Test bench and quality measures for non-intrusive load monitoring algorithms

Non-intrusive load monitoring (NILM) is the process of detecting the operation of electrical appliances from the data provided by a smart meter. This information can be useful in many fields, e.g. energy consumption and cost analysis, load prediction, or even ambient assisted living. Many NILM algorithms have been developed in the past but a well-founded comparison was impossible. This is due to the fact that it is very hard to generate suitable data sets. In this paper we describe one way to achieve this with very low effort. Furthermore we define several numbers for judging on the quality of NILM algorithms allowing for an easier comparison. They are a basis for our future work with NILM systems.