Spatial estimation of nontechnical losses in electric power distribution systems

Abstract

Non-technical losses are mainly caused by clandestine connections, energy thefts, meter fraud, self-connections. These losses generate financial losses to electricity distributors, damage to society, damage to the quality of energy, damage to the reliability of distribution networks, damage to energy efficiency and various other problems. Non-technical losses represent a share of global losses in the electricity distribution system. Numerous previous research has concentrated on detection of these losses. These works determine which consumer units have fraud in the energy meter for example. In general, they use soft computing techniques: neural networks, fuzzy logic, data mining, and text mining. However, the study of the characteristics of the place where losses occur can provide relevant information for a better understanding of the problem. In this context, a new complementary approach is adopted to the problem of losses with a focus on incorporating geographic space into the problem of non-technical losses. Spatial data analysis is used to produce maps that indicate the subareas of the city with a higher probability of non-technical losses. These maps are a tool for quick and easy interpretation. They can be used in conjunction with conventional tools for detection of losses in decision support systems to determine priority regions in the city for combat and prevention of non-technical losses.