关键词: 短期电力负荷预测, 长短期记忆网络, 互补集合经验模态分解, 深度学习

Abstract: With the continuous development of power industry, the importance of load forecasting is becoming more and more obvious. As an important part of load forecasting, short-term load forecasting is of great significance to the dispatching and operation of power system and market transactions. Accurate load forecasting is helpful to improve the utilization rate of power generation equipment and the effectiveness of economic dispatching. Because load data are affected by many random factors and have strong nonlinear characteristics, a short-term power load forecasting method is proposed based on complementary ensemble empirical mode decomposition and long short-term memory. A simulation is made of a city’s power load data using the proposed method, and the simulation results are compared with those of other traditional forecasting methods. It is proved that the long short-term memory model has lower error and higher prediction accuracy. At the same time, the prediction results of complementary ensemble empirical mode decomposition and long short-term memory are compared with those of long short-term memory model under other decomposition methods, which has verified that the complementary ensemble empirical mode decomposition method is effective in improving the prediction accuracy.

Key words: short-term load forecasting, long short-term memory, complementary ensemble empirical mode decomposition, deep learning