BACK PROPAGATION ALGORITHM BASED NEURAL NETWORK FOR WIND SPEED PREDICTION

Abstract

The wind speed forecasting is of extreme importance to aid in the planning studies and scheduled operation of wind power plant. Despite improvements in Wind speed forecasting methods, wind speed forecasts still suffer from relatively high errors, in terms of normalized Mean Squared Error depending on several factors, such as, forecasting horizon, type of forecasting model, size of wind farm and geographic location. In this work, the proposed engine has the structure of Neural Network (NN) with the activation functions of the hidden neurons and the error measure of the training phase. This forecast engine is trained by a new improved back propagation algorithm, which optimizes the free parameters of the NN for wind speed prediction. New forecasting engine for wind speed prediction is proposed to validate the three datasets (solar irradiance(I),wind direction (deg),wind speed(u(mph)) were used for model training and testing.

Author

Ms. K.JENISHA, Dr. T. ARULDOSS ALBERT VICTORIE