An Investigation on Pruned NNARX Identification Model of Hydropower Plant

Abstract

The aim of this paper is to determine an accurate nonlinear system model for identification of dynamics. A small hydropower plant connected as single machine infinite bus (SMIB) system is considered in the study. It is modeled by a neural network configured as a feedforward multilayer perceptron neural network (MLPNN). An investigation is conducted on various NN structures to determine the optimally pruned neural network nonlinear autoregressive with exogenous signal (NNARX) identification model. The structure selection is based on validation tests performed on these network models. The proposed structure identifies the model characteristics, which represent the dynamics of a power plant accurately. The results show an improved performance in identification of power plant dynamics by optimal brain surgeon (OBS) pruned network as compared to un-pruned (i.e., fully connected) network.