GOVERNING AND CONTROL OF HYDROPOWER PLANT USING NEURAL NETWORK

Abstract

The governors employed in hydropower plants are either hydro-mechanical governor or PI, PID governors. In case of hydro-mechanical governors the setting of temporary droop is different for different operating conditions. The two important considerations for governor setting are: - 1. Stable operation during system-islanding conditions or isolated operation. 2. Acceptable speed of response for loading and unloading under normal synchronous operation While in case of PI & PID controllers, their operating parameters are site dependant and may vary from site to site. And also they vary with the passage of time. Therefore the gain (operating parameter) has to be tuned during the commissioning of the plant, and to be modified if any change is required. A neural network based controller is developed which is free from all ill effects. An integrated model of a hydro power plant is used to study its dynamic response to gate input. ,Model of hydraulic governor is used for speed regulation, which is first replaced by a PID controller. The response of PID controller is used to train an ANN model. And finally the power plant model is governed by using ANN model.