Research of Wind Power Prediction Based on PSO-BP Model

The purpose of this article is to explore a new method for electricity prediction. This method is based on wind speed and wind power data from the year 2022. It utilizes a Particle Swarm Optimization (PSO) - backpropagation (BP) neural network model, and two different networks for ultra-short-term prediction. Firstly, this paper aims to sample a sampling point every 15 minutes, then update the collected data every 15 minutes, and finally carry out the prediction of electric power, after analyzing the data from actual wind farms, the PSO-BP neural network algorithm has a very high prediction accuracy, and it can be well applied to the prediction of ultra-short-term wind power generation.