View
214
Download
1
Category
Preview:
Citation preview
Optimal Participation of an EV Aggregator inDay-Ahead Energy and Reserve Markets
Miguel A. Ortega-VazquezMushfiqur R. Sarker
Yury Dvorkin
FERC Software Conference
June 27–29, 2016
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 1 / 21
Overview
1 Introduction
2 Aggregator as a market participantAggregator’s perspectiveEVs’ respective
3 Proposed approachSolution processMathematical model
4 Case StudyTest systemResults
5 Conclusions
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 2 / 21
Introduction
Introduction
Power systems are forever changing:Minimize operating costsMaintain reliable supplyMaximize penetration of emission-free generation
De-carbonization of the road transport sector
Electric vehicles’ (EVs) batteries are poised as excellent candidates toprovide services:
Energy arbitrage (locally or at a system level)Ancillary services (different regulation intervals)
EV Aggregation:As individuals, they cannot participate in wholesale markets (e.g. PJM1 MW minimum capacity)Operation of EVs as an ensemble: coordinated response
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 3 / 21
Introduction
Introduction
Power systems are forever changing:Minimize operating costsMaintain reliable supplyMaximize penetration of emission-free generation
De-carbonization of the road transport sector
Electric vehicles’ (EVs) batteries are poised as excellent candidates toprovide services:
Energy arbitrage (locally or at a system level)Ancillary services (different regulation intervals)
EV Aggregation:As individuals, they cannot participate in wholesale markets (e.g. PJM1 MW minimum capacity)Operation of EVs as an ensemble: coordinated response
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 3 / 21
Introduction
Introduction
Power systems are forever changing:Minimize operating costsMaintain reliable supplyMaximize penetration of emission-free generation
De-carbonization of the road transport sector
Electric vehicles’ (EVs) batteries are poised as excellent candidates toprovide services:
Energy arbitrage (locally or at a system level)Ancillary services (different regulation intervals)
EV Aggregation:As individuals, they cannot participate in wholesale markets (e.g. PJM1 MW minimum capacity)Operation of EVs as an ensemble: coordinated response
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 3 / 21
Introduction
Introduction
Power systems are forever changing:Minimize operating costsMaintain reliable supplyMaximize penetration of emission-free generation
De-carbonization of the road transport sector
Electric vehicles’ (EVs) batteries are poised as excellent candidates toprovide services:
Energy arbitrage (locally or at a system level)Ancillary services (different regulation intervals)
EV Aggregation:As individuals, they cannot participate in wholesale markets (e.g. PJM1 MW minimum capacity)Operation of EVs as an ensemble: coordinated response
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 3 / 21
Introduction
EVs aggregation
Coordination between the system operator (SO) and EV ownersSO seeks to minimize operating costAggregator seeks to maximize profitsEV owner seek to minimize cost of consuming electricity while receivingcompensation for providing services
In this framework, aggregator optimally schedules the EV fleetsconsidering:
EV transportation needs (e.g. motion needs, range anxiety, etc.)EV battery degradation for market participationParticipation into competitive energy and reserve markets
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 4 / 21
Introduction
EVs aggregation
Coordination between the system operator (SO) and EV ownersSO seeks to minimize operating costAggregator seeks to maximize profitsEV owner seek to minimize cost of consuming electricity while receivingcompensation for providing services
In this framework, aggregator optimally schedules the EV fleetsconsidering:
EV transportation needs (e.g. motion needs, range anxiety, etc.)EV battery degradation for market participationParticipation into competitive energy and reserve markets
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 4 / 21
Aggregator as a market participant Aggregator’s perspective
Which Markets?
Energy market:Expectation of energy prices and schedules EVs G2V & V2G
Voluntary up reserves1
Competitive reserve providerExpected revenue obtained in the day-ahead (DA) as a capacitypayment for being on-stand byExpected revenue obtained in real-time (RT) for deployment (if calledby the SO)
Voluntary down reserves:Similar to the up reserve marketNo deployment payment
1Regulating Reserve Service (Up & Down) in the ERCOT market consideredOrtega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 5 / 21
Aggregator as a market participant Aggregator’s perspective
Which Markets?
Energy market:Expectation of energy prices and schedules EVs G2V & V2G
Voluntary up reserves1
Competitive reserve providerExpected revenue obtained in the day-ahead (DA) as a capacitypayment for being on-stand byExpected revenue obtained in real-time (RT) for deployment (if calledby the SO)
Voluntary down reserves:Similar to the up reserve marketNo deployment payment
1Regulating Reserve Service (Up & Down) in the ERCOT market consideredOrtega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 5 / 21
Aggregator as a market participant Aggregator’s perspective
Which Markets?
Energy market:Expectation of energy prices and schedules EVs G2V & V2G
Voluntary up reserves1
Competitive reserve providerExpected revenue obtained in the day-ahead (DA) as a capacitypayment for being on-stand byExpected revenue obtained in real-time (RT) for deployment (if calledby the SO)
Voluntary down reserves:Similar to the up reserve marketNo deployment payment
1Regulating Reserve Service (Up & Down) in the ERCOT market consideredOrtega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 5 / 21
Aggregator as a market participant Aggregator’s perspective
Aggregator’s expectations
The aggregators need to explicitly account for the probability (π) ofhaving their bids/offers accepted
Embed the aggregator’s probability of market outcome into thedecision-making process:
Determine the bids/offers for the DA clearing process based onexpected outcomes (including deployment in RT)
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 6 / 21
Aggregator as a market participant Aggregator’s perspective
Aggregator’s expectations
The aggregators need to explicitly account for the probability (π) ofhaving their bids/offers acceptedEmbed the aggregator’s probability of market outcome into thedecision-making process:
Determine the bids/offers for the DA clearing process based onexpected outcomes (including deployment in RT)
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 6 / 21
Aggregator as a market participant Aggregator’s perspective
Aggregator’s expectations
The aggregators need to explicitly account for the probability (π) ofhaving their bids/offers acceptedEmbed the aggregator’s probability of market outcome into thedecision-making process:
Determine the bids/offers for the DA clearing process based onexpected outcomes (including deployment in RT)
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 6 / 21
Aggregator as a market participant Aggregator’s perspective
Over- and under-committing
The aggregators creates an expectation of the amount of energy thatcould be called upon (pdepl)
In RT, the aggregators’ expectation might be short or long:
The penalty could be adjusted to avoid over-commitments
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 7 / 21
Aggregator as a market participant Aggregator’s perspective
Over- and under-committing
The aggregators creates an expectation of the amount of energy thatcould be called upon (pdepl)In RT, the aggregators’ expectation might be short or long:
The penalty could be adjusted to avoid over-commitments
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 7 / 21
Aggregator as a market participant Aggregator’s perspective
Over- and under-committing
The aggregators creates an expectation of the amount of energy thatcould be called upon (pdepl)In RT, the aggregators’ expectation might be short or long:
The penalty could be adjusted to avoid over-commitments
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 7 / 21
Aggregator as a market participant Aggregator’s perspective
Probabilities curves
Use historical data from market to populate and createProbability-Quantity-Price (PQP) curves:
Uses probabilities to determine bidding strategy in markets:Probability of acceptance (πa)Probability of deployment (πd)
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 8 / 21
Aggregator as a market participant Aggregator’s perspective
Probabilities curves
Use historical data from market to populate and createProbability-Quantity-Price (PQP) curves:
Uses probabilities to determine bidding strategy in markets:Probability of acceptance (πa)Probability of deployment (πd)
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 8 / 21
Aggregator as a market participant EVs’ respective
EV participation modes
Optimal scheduling of these services to maximize profits
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 9 / 21
Aggregator as a market participant EVs’ respective
EV participation modes
Optimal scheduling of these services to maximize profits
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 9 / 21
Aggregator as a market participant EVs’ respective
EV participation modes
Charge
Energy Market
Energy Market Charge
(EMCHG)
Voluntary Down Reserves Market
Reserve Down
(REGDN)
Stop Discharge (STOPDSG)
Discharge
Energy Market
Energy Market
Discharge (EMDSG)
Voluntary Up Reserves Market
Reserve Up (REGUP)
Stop Charge
(STOPCHG)
EV Perspective
Action
Market
Service
Optimal scheduling of these services are necessary to maximize profits
Optimal scheduling of these services to maximize profits
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 9 / 21
Aggregator as a market participant EVs’ respective
EV participation modes
Charge
Energy Market
Energy Market Charge
(EMCHG)
Voluntary Down Reserves Market
Reserve Down
(REGDN)
Stop Discharge (STOPDSG)
Discharge
Energy Market
Energy Market
Discharge (EMDSG)
Voluntary Up Reserves Market
Reserve Up (REGUP)
Stop Charge
(STOPCHG)
EV Perspective
Action
Market
Service
Optimal scheduling of these services are necessary to maximize profitsOptimal scheduling of these services to maximize profitsOrtega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 9 / 21
Proposed approach Solution process
Solution process
Steps:
1) Aggregator performs DA optimization to determine bids/offers in theenergy and reserve markets
2) Aggregator sends bids/offers to the SO/MO
3) SO/MO clears simultaneous energy and reserves markets2
Accepts bids/offers of participants to attain minimum costDetermines energy and reserve clearing pricesAggregator is notified of accepted bids/offers
4) In RT, SO re-dispatches to accommodate deviationsAggregator may be called upon to deploy reserves
2H. Pandzic, Y. Dvorkin, Y. Wang, T. Qiu and D. S. Kirschen, “Effect of time resolution on unit commitment decisions insystems with high wind penetration,” 2014 IEEE PES General Meeting, National Harbor, MD, 2014, pp. 1-5.
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 10 / 21
Proposed approach Solution process
Solution process
Steps:
1) Aggregator performs DA optimization to determine bids/offers in theenergy and reserve markets
2) Aggregator sends bids/offers to the SO/MO
3) SO/MO clears simultaneous energy and reserves markets2
Accepts bids/offers of participants to attain minimum costDetermines energy and reserve clearing pricesAggregator is notified of accepted bids/offers
4) In RT, SO re-dispatches to accommodate deviationsAggregator may be called upon to deploy reserves
2H. Pandzic, Y. Dvorkin, Y. Wang, T. Qiu and D. S. Kirschen, “Effect of time resolution on unit commitment decisions insystems with high wind penetration,” 2014 IEEE PES General Meeting, National Harbor, MD, 2014, pp. 1-5.
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 10 / 21
Proposed approach Solution process
Solution process
Steps:
1) Aggregator performs DA optimization to determine bids/offers in theenergy and reserve markets
2) Aggregator sends bids/offers to the SO/MO
3) SO/MO clears simultaneous energy and reserves markets2
Accepts bids/offers of participants to attain minimum costDetermines energy and reserve clearing pricesAggregator is notified of accepted bids/offers
4) In RT, SO re-dispatches to accommodate deviationsAggregator may be called upon to deploy reserves
2H. Pandzic, Y. Dvorkin, Y. Wang, T. Qiu and D. S. Kirschen, “Effect of time resolution on unit commitment decisions insystems with high wind penetration,” 2014 IEEE PES General Meeting, National Harbor, MD, 2014, pp. 1-5.
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 10 / 21
Proposed approach Solution process
Solution process
Steps:
1) Aggregator performs DA optimization to determine bids/offers in theenergy and reserve markets
2) Aggregator sends bids/offers to the SO/MO
3) SO/MO clears simultaneous energy and reserves markets2
Accepts bids/offers of participants to attain minimum costDetermines energy and reserve clearing pricesAggregator is notified of accepted bids/offers
4) In RT, SO re-dispatches to accommodate deviationsAggregator may be called upon to deploy reserves
2H. Pandzic, Y. Dvorkin, Y. Wang, T. Qiu and D. S. Kirschen, “Effect of time resolution on unit commitment decisions insystems with high wind penetration,” 2014 IEEE PES General Meeting, National Harbor, MD, 2014, pp. 1-5.
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 10 / 21
Proposed approach Mathematical model
Aggregator’s Model
Aggregator strives to maximize their profits:
max{rem + rcap + rdepl − cregup − cregdn − cdeg
}
Revenues3:
rem = ∆t∑t∈T
∑v∈V
λDAt
(ηdsg
v · pemdsgt,v − pemchg
t,v
)
rcap =∑t∈T
∑b∈B
[(wup
t,b λupt,b
)πapup
t +(wdn
t,b λdnt,b
)φapdn
t
]
rdepl = πaπdηdsg∑t∈T
∑v∈V
∑b∈B
(vup
t,b λRTt,b
)(eregup
t,v + 1ηdsg e
stopdsgt,v
)
3M. A. Ortega-Vazquez, F. Bouffard and V. Silva, “Electric Vehicle Aggregator/System Operator Coordination for ChargingScheduling and Services Procurement,” IEEE Transactions on Power Systems, Vol. 28, Issue 2, pp. 1806-1815, May 2013. [url]
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 11 / 21
Proposed approach Mathematical model
Aggregator’s Model
Aggregator strives to maximize their profits:
max{rem + rcap + rdepl − cregup − cregdn − cdeg
}Revenues3:
rem = ∆t∑t∈T
∑v∈V
λDAt
(ηdsg
v · pemdsgt,v − pemchg
t,v
)
rcap =∑t∈T
∑b∈B
[(wup
t,b λupt,b
)πapup
t +(wdn
t,b λdnt,b
)φapdn
t
]
rdepl = πaπdηdsg∑t∈T
∑v∈V
∑b∈B
(vup
t,b λRTt,b
)(eregup
t,v + 1ηdsg e
stopdsgt,v
)
3M. A. Ortega-Vazquez, F. Bouffard and V. Silva, “Electric Vehicle Aggregator/System Operator Coordination for ChargingScheduling and Services Procurement,” IEEE Transactions on Power Systems, Vol. 28, Issue 2, pp. 1806-1815, May 2013. [url]
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 11 / 21
Proposed approach Mathematical model
Aggregator’s Model
Aggregator strives to maximize their profits:
max{rem + rcap + rdepl − cregup − cregdn − cdeg
}Revenues3:
rem = ∆t∑t∈T
∑v∈V
λDAt
(ηdsg
v · pemdsgt,v − pemchg
t,v
)
rcap =∑t∈T
∑b∈B
[(wup
t,b λupt,b
)πapup
t +(wdn
t,b λdnt,b
)φapdn
t
]
rdepl = πaπdηdsg∑t∈T
∑v∈V
∑b∈B
(vup
t,b λRTt,b
)(eregup
t,v + 1ηdsg e
stopdsgt,v
)
3M. A. Ortega-Vazquez, F. Bouffard and V. Silva, “Electric Vehicle Aggregator/System Operator Coordination for ChargingScheduling and Services Procurement,” IEEE Transactions on Power Systems, Vol. 28, Issue 2, pp. 1806-1815, May 2013. [url]
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 11 / 21
Proposed approach Mathematical model
Aggregator’s Model
Aggregator strives to maximize their profits:
max{rem + rcap + rdepl − cregup − cregdn − cdeg
}Revenues3:
rem = ∆t∑t∈T
∑v∈V
λDAt
(ηdsg
v · pemdsgt,v − pemchg
t,v
)
rcap =∑t∈T
∑b∈B
[(wup
t,b λupt,b
)πapup
t +(wdn
t,b λdnt,b
)φapdn
t
]
rdepl = πaπdηdsg∑t∈T
∑v∈V
∑b∈B
(vup
t,b λRTt,b
)(eregup
t,v + 1ηdsg e
stopdsgt,v
)3M. A. Ortega-Vazquez, F. Bouffard and V. Silva, “Electric Vehicle Aggregator/System Operator Coordination for Charging
Scheduling and Services Procurement,” IEEE Transactions on Power Systems, Vol. 28, Issue 2, pp. 1806-1815, May 2013. [url]Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 11 / 21
Proposed approach Mathematical model
Aggregator Model
Costs:cregup = πaπdπshort∑
t∈T
∑b∈B
(vup
t,b λRTt,b
) (pup
t − πaπdpupt
)
cregdn = φaφdφshort∑t∈T
∑b∈B
(vdn
t,b λRTt,b
) (pdn
t − φaφdpdnt
)
cdeg =∑v∈V
Cbatv
∣∣∣∣mv
100
∣∣∣∣∆t
∑t∈T
(pemdsg
t,v + pemchgt,v
)− ξv
BCv
+∑
t∈T
(πaeregup
t,v + φaeregdnt,v
)BCv
{cdeg}4
4M. A. Ortega-Vazquez, “Optimal Scheduling of Electric Vehicle Charging and Vehicle-to-Grid Services at Household LevelIncluding Battery Degradation and Price Uncertainty,” IET Generation, Transmission & Distribution, Vol.8 , Issue 6, pp.1007–1016, Jun. 2014. [url]
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 12 / 21
Proposed approach Mathematical model
Aggregator Model
Costs:cregup = πaπdπshort∑
t∈T
∑b∈B
(vup
t,b λRTt,b
) (pup
t − πaπdpupt
)
cregdn = φaφdφshort∑t∈T
∑b∈B
(vdn
t,b λRTt,b
) (pdn
t − φaφdpdnt
)
cdeg =∑v∈V
Cbatv
∣∣∣∣mv
100
∣∣∣∣∆t
∑t∈T
(pemdsg
t,v + pemchgt,v
)− ξv
BCv
+∑
t∈T
(πaeregup
t,v + φaeregdnt,v
)BCv
{cdeg}4
4M. A. Ortega-Vazquez, “Optimal Scheduling of Electric Vehicle Charging and Vehicle-to-Grid Services at Household LevelIncluding Battery Degradation and Price Uncertainty,” IET Generation, Transmission & Distribution, Vol.8 , Issue 6, pp.1007–1016, Jun. 2014. [url]
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 12 / 21
Proposed approach Mathematical model
Aggregator Model
Costs:cregup = πaπdπshort∑
t∈T
∑b∈B
(vup
t,b λRTt,b
) (pup
t − πaπdpupt
)
cregdn = φaφdφshort∑t∈T
∑b∈B
(vdn
t,b λRTt,b
) (pdn
t − φaφdpdnt
)
cdeg =∑v∈V
Cbatv
∣∣∣∣mv
100
∣∣∣∣∆t
∑t∈T
(pemdsg
t,v + pemchgt,v
)− ξv
BCv
+∑
t∈T
(πaeregup
t,v + φaeregdnt,v
)BCv
{cdeg}4
4M. A. Ortega-Vazquez, “Optimal Scheduling of Electric Vehicle Charging and Vehicle-to-Grid Services at Household LevelIncluding Battery Degradation and Price Uncertainty,” IET Generation, Transmission & Distribution, Vol.8 , Issue 6, pp.1007–1016, Jun. 2014. [url]
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 12 / 21
Proposed approach Mathematical model
Constraints
EVs’ constraints:Minimum energy for motion requirements (ξv)State-of-charge (SoC) dynamicsMinimum/maximum SoC (battery preservation)Charging and discharging ratesEVs’ availabilities
Aggregator’s constraints5:Additional energy requirements for service procurementAggregated bids/offers for energy marketsPQP curve constraints
5M. R. Sarker, Y. Dvorkin and M. A. Ortega-Vazquez, “Optimal Participation of an Electric Vehicle Aggregator inDay-Ahead Energy and Reserve Markets,” IEEE Transactions on Power Systems, Vol. PP, Issue 99, pp. XX, 2106, (earlyaccess). [url]
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 13 / 21
Proposed approach Mathematical model
Constraints
EVs’ constraints:Minimum energy for motion requirements (ξv)State-of-charge (SoC) dynamicsMinimum/maximum SoC (battery preservation)Charging and discharging ratesEVs’ availabilities
Aggregator’s constraints5:Additional energy requirements for service procurementAggregated bids/offers for energy marketsPQP curve constraints
5M. R. Sarker, Y. Dvorkin and M. A. Ortega-Vazquez, “Optimal Participation of an Electric Vehicle Aggregator inDay-Ahead Energy and Reserve Markets,” IEEE Transactions on Power Systems, Vol. PP, Issue 99, pp. XX, 2106, (earlyaccess). [url]
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 13 / 21
Proposed approach Mathematical model
Constraints
EVs’ constraints:Minimum energy for motion requirements (ξv)State-of-charge (SoC) dynamicsMinimum/maximum SoC (battery preservation)Charging and discharging ratesEVs’ availabilities
Aggregator’s constraints5:Additional energy requirements for service procurementAggregated bids/offers for energy marketsPQP curve constraints
5M. R. Sarker, Y. Dvorkin and M. A. Ortega-Vazquez, “Optimal Participation of an Electric Vehicle Aggregator inDay-Ahead Energy and Reserve Markets,” IEEE Transactions on Power Systems, Vol. PP, Issue 99, pp. XX, 2106, (earlyaccess). [url]
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 13 / 21
Case Study Test system
Case study
Aggregator:Fleet of 1000 EVs with driving patterns from the 2009 NHTSAverage battery capacity (Bv) of 24 kWh0.15·Bv ≤ eSoCt,v ≤ 0.95·Bv and random eSoC0Battery degradation characteristic from A1236
V2G & G2V ∈ [0, 3.3] kW and ηR = 90%
System:10-step PQP curves created using historical data from ERCOT for theenergy and reserve markets, assuming π = φModified IEEE RTS-96 three-area system (13620 MW), with 3 windfarms (2400 MW)Optimization horizon of 24 h
MILP solved in GAMS using CPLEX
6A123, ‘AMP20 Lithium Ion Prismatic Cell.’ Available at: http://www.a123systems.com/prismatic-cell-amp20.htmOrtega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 14 / 21
Case Study Test system
Case study
Aggregator:Fleet of 1000 EVs with driving patterns from the 2009 NHTSAverage battery capacity (Bv) of 24 kWh0.15·Bv ≤ eSoCt,v ≤ 0.95·Bv and random eSoC0Battery degradation characteristic from A1236
V2G & G2V ∈ [0, 3.3] kW and ηR = 90%
System:10-step PQP curves created using historical data from ERCOT for theenergy and reserve markets, assuming π = φModified IEEE RTS-96 three-area system (13620 MW), with 3 windfarms (2400 MW)Optimization horizon of 24 h
MILP solved in GAMS using CPLEX
6A123, ‘AMP20 Lithium Ion Prismatic Cell.’ Available at: http://www.a123systems.com/prismatic-cell-amp20.htmOrtega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 14 / 21
Case Study Test system
Case study
Aggregator:Fleet of 1000 EVs with driving patterns from the 2009 NHTSAverage battery capacity (Bv) of 24 kWh0.15·Bv ≤ eSoCt,v ≤ 0.95·Bv and random eSoC0Battery degradation characteristic from A1236
V2G & G2V ∈ [0, 3.3] kW and ηR = 90%
System:10-step PQP curves created using historical data from ERCOT for theenergy and reserve markets, assuming π = φModified IEEE RTS-96 three-area system (13620 MW), with 3 windfarms (2400 MW)Optimization horizon of 24 h
MILP solved in GAMS using CPLEX
6A123, ‘AMP20 Lithium Ion Prismatic Cell.’ Available at: http://www.a123systems.com/prismatic-cell-amp20.htmOrtega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 14 / 21
Case Study Results
Probability of acceptance/deployment (πa/πd)
Monte Carlo (MC) simulations for wind and load scenariosExploration of all combinations of {πa, πd}Total profits for each combination are compared
Total profit (p.u.,$)
CD
F
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.2
0.4
0.6
0.8
1
πa = 0.1,πd = 0.8
πa = 0.4,πd = 0.3
πa = 0.7,πd = 0.3
πa = 0.8,πd = 0.3
πa = 0.9,πd = 0.3
πa = 1.0,πd = 0.3
πa = 0.8,πd = 0.5
πa = 0.9,πd = 0.8
πa = 1.0,πd = 0.8
πa = 0.9 and πd = 0.8, yield the largest profits (CDF)
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 15 / 21
Case Study Results
Probability of acceptance/deployment (πa/πd)
Monte Carlo (MC) simulations for wind and load scenariosExploration of all combinations of {πa, πd}Total profits for each combination are compared
Total profit (p.u.,$)
CD
F
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.2
0.4
0.6
0.8
1
πa = 0.1,πd = 0.8
πa = 0.4,πd = 0.3
πa = 0.7,πd = 0.3
πa = 0.8,πd = 0.3
πa = 0.9,πd = 0.3
πa = 1.0,πd = 0.3
πa = 0.8,πd = 0.5
πa = 0.9,πd = 0.8
πa = 1.0,πd = 0.8
πa = 0.9 and πd = 0.8, yield the largest profits (CDF)
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 15 / 21
Case Study Results
Cost/Benefit analysis
Battery cost sensitivity over {500, 450, 350, 250} $/kWh
Battery cost, Cbatv ($/kWh)
(a)
Ex
pec
ted r
even
ue
($)
2503504505500
200
400
600
Energy arbitrage
DA capacity
RT excercise
Battery cost, Cbatv ($/kWh)
(b)
Exp
ecte
d c
ost
($
)
250350450550−100
−80
−60
−40
−20
0
Degradation: Energy
Degradation: REG
REG penalty
Exp
ecte
d t
ota
l p
rofi
t ($
)
Battery cost, Cbatv ($/kWh)
(c)
2503504505500
100
200
300
400
500
At high BCv the difference in energy market prices is attractiveAt low BCv the degradation costs for participation in regulation are‘affordable’
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 16 / 21
Case Study Results
Offering strategy
0 2 4 6 8 10 12 14 16 18 20 22 2415
20
25
30
35
En
ergy p
rice
($
/MW
h)
Time (h)(a)
Day-ahead (∑
b λDAt,b )
Real-time (∑
b vupt,b · λ
RTt,b )
0 2 4 6 8 10 12 14 16 18 20 22 240
2
4
6
8
10
12
Cap
acit
y p
rice
($/M
W)
Time (h)(b)
Down (∑
b wdnt,b · λ
dnt,b)
Up (∑
b vupt,b · λ
upt,b)
Time (h)(c)
Day
−ah
ead
off
ers
(kW
)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
−3500
−2500
−1500
−500
500
1500
2500
3500
0.2
0.4
0.6
0.8
1
Day
−ah
ead t
ota
l eS
oC
(p.u
., k
Wh
)
EMCHG REGDN STOPDSG EMDSG REGUP STOPCHG∑
eSoC
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 17 / 21
Case Study Results
Benefits to the system
Can EV participation aid the power system?
This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
SARKER et al.: OPTIMAL PARTICIPATION OF AN ELECTRIC VEHICLE AGGREGATOR 9
TABLE ISYSTEM OPERATOR'S COSTS
(37.9%) with a deviation of 0.24 MW (22.1%) was deployed inthe RT. The aggregator favors up regulation because it allowsboth the capacity and exercise revenues to be obtained. Majorreasons for lower quantities of down regulation is caused bythe aggregator's commitment to procuring energy for transporta-tion, which limits the EV fleets capacity for additional chargingfor down regulation, and also because only the capacity rev-enue can be obtained. On the other hand, STOPDSG portionof down regulation can only be activated if energy market dis-charging (EMDSG) occurs. However, because REGUP servicesare profitable, this limits EMDSG from occurring often and soSTOPDSG is limited.For purposes of simplicity, it was assumed the probabilities
of acceptance and deployment were equal, i.e. and. Different values, however, can be chosen for the
down regulation which better resemble the outcome of Fig. 7(b).At the same time, this also shows the SO requires less downregulation.
D. System Operator's PerspectiveTable I uses the MC trials to show the SO's expected oper-
ating cost, standard deviation of cost, and the startup cost ofcommitting additional units in the RT, which is compared to thecost of DA commitments. The base case in Table I presents thecosts in the power system without the aggregator. Next, the casewhen the aggregator participates in energy markets only and fi-nally, the case when aggregator partakes in both markets. FromTable I, the SO's expected costs in the RT are reduced whenthe aggregator provides services to the grid. Also, the startupcosts in the DA and RT decrease. Even though the total quan-titative cost savings seem low, e.g. 0.08% when the aggregatorparticipates in energy market and 0.12% when performing inboth markets the qualitative benefits are of importance [34], andthese would be mirrored as large amounts of money over an op-erating year.The decrease in the start-up costs shows that less cycling of
conventional generation occurs in both the DA and RT [35]. Es-pecially in the RT, the lower start-up costs indicate the SO re-quires less fast-starting units to be on stand-by in the case ofdeviations. This follows because the aggregator obtains energywhen there is an abundance and less need for deployment, andthen supplies it when there is a need, thus making it a viablealternative to conventional generation for reserve provision. Ascompared to the conventional generation, the aggregator has es-sentially no startup costs and also has lower operating costs,
which only include the compensation of the battery degradationto the EVs owners.
V. CONCLUSIONElectric Vehicle (EV) aggregators are the required mediators
between large fleets of EVs and the power system operators.EVs can provide a new stream of services to the power system,however in order to incentivize the EVs' participation in energyand reserve markets, a fair compensation mechanism must inplace to reimburse for degrading batteries beyond their primarypurpose of transportation. This work proposes a framework todetermine the optimal aggregator’s bidding strategy in the en-ergy and regulation reserve markets, which maximizes the ag-gregator's profits while observing the incurred loss of utility forthe EV batteries. In this framework, the aggregator optimizesits strategy as a risk-averse, profit-seeking entity in a competi-tive energy and reserve market, while considering its expectedprobability of acceptances and probability of deployments forup and down regulation.Results show that the aggregator benefits from the reserve
market more than the energy market for two main reasons: 1) itcollects capacity revenue for providing regulation, which doesnot incur degradation, and 2) it gains additional revenue if re-quired to deploy in the real-time. By comparing the total rev-enues with battery costs as of 2012 to costs in the future, theaggregator would obtain a substantial increase in revenue andat the same time would reduce significantly the power systemoperating costs. This is the case because as the cost of the bat-tery decreases, using the EVs as storage and reserve providersbecomes profitable. When the battery costs are high, most of therevenue is obtained from the energy market, however, with lowbattery costs most of the revenues come from regulation reserveprovision. This is because as the battery costs decrease, the pro-vision of regulating reserve would result into two streams ofrevenue: capacity and deployment. The provision of these ser-vices from EVs is also beneficial to the System Operator, sinceit would reduce the total operating costs of the system.
APPENDIX ALINEARIZATION TECHNIQUE
The multiplication of binary variable and continuous vari-able renders an optimization problem non-linear. The lin-earization is performed by introducing a new continuous vari-able that takes on the resulting value of the multiplication, asshown in (25):
(25)
In order to linearize (25), the following constraints are needed:
(26)(27)(28)
where is a large number. For example, if , then ac-cording to (28), . However, if , then equation (28)is non-binding. In addition, in (26), and in (27), ,and thus is obtained. By using constraints (26)–(28), thevariable can either take on the value of 0 or .
With EV participation, total system costs decreaseDecrease in the start-up costs show less cycling of conventionalgeneration occurs in both the day-ahead and real-time
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 18 / 21
Conclusions
Conclusions
Aggregators are required mediators between EV owners and SOOffer new streams of competitive services to the power systemProposed probabilistic framework for optimal participation in energyand reserve markets, influenced by:
Market structure (i.e. revenues and costs)EVs’ battery degradationExpectations of bids/offers acceptance and deployment in the differentmarkets
Mutually beneficial for all players (i.e. SO, Aggregator, EV owners)
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 19 / 21
Conclusions
Conclusions
The participation on the different markets is highly dependent on thebattery cost (BCv):
At high BCv the difference in energy market prices might be attractiveAt low BCv the degradation costs for participation in regulation are‘affordable’
Up reserve provision profitable due to two revenue streams: capacityand deployment
Arbitrage is performed between markets, i.e.Energy is purchased during low-price periodsThen, scheduled for up reserves
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 20 / 21
Conclusions
Source
This presentation is based on:M. R. Sarker, Y. Dvorkin and M. A. Ortega-Vazquez, “OptimalParticipation of an Electric Vehicle Aggregator in Day-Ahead Energyand Reserve Markets,” IEEE Transactions on Power Systems, Vol. PP,Issue 99, pp. XX, 2106, (early access). [url]
Acknowledgment:This work was supported in part by the National Science Foundation(NSF) under Grant No. 1239408In part also by the University of Washington’s Clean Energy Institute(CEI) [url]
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 21 / 21
Conclusions
Source
This presentation is based on:M. R. Sarker, Y. Dvorkin and M. A. Ortega-Vazquez, “OptimalParticipation of an Electric Vehicle Aggregator in Day-Ahead Energyand Reserve Markets,” IEEE Transactions on Power Systems, Vol. PP,Issue 99, pp. XX, 2106, (early access). [url]
Acknowledgment:This work was supported in part by the National Science Foundation(NSF) under Grant No. 1239408In part also by the University of Washington’s Clean Energy Institute(CEI) [url]
Ortega-Vazquez et al. (EE UW) EV Aggregator: Energy & Reserve June 27–29, 2016 21 / 21
Recommended