Anomaly Detection in Smart Meters: Analytical Study

Abstract

Smart grid comprises of various components such as SCADA, AMI that collects a large amount of data at regular intervals of an hour or minute. An ever-increasing population necessitates the monitoring and management of electricity use. It’s not only the problem with excessive power consumption, but also with fluctuations in power supply owing to power theft, leakage, poor infrastructure and incorrect billing. The methodology of data analytics and outcomes of analyzing the dataset available for identifying the improper patterns using anomaly detection algorithms are discussed in this paper. In addition, the study investigated at the tools and platforms for data analytics and simulation environments. Since the estimated data does not show variance with the actual data, it may still be incorrect to judge the dataset is anomaly free. We presented an ICT-solution to simplify smart meter data analyses in this study. Present paper provides an exhaustive simulation based analytical study on smart meters to predict some anomalies like energy leakage, theft etc.