A Short-Term LOAD forecasting Method Based on EEMD-LN-GRU

Abstract

Accurate forecasting of power load is of great significance to economic dispatch and safe operation of power grid. A short-term load forecasting model based on EEMD- LN-GRU is proposed according to the characteristics of uncertainty and nonlinearity of power load. In order to solve the problem of power load fluctuation and mode aliasing caused by empirical mode decomposition (EMD), the original load time series signal is decomposed into multiple intrinsic mode functions (IMF) and residual error component by ensemble empirical mode decomposition (EEMD). Each component signal is predicted by the gating recurrent unit (GRU) after layer normalization (LN). Finally, the components are predicted. The results are recombined, and then the peak valley value of the prediction sequence obtained by peak valley value correction strategy is modified to obtain the final load series. Taking the real load data of power plant of Slovak power company as an example, this method is compared with LSTM, GRU and other methods, and the results show that the proposed method has higher prediction accuracy for load forecasting.