ASSESSMENT OF TRANSMISSION CONGESTION IMPACTS ON ELECTRICITY MARKETS

Size: px

Start display at page:

Download "ASSESSMENT OF TRANSMISSION CONGESTION IMPACTS ON ELECTRICITY MARKETS"

Bethany Warner
5 years ago
Views:

1 ASSESSMENT OF TRANSMISSION CONGESTION IMPACTS ON ELECTRICITY MARKETS presentation by George Gross Department of Electrical and Computer Engineering University of Illinois at Urbana-Champaign University Of New South Wales Kensington, Australia September 23, George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 1

2 OUTLINE Transmission-unconstrained markets Transmission-constrained markets Market performance metrics Measures of congestion impacts Congestion and local market power Congestion impact evaluation examples in various systems 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 2

3 THE TIME FRAME FOR MARKETS We define one hour as the smallest indecomposable unit of time and focus on a specified hour h We discuss the market decisions for that specified hour h - year h - month h - day h h + month time axis monthahead yearahead dayahead specified hour settlement time market decisions 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 3

4 THE CENTRALIZED ELECTRICITY MARKET (CEM) We discuss the structure of the forward market by examining the day-ahead centralized electricity market In fact, the day-ahead market is a collection of 24 separate commodity markets, one for each hour of the day; we focus on the market corresponding to the specified hour h and suppress the hour h in our notation h - year h - month h - day h day-ahead CEM h + month time axis 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 4

5 CEM STRUCTURE seller 1... seller i... seller M MWh $ MWh $ MWh $ CEM operator MWh $ MWh $ MWh $ buyer 1... buyer j... buyer N 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 5

6 CEM PARTICIPANTS The CEM operator is in charge of this market and uses auctions to determine the prices and quantities bought and sold for each hour Sellers are generation entities and brokers/ marketers Buyers are consumers, brokers/marketers, distribution entities and generation entities 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 6

7 THE COMPETITIVE ELECTRICITY MARKET Sellers and buyers in the market submit sealed offers and bids, respectively, describing the price and quantity at which they are willing to sell/buy energy The CEM operator determines the successful offers and bids and the market clearing price by maximizing the social surplus The auction results determine the unit commitment and dispatch of the physical units 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 7

8 SELLER i OFFER AND COSTS $/MWh offer curve α i ( p Si ) C i ( p Si ) = p Si 0 the costs of seller i are given by the area under the curve α i (ξ) dξ min p Si p Si max p Si MWh/h 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 8

9 BUYER j BID AND BENEFITS $/MWh bid curve β j ( p B ) j B j ( p B j ) = p B j β j (ξ) dξ 0 the benefits of buyer j are given by the area under the curve 0 p B j MWh/h 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 9

10 THE SOCIAL WELFARE The social welfare is defined as the total benefits of the buyers minus the total costs of the sellers: S N M B j j= 1 i= 1 B j Si ( P ) C ( P ) i social welfare total benefits total costs 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 10

11 MAXIMIZATION OF THE SOCIAL WELFARE The objective in markets is to maximize the social welfare, so as to determine the maximum net benefits for society We neglect the transmission network constraints The CEM operator solves the resulting optimization problem to determine the successful offers and bids N M max S = B ( p ) C ( p ) st.. N p = j Bj i S j= 1 i= 1 M Bj j= 1 i= 1 p S i 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 11 i supply-demand balance

12 MAXIMIZATION OF THE SOCIAL WELFARE $/MWh demand curve market clearing price social welfare market equilibrium point supply curve market clearing quantity MWh/h 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 12

13 MARKET CLEARING PRICE The market clearing price (system marginal price) is the change in the social welfare for a unit change in the market clearing quantity Each seller receives from the CEM operator for each MWh sold ρ * ρ * Each buyer pays to the CEM for each MWh bought The market clearing price is different from the offer/bid price of nearly every player ρ * 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 13

14 THE THREE-BUS SYSTEM EXAMPLE S 1 ~ lossless system ~ S 2 bus 1 j 0.1 bus 2 B 1 j 0.1 IGO B 2 j 0.1 f 13 max = 200 MW B 3 ~ bus 3 S George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 14

15 THREE-BUS SYSTEM: OFFERS AND BIDS S $/MWh B $/MWh S $/MWh MWh/h B $/MWh MWh/h S $/MWh MWh/h MWh/h B MWh/h $/MWh MWh/h 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 15

17 UNCONSTRAINED SYSTEM REVENUES AND PAYMENTS participant quantity [MWh] price [$/MWh] revenue [$] payments [$] S S S B B B total George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 17

18 MARKET PERFORMANCE BASIC MEASURES The social welfare is a measure of the performance of the market as a whole but it does not provide insights about the performance of the individual players We define two components of social welfare producer surplus consumer surplus 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 18

19 PRODUCER SURPLUS For a seller i, the individual producer surplus measures the difference between the revenues received for the sale at the market clearing price and those that would be received at the offer price S * S = ρ p C ( p ) i = 1,..., M revenues * under ρ i S i S The total producer surplus is S S i = M i=1 S S i 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 19 i revenues under offer

20 CONSUMER SURPLUS For each buyer j, the individual consumer surplus measures the difference between the payments for the commodity at the bid prices of the buyer and those at the market clearing price B * S = B ( p ) ρ p j = 1,..., N actual benefits j j B B The total consumer surplus is S B j = N j=1 S B j 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 20 j payments with ρ*

21 THREE-BUS SYSTEM: TRANSMISSION UNCONSTRAINED EQUILIBRIUM $/MWh 300 S B 2 S B 1 * λ = 29 S 5 S 1 S S B 3 S 1 10 S S S S MWh/h 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 21 S S 3

22 THREE-BUS SYSTEM: TRANSMISSION UNCONSTRAINED DISPATCH S 1 ~ 600 MW ~ 600 MW S 2 bus 1 bus 2 B 1 IGO 300 MW 400 MW lossless system B 2 f 13 max = 200 MW B MW ~ bus 3 S MW λ* = George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 22

24 THREE-BUS SYSTEM: TRANSMISSION UNCONSTRAINED DISPATCH S 1 ~ 600 MW ~ 600 MW S 2 bus 1 IGO bus 2 B B MW f 13 = MW = violation of the line constraint since f 13 > f 13 max B MW ~ bus 3 S MW λ* = George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 24

25 THREE-BUS SYSTEM: TRANSMISSION UNCONSTRAINED DISPATCH The transmission unconstrained dispatch is infeasible because the line flow f 13 violates limit max f 13 The net injections at buses 1 and 2 have to be modified to drive the network to feasibility The only choice of buyer B 3 is to bid sufficiently high to induce seller S 3 to provide supply to meet his load 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 25

26 SOCIAL WELFARE MAXIMIZATION UNDER TRANSMISSION CONSTRAINTS N max S = Bj( pb ) Ci( ps ) j M j= 1 i= 1 s.. t g ( p,, p ; p,, p ) = 0 node n n S S B B 1 M 1 N i set of power flow equations max f ( p,, p ; p,, p ) f line S S B B 1 M 1 N real power line flow limits 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 26

27 CONGESTED LINE AND SYSTEM We call a transmission line congested if the real power line flow violates the line limit, i.e., the corresponding inequality constraint becomes binding : f ( p,, p ; p,, p ) = f S S B B 1 M 1 N max We call the transmission system congested if there are one or more congested lines in the network 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 27

28 CONGESTION Power system reliability considerations require secure operations not only under base case conditions but also under the set of postulated contingency cases Congestion occurs if one or more limit violations are detected either under the base case or in any of the contingency cases The incorporation of transmission considerations requires the representation of the base case and all the postulated contingency cases 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 28

29 THREE-BUS SYSTEM: TRANSMISSION UNCONSTRAINED DISPATCH The transmission unconstrained dispatch is infeasible because the line flow f 13 violates limit f 13 max The net injections at buses 1 and 2 have to be modified to drive the network to feasibility The only choice of buyer B 3 is to bid sufficiently high to induce seller S 3 to provide supply to meet his load 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 29

30 THREE-BUS SYSTEM: ENSURING TRANSMISSION FEASIBILITY (300 x) MW 1 2 (200 2x) MW 3 (500 3x) MW 2 1 max (300 x) + (200 2 x) = 200 MW f George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 30

31 THREE-BUS SYSTEM: REDISPATCH OF SUPPLY x is the amount of redispatch due to the f 13 max impacts of the constraint on seller S 1 2x is the amount of redispatch due to the f 13 max impacts of the constraint on seller S 2 Redispatch calculation: 2 1 (300 x) + (200 2 x) = 200 MW 3 3 so that x = 50 MW 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 31

32 THREE-BUS SYSTEM: REDISPATCH OF SUPPLY Then, the IGO reduces the output of seller S 1 by 50 MW and that of seller S 2 by 100 MW Since there is a willingness to pay by the buyer B 3, the IGO increases the output of seller S 3 by 150 MW The constrained dispatch changes the output of each seller and may impact the load supplied to buyer B George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 32

33 THREE-BUS SYSTEM: TRANSMISSION CONSTRAINED DISPATCH S 1 ~ 550 MW 500 MW ~ S 2 bus 1 bus 2 B 1 IGO 300 MW 400 MW lossless system B 2 f 13 max = 200 MW B MW ~ bus 3 S MW 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 33

34 THREE-BUS SYSTEM: LMPs S 1 ~ MW MW ~ S 2 bus 1 * 10 1 λ * = λ = 20 2 bus 2 B 1 IGO MW MW lossless system B 2 B MW ~ bus 3 S MW 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 34

35 THREE-BUS SYSTEM: LMPs S 1 ~ 550+ P 1 MW 500 P MW + 2 ~ S 2 bus 1 * 10 1 λ * = λ = 20 2 bus 2 B 1 IGO B MW 400 MW lossless system λ 3 * =? bus 3 B MW ~ S MW P + = 1 P George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved P P = 0 3 P = 1 P = 2 1 2

36 THREE-BUS SYSTEM: LMPs S 1 ~ 550 MW 500 MW ~ S 2 bus 1 * 10 1 λ * = λ = 20 2 bus 2 B 1 IGO 300 MW 400 MW lossless system B 2 bus 3 B MW ~ S MW λ * = ( 1)10 = George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 36 3

37 THREE-BUS SYSTEM: LMPs S 1 ~ 550 MW 500 MW ~ S 2 bus 1 * 10 1 λ * = λ = 20 2 bus 2 B 1 IGO 300 MW 400 MW lossless system B 2 f 13 max = bus MW B MW ~ λ * 3 = 30 S MW 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 37

39 SOCIAL WELFARE AND SURPLUSES $/MWh consumer surplus ρ B ρ S producer surplus congestion rents market efficiency loss deadweight loss MWh/h 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 39

40 IMPACTS OF CONGESTION Congestion in the system leads to a change from the single market equilibrium point to different nodal equilibrium points Change in the preferred schedule for the required generation demand balance may lead to possible curtailment in production or consumption The individual surpluses of the players change from the unconstrained market values to those in the markets at each bus under constrained conditions 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 40

41 CONGESTION MEASURES The impacts of congestion may be measured in terms of the energy that needs to be redispatched and/or the financial costs on the various players Measures of congestion impacts in $ redispatch costs congestion rents market efficiency loss 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 41

42 CONGESTION RENTS In the constrained case we have different prices at the different zones, so the players may face different clearing prices depending on their locations The social welfare in this case is given by ˆ ˆS S S Sˆ N B = + + ρ p ρ M p j Bj i S j= 1 i= 1 i congestion rents κ 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 42

43 CONGESTION RENTS In the constrained case, the congestion rents are part of the social welfare ˆ= ˆ B ˆS S S + S + κ The congestion rents are also known as merchandising surplus and correspond to the difference between the amounts paid by buyers and the amounts received by sellers; the congestion rents are collected by the IGO 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 43

44 MARKET EFFICIENCY LOSS Congestion may produce a reduction in the social welfare of the market due to the physical network constraints This reduction is called market efficiency loss and is defined by market efficiency loss ( ) E = S S constrained social welfare In economics, the market efficiency loss is also known as deadweight loss c 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 44 u

45 THREE - BUS SYSTEM: MARKET EFFICIENCY LOSS For the unconstrained case we have S u = 265,600 For the constrained case we have S c = 263,750 The market efficiency loss is ( ) E = S S = 1,850 c u 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 45

46 THREE BUS SYSTEM: CONSTRAINED CASE seller surplus ($) buyer surplus ($) S 1 1,500 B 1 16,000 S 2 2,000 S 3 2,250 total 5,750 B 2 20,000 B 3 216,000 total 252,000 congestion rents ($) 6,000 social welfare ($) 263, George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 46

47 CAUSES OF SOCIAL WELFARE REDUCTION The redispatch of higher-priced units to replace the output of the lower-priced generation The decrease in market efficiency The decrease in the producer surplus of some sellers The decrease in the consumer surplus of some buyers The needs for ancillary services provided by sellers charging higher prices The creation of situations that may lead to the exercise of market power 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 47

48 ADDITIONAL CONGESTION IMPACTS Increase of costs for delayed connection of new generation Reduction in reliability Pollution from older and less efficient plants that must be operated only for reliability purposes 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 48

49 THREE-BUS SYSTEM: LMPs S 1 ~ 550 MW 500 MW ~ S 2 bus 1 * 10 1 λ * = λ = 20 2 bus 2 B 1 IGO 300 MW 400 MW lossless system B 2 f 13 max = bus MW B MW ~ λ * 3 = 30 S MW 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 49

50 SENSITIVITY STUDY IN THE THREE - BUS SYSTEM We investigate the impacts of changing the offer S 3 of seller for his second block by varying the offer price from 29 to 330 $/MWh; the other offers/bids remain unchanged We evaluate the resulting surpluses for the various values of the offers submitted 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 50

51 THREE-BUS SYSTEM: AND * 3 S S λ $ λ 3 * S S 3 interval where increasing price does not impact the market outcomes offer for the last block of seller S 3 in $/MWh $/MWh 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved

54 THREE-BUS SYSTEM: MARKET PERFORMANCE MEASURES 10 3 $ congestion rents market efficiency losses offer for the last block of seller S 3 in $/MWh 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 54

55 LOCAL MARKET POWER Market power is the ability of a firm to profitably raise the price of a product The exercise of market power may be carried out by: the physical withholding of units the financial withholding of units Transmission constraints may create locational market power since they may set up area markets with limited importing capability 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 55

56 SIMULATION STUDIES A seller changes his offer prices by varying the offer price for the last block offered We study the resulting variations of the producer surplus, consumer surplus, congestion rents and market efficiency loss The simulations performed on different systems of various sizes are reported 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 56

57 THE SEVEN-BUS SYSTEM EXAMPLE constrained line offers bids 1 ~ 4 ~ IGO 2 ~ 5 S 1 200@5 600@10 S 3 200@40 100@60 S 4 200@10 300@15 S 5 100@20 300@40 S 7 200@30 200@40 B 1 100@80 100@50 B 2 200@100 - B 3 800@500 - B 4 200@ @120 B 5 100@80 200@50 B 6 200@ @110 B 7 100@90 100@50 3 ~ 6 7 ~ 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 57

58 SEVEN-BUS SYSTEM: PRODUCER SURPLUS impacts of changes in the flow directions S S $ S S 4 S S 7 S S S 1 S offer for the last block of seller S 3 in $/MWh 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 58

60 SEVEN-BUS SYSTEM SENSITIVITY 10 3 $ congestion rents market efficiency losses offer for the last block of seller S 3 in $/MWh George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved

64 SIMULATION RESULTS Congestion situations produce, typically, changes in the consumer and social surpluses, the additional congestion rents component of the social welfare, and the market efficiency loss with respect to the unconstrained case Congestion creates situations which are conducive to the exercise of market power Under price-responsive demand, when a particular seller increases his offer prices, the impacts of congestion on the individual players and the entire market are bounded due to the asymptotic nature of the outcomes 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 64

65 SIMULATION RESULTS We observe the existence of free-riders in the market on both the supply- and demand-sides There are also players that are negatively impacted by the exercise of market power The simulations underline the critical role of the network topology and the relative location of the market players in determining who are the losers and the gainers as a result in such a market power exercise attempt 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 65

66 RTO CONGESTION COSTS congestion costs $ (million) New York PJM NE CAISO CAISO data excludes intra-zonal congestion ISO-NE data represents mitigated congestion costs 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 66

67 FUTURE WORK Modeling incorporation of real power losses detailed representation of additional constraints incorporation of contingency case analysis Parametric analysis demand-side variation multiple players variation of offer/bid prices Study of the market efficiency loss composition 2004 George Gross, University of Illinois at Urbana-Champaign, All Rights Reserved. 67

AN ALTERNATE VIEW OF CONGESTION 2004 George Gross,

Pricing Transmission

1 / 47 Pricing Transmission Quantitative Energy Economics Anthony Papavasiliou 2 / 47 Pricing Transmission 1 Locational Marginal Pricing 2 Congestion Rent and Congestion Cost 3 Competitive Market Model