Variation Characteristics Analysis and Short-Term Forecasting of Load Based on CEEMDAN

With the development of the economy and the improvement of living standards of people, electricity consumption around the world has increased dramatically. However, the load that contains many components with different characteristics is affected by many factors. How to classify and extract load characteristics and improve the accuracy of load forecasting has become a focus of attention. Based on this, this paper proposes using the Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) to decompose the load and get multiple components of different time scales. The hidden characteristics of each components of load are analyzed with the four characteristic indicators. Then, the Pearson coefficient is used to analyze the multi-scale correlation between the load components and the temperature, and at the same time decompose the temperature to dig deep relationship between the load components and the temperature components. Finally, we use the Least Squares Support Vector Machine Optimized by Particle Swarm Optimization (PSO-LSSVM) to forecasting each component of the load, and select the decomposed temperature as part of the input data of the load forecasting model. The forecasting results verify the advantages of the proposed method in the aspects of load characteristic analysis and improvement of load forecasting accuracy.