Research On Short-term Power Load Forecasting Based On Time Series Decomposition And Deep Learning

Power load forecasting is a necessary prerequisite for maintaining the normal operation of the power system and the basis for the economic operation of the power system.With the development of social economy and the improvement of living standards,the electricity consumption in various regions is increasing day by day.In order to ensure the safe operation of the power system,the accuracy of power load forecasting has further requirements.The high-accuracy power load forecasting model can not only improve the flexibility of power load in the power market,but also improve the utilization rate of resources.Therefore,it is very important to establish a suitable forecasting model.The main work of this paper is as follows :Firstly,the research background and current situation of short-term power load forecasting,common forecasting models,data preprocessing methods commonly used in load forecasting process and the development process and variant structure of neural network are introduced.If there are missing values or outliers in the load sequence,the preprocessing method can be used to correct them.Secondly,a short-term load combination forecasting model based on stage decomposition and IWOA-LSTM is proposed.The phase decomposition consists of frequency domain decomposition(FDD)and ensemble empirical mode decomposition(EEMD).The load is decomposed into periodic component and trend component by frequency domain decomposition method,and the high frequency component in the trend component is decomposed by ensemble empirical mode decomposition in the second stage.An improved module of whale optimization algorithm is proposed to balance the global search ability and local search ability of the algorithm.The improved whale optimization algorithm(IWOA)is used to optimize the hyperparameters of long short-term memory(LSTM).Finally,a combined forecasting model is built.Finally,an example analysis is carried out on the real load data set.The results verify the effectiveness of the proposed model.Finally,a short-term load forecasting model based on IWOA optimized CEEMDAttention-Bi LSTM is proposed.Complementary ensemble empirical mode decomposition(CEEMD)is used to eliminate the influence of white noise times.Attention mechanism and bidirectional long short-term memory(Bi LSTM)are used to strengthen the information connection in the process of load forecasting,so as to process long-term load series more efficiently.A short-term load forecasting model of CEEMD-Attention-Bi LSTM optimized by IWOA is established,and an example analysis is carried out on the real load data set to verify the effectiveness of the proposed model.