Development of a Non-Intrusive Load Monitoring (NILM) with Unknown Loads using Support Vector Machine

Abstract

Non-intrusive load monitoring is the process of recognizing and identifying electrical devices and its energy consumption on the entire electrical system through "power signatures". In this process, the aggregated load information is obtained from a single point of measurement. Compared with the traditional way of load identification by setting up multiple devices and sensors, the system uses only one energy measurement device, hence making it more efficient and economical. In this study, the focus was on designing a hardware that can obtain all power quality measurements, data analysis, and appliance identifier, which were analyzed by the microcontroller. The general information and introduction to the system, as well as the past and present literatures about the types of NILM System used by the researchers are presented. It was found that the combined unknown loads can be identified. Three different loads were analyzed at the same time from light bulb, electric fan and heater which gave 8-8.2W, 40-42W, and 238-249W respectively, all determined using a small-scale NILM system equipped with energy metering block and microcontroller that extracts and classifies loads with the use of support vector machine. This has a great significance to the industry and understanding of energy management since the demand for energy is growing rapidly.