Short-term Wind Power Prediction Based on IPSO-GSA Optimized Neural Network

Accurate and reliable short-term wind power forecasting is very important for coping with wind power fluctuation and optimizing wind farm operation. Due to the complexity and variability of wind farm surroundings, traditional forecasting methods are difficult to meet the requirements of refined and dynamic modeling, and there are problems such as easy to fall into local optimum, convergence speed and forecast accuracy to be improved. In response to the above requirements, this paper established an intelligent wind power prediction method based on the IPSO-GSA-BP model Firstly, in view of the defects of the universal gravitation algorithm (GSA) such as the lack of jumping out of the local optimal mechanism and the group memory function, the Inproved Particle Swarm optimization algorithm (IPSO) is used to improve the universal gravitation search algorithm from the aspects of particle inertial mass and global search to improve its optimal precision. Then, the BP neural network is optimized by the optimal parameters obtained by the IPSO-GSA. Finally, a simulation experiment was carried out using the operating data of a wind farm in southern China, and compared with other models. The results showed that the method proposed in this paper has higher accuracy and higher reliability in short-term wind power forecasting, and can realize refined and dynamic modeling, so as to provide a new technological path for optimal scheduling management in wind farm.