Short Term Wind Power Prediction using Feedforward Neural Network (FNN) trained by a Novel Sine-Cosine fused Chimp Optimization Algorithm (SChoA)

Abstract

WIND power generation using wind energy conversion systems (WECS) integrated into the power grid are prone to uncertainty in wind power production. The nonlinear nature of wind, weather conditions and impact of wind speed on the generated power impacts the grid voltage and harmonics. A stable grid operation requires a precise prediction of available electrical power in real-time. As a solution, short-term wind forecasting of wind flow rate is essential to compensate for load variations and improvise wind power generation. To this problem, A hybrid methodology is employed where stochastic optimizer based-Artificial Neural Network (ANN) training is proposed due to a highly effective explorative mathematical model. Stochastic optimizer effectively trains the NN. The performance of the proposed technique is compared to well-known optimization techniques using seasonal case studies. The proposed method has shown better prediction performance as compared to existing techniques. SChoANN achieves up to 94.87% and 97.18% less training error and up to 96.42% and 83.64% less testing error in winter and summer seasons respectively.