Particle Swarm Optimization Based Approach for Estimating the Fuel-cost Function Parameters of Thermal Power Plants with Valve Loading Effects

Abstract

In this article, a new and accurate method for estimating the parameters of thermal power plant fuel-cost function is proposed. The input–output characteristics of thermal power plants are affected by many factors such as the ambient operating temperature and aging of generating units. Thus, periodical estimation of power plant characteristics is very crucial to improve the overall operational and economical practices. The higher the accuracy of the estimated coefficients, the more accurate the results obtained from the economic dispatch and optimal power flow calculations. Different models that describe the input–output relationship of thermal units are considered, including the one that accounts for the valve loading point. The traditional estimation problem is viewed and formulated as an optimization one. The goal is to minimize the total estimation error such that the selected model follows field data measurements as closely as possible. A particle swarm optimization algorithm is employed to minimize the error associated with the estimated parameters. The proposed approach relieves some of the mathematical restrictions typically imposed on system modeling, since it does not require convexity or differentiability, as in the case of many conventional estimation techniques. Various study cases are considered in this work to test the performance of the method. Results obtained are partially compared to those computed by the least error square method. Comparison results are in favor of the particle swarm optimization algorithm in all study cases considered.