NILM Based Conservation Plan Of Residential Load Through AI

Abstract

Managing and tracking electricity usage within homes is crucial to decreasing energy consumption and promoting sustainability. To achieve this, The research on Non-Intrusive Load Monitoring (NILM) methods accurately estimate the power consumption of individual appliances using aggregate measurements of voltage or current in the distribution system. In this study, we evaluate and analyze different learning algorithms used in NILM research and propose criteria to measure their effectiveness. Our aim is to provide a comprehensive assessment of the different approaches and determine the most effective method for monitoring electricity consumption in homes. Work was carried on small scale in which the algorithm is trained on an available residential data set, the testing is done on selected appliances that are commonly used in households. The process initiates with the total aggregated power signal that carries overall voltage and current information. The algorithm separates each appliance's energy consumption and shows the power consumed by them individually.