Look-ahead Coordination of Wind Energy and

Electric Vehicles: A Market-based Approach

Yingzhong Gu, Student Member, IEEE, Le Xie, Member, IEEE Department of Electrical Engineering

Texas A&M University College Station, United States

Abstract-The major subject of this paper is the analysis and simulation of market-based coordination of variable resources with storage systems such as aggregated Plug-in Hybrid Electric Vehicles ( PHEVs). Starting from our recent work on modeling and model predictive scheduling of conventional generation, variable generation, and battery storage systems, we further develop in this paper a look-ahead multi-layered simulation platform which ( 1) takes into account of the transmission congestion; and (2) co-optimizes individual participants' profits from both energy and regulation services markets. In contrast with today's operation software, the proposed scheduling framework could improve the overall system efficiency while observing the transmission network constraints. We illustrate the improved operational efficiency in a modified IEEE 14-bus system.

Keywords: Wind generation; Power markets; Transmission

congestion; Vehicle-to-Grid; Model predictive control; Renewable

energy; Regulation market;

I. INTRODUCTION

Renewable energy resources such as wind and solar are playing a major role in meeting the increasing demand for electricity, reducing reliance on foreign fuel as well as significantly decreasing greenhouse gas (GHG) emissions. Many countries and regions plan to increase the percentage of installed wind power [1]. However, the inter-temporal variability and limited predictability have posed significant challenges to the Regional Transmission Organizations (RTOs) [1 ].

A lot of research has been devoted to more efficient and reliable utilization of large-scale variable resources such as wind power. On the wind generation side, researchers have proposed many strategies to better predict and control the wind output for smoothing out the variations [2-5]. Automatic Generation Control (AGC) of a Wind Farm with Variable Speed Wind Turbines is discussed in [3-4]. The pitch angle control for leveling the wind output power was studied in [2] and [5].

On the demand side, important progress has been made on large-scale deployment of electric vehicles such as Plug-in Hybrid Electric Vehicle (PREV) to serve as energy storage resources in the power grid[6]. The concept of utilizing electric vehicles in cities as battery energy storage to increase the flexibility in power system operation and control has been proposed almost two decades ago [6]. The economic value of

providing "Vehicle-to-Grid (V2G) " services was studied in California [7-9]. In 2008, Turton describes the detailed and global analysis of the potential of V2G technology over the long-term [10]. In 2010, Andersson performs the case studies of Plug-in hybrid electric vehicles as regulating power providers in Sweden and Germany [11].

The main focus of our work is to develop models and software platform to understand the interactions among PREYs, wind farms, conventional power sources, and loads in a market-based approach. Our previous work has studied the model predictive control (MPC)- based coordination of wind farms and battery energy storage systems [12]. However, the individual market participants were assumed to be price takers. In this paper, we propose a look-ahead co-optimization framework which consists of both energy balancing and frequency regulation markets.

The contributions of this paper are suggested as follows:

1) Transmission network constraints are explicitly modeled in this simulation platform.

2) Distributed intelligent decisions are performed at individual participants' level, and the system operator performs security-constrained economic dispatch to obtain the energy and regulation scheduling for all the market participants.

3) All the market participants are eligible to participate into both energy balancing and frequency regulation markets. The potential benefits of participating in both markets can enable a more sustainable pathway towards higher penetration of variable energy resources in future power systems.

The rest of this paper is organized as follows. Section II presents the mathematical model of the proposed two-layered look-ahead coordination framework in a market-based mechanism. In Section III we illustrate the efficacy of the proposed software platform on a modified IEEE 14-bus system. Concluding remarks and future work are discussed in Section IV.

A. Notations

II. MATHEMATIC MODEL

The variables and constants in this section are defined as follows.

Table 1 Notations

Symbols Description Ae' Electricity price in energy market

ArllP Regulation up price in ancillary market

Ardn Regulation down price in ancillary market

Cwc Marginal cost of wind generation

Cbll Battery degradation cost coefficient corresponds to output change

Cbl2 Battery degradation cost coefficient corresponds to regulation service

Cnlg regulation marginal cost for traditional sources

Cc marginal cost for traditional sources

Pvc PHEV discharging power

pcap. rllp Regulation up capacity from PHEV

pcap. rdn Regulation down capacity from PHEV

PDR Power requirement for driving profile

pel dr Power of battery for driving

pgas dr power of rCE for driving

psc" wc Scheduled wind power output

MWC net Net injection of wind power

E Energy storage level of PHEV

PvC_max Maximum output power ofPHEV

EVC_max Maximum energy storage level of PHEV A

PwC_max Predicted maximum available wind energy

pramp wc Ramping rate of wind generation

kc Time index for parking duration

kd Time index for driving duration

B. Aggregated PHEVs coordination model The optimal model of aggregated PHEVs can be

formulated as follows.

min: J = L -Ael (k) PVG (k) -L Arup (k) P;"": (k) k, k,

+LCbI2��;(k)+ LCbI2���(k) k, k,

+ L Cbll [ PVG ( k + 1) - PVG ( k ) r k,

+ LAg"s��"S (kd)+ Ael (T)E(T) kd

Subject to

(1)

E(k)-E(k -1) = -PVG (k )-17PvG (k) (2)

PDR = P:' (kd ) + Pjf."" (kd ) (3)

-PVG_max � Pvdk) � PvG_max (4)

o � P,'G (kd) � min ( Pdr, P"G_max) (5)

o � E( k) � EVG max (6)

o � P;"; � PvG_max (7)

o � �c;: � P,'G _max (8)

o � P,'G (kc) + �,:' ( kc ) � P max (9)

-P max � Py'G (kJ -��� (kJ � 0 (10)

The objective function (1) is to maximize PHEVs' total profit which consists of income in both energy market and regulation market. Then equation (2) defines the state function of PHEV charging and discharging. Equation (3) shows that the total driving profile is balanced by battery energy and gasoline energy. (4)-(10) are inequality constraints for PHEVs. Decision variables of the aggregated PHEVs model are charging/discharging power output PVG , electrical power for driving pd,e' , ICE power for driving Pd/os , and regulation power capacity Pru/ap and PrdncaP.

C. Wind farm model

min: J = L -Ae/ (k) P;� (k)+CwG (k) P;� (k) k (11)

+ LArup (k )M,,�G (k) k

Subject to

o � P;� (k) � PWG max (k) (12)

-�"mp � P;� (k)- P;� (k -1) � �"mp (13)

Equation (11) is the objective function which is the negative profit of wind farm. It consists of market revenue, marginal cost and wind output deviation penalty.

Inequality constraint (12) shows the upper and lower limits for wind generation. The minimum generation limit of wind farm is assumed to be zero. The upper limit for wind generation is the predicted maximum available wind output. Inequality constraint (13) is the ramping rate constraints for wind generators.

Wind turbines are not considered as providers of frequency regulation in this paper, so the only decision variable for wind model is the scheduled output of wind turbines PWGSCh •

D. Traditional Power Plant model In this paper, traditional power plant refers to coal power

plant, nuclear power plant and nature gas power plant. They have dispatchable outputs.

Subject to

min : J = -L [ Ael ( k ) - CG (k ) ] PG ( k ) k

-L [ Arup ( k ) - Crug ( k ) ] P,�,; ( k ) k

-L [ Ardn (k) -Crug (k ) ] P,�": (k ) k

pGrrlln ::; PG ( k ) ::; pGmax

-P,amp ::; PG (k)-PG (k -1)::; P,amp O < peal' < . ( Dmax D P ) - mp _ mm Fe -IC' ramp

O < peal' < . ( n Drrlln p ) - rdn - nun IG - ra , romp

(14)

(15)

(16)

(17)

(18)

Objective function (14) considers the energy market revenue, marginal cost and frequency regulation market revenue. Ramping rates limit the power output and the regulation power capacity. Decision variables are scheduled power output on energy balancing market P G and regulation power capacity Prup"ap and Prdncap on regulation market.

£. Formulation of Bid Curves At each time step, individual market participants performs

model predictive control as discussed in the above three subsections and determines the optimum generation/demand strategy. For each optimization process, the price is fixed. In order to form the bidding curve, we perturb the vector of expected energy and regulation price to obtain the bidding curve. We form the base price information from ERCOT data[13-14]. Then we configure the market and run 24 hours simulation to clear the market. The final market clearing prices information will be selected as the predicted reference prices for model predictive controller.

Prcr�r = [prc1,prc2, ... ,prc24] (19)

The price corresponding to current time point will be perturbed by -10%, -30%, -50%, 10%, 30%, 50%, 70%. Together with the base case price vector, a market participant can output 8 bidding points in the price-quantity plane. These points constitute the piecewise linear bidding curve for individual market participants.

F. Market Clearing Mechanism The market clearing mechanism includes two layers:

energy balancing market and regulation market. Energy balancing market is implemented through smart market code of Matpower 4.03b[15]. A standard DC OPF model is utilized. The system operator level will collect all the bidding curves and clear the market following the principle of maximizing total social welfare when considering all the facilities' constraints and transmission constraints. When determining the market prices, last accepted offer (LAO) mechanism is taken, which takes the uniform price equal to last accepted offer of generators.

The regulation market clearing mechanism is implemented in our own code. The regulation need is determined by load level and wind participation, which is inelastic demand. LAO

mechanism also is taken to set the price as last accepted offer price.

III. ILLUSTRATIVE EXAMPLES

A. Numerical Example The numerical example is modified from the standard IEEE

14-bus system in Figure 1. The generators are assigned as different kinds of power sources. Aggregated PHEVs work as a dispatchable load and generator. Simulation duration is 24 hours. MPC window size also is 24 hours. Market information of 48 hours are collected from ERCOT[13, 16]. Loads are factored out according to the portion of different buses. Regulation Prices and market prices are taken as reference for MPC.

Table 2 Data of PHEV

Para. Description value

N Total number of PHEVs participated 8000

PVG_max Maximum output power of PHEV 20kW

Emax Maximum energy storage level of PHEV 25kWh

Parameters of PHEVs are collected from [17]. 8,000 PHEV s are assumed to participate and perform as a uniformly aggregated entity. For each individual PHEV, rating power is 20kW, and battery capacity is 25kWh. PHEVs participate into energy market, regulation up market and regulation down market. Each day, PHEVs will not be available in power market at 8 a.m. and 16 p.m. because they are assumed In driving status.

Natural Gas

Figure 1 IEEE 14 Test Case

Generator parameters are factored out and modified according to [18]. Ramping rates and marginal costs are applied as is shown in Table 3.

The MPC algorithm is implemented via our own Primal Dual Interior Point Method (PDIPM) program coded in Matlab. The market clearing mechanism is absorbed from Matpower 4.0b3.

The computation environment is shown in Table 4.

Table 3 Generators Configuration

No. Bus Type Cap. Mar. cost Ramp. rate

(MW) ($fMWh) (MWflOmins)

1 I coal 150 22 36

2 1 coal 150 35 36

3 I coal 150 50 36

4 2 nuclear 350 25 10

5 3 wind 50 10 100

6 3 wind 110 15 100

7 3 wind 90 17 100

6 6 PHEVs 200 nfa nfa

7 8 gas 30 30 70

8 8 gas 40 60 70

9 8 gas 40 90 70

Table 4 Computation environment

CPU Intel Core i3 2.13MHz

Memory 4GB

Operating system Windows 7 64bit

Application program Matlab 2009a, 64bit, v7.8.0.347

B. Economic benefit analysis of PHEVs Participation The value and impacts of PHEVs' participation into the

energy balancing market and regulation market have been evaluated in this section.

.� ct

55

50 .... , ... ; ......... ,', .... , ................ ,; .. , .. , ........... ;, ....... , ...... . · . . , · , , , , . · . : : : : 45 ..... ... ;· .......... , .. ·" .. ,·,· .... ··,· .. ,i,·,·, .... · .. , .... ;· .. · .. , · , ...... ·

: :

40 .. .... .. , ............... ,j .... , ....... , .... ; .. , .............. , . ... .. ,. , ...... . , .

35 ........ , ... .. ...... , .. .. , .... , ....... , .... , '. , .............. , ........ ,. , .... .

30

: : : : . . .......... :' ................ � ........ . , ' , ,

25+---T---��--�--�--�--r-�---T--� ·150 ·100 ·50

Quantity o

Figure 2 Bidding Curve of PHEVs

50

Figure 2 shows the bidding curve of aggregated PHEVs at 11 a.m. It is clearly observed that when the price goes lower than 33$fMW, PHEVs are willing to behave as dispatchable load. When the price goes higher than 39$fMW, PHEVs are willing to act as generator to sell energy.

The results of simulation are summarized in the Table 5. It is observed that, with the participation of aggregated PHEV s,

the revenue of wind farm has been increased. The deviation penalty of wind output has been successfully decreased. Therefore, the total profit of wind farm can be increased by 21.8%.

Table 5 Economic Evaluation

Items With PHEVs Without PHEVs

Wind Revenue $55,459.75 $54,485.57

Cost $12,409.62 $12,166.07

penalty $14,367.10 $18,778.63

Profit $28,683.03 $23,540.87

PHEV PHEV Market -$4,436.76 $0

Regulation $4,054.90 $0

battery $2,140.27 $0

degradation

gasoline cost $0 $73,920

Profit -$1,758.41 -$73,920

On the PHEVs side, for 40,000 PHEVs, the gas consumption in one day is 73,920$ in all. The consumption data are calculated from [19-20]. However, if PHEVs participate into energy market and regulation market, their energy cost can be reduced to 4,436.76$. In addition, there is a profit in regulation market at 4,054.90$. Deducted the battery degradation, the total balance for PHEVs is only 1,758.41$. Therefore, with participation into power market, the PHEV s cost can be reduced by 97.62%.

700

650

600

550

500

� 450

� 400

� 350

� 300

Q; 250 � 200 a..

150

100

50

0 0

_Wind Output (Without PHEV) ...... Load Level(With PHEV) _Wind..9utput with PHEV �Load Level Without PHEV

10 15

Hours Figure 3 Load and Wind output profile

20 25

The load profile and wind output with/without PHEVs participation can be observed in Figure 3. Aggregated PHEVs have increased the load level during the night. However, the increased part does not appear at the lowest load point but appears at region near 8 a.m. One reason is that at the early night, due to the ample wind resources, PHEVs are willing to maintain their battery storage at lower level in order to provide regulation down service. Another reason is that because PHEVs' high rating value of charging power, they prefer to charge most of energy at lowest price point. Therefore, there will be load increase impulses rather than flat enhancement.

Wind output under PHEVs participation scene is slightly higher than the situation without PHEVs. This shows wind is not a significant energy provider for the increased PHEV s' load, which is because wind fann's marginal cost is lower than traditional sources, especially during night, wind fann undertakes the base load. The increased part of PHEVs is to be picked up by other traditional sources.

In Figure 4, it is shown that there is no obvious change in energy market price due to PHEVs' participation. Although the total power capacity of aggregated PHEVs is up to 200MW, the inadequate energy storage level (about 250MWh) limits the market power of PHEVs.

40

35

30

� 25 ::2 � Q) 20 g 0: 15 >-e> � 10

W

5

o 5 10 15 20

Hours Figure 4 Energy Market Prices

Figure 5 shows the increase in load level resulting from wind penetration. When PHEV s participate into the regulation market, they can occupy a portion of the market. However, due to their high battery degradation cost compared with low regulation cost of conventional power sources, aggregated PHEVs do not seem to be very competitive.

100 95 90 85 80

� 75

� 70 65

?:- 60 'u 55 ro a. 50 ro 45 () c: 40 0 35 � 30 :::J 25 Ol Q) 20 a:: 15

10 5 0

0 5 10 15 20 25

Hours Figure 5 Regulation Market Pattern

Anyway, the partICIpation of aggregated PHEVs does decrease the regulation price as is shown in Figure 6.

Sometimes the regulation price is higher than situation without PHEV. That is because at the very beginning PHEVs do not have enough energy to provide regulation service, but as a load, they increase the generation output of other power sources especially wind, which increases the total need for regulation service and thus enhances the prices.

20

� :::;: 15 � Q) u (t .2 10 16 "S 0) Q) 0::: 5

o 5 10 15

Hours

Figure 6 Regulation Market Prices

C. Impact of transmission congestion

20

The power flow of the test case which does not include transmission congestions is solved. Some important branches are selected out: Branch 2-4 is the main transmission corridor between load area and traditional power sources (i.e. coal power plant, and nuclear power plant); Branch 3-4 is the major corridor between load area and wind fann; Branch 5-6 is the branch between PHEVs and power sources. The power flows of these branches are shown in Figure 7.

130 120 1 10 100

90 80

� 70 60 � 50

ID 40 > 30 Q)

...J 20 Q; 10 3: 0 0

a.. - 10 -20 -30 -40 -50 -60

0 5 10 15 20 25

Hours Figure 7 Power flow over selected transmission lines

It is easily observed that the wind branch have positive power flow from wind fann to load area during night and negative power flow during daytime. That is to say, during daytime, when wind energy is lacking, other power source will support the load located at wind bus. On the other hand, during

night, when wind energy is ample, wind will support load over other areas.

Table 6 Transmission Constraints Configuration

Branch Transmission Constraints Branch 2-4 lOOMW Branch 3-4 50MW Branch 5-6 80MW

In order to study transmission congestion, constraints are configured on transmission lines of nuclear, wind and aggregated PHEVs respectively, as is shown in Table 6.

1�� _ Price (wihout congestion)

90 - Price at Wind bus

85 _ Price at PHEVs' bus

80

C 75 o 70 t; 65 & 60 § 55 <..l 50 "S 45 ,g 40 �_ 35 - 30 � 25 ct 20

15 10

5 o

5 10

HOURS 15 20

Figure 8 Energy Balancing Prices with Congestion

Figure 8 shows the energy balancing market prices under transmission congestions. Compared with the prices without congestion, it is clearly to see that during night LMP at wind bus is lower and during daytime LMP, at PHEVs' bus where also is heavily loaded area, goes higher.

220 200 180 160 C 140 0 120 �

<1> 100 0) 80 c 0 60 <..l :; 40 0 20 :5 0 � ·20 Qi -40 > ·60 � ,., ·80 e> ·100 <1> ·120 c

UJ ·140 ·160 l�_-+'_�-=i=�=:=;:=�=;:=�=::' ·180

5 10 15 20 25

Hours Figure 9 PHEVs' Charging/Discharging Pattern

In Figure 9, the charging and discharging pattern of PHEVs are shown. It is observed that with transmission congestions PHEVs will store more energy during the night and sell them at

daytime. That is because the congestion will isolate the PHEVs from other sources especially during the daytime with heavy load in the system and then the LMP at PHEVs' bus will be maintained at high level. Therefore, PHEVs can buy and store huge amount of power at low prices during night and sell them when prices go higher.

Figure 10 compares the wind output under non-congestion situation with under congestion situation. It is observed, the transmission congestion will limit the wind output from high level. However, sometimes because congestion limit the power from outer power sources, local load will rely on wind more heavily, which increase the wind output.

250

"0 C "§ 200 ........ <1> :0 .!!l "co >

150 ro """ E :l E

"x ro 100 E

"0

� 15 50 � (L

0 0 5 10 15 20 25

Hours

Figure 10 Wind Output Profile

In Table 7, economic performance of wind farm and aggregated PHEVs are shown. Due to the transmission congestion, the output of wind farm is constrained. Meanwhile, the LMP at wind farm bus is at lower level during night. Hence, wind farm's revenue as well as the profit decreases.

Table 7 Economic Evaluation on Congestion

Items With Without

Congestion Congestion

Wind Revenue $51,942.36 $55,459.75

Cost $11,173.07 $12,409.62

penalty $]7,697.04 $14,367.10

Profit $23,072.25 $28,683.03

PHEV PHEV -$4,057.83 -$4,436.76

Market

Regulation $1,543.68 $4,054.90

battery $748.85 $2,]40.27

degradation

gasoline cost $0 $0

Profit -$3,263.00 -$1,758.4]

On PHEVs side, transmission congestion will increase the LMP at PHEVs bus. Energy during daytime at heavily loaded area is very valuable so PHEVs will sell power and increase its

energy balancing revenue. However, the decrease in wind portion will reduce the total need for regulation, which in turn will reduce PHEVs' regulation profit dramatically. That is mainly because PHEV s have higher regulation marginal cost compared with coal power plant and nature gas power plant so the later entities will take the base need for regulation and PHEVs will be more likely to work as marginal units whose selected capacity will be reduced greatly when total need drops.

IV. CONCLUSION

This paper presents a market-based coordinated scheduling simulation platform for enhanced efficiency of electric power system operations. The two-layered software platform models individual market participants (e.g., wind farms and PHEVs) as look-ahead decision making entities with the objective of maximizing expect profits from both energy and regulation services. At the system operator level, static security constrained economic dispatch is performed with transmission network limits explicitly modeled. The embedded intelligence of short-term available generation and energy/regulation price forecasts at component level extracts the most value from the various energy resources, whereas the coordination of energy and regulation dispatch at the system level ensures the security of the transmission network. Simulations on a modified IEEE 14-bus system demonstrate the potential benefits of such a proposed scheduling framework in electric power systems with many variable resources and PHEVs. Future work will further investigate the performance guarantee of the proposed distributed interactive model predictive scheduling framework. Also the feasibility of treating wind farms as providers of both energy and frequency regulation services will be explored in the future.

V. ACKNOWLEDGEMENT

This work was supported in part by the Department of Electrical and Computer Engineering at Texas A&M University, and in part by Texas Engineering Experiment Station. The authors appreciate the financial support.

VI. REFERENCES

[1] U.S. Department of Energy, 20% Wind Energy by 2030: Increasing Wind Energy's Contribution to u.s. Electricity Supply, 2008.

[2] T. Senjyu, R. Sakamoto, N. Urasaki et al., "Output power leveling of wind turbine generator for all operating regions by pitch angle control, " IEEE

Transactions on Energy Conversion, vol. 21, no. 2, pp. 467-475, 2006.

[3] J. L. Rodriguez-Amenedo, S. Arnalte, and J. C. Burgos, "Automatic generation control of a wind farm with variable speed wind turbines, " IEEE Transactions on Energy Conversion, vol. 17, no. 2, pp. 279-284, 2002.

[4] 1. L. Rodriguez-Amenedo, S. Arnalte, and J. C. Burgos, "Automatic Generation Control of a Wind Farm with Variable Speed Wind Turbines, " IEEE

Power Engineering Review, vol. 22, no. 5, pp. 65-65, 2002.

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[20]

S. M. Muyeen, R. Takahashi, T. Murata et al., "A Variable Speed Wind Turbine Control Strategy to Meet Wind Farm Grid Code Requirements, " IEEE Transactions on Power Systems, vol. 25, no. 1, pp. 331-340, 2010. W. R. Lachs, D. Sutanto, and D. N. Logothetis, "Power system control in the next century, " IEEE

Transactions on Power Systems, vol. 11, no. 1, pp. 11-18, 1996. W. Kempton, and J. Tomic, "Vehicle-to-grid power implementation: From stabilizing the grid to supporting large-scale renewable energy, " Journal of

Power Sources, vol. 144, no. 1, pp. 280-294, 2005. W. Kempton, and J. Tomic, "Vehicle-to-grid power fundamentals: Calculating capacity and net revenue, " Journal of Power Sources, vol. 144, no. 1, pp. 268-279, 2005. W. Kempton, 1. Tomic, S. Letendre et al., "Vehicle­to-grid power: battery, hybrid, and fuel cell vehicles as resources for distributed electric power in California, " Journal of Power Sources, 2001. H. Turton, and F. Moura, "Vehicle-to-grid systems for sustainable development: An integrated energy analysis, " Technological Forecasting and Social

Change, vol. 75, no. 8, pp. 1091-1108, 2008. S. L. Andersson, A. K. Elofsson, M. D. Galus et al., "Plug-in hybrid electric vehicles as regulating power providers: Case studies of Sweden and Germany, " Energy Policy, vol. In Press, Corrected Proof, 2010. L. Xie, and Y. Gu, "Fast Model Predictive Control­based Coordination of Wind Power and Battery Energy Storage Systems, " IEEE Transactions on Industrial Electronics, 2010. (submitted) ERCOT. "Histroric Data of Ancillary Market, " https://pi.ercot.comicontentproxy/publicList?folder_i d=10001730. ERCOT. "Balancing Energy Services Daily Reports, " http://www.ercot.comimktinfo/servicesibal/index.

R. D. Zimmerman, Matpower 4.0b3 User's Manual,

2010 Power Systems Engineering Research Center (P serc), 201O. EROCT. "Load Data, " http://mospublic.ercot.comierc ot/jsp/ca _load _ hist.jsp. N. Rotering, "Optimal Plug-in Electric Vehicle Charge Control In Deregulated Electricity Markets, " EEH-Power Systems Laboratory, Swiss Federal Institute of Technology, Zurich, 2009. L. Xie, J.-Y. Joo, and M. D. Ilic, "Integration of Intermittent Resources with Price-Responsive Loads, " in NAPS, 2009. EIA. "U.S. Gasoline Retail prices, " http://www.eia.d oe.gov/oil_gas/petroleumiinfo �lance/petroleum.htm 1. LisaShea.com. "Car Average MPG Ratings, " http://w ww.lisashea.comimotorcycle/carmileage.html.

VII. BIBLIOGRAPHY

Yingzhong Gu (S'IO) was born in Shanghai, China. He received the B.S. degree in electrical engineering in 2009 from Shanghai Jiao Tong University, Shanghai, China. He is currently pursuing M.S. degree in Electrical Engineering at Texas A&M University. His research interests include power system optimization, Vehicle to Grid technologies, renewable energy, power

system analysis and voltage stability.

I Le Xie (S'05--M'10) is an Assistant Professor in the Department of Electrical and Computer Engineering at Texas A&M University. He received his B.E. in Electrical Engineering from Tsinghua University, Beijing, China. He received M.Sc. in Engineering Sciences from Harvard University in June 2005. He obtained his PhD in

Electrical and Computer Engineering at Carnegie Mellon University in December 2009. His industry experience includes an internship (Jun 2006-Aug 2006) at ISO-New England and an internship at Edison Mission Energy Marketing and Trading (Jun 2007-Aug 2007). His research interest includes modeling and control of large-scale complex systems, smart grids application with renewable energy resources, and electricity markets.