Research on the value mining and behavior pattern analysis technology of low-voltage customers' electricity consumption

Abstract

The current electricity consumption information collection system collects the full amount of electrical parameters for large users with high frequency, while only very little data is collected for low-voltage residential users, which leads to the scale and quality of current low-voltage residential users' electricity consumption data to be improved and makes it difficult to discover phenomena such as electricity theft. In this paper, the electric power big data mining technology is used to model the electricity consumption behavior of low-voltage residential users. By using data extraction, de-duplication, completion, cleaning, conversion and other processing techniques, the value of electricity consumption data can be explored and the quality of data can be improved, so that the patterns of electricity consumption behavior under normal or abnormal conditions can be analyzed. Based on actual residential customers' electricity consumption scenarios, the clustering analysis of low-voltage residential customers' electricity consumption behavior patterns is carried out, and an accurate and stable electricity consumption pattern recognition algorithm is established, which greatly improves the utilization rate of electricity consumption data.