DTW Algorithm Application for Unsupervised Smart Metering system Failure Analysis and End User Consumption Tracking

Abstract

Electric smart meters are small terminals from which plenty of our personal information such as lifestyle, and daily behavior can be collected. With the development of electric power networks and the integration of communication and intelligent facilities, various studies have been conducted to keep the transformation of electric power networks going smart in line with the goals of delivering acceptable power quality. On this journey, as with all technological evolutions, things are not moving the smoothest way they should. Researchers have contributed and ongoing research is helping to ensure stable, reliable, and safe electric power from its generation to the end user. Smart meters have benefited from equal attention and have significantly evolved from their first integration in electric power networks till today. Despite the significant improvements, smart meters are still subject to several malfunctions such as security breaches, operation failures, and inaccurate data collection. This paper presents an unsupervised smart meter operating condition monitoring algorithm based on the Dynamic Time Warping (DTW) algorithm. Our database of three years of data collected on 1010 smart meters is the workbench supporting the variety of tests for the proposed algorithm. The algorithm can detect several types of events giving the possibility to alert on the condition of the meter and the data transmission system.