Performance comparison for feed forward, Elman, and radial basis neural networks applied to line congestion study of electrical power systems

Abstract

This paper presents a comparative analysis of the training performance for three important types of neural networks, namely Feed Forward neural network, Elman neural network, and Radial Basis Function neural network. In order to do this analysis, the authors performed sequential training of all the three neural networks for monitoring the congestion level in the transmission lines of the power system under study. This is accomplished through neural network simulation on the IEEE 30-bus test system under various operating conditions, namely base case, higher loading scenario, and contingency conditions. The findings of this study justify two things. On one hand, the results reveal that all the three neural networks yield successful training and are capable of reducing both the complexity and computational time as compared to the conventional iterative power flow simulation. Furthermore, the comparative analysis justifies that the radial basis function neural network is the fastest of all the three neural networks considered.