Transformation-Driven Hellinger Distance Framework for Electricity Theft Detection in Advanced Metering Infrastructure

View at publisher Summarise with VIRAT
Authors: Sandeep Kumar Singh
Cited by: 0
References: 32

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

The integration of sensing, computing, and communication technologies has improved the efficiency and flexibility of modern power grids. However, it has also increased the vulnerability of advanced metering infrastructure (AMI) to cyber-physical threats, particularly energy theft. This paper proposes a novel, transformation-driven detection framework based on the Hellinger Distance (HD) to identify anomalies in daily electricity consumption at the individual smart meter level. By applying log and antilog transformations to measurement variations, the method enhances distribution separability, improving detection accuracy. The HD metric is then used to quantify statistical deviations between historical and real-time consumption patterns. The proposed scheme operates without requiring system-wide observability or state estimation, making it lightweight and scalable. Experimental validation using real-world datasets from the Irish Social Science Data Archive (ISSDA) and the State Grid Corporation of China (SGCC) smart meter database demonstrates detection rates between 83% and 100% across multiple attack scenarios. Compared with existing statistical and ML-based methods, the proposed approach achieves higher detection rates while remaining computationally efficient, confirming its effectiveness and suitability for practical AMI deployments.

Fields of Study & Subjects

Electricity Theft Detection Techniques Smart Grid Security and Resilience Smart Grid Energy Management

Publication details

Journal / source
IEEE Journal of Emerging and Selected Topics in Industrial Electronics
Volume
7
Issue
3
Pages
1378–1389
Publication date
25 March 2026
Language
EN