price based unit commitment with wind generation and market clearing price variations vaidyanath...

Price based Unit Commitment with Wind Generation and Market Clearing Price

Variations

Vaidyanath RamachandranJunbiao Han

Dr. Sarika Khushalani Solanki, Ph.D.Dr. Jignesh Solanki, Ph.D.

Lane Department of Computer Science and Electrical EngineeringMarch 30, 2015

ACKNOWLEDGEMENTS

• The authors would like to acknowledge partial funding support from NSF#1351201 CAREER grant, NSF# 1232168 for this research work. The authors also would like to thank NETL RUA Grid Technologies Collaborative team.

2

SUMMARY

• Introduction• Market Clearing Price Formulation• Price Based Unit Commitment • Simulation and Results• Conclusions• Future Work

3

INTRODUCTION

• Electric power industry is becoming a deregulated and competitive structure.

• Net cost of electricity is reduced due to increased competition between generation, transmission and distribution entities

• ISO (Independent System operator) establishes rules for energy markets and ensures a fair and non discriminatory market system.

• ISO shields the power market from risks and accumulation of power with single entity.

4

ISO Market Models

• ISO supports different kinds of market models namely: PoolCo, Bilateral Contracts and Hybrid.

• The most prominent one is the PoolCo model – centralized market place that clears the market for power buyers and sellers

• Both buyers and sellers submit bids with information on how much power, what prices, and what time the participant is willing to buy or sell.

5

Power Exchange (PX)

• Power Exchange (PX) – independent non profit entity that conducts the auction for the electricity trades in the market.

• PX calculates the Market Clearing Price (MCP) based on the highest price bid in the market.

• Generation Companies (GENCO) sells electricity to PX, from which Distribution companies (DISCO) and aggregators purchase power.

6

GENCO Role in Power market

• Gencos trade real power, reactive power and operating reserves

• Gencos need innovative strategies to prepare the optimal bid to maximize profits.

• Gencos prepare schedules covering 24 hours (day ahead) - 1 week ahead by Unit Commitment to formulate optimal bid.

• To maximize profits Gencos use Price based Unit Commitment (PBUC).

7

Wind integration by GENCOs

• Apart from innovative bidding strategies, GENCOs have adopted DG resources like Wind farms to their portfolio.

• Helps in supplementing coal/gas fired power plants and meeting green generation mandates.

• Wind farms – clean and innovative source but very intermittent.

• Integrating wind power by GENCOs has many challenges.

8

Challenges in Wind Integration

• Network Constraints – include geographical location of wind farms, capacity of line/cable infrastructure

• Availability – Wind intermittency is a problem with impacts on performance

• Operation – Large wind farms have difficulties like reverse power flow, voltage fluctuations and harmonics

• Pricing – How can we handle wind energy pricing with its variability ?

9

Market Clearing Price Formulation

• Existing Techniques- Electricity Auction based MCP formulation Discriminatory/ pay-as-bid market clearing rules Uniform/ single price market clearing rules Single auction market

• Proposed Technique Optimal Power Flow (OPF) based MCP formulation

with Wind

10

Nomenclature

11

p Price of electricity in $/kWh.

Slope of the linear supply curve.

Slope of the linear demand curve.

N Number of generating units.

D Total demand of the system.

j Index for unit.

Price Axis Intercept of the Demand curve.

C Total Generation Cost

F Total profit of the Genco

Power output of generator i.

Cost function of generator i.

Incremental cost at bus i.

Uniform electricity market price.

P (j, i) Generation of unit j at time i.

R (j, i) Spinning reserve of unit j at time i.

N (j, i) Non-spinning reserve of unit j at time i.

RP (j, i) Energy price at the instant i.

RR (j, i) Spin price at the instant i.

RN (j, i) Non-Spin price at the instant i.

Minimum ON time of unit j.

Minimum OFF time of unit j.

Time duration for which unit j has been ON

at time i.

Time duration for which unit j has been

OFF at time i.

UR (j) Ramp up limit of unit j.

DR (j) Ramp down limit of unit j.

L (t, ON) Lagrangian function at time i for ON status.

(t, ON) Optimal cumulative Lagrangian at hour i

for the ON status.

(t, OFF) Optimal cumulative Lagrangian at hour i

for the OFF status.

Start-up cost for unit j at time i.

Shutdown cost for unit j at time i.

Discriminatory/ pay-as-bid market

• Discriminatory/pay-as-bid market clearing rules – every participant pays or is paid at the price of the winning bid.

• Bidding is made by predicting the cut off price and not marginal cost

• Cost of generation is above the market clearing cost which makes this system less efficient

12

Uniform/Single Price Market

• More efficient and commonly used• After receiving all bids, the ISO aggregates supply

bids to a supply curve (S) and demand bids to a demand curve (D).

• The clearing price is determined from the curves and both sellers and buyers receive same clearing price

• Marginal cost of electricity is used.

13

Single Auction market

• Demand bid is not available and load is assumed to be fixed and only GENCOs participate in bidding.

14

Energy generated by bidder j, represented as

The total combined generation can be calculated by,

The MCP, can be calculated from,

E

If the capacity limits are considered, then the combined supply curve can be represented

OPF based MCP formulation

• Proper pricing mechanism for MCP will improve the market clearing process with wind integration

• A new MCP formulation is developed to handle wind uncertainty

• Time series based OPF is developed and the wind output is considered as and when available.

• Solution of OPF determines MCP for each time interval.

15

OPF based MCP formulation

Objective of OPF is to maximize social welfare. For 24 hour

period, the OPF is formulated as:

Solving this OPF yields the highest value of the bus

incremental cost . The new MCP, defined by , incorporates ρ

the wind generators in the market clearing process.

≥ i 1,2…,N ρ λ ∀∈ 16

Price Based Unit Commitment

• In deregulated market, GENCOs use PBUC to maximize profits by trading energy and ancillary services – satisfying load is no longer an obligation

• PBUC problem can be solved using Lagrangian Relaxation (LR) method and Dynamic programming can be used to determine the Unit commitment status.

17

PBUC problem formulation

The objective of PBUC is to maximize the profit (i.e. revenue minus cost) subject to all prevailing constraints. For unit j at time i, the objective function is given as:

The first part of the equation represents the profit when the unit is ON and the second part represents the profit when the unit is OFF

18


• System ON Constraints

19

• System OFF Constraints


• Minimum ON Time Constraint

20

• Minimum OFF Time Constraint

• Ramp Up Constraint

• Ramp Down Constraint

Dynamic Programming

• After solving PBUC optimization problem, Dynamic programming is used to determine the optimal unit commitment schedule.

• The forward stage of dynamic programming is used to find the optimal cumulative value at every hour for each state described by,

21

Dynamic Programming

• The backward search is used to find out the optimal commitment trajectory

22

Simulations

The IEEE 30 bus system

GENCO 1 • Unit G1 - Bus 1• Unit G2 - Bus 2• Unit G3 - Bus 13• Wind Farm 1 with capacity

of 59.4 MW @ bus 5

GENCO 2 • Unit G4 - Bus 22• Unit G5 - Bus 23• Unit G6 - Bus 27• Wind Farm 2 with capacity

of 35.6 MW @ bus 28

23

Generator Data

24

Parameter Genco I

G1 G2 G3

Unit Type Coal Coal Oil

Pmin (MW) 15 15 10

Pmax(MW) 80 80 50

Ramp Rate(MW/h) 40 40 30

Quick Start (MW) 10 10 1

Minimum ON time (h) 2 2 2

Minimum OFF time 2 2 2

Initial State ON ON ON

Initial Hour (h) 4 4 4

Fuel Price ($/MBtu) 2 2 2

Startup (MBtu) 60 60 30

Cost Coeff. a ($/MWh2) 0 0 0

Cost Coeff. b ($/MWh) 25 24.75 26

Cost Coeff. c ($/h) 0.02 0.0175 0.0250

Parameter Genco II

G4 G5 G6

Unit Type Coal Coal Oil

Pmin (MW) 10 5 10

Pmax(MW) 50 30 55

Ramp Rate(MW/h) 30 15 30

Quick Start (MW) 5 5 1

Minimum ON time (h) 2 2 2

Minimum OFF time 1 1 1

Initial State OFF OFF OFF

Initial Hour (h) 2 2 2

Fuel Price ($/MBtu) 2 2 2

Startup (MBtu) 10 10 10

Cost Coeff. a ($/MWh2) 0 0 0

Cost Coeff. b ($/MWh) 24 26 25.25

Cost Coeff. c ($/h) 0.0625 0.025 0.0083

24-hour Forecasted Wind farm output

25

MCP from OPF solution

26

• From 24 hour OPF solution,the MCP is determined.

• Results show that presence of wind generation decreases the incremental cost of online generators and decreases the MCP

• Wind energy has a positive impact on MCP.

Simulated scenarios

27

Hour Wind PowerWind

Farm1(MW)

WindPower S-1

(MW)

Wind Power S-2

(MW)

Wind Power S-3

(MW)

Wind PowerWind

Farm2(MW)

WindPower S-1

(MW)

Wind Power S-2

(MW)

Wind Power S-3

(MW)

1 1 0.886 0.85 1 0.6 0.531 0.51 0.62 5 4.55 4.25 5 3 2.73 2.55 33 3 2.67 2.55 3 1.8 1.602 1.53 1.84 2 2 1.7 2 1.2 1.2 1.02 1.25 3 3.07 2.55 3 1.8 1.842 1.53 1.86 6 6.082 5.1 6 3.6 3.649 3.06 3.67 11 10.12 9.35 11 6.6 6.077 5.61 6.68 19 18.97 16.15 19 11.4 11.38 9.69 11.49 29 27.50 24.65 29 17.4 16.50 14.79 17.4

10 17 17.60 14.45 17 10.2 10.56 8.67 10.211 17 16.23 14.45 17 10.2 9.738 8.67 10.212 21 20.33 17.85 21 12.6 12.20 10.71 12.613 17 17.21 14.45 17 10.2 10.33 8.67 10.214 6 5.565 5.1 6 3.6 3.339 3.06 3.615 5 5.385 4.25 5 3 3.231 2.55 316 6 5.97 5.1 0 3.6 3.582 3.06 017 0 0 0 0 0 0 0 018 2 0 0 0 1.2 0 0 019 4.5 4.2 3.825 0 2.7 2.52 2.295 020 6 6.72 5.1 6 3.6 4.032 3.06 3.621 19 22.16 16.15 19 11.4 13.29 9.69 11.422 21 21.97 17.85 21 12.6 13.18 10.71 12.623 4 3.91 3.4 4 2.4 2.34 2.04 2.424 0 0 0 0 0 0 0 0

• Three scenarios of low wind volatility, high wind volatility and wind intermittency.

PBUC plans for generators

28

Scenario Hours (0-24)Forecasted

Schedule without Wind

Genco I Unit G1 1111111111111111111111111Unit G2 1111111111111111111111111Unit G3 1111111111111111111111111

Genco II Unit G4 0111111111111111111111111Unit G5 0111111111111111111111111Unit G6 0111111111111111111111111

Forecasted Schedule with

Wind



Scenario 1Low Wind Volatility



Scenario 2 High Wind Volatility



Scenario 3 Brief Wind

Intermittency



Dispatch with forecasted wind power

29

With forecasted wind power, the bidding of GENCO 1 and 2 is shown above.The ancillary services bid for both the cases remains same because all the units of Genco I and Genco II remain “ON” for hours 1-24, therefore not capable of providing non-spinning reserve.

Dispatch with low wind volatility

30

Due to low wind volatility, the Genco I is able to satisfy its energy and ancillary services contract for hours 1-24. For Genco II, there is a decrease in the dispatch from committed value for hours 7-24 as unit G5 turns “OFF” and there is an increase in the ancillary services dispatch from hours 7-24 as the “OFF” unit G5 provides non-spinning reserve.

Dispatch with high wind volatility

31

Due to high wind volatility, from hours 1-6 Genco I is still able to maintain the committed value because the higher capacity units G1 and G2 are able to ramp up to meet the volatility and unable to meet contract from hours 6-23.Highly volatile wind generation results in Genco II violating its contract with the power pool for hours 7-24.

Dispatch with wind intermittency

32

The ramp and quick start constraints of G1 and G2 are such that brief wind intermittency can be met by Genco I and satisfy the contracted valueFor Genco II, it is evident that the ramp up and quick start capabilities of units G4, G5 and G6 are insufficient to meet the wind intermittency in hours 17-20 thereby resulting in the violation of contract

Conclusions

33

• Genco I with units having higher ramping and quick start capabilities is able to meet the contract to the power pool during periods of low volatility and brief wind intermittency

• Genco II is observed to violate its contract during these scenarios.

• For highly volatile wind conditions, both the Gencos fail to satisfy the contract in the hours with high wind volatility

Conclusions

34

• The results of this study compare well with existing literature and provide avenues for future work.

• Presence of wind generation has a positive impact on the electricity prices and leads to reduction of MCP and incremental cost of generators

• Under highly volatile wind conditions, the Gencos may fail to meet their power contracts.

• Wind power can also play a vital role in satisfying the ancillary services contracts to the power market

Future Work

35

• Improvements in Wind forecast techniques• Utilizing battery storage and its effect on PBUC• New strategies for wind power trading.

References

36

[1] Senjyu, T., Yamashiro, H., Shimabukuro, K., Uezato, K. and Funabashi, T., “Fast solution technique for large-scale unit commitment problem using genetic algorithm”, Proceedings of IEE Generation, Transmission and Distribution , Vol. 150, Issue 6,12 Nov 2003, pp. 753-760.

[2] Mantawy, A.H., Abdel-Magid, Y.L. and Selim, S.Z., “A new genetic algorithm approach for unit commitment”, Proceedings of the International Conference on Genetic Algorithms in Engineering Systems: Innovations and Applications , 2-4 Sept 1997, pp.215-220.

[3] Pokharel, B.K., Shrestha, G.B., Lie, T.T., and Fleten, S. E., “Price based unit commitment for Gencos in deregulated markets”, Proceeding of the IEEE Power Engineering Society General Meeting, 12-16 June, 2005, pp. 2159-2164.

[4] Li, T., and Shahidehpour, M., “Price based unit commitment: A case of Lagrangian relaxation versus mixed integer programming”, IEEE Transaction on Power Systems, Vol. 20, No. 4, November 2005, pp: 2015-2025.

[5] Attaviriyanupap, P., Kita, H., Tanaka, E., and Hasegawa, J., “A hybrid LR-EP for solving new profit based UC problem under competitive environment”, IEEE Transactions on Power Systems, Vol. 18, No. 1, Feb. 2003, pp: 229-237.

[6] Xiaohui, Y., Yanbin, Y., Cheng, W. and Xiaopan, Z., “An improved PSO approach for profit based unit commitment in electricity market”, IEEE/PES Transmission and Distribution conference & Exhibition: Asia and Pacific, Dalian, China, 2005.

[7] Watanabe, I., Yamaguchi, N., Shiina, T. and Kurihara, I., “Agent-based simulation model of electricity market with stochastic unit commitment”, International Conference on Probabilistic Methods Applied to Power Systems, 12-16 Sept 2004, pp: 403-408.

[8] Yu, J., Zhou, J., Wu, W. and Yang, J., “Solution of the profit-based unit commitment problem by using the multi agent system”, Proceedings of the 5th World Congress on Intelligent Control and Automation, 15-19 June, 2004, Hangzhou, China.

[9] Fabbri, A., GomezSanRoman, T., RivierAbbad, J., MendezQuezada, V.H., “Assessment of the Cost Associated With Wind Generation Prediction Errors in a Liberalized Electricity Market”, IEEE Transactions on Power Systems, Vol. 20, No. 3, Aug. 2005, pp: 1440- 1446.

[10] Wind Energy Manual, 2002, Iowa Energy Center, USA. http://www.energy.iastate.edu/renewable/wind/wem/wem-02_toc.html.[11] Spencer, R., “Untapped potential of wind power”, IEEE Power Engineering Review, Vol. 22, No. 9, Sep. 2002, pp: 10–11.

References

37

[12] Milligan, M., Porter, K., “Determining the capacity value of wind: A survey of methods and implementation”, Proceedings of Wind Power, Denver, 2005.

[13] Sideratos, G., Hatziargyriou, N.D., “ An advanced statistical method for wind power forecasting”, IEEE Transactions on Power Systems, Vol. 22, No. 1, Feb 2007.

[14] Kariniotakis, G., Waldl, I.H.P., Marti, I., Giebel, G., Nielsen, T.S., Tambke, J., Usaola, J., Dierich, F., Bocquet, A. and Virlot, S., “Next generation forecasting tools for the optimal management of wind generation”, Proceedings of the International Conference on Probabilistic Methods Applied to Power Systems, 11-15 June 2006, pp. 1-6.

[15] DeMeo, E., A., Grant, W., Milligan, M., R., Schuerger, M., J., “Wind plant integration: Cost, status & issues,” IEEE Power and Energy Magazine, Vol. 3, No. 6, Nov/Dec 2005, pp: 38–46.

[16] Zeineldin, H.H.; El-Fouly, T.H.M.; El-Saadany, E.F.; Salama, M.M.A., "Impact of wind farm integration on electricity market prices," Renewable Power Generation, IET, Vol.3, No.1, March 2009, pp.84-95.

[17] Sinha, A., Basu, A.K., Lahiri, R.N., Chowdhury, S., Chowdhury, S.P., Crossley, P.A., “Setting of market clearing price (MCP) in microgrid power scenario”, Proceedings of the IEEE Power and Energy Society General Meeting - Conversion and Delivery of Electrical Energy in the 21st Century, 20-24 July 2008, pp.1-8.

[18] Singh, S.N., Erlich, I., “Strategies for Wind Power Trading in Competitive Electricity Markets”, IEEE Transactions on Energy Conversion, Vol. 23, No. 1, March 2008.

[19] Yan, H., Luh, P.B., “A fuzzy optimization method for integrated power system scheduling and inter utility power transaction with uncertainties”, IEEE Transactions on Power Systems, Vol. 12, No. 2, May 1997.

[20] Martin, S., Smeers Y. and Aguado J.A., “A Stochastic Two Settlement Equilibrium Model for Electricity Markets With Wind Generation,” IEEE Transactions on Power Systems, Vol. 30, No. 1, pp. 233-245, 2015.

[21] Sharma, K.C., Bhakar, R. and Padhy, N.P., Stochastic cournot model for wind power trading in electricity markets,” IEEE PES General Meeting, pp.1-5, 2014.

[22] Ting Dai and Wei Qiao, “Trading Wind Power in a Competitive Electricity Market Using Stochastic Programing and Game Theory,” IEEE Transactions on Sustainable Energy, Vol. 4, No. 3, pp. 805-815 , 2013.

THANK YOU

QUESTIONS ?

38

price based unit commitment with wind generation and market clearing price variations vaidyanath...

Documents