Short-term power load forecasting based on PSO-GRU

Abstract

In order to improve the accuracy of short-term power load forecasting, a short-term power load forecasting model (PSO-GRU) based on gated recurrent unit (GRU) neural network optimized by particle swarm optimization (PSO) is proposed. For the problem of difficult selection of GRU model parameters, PSO is used to optimize model parameters, which avoids the disadvantages of manual parameter adjustment The PSO-GRU prediction model is optimized on the basis of the GRU model, and is better at mining the characteristic information among non-linear and time-series data. The results of a case simulation analysis using power load data from a power company show that the PSO algorithm optimises the GRU model with higher forecasting accuracy compared to the GRU model and the LSTM model.