A novel detector to detect colluded non-technical loss frauds in smart grid

A Non-Technical Loss (NTL) fraud occurs when a fraudster tampers with a smart meter so that the meter registers less electricity consumption than the actual consumed amount, and therefore the utility becomes the victim who suffers the corresponding economic loss. In the literature, many detection schemes have been proposed to detect NTL frauds. However, some NTL frauds are far more complicated than what the existing schemes expect. We recently discovered a new potential type of frauds, a variant of NTL frauds, called Colluded Non-Technical Loss (CNTL) frauds in the Smart Grid. In a CNTL fraud, multiple fraudsters can co-exist or collaborate to commit the fraud. Existing detection schemes cannot detect CNTL frauds since these methods do not consider the co-existing or collaborating fraudsters, and therefore cannot distinguish one from many fraudsters. In this paper, we propose a CNTL fraud detector to detect CNTL frauds. The proposed method can quickly detect a tampered meter based on recursive least squares. After identifying the tampered meter, the proposed scheme can detect different fraudsters using mathematical models. Our experiments show that our method is effective in detecting CNTL frauds.