A Short-term Electric Load Forecasting Method Based on Improved VMD-LSSVR Considering Comprehensive Energy Correlation

摘要

In order to improve the accuracy of short-term electric load forecasting, this paper proposes an improved VMD-LSSVR short-term electric load forecasting method based on Copula theory considering the comprehensive energy correlation. First, Variational Mode Decomposition (VMD) is used to decompose the original load sequence into a series of sub-modal components. Considering the coupling relationship between each component and heat, cooling, and gas loads, Copula theory is used for correlation analysis to determine the input variables of each component model. Secondly, Least Squares Support Vector Regression (LSSVR) model based on mixed kernel function and parameter optimization using cuckoo search algorithm is established for each component. Finally, the final load forecast values are obtained by regression fitting of forecasting results of each component. The model performance is verified by hourly electric load of a certain area from January to June in 2017. The results show that this method has higher forecasting accuracy and provides effective technical support for short-term electric load forecasting compared with other existing models.