Research on Load Forecasting of Power System Based on Deep Learning

In recent years, with the continuous development of deep learning technology, its application in the fields of image recognition, speech recognition and other fields has achieved remarkable results. Therefore, the introduction of deep learning technology into power system load forecasting can improve the accuracy and stability of forecasting and provide more effective decision support for the operation and scheduling of the power system. Electricity load forecasting is an important support for demand-side resource cooperative control, and its technical principle is to forecast the baseline load and adjustable potential of users through suitable forecasting methods based on the user's gateway or sub-historical load data, historical and future meteorological data, and the combination of the user's production, living behaviour habits and external economic situation and other influencing factors. In order to achieve a more accurate prediction of short-term power loads, the study analyses the results of load prediction using recurrent neural networks and their variants, and the experimental results show that the LSTM prediction model can be well applied to the short-term prediction of power loads in a single week's prediction, and has a more stable accuracy and precision rate.