A Convolution Neural Network (CNN) based Deep Learning Neural Network Forecast Model for Wind Energy Prediction

Abstract

Renewable Energy resources are susceptible to the whim and vagaries of nature and are a variable random source of power. Predicting and forecasting the power from these variable power sources define and determine the operation of these system. The forecast of wind energy generation using a deep learning neural network called convolution neural network (CNN) is proposed in this paper. Here, the inputs taken are namely the wind energy, wind speed and angle of wind direction relative to the turbine blades, which are obtained from Sotavento Galicia experimental wind energy farm. The results were verified and validated against an actual generation and compared with ANN and ANFIS based forecast. RMSE, NRMSE and Pearson coefficient are calculated for the same