Electricity metering and classification of tampered meters

In this paper, we identified electric meter tampering as a major problem for DISCOs and developed a technique that determined the status of each customer’s electricity meter using 100 samples data points on electricity consumption data. The data was modeled and the expected electricity consumption for the month was determined. We also determine the consumers meter status and classify them based on either tampered or non-tampered using the sample expected consumption as a benchmark. A Binomial probability distribution model was used to model the meter status and it was found that the probabilities for the non-tampered and tampered meters were 2.79397E-08 and 9.32323E-10 respectively.