Short-Term Load Forecasting Model Based on BWOA-LSSVM

In recent years, advancements in new technologies and the emergence of new scenarios have caused a shift from the traditional pure load demand side to a broader load perspective, posing considerable challenges to conventional load forecasting methods. This transition creates an urgent requirement for power systems to develop regional load forecasting techniques tailored to new scenarios by leveraging contemporary theories and technologies. To enhance load forecasting accuracy, this paper introduces an improved Whale Optimization Algorithm (BWOA) based on the Baleen Search to optimize the Least Squares Support Vector Machine (LSSVM) model. The approach begins by establishing the LSSVM-based forecasting model. Subsequently, the Whale Optimization Algorithm (WOA) is refined by integrating adaptive weights and the Beetle Antennae Search (BAS) strategy. The BWOA is then employed to optimize the penalty factor of the LSSVM, thereby forming the BWOA-LSSVM model for load forecasting. Simulation comparisons demonstrate that the proposed BWOA-LSSVM model significantly improves the precision of power load predictions.