Electricity Anomalies Detection and Automation in Smart Meter System

Abstract

Electricity consumption in day-to-day life has a higher impact on household expenses. This paper focuses on the various machine learning (ML) techniques for the detection of electricity anomalies which are the comparison of supplied electricity power and actual customer’s electricity consumption. Electricity theft leads to economic loss and can be very harmful to grid suppliers. This non-technical loss generates a need for detection of theft or anomalies in power system industries. This paper focuses on the comparative study on detection of electricity theft by the researchers in various aspects and proposed the implementation of this study by ML techniques. This paper first provides an overview of electricity theft and the solution provided by the researchers. Paper surveys the research efforts in a field of energy consumption and analytical overview of data and ML algorithms used. It finally identifies the various problems in anomalies detection of electricity paper also focus on the usage of automation in smart meter system with wireless sensor networks (WSNs) which captures the meter readings and upload the reading data on server of grid provider which helps in generation of bill automatically. From electricity data, two aspects are applied for deep analysis of consumption of electricity. This study highlights use of clustering for data filtering and classification for the identifying the two category of consumption. The study uses various ML algorithms to measure performance of model.