Research on Algorithm and Modification of Elman Network Based on Fuzzy Classification and Short-Term Electricity Consumption Prediction

One of the important issues in the construction of smart grids is power consumption prediction. Accurate electricity consumption prediction is of great significance for the planning and management decisions of the power grid in the economic sector. A new prediction and correction method has been proposed, which utilizes fuzzy clustering and Elman network algorithms. We obtain residential electricity consumption data every 15 minutes using a measurement automation system. This method first uses fuzzy clustering method to classify users based on their electricity consumption behavior, and then calculates the weights of various factors that affect electricity consumption through coefficients. Neural networks are an important method of intelligent information processing, whose task is to obtain knowledge, express knowledge, and infer decision rules from a large amount of observation and experimental data. Neural networks use the idea of nonlinear mapping and parallel processing methods. Then, the short-term electricity consumption prediction of Elman neural network can refer to weighted factors and the past electricity consumption of residents, and finally, correction algorithms are used to correct the predicted values. Compared with other algorithms, the prediction progress of our proposed method is significantly improved.