ANN BASED GOVERNING AND CONTROLLING OF SHP

Abstract

The basic function of a governor in hydro power plant is to control speed and for load. Normally flow control based governing is used and many of the governors used are conventional governors. They suffer from mechanical wear due to their aging effects and the operation based on the droop characteristics. They have to be replaced with the recent advanced governors in order to improve the performance. One of the replacement alternatives would be PID (Proportional Integral Derivative). Further, to improve performance of PID governor Artificial Intelligence controllers can be used. In this dissertation, Hydro power plant simulation has been carried out to design a ANN governor and various components of hydro power plants like Penstock, turbine, Governor and Exciter. In the design of ANN based governor, primarily a normal hydraulic governor is designed and in order to improve the performance results PID based governor is developed. The values of the PID are different for different sites. Generally these are carried out through the manual tuning. In this work, the PID values of governor are optimized through Genetic algorithm in GATOOL Box in MATLAB 7.0. The response of PID governor had been taken to train the Artificial Neural network and after creating ANN network it is replaced with PID governor and the neural network based governor is independent of all the settings