Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable...

42
Unit Commitment for Sustainable Integra4on of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP, Carnegie Mellon Technical Conference on Unit Commitment SoGware Docket AD10‐12 FERC, Washington, DC June 2‐3, 2010

Transcript of Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable...

Page 1: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Unit Commitment for Sustainable Integra4on of  Large‐Scale Wind Power and  Responsive Demand  

©Marija Ilic ECE and EPP, Carnegie Mellon  

Technical Conference on Unit Commitment SoGware Docket AD10‐12 

FERC, Washington, DC June 2‐3, 2010 

Page 2: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Acknowledgment Electric Energy Systems Group (EESG)  

hTp://www.eesg.ece.cmu.edu •  A  mul4‐disciplinary group of researchers from across Carnegie Mellon with common interest in electric energy. 

•  Truly integrated educa4on and research 

•  Interests range across technical, policy,  sensing, communica4ons, compu4ng and much more; emphasis on systems aspects of the changing industry, model‐based simula4ons and decision making/control for predictable performance. 

Page 3: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Outline •  Wind  power and demand side management—new sources of complexity 

•  Problem formula4on for integra4ng wind and elas4c demand—centralized UC/MPC; an example 

•  Scalability problem in centralized UC and UC market implementa4on—common challenge/approach 

•  Temporal vs spa4al LR issue ‐‐basic efficiency/emissions  concern with today’s UC 

•  Managing complexity—Distributed Interac4ve Unit Commitment (DIUC) 

•  Numerical example/comparison with benchmark solu4on‐ Integra4on of >50% wind  

Page 4: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

New source of complexity •  Supply the expected load with whatever produced by 

intermiTent resources combined with other tradi4onal power plants. 

__  = 

 Today: Choose output levels from conven4onal power plants to meet the expected  “net load” at minimum cost. 

Page 5: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Unit commitment for  sustainable integra4on of wind and demand side 

•  Net system demand func4on of system imbalance; greatly complicates the  unit commitment problem 

   ‐number of  states  poten4ally huge in inelas4c demand case; 

 ‐ number of decision variables poten4ally huge in the demand elas4c case.  

•  Two  qualita4vely  different methods”   

    ‐Centralized UC‐‐‐Demand, wind and genera4on cost func4ons short‐run marginal cost;  UC constraints explicit at the ISO level.  

    ‐ Distributed Interac4ve UC (DIUC) ‐When demand and  cost func4ons account for UC constraints (DIUC) [1—6]; computa4onally  manageable and provable lower bound performance under certain condi4ons 

Page 6: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Mo4va4on 

BeTer Predic4on of IntermiTent Resources? 

More efficient u4liza4on of intermiTent resources 

More reliable opera4on of intermiTent resources 

Page 7: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Managing wind power—our approach [5]  

•  Ac4vely control the output of available intermiTent resources to follow the trend of 4me‐varying loads. 

•  By doing so, the need for expensive fast‐start fossil fuel units is reduced. Part of the load following is done via intermiTent renewable genera4on. 

•  The  technique used  for implemen4ng this approach is called model predic4ve control (MPC) [5]. 

•  Implicit value of  storage 

Page 8: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Centralized MPC ‐‐Benchmark 

•  Predic4ve model of load and intermiTent resources are necessary. 

•  Op4miza4on objec4ve: minimize the total genera4on cost. 

•  Horizon: 24 hours, with each step of 5 minutes. 

Predic4ve Model and MPC Op4mizer 

Electric Energy System 

Page 9: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Centralized MPC—Problem Formula4on 

Page 10: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,
Page 11: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

11 

Model Predic4ve Control—Based  

www.jfe‐rd.co.jp/en/seigyo/img/figure04.gif 

•  MPC is receding‐horizon op4miza4on based control. •  At each step, a finite‐horizon op4mal control problem is solved but only one step is implemented. •  MPC has many successful real‐world applica4ons. 

Page 12: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

12 

Numerical Example 

Compare the outcome of ED from both the conven<onal and proposed approaches. 

Page 13: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

13 

Conventional cost over 1 year *

Proposed cost over the year

Difference Relative Saving

$ 129.74 Million $ 119.62 Million $ 10.12 Million

7.8%

*: load data from New York Independent System Operator, available online at hTp://www.nyiso.com/public/market_data/load_data.jsp 

Page 14: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Rethinking the problem— limits to complexity 

•  Clear that today’ MPC algorithms not scalable as of now; 

•  Typical approxima4on—LR  w.r.t.  4me; assumes hierarchical 4me scale separa4on of   SCED and UC 

•   LR w.rt. to 4me  ques4onable  with persistent changes in system inputs;  complexity will grow with new technologies (physics vs. binary decisions);  

Page 15: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

The basic efficiency/emissions concern   

•  Natural “op4mal” schedules of different technologies very different;  horizontal vs. ver4cal  unit commitment  

•  Brute‐forcing all technologies to re‐commit   using common  clock is  generally  very inefficient   

•  Two possible solu4ons:    ‐ centralized  mul4‐stage  UC over several 4me horizons; s4ll very complex 

 ‐interac4ve UC   between the resources and system operators; no LR  wrt 4me; LR wrt units  

Page 16: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

• Decomposi4on over (poruolia) of resources cri4cal; otherwise, the problem is not solvable; this is par4cularly true when combining with  line switching • Proposed  new approach‐Distributed Interac4ve UC (DIUC)  [2‐7] 

 ‐LR w.r.t.   (poruolia of)  units  NOT   w.r.t. 4me   ‐for the  changing industry (distributed, 

interac4ve)  ‐for the regulated industry (centralized 

algorithm) 

Page 17: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Basic idea of minimally coordinated self‐dispatch—Distributed Interac4ve UC (DIUC)  

•  Different technologies perform  look‐ahead decision making  given their unique temporal and spa4al  characteris4cs and system signal (price or  system net demand); they create  bids and are cleared by the  layers of coordinators 

•  Puvng Auc4ons  to  Work in Future Energy  Systems 

•  We illustrate next a  supply‐demand balancing process  in an energy system with wind, solar, conven4onal genera4on, elas4c demand, and PHEVs.  

Page 18: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Our Proposed Framework: DIUC 

18 

Genera4on  Load 

Look‐ahead Dispatch with  Ac4ve Load Management 

Model Predic4ve Control (MPC) 

Page 19: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

At Generator Level—Primal Problem 

Max(Total Profit) 

Predicted price 

Gen Capacity Constraint 

Ramp Rate Constraints 

Available Genera4on Predic4on 

Price 

Power (MW) 

Note: In regulated industry –predicted net system demand 

Page 20: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Main ideas of Adap4ve Load Management (ALM) 

•  Reflec4ng various end‐users’ needs and  preferences into demand response –  End‐users’ info on preference sent to system –  Mapping physical preference into economic preference 

 demand func4on  

•  (current systems) top‐down control of loads   (future systems) two‐way communica4ve and adap4ve control 

•   Load aggregators’ role –  Mediator between system/market and end‐users –  Value of aggrega<ng different resources and risk management 

•  Different load profiles, inelas4c and elas4c demands, distributed energy  resources (DER), etc. 

20 

Page 21: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Today’s  demand response scheme:  Direct load control 

•  One‐way flow of informa4on –  Load management conducted 

by u4li4es –  Top‐down control 

•  Exclusive contracts between supply and demand 

•  Direct load control –  Regardless of end‐users’ 

preferences –  No access to market 

informa4on for end‐users –  End‐users’ informa4on invisible 

to system 

21 

Page 22: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

 Adap4ve Load Management (ALM) –Look‐ahead distributed self‐dispatch 

•  Problem setup –  10 end‐users with different 

temperature preferences –  Op4mizing energy usage 

over 24 hours •  Hourly‐varying electricity price given (real‐4me pricing) 

•  Outdoor weather temperature given 

End­user index 

Temperature setpoints (⁰F) 

1  68  75 

2  70  77 

3  72  75 

4  74  79 

5  75  80 

6  64  75 

7  63  77 

8  72  79 

9  72  77 

10  73  81 

22 

10 

15 

1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 

price (¢/kWh) 

hour 

Hourly real‐<me pricing rates 

Page 23: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Primal problem at the demand level 

•  Obtaining individual demand func4on subject to temperature comfort level 

   where 

   subject to                       for all  

•  Obtain different          s for different       s to infer demand func4ons    PHYSICS AT PLAY 

23 

Timin ≤ Ti[k] ≤ Ti

max

Page 24: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Demand func4on •  Objec4ve 

–  To model price‐responsive loads to integrate into the system op4miza4on 

–  To include informa4on of end‐users’ u4lity (benefit) in system op4miza4on 

–  To see if price‐responsive loads compensate with vola4le intermiTent resources 

•  What it is –  Func4on of end‐users’ willingness‐to‐pay with respect to electricity 

demand quan4ty 

Page 25: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Demand func4on (cont’d) 

•  How to obtain –  Calculate op4mal energy usage by hours with a given electricity price 

–  Perturb the given price by a certain percentage (e.g. ±20%) and re‐calculate op4mal energy usage with new prices 

–  Curve‐fit price‐demand quan4ty pairs to iden4fy the parameters of a demand func4on 

Demand  (kW)  

Unit price (cents/kWh)  

Page 26: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Resul4ng bid curves by demand—Result of solving primal problem 

•  Demand func4ons of end‐user #1 

26 

0 10 20 30 40 50 

0  0.1  0.2  0.3  0.4 

WTP

 (¢/kWh) 

energy usage (kWh) 

hour 13 

0 10 20 30 40 50 

0  0.1  0.2  0.3  0.4 

WTP

 (¢/kWh) 

energy usage (kWh) 

hour 14 

0 10 20 30 40 50 

0  0.1  0.2  0.3  0.4 

WTP

 (¢/kWh) 

energy usage (kWh) 

hour 15 

0 10 20 30 40 50 

0  0.1  0.2  0.3  0.4 

WTP

 (¢/kWh) 

energy usage (kWh) 

hour 16 

0 10 20 30 40 50 

0  0.1  0.2  0.3  0.4 

WTP

 (¢/kWh) 

energy usage (kWh) 

hour 17 

0 10 20 30 40 50 

0  0.1  0.2  0.3  0.4 

WTP

 (¢/kWh) 

energy usage (kWh) 

hour 18 

0 10 20 30 40 50 

0  0.1  0.2  0.3  0.4 

WTP

 (¢/kWh) 

energy usage (kWh) 

hour 19 

0 10 20 30 40 50 

0  0.1  0.2  0.3 0.4 

WTP

 (¢/kWh) 

energy usage (kWh) 

hour 20 

Page 27: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Informa4on flow of ALM—Auc4ons Put to Work (Primal‐Dual Solu4on Interac4ons) 

27 

Secondary level 

Primary level 

Ter4ary level 

Demand func<on  End‐user price 

End‐user 

Load aggregator I

Bid func<on 

Market price 

b(λ) 

λ 

b(λI)  λI

Load aggregator II Load aggregator III

Page 28: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Informa4on flow of ALM 

•   Primary layer (from end‐users to load aggregators)  –  Physical preference  economic preference –  Individual demand func4on 

•   Secondary layer (from load aggregators to market) –  Aggrega4ng end‐users’ energy usage + risk management –  Op4mal energy purchase/market transac4on given system price 

•   Back to primary layer (from load aggregators to end‐users) –  Energy price adjusted according to system/loca4onal price : λLA 

as a func4on of λsystem 

28 

Page 29: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Bid curves for different technologies—result of distributed MPC  

Page 30: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Implementa4on in  Markets—DIUC‐‐‐Primal‐Dual Solu4on Interac4ons (spa4al/not temporal) 

Predic4ve Model ] and MPC Op4mizer 

Generator i 

The System Operator: Maximize Social Welfare While Observing Transmission Constraints   

30 

Aggregated Predic4ve model] and MPC 

Op4mizer 

Load j 

[2] N. Abdel‐Karim and  M. Ilic,"Short Term Wind Speed Predic4on by Finite and Infinite Impulse Response Filters:  A State Space Model Representa4on Using Discrete Markov Process", IEEE PowerTech Conference, Romania June 2009 [3] J. Joo and M.D. Ilic, “A Mul4‐Layered Adap4ve Load Management (ALM) system: informa4on exchange between market par4cipants for efficient and reliable energy use,” submiTed to 2010 IEEE PES Transmission and Distribu4on Conference. 

Page 31: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Integra4ng fossil, wind, solar and demand side  and PHEVs  

Compare the outcome of ED from both the centralized and distributed MPC approaches. 

Page 32: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

BOTH EFFICIENCY AND RELIABILITY MET  

Page 33: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Preliminary Results: 50% Wind 

33 [5] Marija Ilic, Le Xie, and Jhi‐Young Joo, “Dynamic Monitoring and Decision Systems (DYMONDS) for Efficient Integra4on Wind Power and Price‐Responsive Demand: Proof of Concept on the IEEE RTS System”, EESG WP 2009. 

Page 34: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Preliminary Results: 50% Wind 

34 [5] Marija Ilic, Le Xie, and Jhi‐Young Joo, “Dynamic Monitoring and Decision Systems (DYMONDS) for Efficient Integra4on Wind Power and Price‐Responsive Demand: Proof of Concept on the IEEE RTS System”, EESG WP 2009. 

Page 35: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Poten4al Savings 

35 [5] Marija Ilic, Le Xie, and Jhi‐Young Joo, “Dynamic Monitoring and Decision Systems (DYMONDS) for Efficient Integra4on Wind Power and Price‐Responsive Demand: Proof of Concept on the IEEE RTS System”, EESG WP 2009. 

Page 36: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Op4mal Control of Plug‐in‐Electric Vehicles: Fast vs. Smart  

36 

Page 37: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Informa4on flow—Fantas4c Use of Mul4‐layered Dynamic Programming 

37 

Page 38: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Plug‐and‐Play (No Coordina4on)? 

Page 39: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Smart Grid [1]  

Page 40: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Distributed Interac4ve UC‐‐DIUC •  Allow  different technologies to op4mize in an4cipa4on of system condi4ons while taking into account their unique inter‐temporal constraints and sub‐objec4ves;  

•  The distributed op4miza4on done in a look‐ahead dynamic way at the resource level to create cost func4ons (bidding func4ons for both supply and demand) 

•  Result: Lower electricity prices  and higher efficiency; minimized  overall UC cost  

Page 41: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

Our general framework 

•   Use adequate performance measures •  Perform meaningful spa4al  decomposi4on and  

aggrega4on (zoned, poruolios)  •  Iterate to  ensure that the inter‐dependencies are 

accounted for by inter‐ac4vely exchanging  informa4on    ‐Even coordinated soGware  must account for the informa4on that is itera4vely updated from the non‐u4lity resources 

   ‐   In the market environment it becomes necessary to reconcile choice and system‐wide objec4ves  

    ‐Use  predic4ons  create bid  func4ons (instead of numbers) to control  soGware performance 

Page 42: Unit Commitment for Sustainable Integraon of Large‐Scale ......Unit Commitment for Sustainable Integraon of Large‐Scale Wind Power and Responsive Demand ©Marija Ilic ECE and EPP,

References •  [1] Electric Energy Systems Group (EESG) at CMU  hTp://eesg.ece.cmu.edu •   Ilic, Marija “Dynamic Monitoring and Decision Systems (DYMONDS) and 

Smart Grids: One and the Same, CMU EESG WP 019, October 2009.  •  [2] Ilic, Marija “IT‐enabled Rules, Right and Responsibili4es (3Rs) for 

Efficient Integra4on of Wind and Demand Side Response”, Public U4lity Fortnightly Magazine, Dec 2009.  

•  [3] M. Ilić, J.Y. Joo, L. Xie, M. Prica, and N. Rotering, A Decision Making Framework and Simulator for Sustainable Electric Energy Systems, IEEE Transac4ons on Sustainability, under review – submi=ed Jan 2010  

•  [4] M. Ilić, L. Xie, and J.Y. Joo, Efficient CoordinaDon of Wind Power and Price‐Responsive Demand Part I: TheoreDcal FoundaDons, Parts I and  II: Case Studies, IEEE Transac4ons on Power Systems, under review – accepted Dec 2009 

•  [5] L. Xie and M. Ilić, Model PredicDve Economic/Environmental Dispatch of Power Systems with Intermi=ent Resources, IEEE PES General Mee4ng, July 2009 

•  [6] J.Y. Joo and M. Ilić, A MulD‐Layered AdapDve Load Management (ALM) System, IEEE PES Transmission and Distribu4on Conference, April 2010. 

•  [7]  N. Rotering and M. Ilić, OpDmal Charge Control of Plug‐In Hybrid Electric Vehicles In Deregulated Electricity Markets, IEEE Transac4ons Power Systems, accepted 2010