Research on Electricity Load Based on Deep Clustering Model

Abstract

With the development of data integration of distribution, the scale of electricity consumption data has increased. How to use the current advanced intelligent management technology to carry out in-depth information mining of electricity big data has become the key to the management of distribution data. Based on the deep clustering algorithm, this paper establishes a clustering model of distribution electricity data. The model not only takes advantage of the K-Means clustering algorithm's simplicity and fast operation speed, but also uses a special RNN model to automatically extract clustering features based on the timing characteristics of electricity data. The comparison with the traditional clustering model shows that the model has a significant increase in clustering accuracy. Finally, the article comprehensively uses the clustering results to automatically detect the abnormal electricity consumption data of the users, thereby improving the work efficiency of the staff.