Abstract
Non-technical losses (NTL) is a problem that many utility companies try to solve, often using black-box supervised classification algorithms. In general, this approach achieves good results. However, in practice, NTL detection faces technical, economic, and transparency challenges that cannot be easily solved and which compromise the quality and fairness of the predictions. In this work, we contextualise these problems in an NTL detection system built for an international utility company. We explain how we have mitigated them by moving from classification into a regression system and introducing explanatory techniques to improve its accuracy and understanding. As we show in this work, the regression approach can be a good option to mitigate these technical problems, and can be adjusted in order to capture the most striking NTL cases. Moreover, explainable AI (through Shapley Values) allows us to both validate the correctness of the regression approach in this context beyond benchmarking, and improve the transparency of our system drastically.
Highlights
The services provided by energy companies are essential to societies, but are rather expensive: the necessary infrastructure to provide them includes power plants, kilometres of pipes and lines, and millions of meters, whose economic cost is covered by the bills paid by the companies’ customers and, in many cases, taxes.Another less visible cost that these companies face are the energy losses, i.e., the gap between the energy provided and the energy billed to the customers
Our analysis beyond benchmarking confirms the correctness of the regression model in terms of explainability: we report that regression learns better and more reliable patterns than our previous classification system
In this work we propose a novel approach to detecting Non-technical losses (NTL): to use as a label the amount of energy recovered in an NTL
Summary
The services provided by energy companies are essential to societies, but are rather expensive: the necessary infrastructure to provide them includes power plants, kilometres of pipes and lines, and millions of meters, whose economic cost is covered by the bills paid by the companies’ customers and, in many cases, taxes. Another less visible cost that these companies face are the energy losses, i.e., the gap between the energy provided and the energy billed to the customers. The customers’ pre-selection was based on simple rules indicating an abnormal consumption behaviour according to the stakeholder’s knowledge (e.g., an abrupt decrease of consumption). In the era of big data and machine learning, utility companies exploit the data available in their information systems and combine them with other contextual information to design more accurate campaigns, including statistical and machine learning-based techniques
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
