Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Energy Disaggregation & Appliance Identification in a Smart Home: Transfer Learning enables Edge Computing

M. Hashim Shahab, Hasan Mujtaba Buttar, Ahsan Mehmood, Waqas Aman, M. Mahboob Ur Rahman, M. Wasim Nawaz, Haris Pervaiz, Qammer H. Abbasi

TL;DR

A novel deep-learning and edge computing approach to solve the NILM problem and a few related problems as follows and achieves a maximum accuracy of 94.6 % for home-NILM, 81 % for site-NILM, and 88.9 % for appliance identification.

Abstract

Non-intrusive load monitoring (NILM) or energy disaggregation aims to extract the load profiles of individual consumer electronic appliances, given an aggregate load profile of the mains of a smart home. This work proposes a novel deep-learning and edge computing approach to solve the NILM problem and a few related problems as follows. 1) We build upon the reputed seq2-point convolutional neural network (CNN) model to come up with the proposed seq2-[3]-point CNN model to solve the (home) NILM problem and site-NILM problem (basically, NILM at a smaller scale). 2) We solve the related problem of appliance identification by building upon the state-of-the-art (pre-trained) 2D-CNN models, i.e., AlexNet, ResNet-18, and DenseNet-121, which are fine-tuned two custom datasets that consist of Wavelets and short-time Fourier transform (STFT)-based 2D electrical signatures of the appliances. 3) Finally, we do some basic qualitative inference about an individual appliance's health by comparing the power consumption of the same appliance across multiple homes. Low-frequency REDD dataset is used for all problems, except site-NILM where REFIT dataset has been used. As for the results, we achieve a maximum accuracy of 94.6\% for home-NILM, 81\% for site-NILM, and 88.9\% for appliance identification (with Resnet-based model).

Energy Disaggregation & Appliance Identification in a Smart Home: Transfer Learning enables Edge Computing

TL;DR

A novel deep-learning and edge computing approach to solve the NILM problem and a few related problems as follows and achieves a maximum accuracy of 94.6 % for home-NILM, 81 % for site-NILM, and 88.9 % for appliance identification.

Abstract

Non-intrusive load monitoring (NILM) or energy disaggregation aims to extract the load profiles of individual consumer electronic appliances, given an aggregate load profile of the mains of a smart home. This work proposes a novel deep-learning and edge computing approach to solve the NILM problem and a few related problems as follows. 1) We build upon the reputed seq2-point convolutional neural network (CNN) model to come up with the proposed seq2-[3]-point CNN model to solve the (home) NILM problem and site-NILM problem (basically, NILM at a smaller scale). 2) We solve the related problem of appliance identification by building upon the state-of-the-art (pre-trained) 2D-CNN models, i.e., AlexNet, ResNet-18, and DenseNet-121, which are fine-tuned two custom datasets that consist of Wavelets and short-time Fourier transform (STFT)-based 2D electrical signatures of the appliances. 3) Finally, we do some basic qualitative inference about an individual appliance's health by comparing the power consumption of the same appliance across multiple homes. Low-frequency REDD dataset is used for all problems, except site-NILM where REFIT dataset has been used. As for the results, we achieve a maximum accuracy of 94.6\% for home-NILM, 81\% for site-NILM, and 88.9\% for appliance identification (with Resnet-based model).
Paper Structure (32 sections, 4 figures, 3 tables, 1 algorithm)

This paper contains 32 sections, 4 figures, 3 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Contributions
  3. Outline
  4. Related Work
  5. NILM Datasets and Data Preprocessing
  6. Datasets used in this work
  7. REDD Dataset kolter2011redd
  8. REFIT Dataset 278e1df91d22494f9be2adfca2559f92
  9. Pre-processing on REDD dataset
  10. Pre-processing on REFIT dataset
  11. Home-NILM & Site-NILM
  12. Home-NILM
  13. Site-NILM
  14. Appliance identification
  15. Custom 2D dataset generation
  16. ...and 17 more sections

Figures (4)

  • Figure 1: Architecture of the proposed seq2-[3]point CNN model. It consists of five convolutional layers, two custom FC layers, and outputs a sequence/vector of size three.
  • Figure 2: Starting from left, first column shows (time-domain) load profile of three appliances (refrigerator, microwave, kitchen outlet); second column shows STFT spectrograms of the three appliances; third column shows the Wavelet of the three appliances; fourth column shows the fusion of two transforms (Wavelet+STFT) for the three appliances. Overall, it can be appreciated that Wavelets and STFT for three appliances are distinct, and thus, could help in appliance identification.
  • Figure 3: The histograms of (refrigerators across five homes): (a) Max and avg. power consumption (b) five power states.
  • Figure 4: Home-NILM (a) on Dishwasher (b) on Microwave. GT is acronym for ground truth, PD is acronym for prediction by the seq2-[3]point model.