AI-Based Electricity Theft Detection System for Distribution Line, Auto Cutt-off

Authors:

Supriya Mhaske, Vaishnavi Jadhav, Dhiraj Shinde, Prajwal Narute

Department of Electrical Engineering

SVPM’S College of Engineering, Malegaon (Bk), Tal: Baramati, Dist: Pune, INDIA. Sub Author – Mr. Yadav S.N.

Assistant Professor, Department of Electrical Engineering College of Engineering, Malegaon (Bk), India Sny837@gmail.com

Abstract—Energy theft in transmission lines causes significant power losses and affects the overall efficiency and reliability of electrical power systems. In many practical cases, theft detection is carried out using manual inspection and simple monitoring techniques, which are time-consuming and often fail to detect hidden or irregular theft activities. Hence, there is a strong need for an automated and intelligent approach to identify energy theft in real time.

This paper presents an AI-based energy theft detection system for transmission lines using real-time electrical data analysis. The proposed system collects parameters such as voltage, current, power flow, and load variation from monitoring units installed at different points on the transmission line. The collected data is processed to understand the normal operating behavior of the system, and a machine learning–based model is used to detect abnormal patterns that may indicate energy theft.

The system focuses on identifying unusual changes in power flow and load patterns that cannot be explained by normal operating conditions. By learning from historical and real-time data, the AI model improves its ability to distinguish between genuine load variations and suspicious activities related to theft. This makes the detection process more reliable compared to traditional rule-based methods.

Based on the output of the AI model, the system classifies operating conditions as normal or suspicious and generates alerts for further investigation. This method reduces dependency on continuous manual supervision and improves the speed and accuracy of theft detection. The proposed system is simple to implement, scalable in nature, and suitable for integration with smart grid and modern energy management systems to reduce non-technical power losses..

Keywords— Energy theft detection, transmission line monitoring, artificial intelligence, machine learning, anomaly detection, non-technical losses, Arduino-based system, smart grid