Energy management supporting high penetration of solar photovoltaic generation for smart grid using solar forecasts and pumped hydro storage system

Abstract

Abstract The growing penetration level of solar photovoltaic technology is becoming a challenging task in the smart energy management systems. The power generated from the solar photovoltaic (SPV) systems is intermittent. Therefore, it is imperative to best predict the incoming solar energy and estimate the power generated from SPV systems. In this paper, the solar energy forecasting is performed using a hybrid model consisting of neural networks and wavelet transform. The performance of the proposed model is evaluated based on both root mean square error (RMSE) and mean absolute error (MAE). To validate the proposed method the above results are compared with other existing approaches like ANN and found better within desired limits. There is a pumped hydro storage (PHS) in the configuration under study to meet the grid requirements. In order to obtain more accurate and practical results, demand response (DR) program has been also integrated in the formulation of the problem. An adequacy analysis is also carried out under various consumer flexibility scenarios. Performance analysis of the proposed energy management system has been done using MATLAB/Simulink platform, and the same is validated on 5 kW SPV system. Further, the proposed model can be applied to large-scale systems.