Low-Frequency Energy Disaggregation based on Active and Reactive Power Signatures

Abstract

Non-Intrusive Load Monitoring aims to extract the energy consumption of individual electrical appliances through disaggregation of the total power consumption as measured by a single smart meter at a household. Deep neural networks and especially Convolutional Neural Networks (CNNs) have become popular in solving the Non-Intrusive Load Monitoring problem. However, since NILM is a time series problem mostly 1-D CNNs have been utilized, thus not fully exploiting the capability of CNNs which are advantageous mostly in 2-D data such as images. Therefore, in the proposed architecture 2-D signatures of low frequency active and reactive power are utilized. The proposed architecture was evaluated on the AMPds2 dataset reporting performances up-to 96.1% in terms of estimation accuracy outperforming all previously reported approaches on the same dataset by 1.1%, in terms of absolute improvement.