Statistical assessment of abrupt change detectors for non-intrusive load monitoring

Non-Intrusive Load Monitoring (NILM) is the process of monitoring Home Electrical Appliances (HEAs) from voltage and current measurements at the Point of Common Coupling (PCC). The resulting information can be useful to raise greater awareness among customers about their HEAs' usage for a more efficient energy behaviour. NILM methods can be roughly categorized into event-based and non event-based approaches, depending on whether or not they rely on the detection of HEAs' significant state transitions (e.g. On/Off or state change). In this paper, we focus on event-based methods and describe a test bench to statistically assess three abrupt change detectors. Then, we define several evaluation metrics for judging on performance and quality of the detection algorithms in order to carry out a comparison. Finally, the three abrupt change detectors are applied to real data coming from a scenario of HEAs switched on and off by hand.