A month ahead micro-hydro power generation scheduling using artificial neural network

The government has a continuing effort to reduce the utilization of fossil fuels which emit substantial amount of green house gases. A suitable means of tapping energy which is clean and environmentally friendly is sought. One such energy source is micro-hydro power. In this paper, an intelligent method of scheduling the generation of a micro-hydro power plant (MHPP) for a period of one month is done. A monthly forecasting of hydro power discharge was conducted first with the aid of artificial neural network and then followed by the computation of the power output of MHPP. Simulation results show that a multilayer perceptron (MLP) employing back-error propagation (BP) is successful in forecasting of the discharge as exemplified by a less than 5% error of the test data. The statistical result of interpretation of data conform the validity of forecasted result from the desired values of water discharge from a reservoir.