Hybrid Learning Paradigm for Non-Intrusive Load Monitoring of Residential Customers

Residential and commercial customers are the major energy consumers in the world and their consumption is constantly increasing. Effective demand-side management (DSM) schemes are developed to control the consumption of consumers through appliance load monitoring (ALM). ALM can be done through smart sensors or plugs, but due to their high cost, load disaggregation techniques are gaining more popularity. In this paper, a hybrid adaptive-neuro fuzzy inference system (ANFIS) has been developed for load monitoring. In the proposed technique, event detection is used to detect load activation events. Then a fine tree classifier is used for the classification of target appliances. Moreover, energy consumed by the target appliances is also estimated. The technique is applied to the publicly available UK Domestic Appliance-Level Electricity UK-DALE) dataset for validation purposes and the results are compared with the existing literature.