An Improved Method for Modeling the Island Photovoltaic Power Generation System With MPPT

The natural environment of the island is characterized by large uncertainties, such as temperature, humidity, and irradiance, which may affect the power generation efficiency of renewable energy. This paper takes the photovoltaic power generation as the research background, studies the direct relationship between the photovoltaic power generation power and the external environment (irradiance and temperature), and proposes a generalized regression neural network prediction (GRNN) based MPPT control strategy. Firstly, starting from the mathematical model of photovoltaic power generation system, under the premise of neglecting the power loss of DC/DC converter, the relationship between the maximum output power of the photovoltaic system and the optimal duty cycle is obtained. Then, the GRNN is trained with historical simulation data to obtain a direct relationship between the maximum output power of the photovoltaic system, the external environment (irradiance and temperature), and the internal parameters of the photovoltaic cell that affect the power generation efficiency. Finally, some simulation experiments are done in simulink. The results verify that the MPPT strategy proposed in this paper has better transient performance than traditional P&O method in photovoltaic power generation system.