Short-Term Power Load Forecasting Based on VMD-SSA-LSTM

The essence of power load forecasting is time series forecasting, which has nonstationarity and complexity of influencing factors. In order to improve the accuracy of prediction and solve the problems of randomness and difficulty in selecting parameters of Long-Short Term Memory neural networks, this paper proposes a VMD-SSA-LSTM algorithm which is combined with Variational Modal Decomposition(VMD), Genetic Algorithm(GA), Sparrow Search Algorithm(SSA) and Long-Short Term Memory(LSTM) neural networks. First, VMD is used to decompose the power load data into intrinsic modal functions with different characteristics and frequencies, and then the processed data is used to train the LSTM model, with the help of sparrow search algorithm. The experiment results show that the proposed model has a great improvement in prediction accuracy compared with other traditional prediction models, and can be effectively applied to short-term power load forecasting.