Enel and C3 have been working together since 2013. Two of Enel's enterprise-wide digital transformation efforts with C3 are fraud detection and predictive maintenance of distribution assets. With C3, Enel transformed its approach to identifying and prioritizing electricity theft (non-technical loss), with a goal to double the recovery of unbilled energy while improving productivity. The effort required building AI/machine learning algorithms to match the performance delivered by Enel experts using a process honed over 30 years of experience. To accomplish this, the teams worked together to replace traditional non-technical loss identification processes with the C3 Fraud Detection application.

Non-technical losses (NTL) such as electricity theft cause significant harm to our economies, as in some countries they may range up to 40% of the total electricity distributed. Detecting NTLs requires costly on-site inspections. Accurate prediction of NTLs for customers using machine learning is therefore crucial. To date, related research largely ignore that the two classes of regular and non-regular customers are highly imbalanced, that NTL proportions may change and mostly consider small data sets, often not allowing to deploy the results in production. In this paper, we present a comprehensive approach to assess three NTL detection models for different NTL proportions in large real world data sets of 100Ks of customers: Boolean rules, fuzzy logic and Support Vector Machine. This work has resulted in appreciable results that are about to be deployed in a leading industry solution. We believe that the considerations and observations made in this contribution are necessary for future smart meter research in order to report their effectiveness on imbalanced and large real world data sets.