A photovoltaic power prediction method using backpropagation neural network with an improved osprey optimization algorithm

In recent years, photovoltaic power generation has been widely applied due to its carbon-free emission. However, photovoltaic power generation is directly influenced by solar irradiation intensity, and weather factors can directly affect photovoltaic power generation. To improve the efficiency of photovoltaic power generation and avoid its impact from weather factor, photovoltaic prediction technology is used to forecast the amount of photovoltaic power generation. As a data fitting model with high accuracy, the backpropagation (BP) neural network is employed for photovoltaic power generation prediction. However, the prediction accuracy of the BP neural network is influenced by its model structure, and selecting optimized model parameters can enhance its prediction accuracy. To improve the prediction accuracy of the BP photovoltaic prediction model, a parameter-optimized BP neural network model is proposed and applied into photovoltaic power generation prediction. The specific work is as follows: Firstly, based on the Osprey Optimization Algorithm (OOA), several improvement strategies are utilized to enhance its optimization performance. Secondly, the improved OOA (IOOA) algorithm is applied to optimize the parameters of the BP neural network model to obtain the IOOA-BP prediction model. Finally, the IOOA-BP prediction model is used for photovoltaic power generation prediction. The simulation experiment results of photovoltaic power generation prediction demonstrate that this model effectively improves the accuracy of photovoltaic prediction results.