The impact of Electrical Signature Quality in NILM process: a preliminary analysis based on unsupervised approch

Non-Intrusive Load Monitoring NILM - has always presented itself as a promising technology for optimising energy management. However, its success is far from obvious due to some inherent limitations. This paper preliminarily analyses the crucial role played by Electrical Signature Quality in the optimisation of NILM processes using a non-supervised approach based on clustering and statistical methods. A simple and effective disaggregation mode is proposed, in which the only a priori knowledge required is that of the nominal nameplate characteristics of the monitored system. The study is based on a realistic context that emulates a real office scenario, with three multi-state electrical equipment, monitored through an appropriate experimental setup.