Effects of the Transmission Network on Electricity Markets © 2011 D. Kirschen and the University of Washington 1 Introduction • No longer assume that all generators and loads are connected to the same bus • Need to consider: – Congestion, constraints on flows – Losses • Two forms of trading – Bilateral or decentralized trading – Pool or centralized trading © 2011 D. Kirschen and the University of Washington 2 Bilateral or decentralized trading • Transactions involves only buyer and seller • Agree on price, quantity and other conditions • System operator – Does not get involved directly in trading – Maintains balance and security of the system • Buys or sells limited amounts of energy to keep load and generation in balance • Limits the amount of power that generators can inject at some nodes if security cannot be ensured by other means © 2011 D. Kirschen and the University of Washington 3 Example of bilateral trading Bus A Bus B G1 L1 G2 L2 • • • • G3 G1 sold 300 MW to L1 G2 sold 200 MW to L2 Prices are a private matter Quantities must be reported to system operator so it can check security © 2011 D. Kirschen and the University of Washington 4 Example of bilateral trading Bus A Bus B G1 L1 G2 L2 • • • • G3 G1 sold 300 MW to L1 G2 sold 200 MW to L2 If capacity of corridor ≥ 500 MW No problem If capacity of corridor < 500 MW some of these transactions may have to be curtailed © 2011 D. Kirschen and the University of Washington 5 But curtail which one? • Could use administrative procedures – These procedures consider: • Firm vs. non-firm transactions • Order in which they were registered • Historical considerations – Do not consider relative economic benefits – Economically inefficient – Better to let the participants themselves decide what makes sense • Participants should purchase right to use the network when arranging a trade in energy – Physical transmission rights – Support actual transmission of power over a given link © 2011 D. Kirschen and the University of Washington 6 Physical transmission rights Bus A Bus B G1 L1 G2 L2 • • • • • G3 G1 sold 300 MW to L1 at 30 $/MWh G2 sold 200 MW to L2 at 32 $/MWh G3 selling energy at 35 $/MWh L2 should not pay more than 3 $/MWh for transmission rights L1 should not pay more than 5 $/MWh for transmission rights © 2011 D. Kirschen and the University of Washington 7 Problems with physical rights • Parallel paths • Market power © 2011 D. Kirschen and the University of Washington 8 Parallel paths xA P FA 1 2 P FB xB F A = xB xA + xB © 2011 D. Kirschen and the University of Washington P F B = xA xA + xB P 9 Parallel paths C A B Branch Z 1 Reactance Capacity [p.u.] [MW] 1-2 0.2 126 1-3 0.2 250 2-3 0.1 130 2 3 Y D © 2011 D. Kirschen and the University of Washington 10 Parallel paths C A B I Z II 1 2 3 Y D 400 MW transaction between B and Y Need to buy transmission rights on all lines © 2011 D. Kirschen and the University of Washington 11 Parallel paths Branch C A B I Z II 1 2 Reactance Capacity [p.u.] [MW] 1-2 0.2 126 1-3 0.2 250 2-3 0.1 130 3 Y D F I = 0.2 0.2 + 0.3 400 MW transaction between B and Y ´ 400 = 160 MW F II = 0.3 0.2 + 0.3 ´ 400 = 240 MW Not possible because exceeds capacities of lines 1-2 and 2-3 © 2011 D. Kirschen and the University of Washington 12 Counter-flows C A B 200 MW transaction between D and Z III Z 1 2 IV F III F IV = 3 Y D © 2011 D. Kirschen and the University of Washington = 0.2 0.2 + 0.3 0.3 0.2 + 0.3 ´ 200 = 80 MW ´ 200 = 120 MW 13 Resultant flows Branch C A B Z 1 2 Reactance Capacity [p.u.] [MW] 1-2 0.2 126 1-3 0.2 250 2-3 0.1 130 3 Y I D F12 = F23 = F - F F13 = F II -F IV III = 160 - 80 = 80 MW = 240 - 120 = 120 MW The resultant flows are within the limits © 2011 D. Kirschen and the University of Washington 14 Physical rights and parallel paths • Counter-flows create additional physical transmission rights • Economic efficiency requires that these rights be considered • Decentralized trading: – System operator only checks overall feasibility – Participants trade physical rights bilaterally – Theory: • Enough participants market discovers optimum – Practice: • Complexity and amount of information involved are such that it is unlikely that this optimum can be found in time © 2011 D. Kirschen and the University of Washington 15 Physical rights and market power Bus A • • • • • • Bus B G1 L1 G2 L2 G3 G3 only generator at bus B G3 purchases transmission rights from A to B G3 does not use or resell these rights Effectively reduces capacity from A to B Allows G3 to increase price at B “Use them or loose them” provision for transmission rights: difficult to enforce in a timely manner © 2011 D. Kirschen and the University of Washington 16 Centralized or Pool Trading • Producers and consumers submit bids and offers to a central market • Independent system operator selects the winning bids and offers in a way that: – Optimally clears the market – Respects security constraints imposed by the network • No congestion and no losses: uniform price • Congestion or losses: price depend on location where generator or load is connected © 2011 D. Kirschen and the University of Washington 17 Borduria-Syldavia Interconnection Borduria DB= 500MW Syldavia DS= 1500 MW • Perfect competition within each country • No congestion or losses within each country – Single price for electrical energy for each country – Price = marginal cost of production © 2011 D. Kirschen and the University of Washington 18 Borduria-Syldavia Interconnection Borduria Syldavia DB= 500MW DS= 1500 MW p B = MC B = 10 + 0.01PB [$ / MWh] p S = MC S = 13 + 0.02 P S [$ / MWh] $/MWh $/MWh 43 15 10 13 500 MW p B = MC B = 10 + 0.01´ 500 = 15 $ / MWh © 2011 D. Kirschen and the University of Washington 1500 MW p S = MC S = 13 + 0.02 ´ 1500 = 43 $ / MWh 19 Borduria-Syldavia Interconnection Borduria DB= 500MW Syldavia DS= 1500 MW Economic effect of an interconnection? © 2011 D. Kirschen and the University of Washington 20 Can Borduria supply all the demand? Borduria DB= 500MW PB = 2000MW PS = 0MW Syldavia DS= 1500 MW MC B = 30$ / MWh MC S = 13$ / MWh • Generators in Syldavia can sell at a lower price than generators in Borduria • Situation is not tenable • Not a market equilibrium © 2011 D. Kirschen and the University of Washington 21 Market equilibrium Borduria Syldavia DB= 500MW DS= 1500 MW p =pB =p S PB + PS = D B + D S = 500 + 1500 = 2000MW p B = MC B = 10 + 0.01PB [$ / MWh] p S = MC S = 13 + 0.02 P S [$ / MWh] p = p B = p S = 24.30$ / MWh PB = 1433MW PS = 567MW © 2011 D. Kirschen and the University of Washington 22 Flow at the market equilibrium Borduria Syldavia DB= 500MW DS= 1500 MW PB = 1433MW PS = 567MW FBS = PB - D B = DS - PS = 933MW © 2011 D. Kirschen and the University of Washington 23 Graphical representation p S = MC S p B = MC B Supply curve for Syldavia Supply curve for Borduria 24.3 $/MWh 24.3 $/MWh PB= 1433 MW PS = 567 MW FBS= 933 MW D B = 500 MW D S = 1500 MW D B + D S = 2000 MW © 2011 D. Kirschen and the University of Washington 24 Constrained transmission • What if the interconnection can carry only 400 MW? – PB = 500 MW + 400 MW = 900 MW – PS = 1500 MW - 400 MW = 1100 MW p B = MC B = 10 + 0.01´ 900 = 19 $ / MWh p S = MC S = 13 + 0.02 ´ 1100 = 35 $ / MWh • Price difference between the two locations • Locational marginal pricing or nodal pricing © 2011 D. Kirschen and the University of Washington 25 Graphical representation p S = MC S p B = MC B 35 $/MWh 16 $/MWh PB= 900 MW PS= 1100 MW FBS= 400 MW D B= 500 MW D S = 1500 MW D B + D S = 2000 MW © 2011 D. Kirschen and the University of Washington 26 Summary Separate markets Single market PB [MW] 500 1,433 Single market with congestion 900 p B [$/MWh] 15 24.33 19 R B [$/h] 7,500 34,865 17,100 E B [$/h] 7,500 12,165 9,500 PS [MW] 1500 567 1100 43 24.33 35 R S [$/h] 64,500 13,795 38,500 E S [$/h] 64,500 36,495 52,500 0 933 400 R TOTAL = R B + R S 72,000 48,660 55,600 E TOTAL = E B + E S 72,000 48,660 62,000 p S [$/MWh] F BS [MW] © 2011 D. Kirschen and the University of Washington 27 Winners and Losers • Winners: – Bordurian generators – Syldavian consumers • Losers – Bordurian consumers – Syldavian generators • Congestion in the interconnection reduces these benefits © 2011 D. Kirschen and the University of Washington 28 Congestion surplus Consumer payments: E TOTAL = p B × DB + p S × D S Producers revenues: RTOTAL = p B × PB + p S × PS = p B ×( D B + FBS ) + p S ×( D S - F BS ) Congestion or merchandising surplus: E TOTAL - R TOTAL = p S × D S + p B × DB - p S × P S - p B × PB = p S ×( D S - P S ) + p B × ( DB - P B ) = p S × F BS + p B ×( - F BS ) = ( p S - p B ) × F BS © 2011 D. Kirschen and the University of Washington 29 Congestion surplus © 2011 D. Kirschen and the University of Washington 30 Congestion surplus • Collected by the market operator in pool trading • Should not be kept by market operator in pool trading because it gives a perverse incentive • Should not be returned directly to network users because that would blunt the economic incentive provided by nodal pricing © 2011 D. Kirschen and the University of Washington 31 Pool trading in a three-bus example C A Branch Reactance Capacity [p.u.] [MW] 1-2 0.2 126 1-3 0.2 250 2-3 0.1 130 B 1 2 50 MW 60 MW 3 Generator Capacity Marginal Cost [MW] [$/MWh] A 140 7.5 B 285 6 C 90 14 D 85 10 D 300 MW © 2011 D. Kirschen and the University of Washington 32 Economic dispatch A C 125 MW 0 MW B 285 MW F12 1 F13 2 F 23 50 MW 60 MW 3 0 MW D 300 MW © 2011 D. Kirschen and the University of Washington 33 Superposition 60 MW 360 MW 1 2 3 300 MW 300 MW 1 2 3 300 MW 60 MW 60 MW 1 2 3 © 2011 D. Kirschen and the University of Washington 34 Flows with economic dispatch A C 125 MW 0 MW B 285 MW 156 MW 1 2 204 MW 96 MW 50 MW 60 MW 3 0 MW D 300 MW © 2011 D. Kirschen and the University of Washington 35 Overload! A C 125 MW 0 MW FMAX = 126 MW B 285 MW 156 MW 1 2 204 MW 96 MW 50 MW 60 MW 3 0 MW D 300 MW © 2011 D. Kirschen and the University of Washington 36 Correcting the economic dispatch Additional generation at bus 2 0.6 MW 1 MW 1 MW 1 2 0.4 MW 3 © 2011 D. Kirschen and the University of Washington 37 Superposition 1 360 MW 2 60 MW 156 MW 96 MW 204 MW 3 300 MW 50 MW 50 MW 30 MW 1 2 20MW 1 3 310 MW 2 10 MW 126 MW 116 MW 184 MW 3 300 MW © 2011 D. Kirschen and the University of Washington 38 Correcting the economic dispatch Additional generation at bus 3 1 MW 0.4 MW 1 2 0.6 MW 3 1 MW © 2011 D. Kirschen and the University of Washington 39 Superposition 1 360 MW 2 60 MW 156 MW 96 MW 204 MW 3 300 MW 75 MW 30 MW 1 45 MW 2 1 2 3 75 MW 285 MW 60 MW 126 MW 66 MW 159 MW 3 225 MW © 2011 D. Kirschen and the University of Washington 40 Cost of the dispatches • • • • Economic dispatch: Redispatch generator 2: Redispatch generator 3: Cost of security: © 2011 D. Kirschen and the University of Washington 2,647.50 $/h 2,972.50 $/h 2,835.00 $/h 187.50 $/h 41 Security constrained dispatch A C 50 MW 0 MW B 285 MW 126 MW 1 2 159 MW 66 MW 50 MW 60 MW 3 75 MW D 300 MW © 2011 D. Kirschen and the University of Washington 42 Nodal prices • Cost of supplying an additional MW of load at a particular node without violating the security constraints • Start from the security constrained dispatch © 2011 D. Kirschen and the University of Washington 43 Nodal prices A B C 0 MW 50 MW 285 MW 126 MW 1 2 159 MW 66 MW 50 MW 60 MW 3 75 MW D 300 MW © 2011 D. Kirschen and the University of Washington • Node 1: • A is cheapest generator • p 1 = MC A = 7.50 $ / MWh 44 Nodal prices A B C 0 MW 50 MW 285 MW 126 MW 1 2 159 MW 60 MW 66 MW 50 MW 3 75 MW D 300 MW © 2011 D. Kirschen and the University of Washington • Node 3 • A is cheaper than D • Increasing A would overload line 1-2 • D is cheaper than C • Increase D by 1 MW p 3 = MC D = 10 $ / MWh 45 Nodal prices A B C 0 MW 50 MW 285 MW 126 MW 1 2 159 MW 66 MW 50 MW 60 MW 3 75 MW D 300 MW © 2011 D. Kirschen and the University of Washington • Node 2 • C is very expensive • Increasing A or D would overload line 1-2 • ? 46 Nodal price at node 2 0.6 MW 1 MW 1 MW 1 0.4 MW 2 3 1 MW 0.2 MW 1 0.8 MW 2 3 1 MW © 2011 D. Kirschen and the University of Washington 47 Nodal price at node 2 • Increase generation at node 3 AND decrease generation at node 1 DF12 =0 DP2 = DP1 1 1 MW 2 3 © 2011 D. Kirschen and the University of Washington DP3 48 Nodal price using superposition 0.6 MW 1 MW 1 MW 1 0.4 MW DP1 + DP 3 = DP2 = 1 MW 0.6 DP1 + 0.2 DP3 = DF12 = 0 MW 2 DP1 = -0.5 MW 3 1 MW DP3 = 1.5 MW 0.2 MW 1 2 0.8 MW 3 1 MW © 2011 D. Kirschen and the University of Washington p 2 = 1.5× MC D - 0.5 × MC A = 11.25 $ / MWh 49 Observations • Generators A and D are marginal generators because they supply the next MW of load at the bus where they are located • Generators B and C are not marginal • Unconstrained system: 1 marginal generator • m constraints: m+1 marginal generators • Prices at nodes where there is no marginal generator are set by a linear combination of the prices at the other nodes © 2011 D. Kirschen and the University of Washington 50 Summary for three-bus system Bus 1 Consumption [MW] Bus 2 Bus 3 System 50 60 300 410 Production [MW] 335 0 75 410 Nodal marginal price [$/MWh] 7.50 11.25 10.00 - 375.00 675.00 3,000.00 4,050.00 2,512.50 0.00 750.00 3,262.50 Consumer payments [$/h] Producer revenues [$/h] Merchandising surplus [$/h] 787.50 (= congestion surplus) © 2011 D. Kirschen and the University of Washington 51 Counter-intuitive flows A B 50 MW C π1=7.50 $/MWh 0 MW 285 MW π2=11.25 $/MWh 126 MW 1 2 159 MW 66 MW 50 MW 60 MW π3=10.00 $/MWh 3 75 MW Power flows from high price to low price! D 300 MW © 2011 D. Kirschen and the University of Washington 52 Counter-intuitive prices • Prices at nodes without a marginal generator can be higher or lower than prices at the other nodes • Nodal prices can even be negative! • Predicting nodal prices requires calculations • Strategically placed generators can control prices • Network congestion helps generators exert market power © 2011 D. Kirschen and the University of Washington 53 Effect of losses on prices 1 2 G D L variable æ Sö 2 =I R» èV ø 2 R= P2 +Q2 V G( D) = D + L = D + K × D 2 ×R » R V 2 ×P = K× P 2 2 2 DG = G( D + DD) - G( D) = DD + 2 DD× D×K = (1+ 2 D× K ) DD DC = c (1 + 2D× K ) DD DC DD = c (1 + 2D× K ) © 2011 D. Kirschen and the University of Washington p1 = c p 2 = p 1 (1 + 2 D×K ) 54 Losses between Borduria & Syldavia PS = D S - FBS PB = D B + FBS + K × FBS 2 Minimization of the total cost 37500 With losses No losses 37000 Generation Cost [$/h] 36500 36000 35500 35000 34500 80 10 60 10 40 10 20 10 0 00 10 98 0 96 0 94 0 92 0 90 0 88 0 86 0 84 0 82 0 80 0 78 0 76 0 74 0 72 70 0 34000 Power Transfer [MW] © 2011 D. Kirschen and the University of Washington 55 Mathematical Formulation of Nodal Pricing © 2011 D. Kirschen and the University of Washington 56 Introduction • Independent System Operator needs systematic method to calculate prices • Constrained optimization problem – Maximization of global welfare • Assume perfect competition © 2011 D. Kirschen and the University of Washington 57 One-bus network D Total demand of the consumers P Total production of the generators B(D) C(P) Consumers’ benefit function Producers’ cost function B(D) - C(P) Global welfare Maximize B(D) - C(P) Subject to: P-D= 0 © 2011 D. Kirschen and the University of Washington 58 One-bus network Lagrangian: (D, P, p ) = B(D) - C(P) + p (P - D) Optimality conditions: ¶ dB º -p = 0 ¶D dD ¶ dC º+p = 0 ¶P dP ¶ º P-D=0 ¶p © 2011 D. Kirschen and the University of Washington dB dC = =p dD dP Consumption and production increase up to the point where marginal value = marginal cost = price 59 Network of infinite capacity with losses I k = Pk - Dk : net injection at bus k Network creates economic welfare by allowing trades between nodes with positive injections and nodes with negative injections Wk (I k ) Ik < 0 Benefits of consumers at node k Ik > 0 - Cost of producers at node k n W = åWk (I k ) : Global welfare k=1 © 2011 D. Kirschen and the University of Washington 60 Network of infinite capacity with losses Welfare maximization: é n ù max (W ) = max êåWk (I k )ú Ik Ik ë k=1 û Alternative formulations: ìn ü ìn ü min ( -W ) = min íå[ -Wk (I k )]ý = min íå Ck (I k )ý Ik Ik î k=1 þ I k î k=1 þ Assumes that: • Demands are insensitive to prices • Loads are constant • Hence consumers’ benefits are constant Equivalent to Optimal Power Flow problem © 2011 D. Kirschen and the University of Washington 61 Network of infinite capacity with losses ì n ü min íå Ck (I k )ý Ik î k=1 þ Constraints: • No constraints on network flows because infinite capacity • Total generation = total load + losses or • Net injection = total losses in the branches of the network n åI k = L(I1 , I 2 , .. I n-1 ) k=1 (Bus n is the slack bus) © 2011 D. Kirschen and the University of Washington 62 Network of infinite capacity with losses é ù = å Ck (I k ) + p ê L ( I1 , I 2 , .. , I n-1 ) - å I k ú ë û k=1 k=1 n æ ¶L ö ¶ dCk º +p ç -1÷ = 0 ¶I k dI k è ¶I k ø ¶ dCn º -p = 0 ¶I n dI n n æ dCk dCn æ ¶L ö ¶L ö = 1= p ç 1ç ÷ dI k dI n è ¶I k ø è ¶I k ÷ø k = 1, .. n - 1 n ¶ º L ( I1 , I 2 , .. , I n-1 ) - å I k = 0 ¶p k=1 © 2011 D. Kirschen and the University of Washington 63 Network of infinite capacity with losses æ dCk dCn æ ¶L ö ¶L ö = 1= p ç 1ç ÷ dI k dI n è ¶I k ø è ¶I k ÷ø k = 1, .. n -1 Nodal price at bus k is related to the nodal price at the slack bus ¶L >0 ¶I k dCk dCn < dI k dI n If the injection at a bus increases the losses, the price at that node will be less than the price at the slack bus • Penalizes the generators at that bus • Encourages consumers at that bus © 2011 D. Kirschen and the University of Washington 64 Network of finite capacity Limits on line flows: Fl ( I1, I 2 , .. I n-1 ) £ Fl l = 1, .. m max é ù = å Ck (I k ) + p ê L ( I1 , I 2 , .. , I n-1 ) - å I k ú k=1 ë k=1 û n n m + å ml éë Fl l=1 © 2011 D. Kirschen and the University of Washington max - Fl ( I1 , I 2 , .. , I n-1 ) ùû 65 Network of finite capacity n é ù m = å Ck (I k ) + p ê L ( I1 , I 2 , .. , I n-1 ) - å I k ú + å ml éë Fl max - Fl ( I1 , I 2 , .. , I n-1 ) ùû ë û l=1 k=1 k=1 n æ ¶L ö m ¶Fl ¶ dCk º +p ç - 1÷ - å ml =0 ¶I k dI k è ¶I k ø l=1 ¶I k ¶ dCn º -p = 0 ¶I n dI n k = 1, .. n - 1 n ¶ º L ( I1 , I 2 , .. , I n-1 ) - å I k = 0 ¶p k=1 ¶ º Fl max - Fl ( I1 , I 2 , .. I n-1 ) ³ 0 ¶ ml l = 1, .. m ml × éë Fl max - Fl ( I1, I 2 , .. I n-1 ) ùû = 0 ; ml ³ 0 © 2011 D. Kirschen and the University of Washington l = 1, .. m 66 Assume that only line i is congested mi ³ 0; m j = 0 for j ¹ i æ dCk ¶L ö ¶Fi = p ç 1+ mi ÷ dI k ¶I k è ¶I k ø k = 1, .. n -1 dCn =p dI n n åI k=1 k Price at all buses (except slack bus) is affected by the congestion on one line. = L ( I1 , I 2 , .. , I n-1 ) Fi ( I1, I 2 , .. I n-1 ) = Fi © 2011 D. Kirschen and the University of Washington max ; mi > 0 67 Network of finite capacity: DC model n DC power flow: ( j=1 ( Fij = yij q i - q j Line flows: ) i = 1, .. n ) i, j = 1, .. n I i = å yij q i - q j Line flow constraints: ( ) yij q i - q j £ F max ij i, j = 1, .. n Lagrangian function é n ù n n = å Ci (I i ) + å p i ê å yij (q i - q j ) - I i ú + å å mij éë Fijmax - yij (q i - q j )ùû i=1 i=1 ë j=1 û i=1 j=1 n n © 2011 D. Kirschen and the University of Washington 68 Network of finite capacity: DC model é n ù n n max é = å Ci (I i ) + å p i ê å yij (q i - q j ) - I i ú + å å mij ë Fij - yij (q i - q j )ùû i=1 i=1 ë j=1 û i=1 j=1 ¶ dCi º - p i = 0 i = 1, .. n ¶I i dI i n ¶ º å yij p i - p j + mij - m ji = 0 i = 1, .. n -1 (Slack at bus n) ¶q i j=1 n ¶ º å yij q i - q j - I i = 0 i = 1, .. n ¶p i j=1 ¶ º Fijmax - yij (q i - q j ) ³ 0 i, j = 1, .. n ¶ mij n n ( ) ( ) mij × éë Fijmax - yij (q i - q j ) ùû = 0 ; mij ³ 0 © 2011 D. Kirschen and the University of Washington i, j = 1, .. n 69 Implementation • m binding constraints m+1 marginal generators – Price at these buses determined using – m+1 known prices dCi - pi = 0 dIi • n-m-1 unknown prices • m unknown Lagrange multipliers mij • Use the second optimality condition to determine these prices and shadow prices: n ¶ º å yij p i - p j + mij - m ji = 0 ¶q i j=1 ( © 2011 D. Kirschen and the University of Washington ) i = 1, .. n - 1 (Slack at bus n) 70 Implementation n å y (p ij i ) - p j + mij - m ji = 0 j=1 i = 1, .. n -1 • K: set of buses where the price is known • U: set of buses where the price is unknown n Yiip i - å yijp j + å yij ( mij - m ji ) = å yijp j jÎU j=1 jÎK n -å yijp j + å yij ( mij - m ji ) = -Yiip i + å yijp j jÎU j=1 © 2011 D. Kirschen and the University of Washington i ÎU; i ¹ slack bus i ÎK; i ¹ slack bus jÎK 71 Example C A B 1 2 3 D Marginal generators at buses 1 & 3 Price at bus 2 is unknown m12 is also unknown © 2011 D. Kirschen and the University of Washington dC A p1 = = 7.5$/MWh dPA dC D p3 = = 10.0 $/MWh dPD K = {1, 3} U = {2} 72 Example Choose bus 3 as the slack bus n Yiip i - å yijp j + å yij ( mij - m ji ) = å yijp j jÎU j=1 i ÎU; i ¹ slack bus jÎK K = {1, 3} U = {2} i = 2 : Y22p 2 - y12 m12 = y21p 1 + y23p 3 n -å yijp j + å yij ( mij - m ji ) = -Yiip i + å yijp j jÎU i = 1: j=1 i ÎK; i ¹ slack bus jÎK - y12p 2 + y12 m12 = -Y11p 1 + y13p 3 © 2011 D. Kirschen and the University of Washington 73 Pool trading in a three-bus example C A Branch Reactance Capacity [p.u.] [MW] 1-2 0.2 126 1-3 0.2 250 2-3 0.1 130 B 1 2 3 D æ -10 5 5 Y = ç 5 -15 10 ç 10 -15 è 5 © 2011 D. Kirschen and the University of Washington ö ÷ ÷ ø 74 Example Y22p 2 - y12 m12 = y21p 1 + y23p 3 -y12p 2 + y12 m12 = -Y11p 1 + y13p 3 æ -10 5 5 Y = ç 5 -15 10 ç 10 -15 è 5 ö ÷ ÷ ø ì5p 2 - 5 m12 = 25 í î-15p 2 + 5 m12 = -137.5 ìp 2 = 11.25 $/MWh í î m12 = 6.25 $/MWh © 2011 D. Kirschen and the University of Washington 75 Financial Transmission Rights © 2011 D. Kirschen and the University of Washington 76 Managing transmission risks • • • • Congestion and losses affect nodal prices Additional source of uncertainty and risk Market participants seek ways of avoiding risks Need financial instruments to deal with nodal price risk © 2011 D. Kirschen and the University of Washington 77 Contracts for difference • Centralized market – Producers must sell at their nodal price – Consumers must buy at their nodal price • Producers and consumers are allowed to enter into bilateral financial contracts – Contracts for difference © 2011 D. Kirschen and the University of Washington 78 Example of contract for difference Borduria 400 MW Borduria Power Syldavia Syldavia Steel 400 MW • Contract between Borduria Power and Syldavia Steel – Quantity: 400 MW – Strike price: 30 $/MWh • Other participants also trade across the interconnection © 2011 D. Kirschen and the University of Washington 79 No congestion market price is uniform Borduria 400 MW Borduria Power πB = 24.30 $/MWh • • • • • • Syldavia Syldavia Steel 400 MW πS = 24.30 $/MWh Borduria Power sells 400 at 24.30 gets $9,720 Syldavia Steel buys 400 at 24.30 pays $9,720 Syldavia Steel pays 400 (30 - 24.30) = $2,280 to Borduria Power Syldavia Steel’s net cost is $12,000 Borduria Power’s net revenue is $12,000 They have effectively traded 400 MW at 30 $/MWh © 2011 D. Kirschen and the University of Washington 80 Congestion Locational price differences Borduria 400 MW Borduria Power πB = 19 $/MWh • • • • • • Syldavia Syldavia Steel 400 MW πS = 35 $/MWh Borduria Power sells 400 at 19.00 gets $7,600 Syldavia Steel buys 400 at 35.00 pays $14,000 Borduria Power expects 400 (30 -19) = $4,400 from Syldavia Steel Syldavia Steel expects 400 (35 -30) = $2,000 from Borduria Power Shortfall of $6,400 Basic contracts for difference break down with nodal pricing! © 2011 D. Kirschen and the University of Washington 81 Financial Transmission Rights (FTR) • Observations: – Shortfall in contracts for difference is equal to congestion surplus – Congestion surplus is collected by the system operator • Concept: – System operator sells financial transmission rights to users – FTR contract for F MW between Borduria and Syldavia entitles the owner to receive: F ×( p S - p B ) – Holders of FTRs are indifferent about where they trade energy – System operator collects exactly enough money in congestion surplus to cover the payments to holders of FTRs © 2011 D. Kirschen and the University of Washington 82 Example of Financial Transmission Rights Borduria 400 MW Borduria Power Syldavia Syldavia Steel 400 MW • Contract between Borduria Power and Syldavia Steel – Quantity: 400 MW – For delivery in Syldavia – Strike price: 30 $/MWh • To cover itself against location price risk, Borduria Power purchases 400 MW of financial transmission rights from the System Operator © 2011 D. Kirschen and the University of Washington 83 Example of Financial Transmission Rights Borduria 400 MW Borduria Power πB = 19 $/MWh • • • • • • • Syldavia 400 MW Syldavia Steel 400 MW πS = 35 $/MWh Borduria Power sells 400 at 19.00 gets $7,600 Syldavia Steel buys 400 at 35.00 pays $14,000 The system operator collects 400 (35 -19) = $ 6,400 in congestion surplus Borduria Power collects 400 (35 -19) = $6,400 from the system operator Borduria Power pays Syldavia Steel 400 (35 -30) = $2,000 Syldavia Steel net cost is $12,000 The books balance! Borduria power net revenue is $12,000 © 2011 D. Kirschen and the University of Washington 84 Financial transmission rights (FTR) • FTRs provide a perfect hedge against variations in nodal prices • Auction transmission rights for the maximum transmission capacity of the network – The system operator cannot sell more transmission rights than the amount of power that it can deliver – If it does, it will lose money! • Proceeds of the auction help cover the investment costs of the transmission network • Users of FTRs must estimate the value of the rights they buy at auction © 2011 D. Kirschen and the University of Washington 85 Financial transmission rights • FTRs are defined from point-to-point • No need for a direct branch connecting directly the points between which the FTRs are defined • FTRs automatically factor in the effect of Kirchoff’s voltage law • Problem: – There are many possible point-to-point transmission rights – Difficult to assess the value of all possible rights – Difficult to set up a market for point-to-point transmission rights © 2011 D. Kirschen and the University of Washington 86 Flowgate rights • Observation: – Typically, only a small number of branches are congested • Concept: – Buy transmission rights only on those lines that are congested – Theoretically equivalent to point-to-point rights • Advantage: – Fewer rights need to be traded – More liquid market • Difficulty: – Identify the branches that are likely to be congested © 2011 D. Kirschen and the University of Washington 87