Diagnosis of nontechnical energy losses using Bayesian Networks

Nontechnical losses are a major problem in Brazil as in many developing countries. This paper proposes a methodology that quantifies sources of nontechnical losses and their regional distribution using Bayesian Network. This is a probabilistic graphical model in which a problem is structured as a set of variables linked through probabilistic relationships. Five groups were inspected through random sampling, and the diagnosis with likelihoods for each variable was calculated. Then, situations of conformity and nontechnical losses were estimated. The model has been applied in a Brazilian distribution concessionary, showing its utility in planning efficient actions to reduce nontechnical losses.