The Impact of Electric Vehicles on the Grid · Electric Vehicle Research Project: Goals • Study impact of electric vehicles on the grid –at distribution level • Design an optimal

The Impact of Electric Vehicles on the Grid

Dr.JuliandeHoog

PresentationtoEconomyandInfrastructureCommittee,StateParliamentofVictoria,Australia

9November2017

Feb 2012 – Feb 2015 University of Melbourne Electric Vehicle Research Project: Goals

• Study impact of electric vehicles on the grid – at distribution level

• Design an optimal charging policy, informed by:– Electricity market spot price– State of charge in every battery– Present and anticipated constraints of the distribution network

Ø Maximal uptake of electric vehicles with minimal capacity upgrade requirements

Simulation Approach

Networks

• Models of several neighbourhoods (Melbourne, Townsville)• Real line and transformer specs• Accurate phase allocation (although not for all)

House on Phase AHouse on Phase BHouse on Phase C

Pole

Transformer

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17 18 19 222120 23 24

252627

28

2930

31

3233343536

37

38

39

4041 42 43 44 45 46 47 48 49 50 51 52 53

54

55

56

57585960616263646566

67 68 69 70 71 72

73

74

75

767778

79

80

81

828384858687

88

89 90 91 92 93

94

95

96979899100101102103

104

105

106

107

108

109

110

111

112113

House on Phase AHouse on Phase BHouse on Phase C

Pole

Transformer

Simulation Approach

Networks

Household Demand

• Household demand as measured at distribution transformer / or at individual houses

• Range of months

0 2 4 6 8 10 12 14 16 18 20 220

20

40

60

80

Time of Day

Tota

l Dem

and

(kVA

)

Phase APhase BPhase C

Simulation Approach

Networks

Household Demand

Travel Patterns and Charging Needs

• 2009 VISTA Travel Survey (Victoria Department of Transport)• 13,000 records of 24-hour vehicle travel profiles• Know when vehicles home, and how far they have travelled

10 20 30 40 50 60 70 80 90 100 >1000

5

10

15

20

Daily maximum travel distance (km)

% o

f tot

al

0 2 4 6 8 10 12 14 16 18 20 22 240

5

10

Time of day

Hom

e ar

rival

s, %

Simulation Approach

Networks

Household Demand

Travel Patterns and Charging Needs

Load Flow and Simulation

• Load flow using MATLAB SimPowerSystems• Optimisation using MATLAB Optimisation Toolbox• All else using POSSIM (POwer Systems SIMulator)

Simulation Approach

Networks

Household Demand

Travel Patterns and Charging Needs

Load Flow and Simulation

Validation

• Extensive comparison of simulator outputs to real measurements• Voltages within 1.5 V on average (0.6%)• Currents within 6 A on average (7%)

0.6

0.9

1.2

V B (pu)

RealSimulated

0 2 4 6 8 10 12 14 16 18 20 22 240

0.3

0.6

I B (pu)

Time of day

Early Observations 1: Network Impacts

Peak Demand

8:00 12:00 16:00 20:00 0:00 4:00 8:000

0.5

1

1.5

2

Dem

and

(pu)

VehicleHousehold

• Thermal overload• Increased ageing / failure of network assets

Early Observations 1: Network Impacts

Peak Demand

Voltage Drop

8:00 12:00 16:00 20:00 0:00 4:00 8:000.9

1

1.1Vo

ltage

(pu)

• Voltage below minimum required levels at certain houses in the network

• Appliances may run inefficiently / reduced lifetimes

Transformer234V 230V 226V

231V 224V 220V

Early Observations 1: Network Impacts

Peak Demand

Voltage Drop

Phase Unbalance

IA

IB

IC IA • Neutral current• More losses• Increased voltage drop

8:00 12:00 16:00 20:00 0:00 4:00 8:000

2

4

Unb

alan

ce (%

)

Early Observation 2: Importance of Location

A: 50B: 43C: 21

House on Phase AHouse on Phase BHouse on Phase C

Pole

Most sensitive!

Early Observation 2: Importance of Location

• Adding a single vehicle at the red house has the same impact on minimum voltage as adding vehicles at all (45) green houses!

• Reason: voltage drop and unbalance in the network

• Vulnerable: houses distant from the transformer on highly loaded phases

• Significant implications for fairness of charging. Should all customers have equal rights?

House on Phase AHouse on Phase BHouse on Phase C

Pole

Early Observation 3: Flexibility

8:00 12:00 16:00 20:00 0:00 4:00 8:00Time of day

Indi

vidu

al v

ehic

les

Home, fully charged Home, charging Away

Early Observation 3: Flexibility

8:00 12:00 16:00 20:00 0:00 4:00 8:00Time of day

Indi

vidu

al v

ehic

les

Home, fully charged Home, charging Away

Coincident charging

Early Observation 3: Flexibility

8:00 12:00 16:00 20:00 0:00 4:00 8:00Time of day

Indi

vidu

al v

ehic

les

Home, fully charged Home, charging Away

Coincident charging

Flexibility

Load Shifting for Electric Vehicles

UtilityOperator SmartMeter ChargePoint ElectricVehicle

Z.Angelovski and K.Handberg (DiUS Computing), together with Raman Jegatheesan (United Energy)Demand management of electric vehicle charging using Victoria’s Smart Grid. Technical report, DiUS Computing (May 2013).

By controlling timing of charging, negative impacts can be avoided

: current supplied to vehicle k at time t

Provide as much current to vehicle charging as the network will allowmaxxk ,t k=1

K

∑ xk,tt=0

T

∑xk,t

Uncontrolled Controlled

Total Demand:

Voltage:

Charging Profiles:

8:00 12:00 16:00 20:00 0:00 4:00 8:000

0.5

1

1.5

2D

eman

d (p

u)

VehicleHousehold

8:00 12:00 16:00 20:00 0:00 4:00 8:000.9

1

1.1

Volta

ge (p

u)

8:00 12:00 16:00 20:00 0:00 4:00 8:0050

100

Time of Day

SOC

(%)

8:00 12:00 16:00 20:00 0:00 4:00 8:0050

100

Time of Day

8:00 12:00 16:00 20:00 0:00 4:00 8:000.9

1

1.1

8:00 12:00 16:00 20:00 0:00 4:00 8:000

0.5

1

1.5

2

VehicleHousehold

Conclusions

• Uncontrolled chargingallowsEVpenetrationofonly10-15%;typicallythefirst“pointoffailure”isvoltagedrop

• Optimal chargingallowsEVpenetrationof80%ormoreinthenetworkswestudied,usingonlyexistinginfrastructure

• Installinganewtransformercosts~$100,000- $150,000…Savings(oratleastdelayedupgrades)of$1000- $2000perhousehold.

• Price-basedoptimisation leadstosavingsof10-20%(andconsiderablymoreondayswithgreaterfluctuationsinprice)

Distributed Charge Control (Lu Xia)

Local voltage is a good indicator of total network demand

Use local voltage as a demand response signal

0

0.5

0.9

1.3

Tota

l dem

and

(per

uni

t)

0:00 4:00 8:00 12:00 16:00 20:00 24:000.9

1.0

1.1

1.2

Loca

l Vol

tage

(per

uni

t)

Total DemandLocal Voltage

Measure local voltage

Increase/Decrease

Charge Rate

Apply for discrete interval

Distributed DSM solutions

Uncontrolled Centralised Distributed

8:00 12:00 16:00 20:00 0:00 4:00 8:000

0.5

1

1.5

2

Dem

and

(pu)

VehicleHousehold

8:00 12:00 16:00 20:00 0:00 4:00 8:000.9

1

1.1

Volta

ge (p

u)

8:00 12:00 16:00 20:00 0:00 4:00 8:0050

100

Time of Day

SOC

(%)

8:00 12:00 16:00 20:00 0:00 4:00 8:0050

100

Time of Day

8:00 12:00 16:00 20:00 0:00 4:00 8:000.9

1

1.1

8:00 12:00 16:00 20:00 0:00 4:00 8:000

0.5

1

1.5

2

VehicleHousehold

8:00 12:00 16:00 20:00 0:00 4:00 8:0050

100

Time of Day

8:00 12:00 16:00 20:00 0:00 4:00 8:000.9

1

1.1

8:00 12:00 16:00 20:00 0:00 4:00 8:000

0.5

1

1.5

2

VehicleHousehold

© Copyright The University of Melbourne 2017

Thank you!

Questions?