Cognitive Smart Plugs for Signature Identification of Residential Home Appliance Load using Machine Learning: From Theory to Practice

Abstract

The identification of individual household appliances in the residential power grid can provide better consumption control and detection of anomalies present in some of these appliances. This identification is only possible if each electric appliance has an Electric Load Signature (ELS). The generation of ELS occurs through the Internet of Things (IoT) equipment, such as Smart Meter (SM) or Smart Plugs (SPs), which provides information necessary for this purpose. The proposed work allows the reading and detection of residential household appliances in the network, through the individual ELS of each load, using SPs together with the Machine Learning Algorithm. Some important electrical parameters will be analyzed and detected individually. With the aid of the Decision Tree and Naive Bayes algorithms, the creation data of each ELS is stored in a centralized database present in the Home Energy Management System and trained so that the identification in each SP is possible. A visual application is provided to the consumer at the HEM to be able to see which appliances are operating, consumption history, as well as anomalies and unwanted changes, present in the residential network.