A Scheme for Charging Load Prediction of EV Based on Fuzzy Theory

Abstract

With the development of the electric vehicle industry, the increasing charging load of electric vehicles has brought enormous pressure to urban distribution networks, affecting their safety and efficiency. Therefore, the power system needs a better load forecasting model to predict the load value more accurately. Considering that traditional electric vehicle charging load prediction models still have many shortcomings, this article uses fuzzy theory to deal with uncertain influencing factors and combines fuzzy clustering method to analyze the charging habits of local residents. In order to make the power load prediction results more effective and reliable, this paper proposes a fuzzy neural network prediction model that takes into account user habits. The key influencing factors are fuzzy processed, and the fuzzy c-means clustering method is introduced for mining. The law of charging time for most car owners is analyzed. The superiority of the prediction model incorporating fuzzy theory was verified through comparative experiments. The accuracy of the final time-sharing prediction is above 90%.