Commercial losses estimations through consumers´ behavior analysis

The recovery of lost incomes caused by commercial losses represents, to the electric sector, an important resource for improving the electric power distribution companies´ results by increasing their revenues. Attempts to combat fraud are regularly carried through, but however efficient they are to prevent future losses, they are not capable of fully recovering the losses that may have already occurred. Once the income is lost, there is the necessity to esteem the amount of lost energy to charge the consumer who has committed the fraud. The traditional methods, such as applying correction factors (when it is possible to evaluate measurement errors) or analyzing the consumption history (when the fraud is recent), are not viable for the cases where previous measurements of consumption do not exist or are invalid. This work developed a methodology to estimate consumption through the consumers´ behavior analysis, with the definition of demand and load characteristic factors, endorsed by measurements. With the installed power (obtained through load inspection at the moment of the fraud finding), the customers with irregular installation (that have not being charged and do not have consumption history) can have their energies estimated with precision. This paper presents concepts, the description of the methodology adopted to estimate customers´ past consumption and the results obtained, suggesting the procedures for future data updates on the support software program developed and the alternatives for future improvements. Between the main stages of the methodology, it is important to highlight the generation of typical factors, where customers demand and load characteristic factors are calculated, and consumption estimated through statistical regression models, defining models for each consumption class and activity. The work developed provided the energy consumption estimations for fraudsters with no consumption history, using only the installed power. Comparisons b- tween the methods developed and the techniques previously available had also been carried through.