Short-term load forecasting of multi-layer LSTM neural network considering temperature fuzzification

The popularity of smart meters provides a large amount of data for short-term load forecasting (STLF), making it possible to make detailed forecasting of load, among which temperature is an important factor affecting summer loads. A Multi-Layered LSTM short-term load forecasting method considering fuzzy temperature is proposed in this paper. Firstly, the membership function is used to blur the temperature at the forecasting time and the average temperature of the day, so as to reduce the influence of summer temperature fluctuation on load forecasting. Secondly, an LSTM forecasting network with three hidden layers is established, and Adam optimization algorithm is used to improve the adaptive learning ability of LSTM gradient parameters. Finally, the actual-measured temperature and load data from June to August of 2018 in a southwest region were put into BP neural network, support vector machine and the proposed model. The forecasting results show that the proposed model has the highest accuracy, and temperature fuzzification improves the generalization of the model.