1. 22 - 24 September 2010
Lyngby - Denmark
EES-UETP Electric Vehicle Integration into Modern Power Networks
Smart EV grid interfaces responding to
frequency variations to maximize
renewable energy integration
João A. Peças Lopes
INESC Porto / FEUP
(jpl@fe.up.pt)

2. Introduction
Large scale deployment of EV
• Steady-state impacts related with
voltage drops and branch overloads
Grid restrictions may limit the growth of
EV penetration, if no additional measures
are adopted. Solution:
Active management of EV batteries
• Dynamic issues
 EV participating in primary
frequency control
 EV participating in AGC (secondary
frequency control)
o

3. Introduction
• Renewable energies need to increase their
penetration in the generation mix in order to
reduce CO2 emissions
• There are renewable power sources
characterized by some variability
• In isolated Grids if EVs participate in primary
frequency control, major benefits to the
integration of RES in large scale are expected
• When parked and plugged-in, EVs will either
absorb energy (and store it) or provide
electricity to the grid when (the V2G concept).
• Existing EV grid interfaces are passive
devices that do not allow the required flexibility

4. The MERGE control concept
• A two level hierarchical control approach needs to be adopted:
• Local control housed at the EV grid interface, responding locally to grid
frequency changes and voltage drops;
• Upper control level designed to deal with:
• “short-term programmed” charging to deal with branch congestion,
voltage drops
• Delivery of reserves (secondary frequency control);
• Adjustments in charging acoording to the availability of power
resources (renewable sources).

5. EV Voltage / Frequency support modes
Local Control
Voltage Control
Coordinated Control
Primary Control
(local control)
Frequency Control
Secondary Control

6. Conceptual Framework For EV Integration
• EV must be an active element
within the power system
• The Upper Level control
requires interactions with:
• An Aggregating entity to
allow:
 Reserve management
Electricity Market
Operators
 Market negotiation

7. Delivery of Primary Reserve / Local Frequency Control
Methodology
Primary domain of application: Islanded grids (islands or networks
operated in islanding conditions)
1. An isolated system has been characterized in terms of available generation
and load. These components were modeled connected to a single bus system,
where the several types of generation are then modeled individually together
with the load.
2. A sudden change on wind power generation was simulated in order to assess
its impact on the system’s frequency. Several scenarios were created for this
purpose.
3. EV penetration was then characterized and the model for EV connections,
featuring V2G, has been developed. This model was included in the single bus
system and, finally, its effects on the system’s dynamic behaviour were
evaluated running simulations in the same conditions as defined in 2.

8. Primary Reserve
EV Electronic Grid Interface Modelling
• For frequency control the envisioned
response from EVs is shown in the figure:
P
 When facing frequency deviations Pmax
EVs may slow down/speed up their
charging or even inject active power
into the grid
 A dead band for battery premature
exhaustion prevention is required EV consumption
 Prated MW/Hz proportional gain
f
controls the reaction to frequency
deviations Dead
Band
Pmin
PInjection PConsumption
Droop control for EVs
V2G mode

9. Primary Reserve
EV Electronic Grid Interface Modelling
• A PQ inverter control logic was adopted
• Set-points for active power controlled by
a proportional gain that reacts to
frequency deviations
v,i
v,i
v  v  k( iref i )
*
iact
ireact
P, Q
1
 
TQ s  1
Control loop for EVs active power set-point

PQ inverter control system

10. Primary Reserve
Evaluation of the performance of grid
Case Study: small island normally fed from Diesel generation

11. Primary Reserve
Scenarios characterization
Scenario 1 Scenario 2
PDiesel1,2 (kW) 1500 1500
• Isolated system composed by:
 4 diesel units PDiesel3,4 (kW) 1800 1800
 2 wind turbines (1 more for scenario PWind (kW) 1320 1980
2)
PPV (kW) 100 100
 Mild PV penetration
 Load ranging from 1770kW to Installed power
4200kW
Scenario 1 Scenario 2
• Vehicles: PTotal load (kW) 2172 2172
 1 vehicle per household
Pload (kW) 1770 1770
 2150 vehicles
 323 (15%) EVs PEV load (kW) 402 402
 3 EV types: PEV available (kW) 851 851
o 1xPHEV: 1.5kW
o 2xEVs: 3kW and 6kW Pwind (kW) 900 1272
o Charging time: 4h Psync1 (kW) 636 450
Psync2 (kW) 636 450
Valley hour operation (load plus generation
dispatch)

12. Primary Reserve
Scenarios characterization
• Sudden shortfall on wind speed may jeopardize current power quality
standards under EN 50.160 for isolated systems
• Large frequency excursions due to wind power changes become a limiting
factor to the integration of Intermittent Renewable Energy Sources like wind
power)
10
9
Wind Speed (m/s)
8
7
6
5
0 1 2 3 4
Time (s)
Disturbance applied to the case study

13. Primary Reserve
Grid Modelling
• A single bus model of the system was
developed using Matlab/Simulink:
 Wind speed suffers time domain
changes
 Electrical component and their links
in a steady state frequency domain
model
• To each generation a dynamic model was
assigned:
 diesel generator  4th order
model, with frequency regulation
performed through conventional
proportional and integral control
loops
 Wind generator  simple induction
machine
Isolated system single-line diagram

14. Primary Reserve
Results – Scenario 1
50.5 3
2.5
System Frequency (Hz)
50
PDiesel (MW)
2
1.5
49.5
1
49 0.5
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
Time (s) Time (s)
PW = 1.3 MW; EV - charge mode
PW = 1.3 MW; EV - freq. control
1.5 0.1
0
1
-0.1
PWind (MW)
0.5 PEV (MW) -0.2
-0.3
0
-0.4
-0.5 -0.5
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
Time (s) Time (s)

15. Primary Reserve
Results – Scenario 2
50.5 3
2.5
System Frequency (Hz)
50
PDiesel (MW)
2
1.5
49.5
1
49 0.5
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
Time (s) Time (s)
PW = 2.0 MW; EV - charge mode
PW = 2.0 MW; EV - freq. control
1.5 0.1
0
1
-0.1
PWind (MW)
0.5 PEV (MW) -0.2
-0.3
0
-0.4
-0.5 -0.5
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
Time (s) Time (s)

16. Primary Reserve
Conclusions
• It is possible to verify that system dynamic performance was improved dramatically
when EVs are participating in frequency control
• Further sensitivity analysis is still needed to identify the best control parameters for the
droop control mode of the electronic grid interface used by the EVs
50.3
PW = 1.3 MW; EV - charge mode
50.2 PW = 1.3 MW; EV - freq. control
System Frequency (Hz) 50.1 PW = 2.0 MW; EV - freq. control
50
49.9
49.8
49.7
49.6
49.5
49.4
49.3
0 1 2 3 4 5 6 7 8 9 10
Time (s)
• The presence of a considerable amount of storage capability connected at the
distribution level also allows the operation of isolated distribution grids with large
amounts of IRES and/or microgeneration units connected to it

17. Implementation of EV Grid Interfaces
Power Electronic Converter:
The “Black Box” interface between the Low Voltage Grid (AC) and EV Battery
(DC) Design Requirements Converter
Functions
Three-Phase
Three Leg
Grid physical ⁄ v ►
Converter
connection Single-Phase
Two Leg Converter
Battery charge AC/DC conversion Rectifier
+ ⁄ + ► +
V2G capability DC/AC conversion Inverter
Grid “clean” interface Low harmonic content Controlled Three
⁄ ► Level Converter
Small displacement
factor
17

18. Implementation of EV Grid Interfaces
POWER CONVERTER – SINGLE & THREE PHASE TOPOLOGIES
Three-Phase, Three-level, Bidirectional Converter:
Power Matrix
Convert of switches
er
Time Variant
Power
Non-linear
Convert
System
er
18

19. EV Grid Interfaces
Low Level Control:
Closed loop control which outputs high frequency signals for each switch
Three‐Phase currents control : Sliding‐Mode Vectorial Control
‐ Nearly sinusoidal phase currents = Low harmonic distortion
‐ Currents in phase with voltages = Small displacement factor
‐ Static and dynamic phase current following
‐ Capacitor voltage equalization
‐ Robustness = immunity to disturbances
Grid/Battery charging current control: Proportional‐Integral external loop
‐ “Current source” converter behaviour
‐ Dynamic current following and near to zero static error
High Level Control:
Defines a current reference to Low Level Control
Charge Control  Grid/battery requirements: charging current, end of charge, Minimum and
maximum SOC levels …
Droop‐control  Grid frequency or voltage control: set‐point, dead‐band and slope
19

20. EV Grid Interfaces
High Level Control: outputs the battery charging current reference for the Low
Level Control
Charge Control: provides the charging current reference within the battery constraints
20

21. EV Grid Interfaces
Droop Control: outputs the droop charging current reference to the Charge
Control
Reacts to Voltage and Frequency local deviations according to respective droop
functions Central control units establish and communicate droop defining
parameters
Charging
Current Reference
=
output of
Frequency Droop
or
Voltage Droop
within
Battery
Charge
Constraints
21

22. Secondary frequency control
• Load variations or changes in generation output (namely from variable
generation units) provoke load / generation imbalances that lead to:
1. frequency changes and
2. inter-area power unbalances regarding scheduled power flows
• EV battery charging can be considered as very flexible loads, capable of
providing fast reserves (through the aggregators)
• An increased robustness of operation can be achieved
• The reserve levels can be reduced (depending on the hour of the day, taking
into account that the number of grid plugged vehicles)
22

23. Secondary Reserve
AGC operation with EV
• Modification of the active power set-points of generators and EV
• Some modifications need to be introduced in conventional AGC systems:
 redefinition of the partipation factors and
 introduction of an additional block to communicate with EV aggregators
• These control functionalities to be provided by EV are intended to keep the
scheduled system frequency and established interchange with other areas within
ini
predefined limits, enabling further deployment of IRES Pe 1
+
fp1 + Pref1
fi
fi + B fpm + Prefm
- +
fREF Pe ini
+ ACE
m
-KI/s +
Pif1 + -
ini
Pa 1
+ m k -
+ +
-  Pe
i 1
ini
i   Pa ini
i 1
i
fpA1 + Prefa1
Pifn Aggregators
PifREF
fpAk + Prefak
-
ini
Pa k

24. Secondary Reserve
Evaluating the Contribution of EV for Secondary Frequency Control
• Definition of a case-study: Portugal /Spain (European interconnected system)
 Grid selection
 Modeling
• Setting up a contingency / disturbance
• Evaluating the system dynamic performance:
 Without the participation of EV
 With the participation of EV

25. Secondary Reserve
Case-Study – Definition
12
10
• Portuguese transmission/generation 8
% of EV Cha rging
network, including existing tie lines with 6
Spain (equivalent) 4
• Technical constraints  Portugal will 2
not export more than 1500 MW or 0
1 5 9 13 17 21
import more than 1400 MW
Hours
Percentage of EV charging during a typical day, under a
smart charging strategy (EV  30% of total fleet)
Installed capacity
• 30% EV penetration
 20% PHEV  1.5 kW
 40% EV1  3 kW
 40% EV2  6 kW
• EV load was following a smart
charging scheme

26. Secondary Reserve
Case-Study – Definition
• Example of a windy day in the Portuguese system in the Autumn of 2009

27. Secondary Reserve
Case-Study – Dynamic Modeling
• Transmission system with 2 control areas (Portugal and Spain)
• 5 tie lines interconnecting areas 1 and 2 at 400 kV
• Generator equivalents per technology at each substation node:
 Conventional generator  4th order model synchronous machine
o Thermal units  simple governor and a three stage thermal turbine with reheat
o Hydro units  governor with transient droop compensation and a typical hydro
turbine
o IEEE type 1 voltage regulator was used
 Wind generators  3rd order model squirrel cage simple induction machine
o undervoltage relay setting  0.9 p.u.
• Voltage levels: 150 kV, 220 kV and 400 kV
• One AGC per area
Proportional
Control
1 R
Cvopen Pmecmax Pe
- - Synchronous
+ Governor Turbine +
Pref Pmec Generator
(AGC signal) Cvclose 0

28. Secondary Reserve
Case-Study – Disturbance and Scenario Definition
Winter valley period (6 a.M.)
Simplified Portuguese Transmission
C1
Network Control Area 1 Control Area 2
C15 C16 C17
H W11
~ C2 H ~
T
~
N
~
W2 H
~ W1
C7
H 1
2 ~ W6
11 10
150 kV
400 kV C5 C6 17
12 ~ W4 ~ W5
H H
13 400 kV
220 kV
14 15
220 kV
400 kV
16 4 3
C8
H
~ W7
~ ~
W3
C3 C4
H TG
220 kV
400 kV
18
6 5
C9 22
220 kV
TG
~ W8 150 kV
C11 23
TC
~
19 ~ 8
220 kV W9 150 kV
C10
400 kV 400 kV
20 21 H 7
• Event  300 ms fault at line 15-16 C12
TC
C13
TC
C14
H
~ ~ W10 ~
• Impact of EV in the AGC operation:
1. EV are not used for AGC 24 25 9
operation Equivalent Generator Types
~
C(TC): Conventional Fuel or Coal
~
C(TG): Conventional Gas
~
C(H): Conventional Hydro
2. EV are obtaining active power
~
N: Conventional Nuclear W: Wind
set-points from the AGC,
through the aggregation units

29. Secondary Reserve
Results – Interconnection active Power Flow
1000
With participation of EV
Without participation of EV
500
0
-500
(MW)
-1000
interconnection
-1500
P
-2000
-2500
-3000
-3500
0 100 200 300 400 500 600 700 800 900
Time (s)

30. Secondary Reserve
Results – Used Reserve Levels
Reserve Used Without EV Participating in Secondary Control
Used Reserve (MW)
Reserve (MW)
t=2min t=15min
Hydro 461 461 461
Thermal 590 211 256
EV 0 0 0
Total 1049 672 717
Reserve Used With EV Participating in Secondary Control
Used Reserve (MW)
Reserve (MW)
t=2min t=15min
Hydro 461 192 316
Thermal 590 31 74
EV 581 581 581
Total 1630 804 971

31. Secondary Reserve
Results – Frequency Evolution
With participation of EV
50.2 Without participation EV
50.1
Frequency (Hz)
50
49.9
49.8
49.7
-10 -5 0 5 10 15 20 25 30
Time (s)

32. Secondary Reserve
Results – Electrical Current in the Line 16-18
0.75
With participation of EV
Without participation of EV
0.7
0.65
0.6
(p.u.)
0.55
16-18
0.5
I
0.45
0.4
0.35
0 20 40 60 80 100 120
Time (s)

33. Secondary Reserve
Results – Electrical Current in the Line 20-21
1
With participation of EV
Without participation of EV
0.95
0.9
0.85
(p.u.)
0.8
20-21
I
0.75
0.7
0.65
0 20 40 60 80 100 120
Time (s)

34. Secondary Reserve
Results – Area Control Error for Portugal
3000
With participation of EV
Without participation of EV
2000
1000
ACE (MW)
0
-1000
-2000
-3000
0 100 200 300 400 500 600 700 800 900
Time (s)

35. Secondary Reserve
Conclusions
• Three main conclusions that can be drawn from these studies:
1. Improvement of the system robustness of operation
2. Increase of the system reserve levels that can be effectively mobilized for secondary control use
3. Increase safe integration of renewable power sources in the system
• Fast reaction of EV + communication + control architecture = fast and effective AGC operation
• When EV are participating in secondary frequency control, further integration of IRES in interconnected
grids is possible
• Additional economical and environmental benefits are expected from the adoption of EV smart control
strategies, mainly due to avoided start-up of expensive and highly pollutant generation units that compose
the tertiary control
• As a counterpart EV owners must be properly remunerated when participating in the provision of this
type of ancillary services in order to make this concept efficient and with sufficient adherence

36. Final Conclusions
• A specific EV grid interface needs to be adopted in order to allow EV to participate
in the provision of ancillary reserve services;
• This on board device can be integrated with the EV battery management system
• The adoption of such control approach allows increased dynamic robustness of
operation to the system
• Large penetration levels of renewable variable power generation are feasible,
specially in isolated grids.
.