Energy theft and defective meters detection in AMI using linear regression

Abstract

Electricity theft is always a ticklish problem faced by utilities around the world. To mitigate and detect energy theft, utilities are leveraging on the consumers' energy consumption dataset obtained from advanced metering infrastructure to identify anomalous consumption patterns. However, real energy theft sample as well as the distribution station smart meter readings do not exist in Malaysia because smart grid is not fully implemented. Therefore, we design and construct a small-scaled advanced metering infrastructure test rig in the laboratory to evaluate the performance and reliability of our previously proposed linear regression-based detection schemes for energy theft and defective meters in small grid environment. Simulations and electrical tests are conducted and the results show that the proposed algorithms can successfully detect all the fraudulent consumers and discover faulty smart meters in smart grids.