Interpolation and fraud detection on data collected by automatic meter reading

Abstract

Automatic and remote reading systems of energy meters are spreading more each day. However, electricity meter data sometimes bear missing elements and outliers, due to communication faults, device faults, or energy fraud. We set up mathematical models in order to be able to interpolate missing data and detect fraud. In this work, two models are developed and compared in terms of performance, using long-term real consumption data.