Short-term Power Load Forecasting Model Based on Model Fusion

Abstract

Power load forecasting plays a vital role in ensuring the safe and stable operation of power system, ensuring the reliability of power supply and improving social and economic benefits. Short-term load forecasting is the basis of power grid operation and is generally used to make generation and dispatching plans. At the same time, accurate short-term load forecasting results can reduce the cost of power generation and improve economic benefits. Short-term load forecasting is of great significance to the optimization of generation side and user side two-way dispatching and the improvement of power energy efficiency. In this work, a short-term power load forecasting model based on blending model fusion was established by combining machine learning algorithms such as extreme learning machine, random forest, XGBoost, wavelet neural network, etc. The forecasting model is expected to offer accurate short-term load forecasting and ensure the balance between supply and demand of power grid, thus ensuring the economic, reliable, high-quality and efficient power supply of power grid.