ANN Based Approach for Electricity Theft Detection in Single Phase Energy Meters

The major challenges being faced by Power Distribution Utility today are high Aggregate Technical (AT&C) Losses, bad debts, financial liabilities and aging infrastructures. Electricity theft remains a significant challenge in power sector resulting substantial financial losses for utilities and adverse social consequences. Although the AT&C losses have been reduced significantly over past two years, there is a need to further minimize these losses. Due to large database, the utilities find it difficult to implement conventional methods of theft detection. It is therefore essential to develop Artificial Intelligence and Machine Learning based algorithms to tackle this situation. This research paper presents an innovative solution utilizing Artificial Neural Networks (ANN) to detect theft of electricity in case of a single phase meter. The real world data taken from the regions of Maharashtra state in India and consumer profiles are employed to validate the effectiveness of the approach. The proposed methods are novice, simple, cost effective and feasible.