Research on Cold Load Forecasting Model Based on Long Short-Term Memory

Abstract

In order to avoid the peak load of national power system and realize the peak shift and valley filling of power system, a cooling load forecasting method based on long short-term memory is proposed. Firstly, the multi-dimensional external information such as outdoor temperature, environmental humidity and wet bulb temperature are modeled. Then, long short-term memory is used to obtain the historical cooling load information, which can effectively predict the future cooling load demand. The experimental results show that the prediction accuracy of this method is significantly higher than that of manual prediction, and has good mobility. At present, this method has played a good application value in real scenes.