Search

With the development of smart power distribution networks, the demand of smart grid user-side management is increasingly urgent. To improve the accuracy of short-term electric load forecasting for individual users, this study proposes a short-term power load forecasting model based on K-means and FCM–BP. Firstly, by analyzing the users’ electricity consumption features, K-means is applied to group users into two clusters. Secondly, for users with strong correlation at adjacent moments, local similar data is filtered out with the help of improved Fuzzy C-Mean clustering (FCM), integrating the load value of the adjacent moments into new input features. For users with weak correlation at adjacent moments, the local similar daily data are utilized as features. Finally, the feather vectors are used as input data for BP Neural Network, which is utilized to forecast the short-term load. Experimental results show that the clustering method proposed concurs with the characteristics of users’ electricity consumption behavior. With the same forecasting method, the accuracy of clustering-based load forecasting is higher than that of non-clustering load forecasting. Compared with traditional BP, RBF and GRNN Neural Network, this model has higher forecasting accuracy.