A Prediction Model for Wind Power Reserve Capacity Based on PSO-BP Neural Network Algorithm

This study presents a model for estimating wind power reserve capacity using a neural network method, with a focus on the effects of wind power volatility and randomness on power grid dispatching. The proposed model employs the particle swarm algorithm to optimize the weights between the connection layers in the backpropagation neural network. This optimization aims to enhance the accuracy of predicting future power values in wind power systems. Additionally, the model utilizes the Pearson correlation coefficient to identify the influencing factors that exhibit a positive relationship with the prediction error. Subsequently, a multivariate regression method is employed to establish the correlation between the extracted influencing factors and the calculation of wind power reserve capacity. This study presents an analysis of a computation example utilizing the real operational data obtained from a wind farm located in Belgium. The simulation results indicate that 80% of the prediction errors fall within the estimation interval of the model, providing more evidence to support the efficacy of the approach.