A PSO-ANFIS based Hybrid Approach for Short Term PV Power Prediction in Microgrids

Abstract

This paper proposes a hybrid approach based on a combination of particle swarm optimization (PSO) and adaptive neuro-fuzzy inference systems (ANFIS) for one-day-ahead hourly photovoltaic (PV) power generation prediction in microgrids. The increasing penetration of solar PV energy into electric power generation systems imposes important issues to address resulting from its intermittent and uncertain nature. These challenges necessitate an accurate PV power generation forecasting tool for planning efficient operation of power systems and to ensure reliability of supply. In this paper, a combination of PSO and ANFIS is used to develop a PV power prediction model. To demonstrate the effectiveness of the proposed method, it is tested based on practical information of PV power generation data of a real case study microgrid in Beijing. The proposed approach is compared with two other prediction methods. Evaluation of forecasting performance is made with the persistence forecasting method as a reference model, and results are compared with actual scenario. The proposed approach outperformed back propagation neural network and persistence based forecasting methods, demonstrating its favorable accuracy and reliability.