The detection of non-technical losses and electricity theft by smart meter data and Artificial Intelligence in the context of electric distribution utilities: A comprehensive review

Abstract

Across the world, electric distribution utilities are facing two major challenges i.e. non-technical losses and electricity theft. The amount of these losses is very high (around 40% of the total electric transmission and distribution power) which cannot be ignored because the performance of electric distribution networks adversely falls down due to these losses. To minimize these losses, many traditional methods are in practice but these are not so effective and also very time consuming. Hence, keen researchers are looking forward about recent technology i.e. in the field of Artificial Intelligence because NTL detection by Artificial Intelligence is superior to the traditional techniques in terms of accuracy, efficiency, time-consumption, precision, and labor requirement and can significantly reduce this loss through data analysis techniques. As the smart distribution network generates a massive amount of data i.e. the power consumption of individual users and so on. Therefore, by using this data, machine learning and deep learning techniques can accurately identify electricity theft users. The existing literature only shows the detection of NTL using Artificial Intelligence (AI) so far. This paper provides the causes of NTL followed by an impact on economies, a variation of NTL in different countries. Further, we have studied thoroughly the exercise of technical surveys. On the basis of different AI techniques and essential parameters, a comparison with the existing work has been analyzed. Different simulation tools and compatible environments have been explained. Multiple challenges occur during AI-based detection of NTL, and their possible solutions are also being discussed.