Electricity Theft Detection for Smart Grid Security using Smart Meter Data: A Deep Learning - CNN based Approach

Abstract

Not only can theft of energy result in monetary losses, but it also results in expenses that are not technical for energy providers and even for the power infrastructure. Theft of energy has a detrimental effect on both the financial viability and the quality of the electricity. Through the integration of information and energy flows, smart grids have the potential to avoid power theft. The analysis of data from smart grids makes it easier to identify instances of power theft. In contrast, previous systems fared badly when it came to detecting instances of energy theft. To help and assess energy supply businesses in decreasing the barriers of low energy, unexpected power use, and poor power management, we presented in this study a method to detect electricity theft based on consumption data from smart meters. This was done in order to assist and evaluate energy supply businesses. More specifically, the Deep CNN model is able to successfully accomplish two tasks: it differentiates between periodic and non-periodic energy while maintaining the overall features of the power consumption dataset. When it comes to detecting instances of energy theft, the results of the tests indicate that the deep CNN model displays the highest level of accuracy and exceeds prior versions.

Keywords

• Electricity theft
• Economic losses
• Smart meter
• Convolutional neural networks
• Power consumption