Research on Short-Term Load Forecasting Based on the Integrated Model

With the transformation of the traditional power system to the new-type power system, the load shows higher fluctuation, and the traditional short-term load forecasting model faces the challenge of prediction accuracy. In this paper, a short-term load forecasting method based on the integrated model is proposed. Firstly, load feature extraction and feature selection are carried out by considering real-time meteorological features, seasonal features, load type, and other influencing features, and the MIC coefficient is employed to perform feature selection, so as to eliminate features less correlated with the load. On this basis, in order to explore the load characteristics fully, cluster analysis is conducted based on the improved K-means algorithm, and typical load clustering results can be obtained. Then, the optimal combination forecasting model based on SVM and multi-layer LSTM is established for each type of load, and the short-term load forecasting based on the integrated model is ultimately realized. The case study shows that compared with the traditional short-term load forecasting model, the integrated model proposed in this paper can better capture the pattern of load change and has higher forecasting accuracy, which has promising application value in short-term load forecasting situations.