Volume 19: Sustainable Energy Solutions for a Post-COVID Recovery towards a Better Future: Part II

Short-term Load Forecasting Based on Slime Mould Algorithm Optimized Least Square Support Vector Machine Combined with Variational Modal Decomposition Can Ding, Qingchang Ding

Abstract

To improve the accuracy of power system short-term load forecasting and energy utilization efficiency, based on Variational Mode Decomposition (VMD), Least Squares Support Vector Machine (LSSVM), and Slime Mould algorithm (SMA), a combined load forecasting model of VMD-SMA-LSSVM is proposed. First, the load signal was decomposed by VMD. For the decomposed sub-sequences, a combined algorithm based on slime mold optimization algorithm and least square support vector machine algorithm is used to predict respectively. Then, the predicted results of each sub-sequence were superposed and reconstructed to get the final predicted value. By comparing it with other machine learning models and other decomposition methods., this research results show that the load prediction results based on this method have more excellent prediction effects than methods.

Keywords Short-term power load forecasting, Variational modal decomposition, Slime mold optimization, Least squares support vector machine