STRATEGIC BIDDING IN ELECTRICITY GENERATION SUPPLY MARKETS George Gross David J. Finlay George Deltas Dept. of Electrical Computer Engineering OOCL Transportation Co Dept, of Economics University of Ilinois at Urbana-Champaign 300 Central Ave Urbana IL 6180 University of Illinois at Urbana-Champaign Mountain view, CA 94043 Champaign, IL 61820 Abstract We provide a brief review of the EWPP operation [1].The Pool dispatcher is charged with determining on a daily basis We formulate a general framework of a competitive electric- the schedule for the so-called availability declaration period ity generation supply market(CEM, embodying the salient (ADP),a 39-bour period running from 9: 00 p. m. on day 1, attributes of the poolco concept. This framework serves two the bid submittal day, to 12: 00 noon on day &. The gener principal purposes: to solve the selection by the CEM opera- tion schedule for the period known as the schedule day, running tor of the winners in a sealed bid auction for the right to serve from 5: 00 a. m. on day 1 to 5: 00 a.m. on day z, is then accepted load in each period of the auction horizon; and, to determine as the actual schedule for the next day. By 10: 00 a m. each the profit-maximising strategic bids of a generation supplier. day, the dispatcher produces a forecast of national demand for The formulation represents the physical and operating con- every half hour of the ADP. Also by 10: 00 a m, each bidder siderations of the electric generation system, the multi-period must submit an ofer file for each of his gensets. a genset is a nature of the auction as well as the market economics. The re- unit or a group of units which are considered together for the sulting large-scale nonlinear programming model has a struc- purposes of the dispatch. The offer file contains informatio are that is effectively exploited for solution by Lagrangian on the availability[maximum capacity] of the genset for each relaxation, Under conditions of a perfectly competitive mar- of the 78 ADP half hours; the offer price of the genset;the ket, the strategic bids of a player can be derived analytically. genset start-up prices; and, the genset operational characteris- Numerical results illustrate the effectiveness of the strategic tics. The prices charged from the Pool for operation and start bids. Directions for future research are discussed gation on a bidder to reveal its genset's true costs. The genset Introduction ffer price is specified as piece-wise linear function known as the Willans line[2]. A maximum of three segments can be The Poolco concept is based on the England and Wales submitted per genset Figure l shows an example. Power Pool(EWPP)[1]. The Pool is ontrols the scheduling and dispatch of generation to meet load around the clock and operates the electricity spot market. vir ofer Price tually all power is transacted through the pool and the multi ple buyers and sellers have set up what has become the largest Proe io ch itive mark in to such a structure. In particular, we consider the task such take best advantage of the sealed bid auction given the g ating resources, costs and constraints. We refer to this as the optimal bidding strategy problem. To formulate and attack this problem, we develop a mathematical framework of the op eration of a competitive electricity generation supply market (CEM). Our work explicitly considers the competitive bidding Output Power mechanism in electricity and takes into account the unique fea tures and problems associated with the generation of electrical Figure 1: Example of a willans line The Willans line is completely specified by at most 8 parame- two elbow points and the minimum and maximum output of the genset pi and prse Patcher determines the schedule of generation to meet the fore asted demand at minimum cost to the pool. This problem is ntially the unit commitment. Basically, the dispatcher ar- 0-7803-4403-098s10.00@1998IEEE 309