Development of an Intelligent System for Smart Home Energy Disaggregation Using Stacked Denoising Autoencoders

Abstract

Energy sustainability remains one of the biggest challenges for the Philippines’ energy sector with 51% of the demand coming from the residential and commercial sectors. Intelligent energy monitoring systems play a key role with the opportunity to contribute sizeable amount of energy savings by providing meaningful consumption feedback to home owners. While smart meters provide an ideal ubiquitous energy monitoring solution, these devices lack appliance-level feedback. In this study, we present a concept of augmenting smart utility meters with an energy disaggregation method using Stacked Denoising Autoencoders to provide appliance-level feedback to home owners. We evaluate the results using absolute mean loss and proportion of energy correctly assigned as metrics for the signal disaggregation. The results show that the model was able to decompose an aggregate appliance signal and provide an itemized appliance-level power consumption.