Electric Vehicles Aggregator/System

Operator Coordination for Optimal

Charging Scheduling and Services
Procurement

Miguel Ortega-Vazquez (UW)

Francois Bouffard (McGill)
Vera Silva (EDF)

UNIVERSITY of WASHINGTON
Seattle, USA
 Introduction
• Decarbonisation of the transportation sector:
– Electricity as energy carrier
– Ideally, energy from RES
– Large fleets of EVs will constitute a significant share of the
wide system demand

• Advent of the Smart Grid

– Demand will feature higher degrees of communication and
control

. Miguel Ortega-Vazquez - Electrical Engineering - UNIVERSITY of WASHINGTON 2

 Electric vehicles (EVs)
• Unlike most existing loads, EVs are equipped with a
battery:
– Waive their energy requirements in time
– Energy storage
– Provide ancillary services, e.g. reserve

. Miguel Ortega-Vazquez - Electrical Engineering - UNIVERSITY of WASHINGTON 3

 Impacts
• The pliant nature of the EVs as demand makes them
good candidates to impact the system:

– Technically: provide flexibility to the system

– Economically: trading that flexibility

. Miguel Ortega-Vazquez - Electrical Engineering - UNIVERSITY of WASHINGTON 4

 EVs operating modes
EV operating mode Schedule &
p

• Vehicle charging
$
p $
• Vehicle discharging (V2G)
$ p

• Vehicle plugged-in $
$ p

• Vehicle unused $

• Vehicle circulating

. Miguel Ortega-Vazquez - Electrical Engineering - UNIVERSITY of WASHINGTON 5

 Power system + EVs

Wholesale

Energy

Co-ordination
Retail

Aggregator

. Miguel Ortega-Vazquez - Electrical Engineering - UNIVERSITY of WASHINGTON 6

 Aggregator
• Bridge between EVs and power system players:

ct tt
lt Aggregator et

gT(x,u,p) ≤ 0

gD(x,u,p) ≤ 0

. Miguel Ortega-Vazquez - Electrical Engineering - UNIVERSITY of WASHINGTON 7

 Interaction Aggregator & SO
• Generation scheduling
– Minimize total operating costs
• Power balance constraint (including EVs’ charge and discharge)
• Security: system reserve requirements (unforeseen events)
• Units technical operating constraints:
– Minimum up- and down-times
– Up and down ramp rate limits
• EVs scheduling
– Sell flexibility services to the system
• EVs’ energy requirements
• EVs operating constraints:
z V2G p
Cup pCdn
– State of charge
– Charging and discharging rates

. Miguel Ortega-Vazquez - Electrical Engineering - UNIVERSITY of WASHINGTON 8

 From day-ahead to real-time
• Schedules are performed a day-ahead on the basis of:
– Wind power generation forecast
– Demand forecast
– Forecasted EVs availabilities
• Deviations between forecasts and actual values
materialize in real-time
– Conventional generation deployment to meet net demand
deviations
– Deployment of EVs’ flexibility services as a function of the
agreed prices the day before

. Miguel Ortega-Vazquez - Electrical Engineering - UNIVERSITY of WASHINGTON 9

 Test system
• IEEE-RTS omitting hydro generation
– 26 Units, 3105 MW installed Cost functions
5000 A
Group # Units Capacity (MW) Fuel
B
A 5 12 oil/steam 4000
C
B 4 20 oil/CT

Cost, $/h
C 4 76 coal/steam 3000 D

D 3 100 oil/steam 2000 E

E 4 155 coal/steam F
F 3 197 oil/steam 1000
G
G 1 350 oil/steam 0 H
H 2 400 nuclear 0 100 200 300 400
Composition Power, MW
A 40
H
2% B
Price, $/MWh

26% 2%
G 30
11% C 20
10%
F 10
19% D
E 10% 0
0 500 1000 1500 2000 2500 3000
20%
Power, MW
Market Clearing at Period 20
. Miguel Ortega-Vazquez - Electrical Engineering - UNIVERSITY of WASHINGTON 10
 Demand and wind profiles
• Forecasted and actual values…
300

wind power, MW
demand, MW

2500
200

2000
100

1500 0
5 10 15 20 5 10 15 20
time, hrs time, hrs
demand, MW

2500

2000 Actual
Forecasted
1500
2 4 6 8 10 12 14 16 18 20 22 24
time, hrs

. Miguel Ortega-Vazquez - Electrical Engineering - UNIVERSITY of WASHINGTON 11

 40

Price, $/MWh
EVs’ characteristics
30

20

10

0
EV
0 parameter
500 EV
2000type 2500
1000 1500 3000
Power,
BEVMW City-BEV PHEV Value
EV fleet composition (%) Market Clearing
37 at Period1020 53 100
Battery capacity
40 (kWh) 35 16 18 24.1
Clearing Price Power Reserve
price, $/MWh
Energy consumption (kWh/km) 0.2 0.12 0.212345 0.192
30 21 22
EV range (km) 175 133
141516
90 125.5
Average distance (km/day) 24 17 18 19 20 10111213
20 40.0
Daily consumption
10
25 (kWh/day)
26 7.68
Charge/discharge rate (kW/h) 3.70
0
Total number
0 of vehicles
500 1000 1500 2000 2500 689,475
3000
power, MW
25 25
a) b)
20 20
EVs, %

EVs, %
15 15

10 10

5 5

0 0
5 10 15 20 5 10 15 20
time, hrs time, hrs

. Miguel Ortega-Vazquez - Electrical Engineering - UNIVERSITY of WASHINGTON 12

 demand

wind pow
2000
100

System with no EVs 1500

5 10 15
time, hrs
20
0
5 10
time

40
30

Price, $/MWh
demand, MW
a) 2500 30
$/MWh

20
20
2000 Actual
10 10
Forecasted
0
0 1500 0 1000 2000 3000
2 4 6 8 10 12 14 16 18 20 22 24 5 10 15 20
Power,
time, hrsMW
time, hrs 21 22
Market Clearing at Period 12
25
40
Power 15 Reserve
16

$/MWh
14
demand, MW

3000

Price,$/MWh
30
20
2500 b) 10 11 12 13
20

price,
2000 15
10

1500 0
2 4 6 8 10 12 14 16 18 20 22 24 100 1000 2000 3000
2000 2200 2400 2600
time, hrs Power, MW 21 22
power, MW
1000
Market Clearing at Period 12
25
25
Power
14 1525 16
Reserve
a) b)

EVs,$/MWh
SR, MW

c) 20 20
20
500 10 11 12 13

EVs, %
15 15
price,
15
10 10
0
2 4 6 8 10 12 14 16 18 20 22 24 5 5
10
time, hrs 2000 0 2200 0 2400 2600
. Miguel Ortega-Vazquez - Electrical Engineering - 5 10 15 power,
20 MW
UNIVERSITY of WASHINGTON5 10 15 20 13
 Uncontrolled EVs – 30% penetration
30
a)
$/MWh

20

100 100
10 a) b)

Charging, %

V2G, %
50 50
0
2 4 6 8 10 12 14 16 18 20 22 24
time, hrs 0
5 10 15 20
0
5 10 15 20
time, hrs time, hrs
100 100
demand, MW

3000 c) d)

Circulating, %
Reserve, %
2500 b) 50 50

2000 0 0
5 10 15 20 5 10 15 20

1500 time, hrs time, hrs

2 4 6 8 10 12 14 16 18 20 22 24 100
time, hrs e)

Unused, %
1000 50

0
SR, MW

c) 5 10 15 20
500 time, hrs

0
2 4 6 8 10 12 14 16 18 20 22 24
time, hrs
. Miguel Ortega-Vazquez - Electrical Engineering - UNIVERSITY of WASHINGTON 14
 Uncontrolled EVs – 30% penetration
30
a)
$/MWh

20

10 100 100
a) b)

Charging, %

V2G, %
50 50
0
2 4 6 8 10 12 14 16 18 20 22 24
time, hrs 0 0
5 10 15 20 5 10 15 20
time, hrs time, hrs
demand, MW

3000 100 100

c) d)

Circulating, %
Reserve, %
2500 b) 50 50

2000
0 0
5 10 15 20 5 10 15 20
1500 time, hrs time, hrs
2 4 6 8 10 12 14 16 18 20 22 24
100
time, hrs e)

Unused, %
1000 50
SR, MW

c) 0
5 10 15 20
500 time, hrs

0
2 4 6 8 10 12 14 16 18 20 22 24
time, hrs
. Miguel Ortega-Vazquez - Electrical Engineering - UNIVERSITY of WASHINGTON 15
 Effect of the EVs’ penetration
Optimized Uncontrolled

3000 3000
net demand, MW

net demand, MW
2500 2500

2000 2000

1500 1500
5 10 15 20 5 10 15 20
time, hrs time, hrs

40 40
20% EVs 20% EVs
35 35
60% EVs 31% EVs
price, $/MWh

30 100% EVs
price, $/MWh 30

25 25

20 20

15 15

10 10
5 10 15 20 5 10 15 20
time, hrs time, hrs

. Miguel Ortega-Vazquez - Electrical Engineering - UNIVERSITY of WASHINGTON 16

 400
a)
200

MW
0

Effect of the V2G markup -200

-400
0 5 10 15 20 25
30 time, hrs
a)

charge, kWh
$/MWh

20 20
3000
10
10
2000
0
2 4 6 8 10 12 14 16 18 20 22 24
0 1000 time, hrs
2 4 6 8 10 12 14 16 18 20 22 24
time, hrs 0
0
10

100

MW MW
demand, MW

3000
1500
80

capacity,
Injected in V2G
2500 b) 60
1000 Discharging losses

power,
40
2000

Reserve
500
20

1500 00
2 4 6 8 10 12 14 16 18 20 22 24 1 2 42 6 38 10 4 12 14
5 16 6 18 20 7 22 24
8
time, hrs time,
zV2G hrs
, $/MWh

1000
SR, MW

c)
500

0
2 4 6 8 10 12 14 16 18 20 22 24
time, hrs
. Miguel Ortega-Vazquez - Electrical Engineering - UNIVERSITY of WASHINGTON 17
 Effect of the pricing of the reserve
30
from EVs
a)
- 20 The price of the reserve from the supply
$/MWh

10 side is 20% of the bids of energy

100 100
a) b)

Charging, %
- Reserve from EVs at 2 $/MW

V2G, %
50 50
0
2 4 6 8 10 12 14 16 18 20 22 24
time, hrs 0
5 10 15 20
0
5 10 15 20
time, hrs time, hrs
100 100
c) d)

Circulating, %
demand, MW

3000

Reserve, %
2500 b) 50 50

2000 0 0
5 10 15 20 5 10 15 20
time, hrs time, hrs
1500
2 4 6 8 10 12 14 16 18 20 22 24 100
e)
time, hrs

Unused, %
50
1000

0
SR, MW

c) 5 10 15 20
500 time, hrs

0
2 4 6 8 10 12 14 16 18 20 22 24
time, hrs
. Miguel Ortega-Vazquez - Electrical Engineering - UNIVERSITY of WASHINGTON 18
 Reducing the exercise price for reserve
30
deployment
a)
$/MWh

- 20 Price of the reserve from the supply side at

10 30% of the bids of energy
100 100
a) b)

Charging, %
Scheduled

V2G, %
- Reserve from EVs at 2 $/MW
Deployed
50 50
0
- Exercise
2 4 price
6 8 at 5
10 $/MWh
12 14 16 18 20 22 24
time, hrs 0
5 10 15 20
0
5 10 15 20
time, hrs time, hrs
100 100
c) d)

Circulating, %
demand, MW

3000

Reserve, %
400
2500 b) 50 50
a)
200
2000

MW
0 0
5 10 15 20 5 10 15 20
0 time, hrs time, hrs
1500
2 4 6 8 10 12 14 16 18 20 22 24 100
e)
time, hrs

Unused, %
-200
0 5 10 15 20 25
1000 50 time, hrs
charge, kWh
0
SR, MW

c) 20 5 10 15 20
500 time, hrs
10

0
2 4 6 8 10 12 14 16 18 20 22 24
0 time, hrs
2 4 6 8 10 12 14 16 18 20 22 24
time, hrs
. Miguel Ortega-Vazquez - Electrical Engineering - UNIVERSITY of WASHINGTON 19
 Summary
• Presented the necessary adaptations to a typical US-
style short-term forward electricity markets including
EVs fleet aggregators
• The operating constraints of the EVs are explicitly
modeled
• Represents adequately the collective operation of the
EVs to attain benefits while reducing operating costs
• Several ancillary services and their deployment are
represented: V2G, Reserve capacity as additional charge
and interruption of scheduled charging/discharging

. Miguel Ortega-Vazquez - Electrical Engineering - UNIVERSITY of WASHINGTON 20

 Conclusions
• If no coordination is implemented, only a limited
amount of EVs can be successfully integrated in the
system
• Implementing an approach such as the proposed,
allows the integration of large volumes of EVs without
needing to invest in generation assets
• The participation of the aggregator(s) as service
providers determined by the services bids

. Miguel Ortega-Vazquez - Electrical Engineering - UNIVERSITY of WASHINGTON 21

 Conclusions
• Charging process is scheduled at the cheapest periods
• Services from EVs are scheduled only if these are
competitive with those from conventional sources
• V2G is attractive to avoid synchronization of expensive
generation
• Reserve is in the form of interruption of charging is
attractive, but it can only be offered when EVs are
charging  light loads…
• Reserve as additional charge are attractive, but only
when their supply equivalent is expensive
. Miguel Ortega-Vazquez - Electrical Engineering - UNIVERSITY of WASHINGTON 22