An Intelligent Operation and Maintenance System for Power Consumption Based on Deep Learning

2019 4th Asia Conference on Power and Electrical Engineering

能源学;电工学

Huang, Zengrui^1 ; Mao, Wei^1 ; Chen, Ming^2 ; Wu, Qiang^2 ; Xiong, Boyue^2 ; Xu, Wei^2

School of Computer Science, Fudan University, Yangpu District, Shanghai

200433, China^1

State Grid Shanghai Municipal Electric Power, LTD., Minghang District, Shanghai

201199, China^2

关键词: Abnormal behaviours;    Anomaly identification;    Data informations;    Intelligent operations;    Massive data sets;    Mathematical statistics methods;    Multiple dimensions;    Subsequent data analysis;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/486/1/012107/pdf
DOI  :  10.1088/1757-899X/486/1/012107

来源: IOP

【 摘 要 】

With the advent of the era of power big data, power companies can easily access more and more user power data, but they are not making full use of such massive datasets to reflect their value. Due to the huge amount of data collected by the power system, the total amount of abnormal data collected cannot be ignored. These abnormal data cause great interference to subsequent data analysis and maintenance; It's also important to find how to identify user's abnormal behaviour based on the power dataset. In this paper, a smart operation and maintenance system based on deep learning is proposed. Based on the composite mathematical statistics method, the data is efficiently reviewed and cleaned. The isolated forest algorithm is used to quickly identify the users with abnormal power consumption and GAN is used to help evaluate this model. This system can help power companies efficiently complete data review and anomaly identification, and can integrate data information from multiple dimensions to guide the actual maintenance and verification.