| Deregulation of electricity industry brings about competitive markets and decentralized control of the generating units into the electricity industry. Decentralized control of generating units lets power producers (GENCOs) schedule their own units to participate in the market. At the same time, it also changes the way some operating constraints are perceived from the generation side that obligation to meet certain load demand and system requirements are no longer in place. In a deregulated environment, power producers participate in the market through bilateral contracts and/or selling its capacity through competitive bidding at the Power Exchange in order to maximize its profit. A new unit commitment algorithm, namely price-based unit commitment (PBUC), is therefore necessary for the GENCOs to schedule their units. As market price of power is unknown beforehand, a GENCO would need to predict the price to anticipate possible profit in order to make the unit commitment decision.; This dissertation proposes a new formulation for the PBUC problem along with the solution method that incorporates the stochastic nature of the market-clearing price (MCP) of power, and considers only constraints that are local to the GENCO itself including environmental constraints. A Lagrangian Relaxation (LR)-based method is proposed as solution technique that features a probabilistic dynamic programming in conjunction with Newton-Raphson method to solve the environmentally constrained price-based unit commitment (EC-PBUC) problem. The hourly MCP is modeled using a stochastic process of the marginal unit that clears the market at a particular hour, which depends on aggregate load and generating unit availabilities. Several test cases with increasing complexity are used to test the proposed solution method. Test results indicate the viability of the proposed technique. |
