Remote Appliance Load Monitoring and Identification in a Modern Residential System With Smart Meter Data

In this article, an innovative procedure for residential electrical load monitoring applicable to smart meters is proposed based on a modified decisive multi-objective optimization. In this procedure, different load features are extracted from household loads for optimization-based monitoring. Each load monitoring feature reflects an objective function, which is also simultaneously minimized using modified Artificial Bee Colony (ABC) algorithm. This technique is not heavily dependent on high amount of training data which is essential for the most machine learning techniques with some modification in choosing the initial search space. The proposed technique uses with real life raw data using low sampling rates. An Internet of Things (IoT) based approach is also implemented for remote monitoring of connected loads in the system along with monitoring. The proposed technique is user friendly, requires only load signature data for verification purpose which can be easily extracted. The paper demonstrates event-based appliance load monitoring and comparison with benchmark dataset which shows marked improvement over existing state-of-the-art-techniques.