Abnormal Operations Detection of Residential Electric Appliances using Non-Intrusive Load Monitoring

A real-time abnormal condition monitoring system for household appliances will help save energy, enhance safety, and extend the lifetime of appliances. Most solutions proposed in the past involve individual appliance-level monitoring and, thus, cannot be deployed directly on residential premises with a large number of appliances due to excessive cost and complexity. In this paper, we present a real-time abnormal condition detection mechanism that utilizes the principles of non-intrusive load monitoring to overcome the issues of existing work. A detection accuracy of 99 percent was achieved in experiments conducted with the generated custom dataset and two publicly available datasets.