Neural Network Approach to Economic Dispatch: Case Study: Nigerian Thermal Stations

Abstract

This paper presents an application of a new Hopfield model with simulated annealing in solving the economic dispatch problem of power systems. Factors such as power mismatch, total fuel cost and the transmission line losses are used in defining the energy function to solve the economic dispatch problem using the Hopfield model. Each term of the energy function is multiplied by a weighting factor which can either be logically selected or directly estimated according to the specified power mismatch. A linear input-output model that takes into account the transmission losses was used here instead of the usual sigmoidal neuron model. It is observed that economic scheduling is an optimization problem. In Hopfield network approach the objective function of the economic dispatch problem is transformed into a Hopfield energy function, which is then minimized through iterations. Computational results based on the proposed method are compared with well-known classical solution approach and found satisfactory. The result reveals that the new algorithm is relatively simple and fast in respect of computational requirement when compared with conventional.