Energy Disaggregation Using Two-Stage Fusion of Binary Device Detectors

Abstract

A data-driven methodology to improve the energy disaggregation accuracy during Non-Intrusive Load Monitoring is proposed. In detail, the method uses a two-stage classification scheme, with the first stage consisting of classification models processing the aggregated signal in parallel and each of them producing a binary device detection score, and the second stage consisting of fusion regression models for estimating the power consumption for each of the electrical appliances. The accuracy of the proposed approach was tested on three datasets—ECO (Electricity Consumption & Occupancy), REDD (Reference Energy Disaggregation Data Set), and iAWE (Indian Dataset for Ambient Water and Energy)—which are available online, using four different classifiers. The presented approach improves the estimation accuracy by up to 4.1% with respect to a basic energy disaggregation architecture, while the improvement on device level was up to 10.1%. Analysis on device level showed significant improvement of power consumption estimation accuracy especially for continuous and nonlinear appliances across all evaluated datasets.

A Smart Home Energy Management System Using Two-Stage Non-Intrusive Appliance Load Monitoring over Fog-Cloud Analytics Based on Tridium’s Niagara Framework for Residential Demand-Side Management

“…A ROC curve consisting of the false-positive rate (FPR) and true-positive rate (TPR) produces a graph showing the performance, in terms of recognition, of a recognizer where it is examined at all recognition thresholds or with its different configuration settings [ 47 ]. In an ROC curve, an error to a trained and validated recognizer can be computed through the Euclidean distance, from obtained (FPR, TPR) to the perfect recognition (FPR = 0, TPR = 1) [ 29 , 30 ]. Equations (2) and (3) define FPR and TPR, respectively.…”

“…Moreover, the HEMSs that were implemented in [ 7 , 8 , 9 ] monitor relevant electrical appliances in an intrusive way. NIALM approaches that monitor relevant electrical appliances in a non-intrusive fashion have been developed significantly in [ 18 , 25 , 28 , 29 , 30 , 31 , 32 , 33 , 34 ]. These approaches can provide accurate energy disaggregation in many practical cases.…”

Non intrusive load monitoring for demand side management

“…NILM algorithms can be classified as proposed by Schirmer et al (2020) who distinguish between algorithms using source separation, and those who does not. Source separation, or blind signal separation, is a known filtering problem where an aggregate signal composed by a mixture of a set of other signals is processed in order to obtain the individual contributions.…”

Multivariate Non-Negative Matrix Factorization with Application to Energy Disaggregation

“…Non-Intrusive Load Monitoring (NILM) aims to extract the energy consumption per appliance using the measurement of the aggregated consumption of a household or building [1]. Aiming to extract multiple estimations using only a single observation, NILM can be considered as a single-channel source separation problem, which is practically impossible to be solved analytically [2]. The approaches that have been investigated in order to solve the NILM problem and provide estimates of appliances' consumptions can briefly be classified into three categories.…”