Short- and medium-term forecasting of distributed PV output in plateau regions based on a hybrid MLP-FGWO-PSO approach
Chongyu Cui, Hongmei Wu, Xiaoyan Jiang, Liupan Jing
Abstract
Photovoltaic (PV) power generation is highly valued for its straightforward implementation and use of sustainable solar resources. However, PV systems in highland regions face greater variability and unpredictability due to complex, fluctuating weather patterns, making accurate forecasting of PV output difficult. Current PV forecasting methods fail to adequately account for the numerous factors impacting PV productivity and do not reliably optimize prediction accuracy and convergence speed. This paper puts forth a hybrid MLP-FGWO-PSO approach to PV output forecasting in highlands, utilizing a combination algorithm to leverage similarities across days. The proposed method refines the grey wolf optimization (GWO) technique and integrates it with particle swarm optimization (PSO) and multilayer perceptron (MLP) models. Through standardized evaluation, the hybrid model demonstrates improved prediction accuracy, stability, and precision by fully accommodating variable meteorological conditions.