Application of Statistical Theory for Analyzing Real TIME Abnormal Electricity Consumption Behavior From Automatic Meter Reading

The authors researched and found an algorithm and built a software program to analyze real time abnormal electricity consumption behavior from Automatic Meter Reading (AMR) combining the theoretical basis of correlation statistics and linear regression, which helps to quickly detect electrical leakage from 30 days down to 2-3 days, at the same time shorten the search scope from 18% to only 0.44% of total electricity customers. The research has contributed to limiting the situation of petitioning for high electricity bills as well as avoiding electric shock accidents due to electrical leakage, proactively providing information to customers with spikes via Email, Zalo, SMS,..., contributing to the satisfaction of electricity customers; At the same time, supporting the Electricity sector in carrying out the task of digital transformation in business and customer service