An Chaotic Pseduo Inverse Polynomial Perceptron Network for Short Term Solar Power Prediction

High precision prediction of solar power generation is very much necessary with the continuous increase of grid connected solar electricity. The accurate power prediction is extremely important for the optimal scheduling and safe operation of the grid. In this paper, an Chaotic Water Cycle Algorithm (CWCA) based Pseudo Inverse Polynomial Perceptron Network (PIPPN) is proposed to accurately predict the solar power for different weather condition and for different time horizon. The random input layer weights of the PIPPN are optimized using the CWCA. Here, a sinusoidal chaotic map is applied to diversify the populations to improvise the performance of the basic PIPPN. The chaos in proposed Chaotic PIPPN (CPIPPN) helps to predict the future solar power very efficiently. The performance of the proposed CPIPPN model is verified through various performance measures. The dominance and diversity of the proposed CPIPPN method is verified against the basic Polynomial Perceptron Network (PPN) and PIPPN for 5 minute and 1 hour ahead time horizon.