Power Generation Scheduling of Hydropower Plants Using an Artificial Neural Network (ANN)

Abstract

This paper presents the determination of power generation in hydropower plants using an artificial neural network ANN in the Central-1 region power grid of EDL. The goal aims to reduce the generation cost in terms of the total cost of each power plant to generate the electricity at its lowest point and to maximize the power generation to balance the supply and demand sides. The proposed ANN is applied to solve an optimal generation scheduling, an optimization problem, and an economical dispatching problem. In addition, a quadratic function uses the Lambda iteration method to consider an optimal dispatch problem in a hydropower plant system. Use the ANN tool in MATLAB to solve the power plant generation problem and train it with the back-propagation algorithm considered as 10 power plants in EDL's Central-1 region power grid. The results of the studies show the best-operating costs in comparison between the proposed ANN and the lambda iteration method, which are significantly less than the operating costs of the current system. For the ANN accuracy is measured using the Root Mean Square Error RMSE of the input-output relationship. It shows that the ANN is highly efficient and has an accuracy of better than 0.90.