Reviewing Non-intrusive Load Monitoring Using a Pilot Study of an IoT Device to Disaggregate Energy Usage

Non-intrusive load monitoring (NILM) disaggregates energy consumption data collected from a single measurement point into appliance-level data. This process facilitates energy savings. Most studies treat NILM as a residential task with few considering its application in industry. By chronologically reviewing existing literature, this paper presents a review of the latest research in NILM, focusing on its potential employment within a utility company, Northern Ireland Water. A practical example of NILM is also provided using data collected by a pilot IoT device where the benefits of NILM are exhibited via a cost analysis. Results from the literature review show deep learning models to be the most recent preferred disaggregation approach. Furthermore, the standardization of evaluation metrics is deemed essential to facilitate the comparison of different disaggregation models. Finally, the NILM tool kit is outlined as a useful platform for Northern Ireland Water to practically implement NILM.