed
in this light.
The results show that the renewable energy uptake increases with the new strategy for charging stations.
This is a significant improvement considering the fact the energy share from switching station is small
compared to that of the overall electrical demand.
74
Figure 8-14 Renewable Energy Direct Usage with Optimized Battery Switching
The values of renewable energy direct usage is provided in tabular form for comparison.
Renewable Energy Directly Used without optimization at swap stations
PEM/RER
70
80
90
100
110
120
0Sun – 100 Wind
25 Sun – 75 Wind
50 Sun – 50 Wind
75 Sun – 25 Wind
100 Sun – 0 Wind
68.35
69.14
63.34
54.74
45.23
75.77
76.20
68.64
58.26
47.32
81.72
81.38
73.08
61.57
48.98
86.43
85.40
77.38
64.38
50.59
89.47
88.42
80.68
67.07
51.98
91.84
90.71
83.80
69.90
53.14
Table 8-2 Renewable Energy Direct Usage with optimized charging at swap stations
Renewable Energy Directly Used with optimization at swap stations
PEM/RER
70
80
90
100
110
120
0Sun – 100 Wind
25 Sun – 75 Wind
50 Sun – 50 Wind
75 Sun – 25 Wind
100 Sun – 0 Wind
68.50
69.17
63.60
55.05
45.90
75.90
76.24
68.75
58.73
48.02
82.15
81.75
73.61
62.19
49.64
86.97
85.97
77.94
65.06
51.15
90.08
89.11
81.36
67.76
52.65
93.15
91.81
84.67
70.34
53.78
Table 8-3 Renewable Energy Direct Usage without optimized charging at swap stations
The other parameters remain roughly same as before and they are shown here for comparison.
75
Figure 8-15 Energy share of Battery Switching with optimized battery charging
Figure 8-16 EV Charging by Grid Control with Optimized Battery Switching
76
Figure 8-17 EV Charging Requested with Optimized Battery Switching
We can conclude that battery switching is a viable alternative to fast charging considering the benefits to
the grid. The economic aspects of the same needs to be worked out to implement battery swapping in the
mobility network of the future.
With the optimized battery charging method, the effect of penetration of electric vehicles in the vehicle
fleet on direct usage of renewable energy can be shown here. The system under consideration is of the
configuration PEM 120 (120 percent of installed renewable energy compared to average demand) and 75
% wind and 25 % solar system.
Figure 8-18 Effect of EV penetration on direct Renewable energy usage in wind majority system
77
There is a 4 percent increase in direct consumption of renewable energy produced as the electric vehicle
penetration increases from 20 percent of the fleet to 100 percent of the fleet. This result shows that a PEM
120 system with 25% solar and 75% wind is already suited for the current demand, while increasing the
EV penetration can enhance the direct consumption of renewable energy generation. It should be noted
that, the net demand increases as the electric vehicle penetration increases and despite that the direct
usage of renewable energy increases.
The positive impact of electric vehicle introduction is more pronounced in systems with higher solar
penetration. In the following figure a PEM 120 system with 75% solar and 25% wind and the effect of EV
penetration is showed. The renewable energy direct consumption improves by 7.4% as the electric vehicle
penetration increases from 20 percent of the fleet to 100 percent.
Figure 8-19 Effect of EV penetration on direct Renewable energy usage in solar majority system
8.5 Effect of Fleet Range on Micro-grid Energy Usage
The micro-grid was simulated with different ranges for the fleet to see the effects on the various
parameters described in the other simulations. All the results are for systems with 25% installed solar
energy and 75% wind energy with battery switching as the mode of Quick charge replenishment.
The figure below shows the effect of the range of fleet as well as PEM on direct consumption of
renewable energy in the micro-grid. We can see that as the range changes for the same PEM, the
improvement in direct green energy usage is not substantial. There is also a saturation of direct renewable
energy usage as the PEM increases beyond 100%.
78
Figure 8-20 Renewable Energy Direct Usage in relation to range of the fleet
The energy share of battery switching in the total energy usage changes with the range of the fleet as well
as with PEM ratio. This result follows from the results of 6.3.1 on energy share from quick charge
replenishment.
Figure 8-21 Energy Share of Battery Switching Stations in relation to range of fleet
For the same PEM ratio, the energy share of battery switching reduces with increasing range of fleet. The
energy share of battery switching reduces as the PEM increases for the same range of fleet.
This is due to the fact that EVs are being charged by grid control in the presence of excess renewable
energy.There is a steady increase in extra EV charging by grid control as the PEM increases.
79
Figure 8-22 EV charging by grid control in relation to range of fleet
The effect of PEM on EV charging by grid control is shown here for the fleet of range 80 km. There is a
clear increase in the EV charging by grid control while the total energy share from slow charging does not
change as it depends strongly on the range of the fleet.
Figure 8-23 Effect of PEM on EV Charging by Grid Control
80
9
Conclusion and Discussion
The transition of mobility from fossil fuels to an electrical energy based system needs major overhaul of
the electricity network. It presents a unique opportunity to integrate and utilize the synergy of renewable
energy production with electric vehicles. From this thesis study some of the conclusions that can be
drawn are as follows.
The mobility pattern follows a pattern which has similarity to the demand on the electric grid. This further
illustrates that if the electric vehicle charging is not properly managed the evening peak demand on the
grid will be exacerbated by the electric vehicle introduction. This is illustrated from the results of dumb
charging scenario.
There is a minimum level of energy requirement on quick charge replenishment which can be established
with dumb charging as the mode of charging. The aim of smart/green charging is to make sure the
requirement on QCR will be as close to the requirements with dumb charging. Otherwise the problem is
shifted from low voltage network to QCR stations which would be limited by the connection capacity.
The range of the vehicle has a strong impact on QCR requirements which is an expected result as higher
range indirectly means more energy buffering in the vehicles and the demand is shifted.
Electric vehicles can improve the direct renewable energy usage if the EV charging is following the
available renewable energy on the grid. If the installed renewable energy is higher, the demand on QCR is
markedly lower. Hence electric vehicles and renewable energy sources can facilitate the deployment of
each other. The usage profile of EVs is more suited for the profile of wind energy and hence from an
energy flow perspective a wind majority system matches the EV demand better than a solar majority
system.
Some of the key results established with this study are listed here.
-
-
Vehicles will be idle for a large portion of the day (usage a day on average is 2 hours)
Vehicles are present at home or at work the maximum amount of time
Providing charging infrastructure at home as well as at work is very important to improve the
utility of the EVs, to reduce the peak load and also to reduce Quick Charge Requirements.
Uncontrolled charging certainly worsens the peak demand and hence a combination of delayed
charging with renewable energy production based grid controls are required along with
introduction of EVs.
Battery switching with smart charging is a viable alternative to fast charging.
Wind energy suits the electric mobility profile and hence would result in higher direct
consumption of renewable energy.
Increasing the EV penetration in vehicle fleet increases the direct consumption of renewable
energy despite increase in overall load.
Increasing the range of the fleet and the installed renewable energy reduces the energy demand
from quick charge replenishment.
The results on QCR requirements are used to estimate the infrastructure requirements and battery
switching results in lower peak requirements as well as lesser number of switching lanes at
additional number of batteries required as the downside.
81
-
Increasing the fast charge rate decreases the time spent for charging and hence reduces the
number of serving units required without affecting the peak power requirements.
9.1 Suggested Future Work
The mobility pattern is derived based on MON survey and it yields travel profile that is quite
representative to actual usage. For a comprehensive analysis it is necessary to log the travel pattern of a
representative population as they use the vehicles normally and model the charging load. The resolution
of MON survey is one hour and that should improve to obtain finer per minute grid requirements. For
estimating usage of EVs for V2G regulation this is very important. The predicted EV charging loads are
based on the assumption that the mobility pattern does not change with the introduction of electric
vehicles. This is not always true as the introduction of disruptive technology changes the behaviour of
users as is evident from the smart phone induced behaviour changes.
One of the major drawbacks of the modelling with localized generation and demand profile is that there is
no possibility to use renewable generation shifted in space. The concept of “desertec” is based on the fact
that the peak solar production is North Africa can be used by transporting it to industry centres in Europe
to match the peak demand. For a sustainable electric supply we cannot ignore this effect and hence the
result of solar faring poorly compared to wind will not hold true if there is a transmission infrastructure in
place to transfer power over long distances.
Effects of increasing weight of the car with increasing range are not modelled in the current simulation.
The fleet is always modelled as uniform vehicles. The provision is provided in the model to define
various subclasses of vehicles. With exact data this can be modelled. The charging efficiency variation
with increasing rate of charging is not modelled as well. These are easier to implement with exact data for
the mobility profile.
The battery switching mode of quick charging can be studied as a topic in itself with its many conflicting
aspects. The optimization of capital costs with running costs and battery ageing analysis for estimating the
correct pricing structure and the exact required batteries at each switching stations based on local traffic
conditions are topics to be studied in detail. A study can be done to locate the switching stations
geographically based on the layout of the highway network.
82
10 Bibliography
[1] Ashtari, A., Shahidinejad, S., & Molinski, T. (n.d.). PEV Charging Profile Prediction and Analysis
Based on Vehicle Usage Data. IEEE Transactions on Smart Grids .
[2] BetterPlace. (2012). Betterplace. Retrieved March 2012, from http://www.betterplace.com/thesolution-switch-stations
[3] Boston Consulting Group. (2010). Batteries for Electric Cars: Challenges,Opportunities and the
Outlook to 2020.
[4] Brooker, A., Thornton, M., & Rugh, J. (2010). Technology Improvement Pathways to Cost-Effective
Vehicle Electrification. SAE 2010 World Congress. Detroit, Michigan.
[5] Chopra, S., & Bauer, P. (2011). Driving Range Extension of EV with On-road Contactless Power
Transfer - A Case Study. IEEE Transactions on Industrial Electronics .
[6] DeForest, N., Funk, J., Lorimer, A., Ur, B., Sidhu, I., Kaminsky, P., et al. (2009). Impact of
Widespread Electric Vehicle Adoption on the Electrical Utility Business – Threats and Opportunities.
[7] Department for Business Enterprise and Regulatory Reform & Department of Transport UK. (2008).
Investigation into the Scope for the Transport Sector to Switch to Electric Vehicles and Plug-In
Hybrid Vehicles.
[8] Dorresteijn, S. (2011, November). EV Charging Infrastructure DC Fast Charge Solutions. Retrieved
April 2012, from ABB:
http://www02.abb.com/global/huabb/huabb008.nsf/0/b6f99174f9374993c125795e003c0f2a/$file/AB
B-EVCI+Company+Presentation+2011-11-24.pdf
[9] Duvall, M., & Knipping, E. (. (2007). Environmental Assessment of Plug-In Hybrid Electric Vehicles.
Volume 1: Nationwide Greenhouse Gas Emissions.
[10]
Ecotality. (2012). The EV Project Q4 Report 2011.
[11]
Farmer, C. e. (2010). Modeling the Impact of Increasing PHEV Loads on the Distribution
Infrastructure. 43rd Hawaii International Conference on System Sciences (HICSS).
[12]
IEA. (2009). World Energy Outlook 2009, International Energy Agency (OECD/IEA).
[13]
Kaas, A. (2011). Sustainable off-grid power for rural areas. Delft: TU Delft.
[14]
Lampropoulos, I., Vanalme, G., & Kling, W. (2010). A methodology for modelling the behaviour
of electricity prosumers within the smart grid. IEEE PES Conference on Innovative Smart Grid
Technologies Europe. Gothenburg.
[15]
Leake, E. (2010). Genset-solar-wind hybrid power system of off-grid power station for rural
applications. Delft: TU Delft.
[16]
MacKay, D. J. (2009). Sustainable Energy :Without the Hot Air. UIT Cambridge,UK.
83
[17]
NU.nl. (2012, May). New fast charging standard. Retrieved May 29, 2012, from NU:
http://www.nu.nl/auto/2802264/nieuwe-snellaadstandaard-ontwikkeld.html
[18]
Schwartz, A. (2010, October). Fast Company. Retrieved March 2012, from Robotic BatterySwitching Electric Taxi Program Coming to Bay Area: http://www.fastcompany.com/1698338/betterplace-bringing-switchable-battery-electric-taxi-program-to-the-bay-area
[19]
Senan, M. (2009, June). Fp7ireland. Retrieved March 2012, from
http://www.fp7ireland.com/cms/Documents/Electrification%20of%20Road%20Vehicles_947.pdf
[20]
Siemens. (n.d.). Retrieved March 2012, from Energy Siemens:
http://www.energy.siemens.com/us/pool/us/powertransmission/transformers/downloads/TRANSFORMERS-PowerTransfrs10to100MVA-1.pdf
[21]
The Ural Company. (2006). OIL-COOLED POWER TRANSFORMERS. Retrieved May 2012,
from The Ural Company for Transformer Technologies.
[22]
Thomas, B. (2012, June 13). What's The Real Cost of a Better Place Electric Car in Israel?
Retrieved June 15, 2012, from Green car reports:
http://www.greencarreports.com/news/1076914_whats-the-real-cost-of-a-better-place-electric-car-inisrael
[23]
TNO - Science and Industry,RWTH - Institute for High Voltage Technology,ECN - Policy
Studies. (2010). Opportunities for the Use of Renewable Energy in Road Transport Road Transport.
[24]
Tuffner, F., & Kintner-Meyer, M. (2011). Using Electric Vehicles to Meet Balancing
Requirements Associated with Wind Power.
[25]
Vereczki, G. (2012). Effects and Opportunities of Supplying Electric Vehicles by Public Charging
Stations. Delft.
[26]
Wagner, F., Lakshmanan, B., & Mathias, M. (2010). Electrochemistry and the Future of the
Automobile. Journal of Physical Chemistry Letters , 2204-2219.
[27]
Wu, H., Gilchrist, A., Sealy, K., Israelsen, P., & Muhs, J. (2011). A Review on Inductive
Charging for Electric Vehicles. IEEE International Electric Machines & Drives Conference.
[28]
Zhao, L., Prousch, S., Hubner, M., & Moser, A. (2010). Simulation methods for assessing electric
vehicle impact on distribution grids. Transmission and Distribution Conference and Exposition,IEEE
PES, (pp. 1-7).
84
11 Appendix
11.1 Appendix A EV Load Modelling Flow Charts
This section shows the detailed flow charts for the various modules in the EV Fleet modelling code.
11.1.1 Trip Scheduling and Tracking Flow Chart
85
11.1.2 EV Power Estimator Flowchart
The LV power estimator sub-routine estimates the LV power using the charging algorithm described by
the dumb charging and green charging algorithms as described in Figure 4-3 and Figure 4-5. The subroutine is shown graphically below.
86
11.1.3 LV Power Estimator Flow chart
87
11.1.4 SOC Update Sub-Routine
88
11.1.5 SOC LV Sub-Routine Flow Chart
This sub module follows the same as that of the LV power estimator sub module described in 4.3.2.
Hence further explanation is not provided here. The graphical representation of the same is provided
below.
11.2 Appendix B Queuing Theory and QCR Infrastructure Requirements
Queuing theory and the estimation of QCR requirements is described in the following paper submitted for
the IECON 2012 conference.
89
90
91
92
93
94
95
96
97
98
99
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