COLD: Concurrent Loads Disaggregator for Non-Intrusive Load Monitoring (2021-06-04T00:00:00.000000Z)

TL;DR

COLD, a transformer-based model specifically designed to address the challenges of disaggregating high-frequency data with multiple simultaneously working devices is proposed, achieving 95% load identification accuracy and 82% disaggregation performance on the test data.

Abstract

Non-intrusive load monitoring (NILM) enables detailed analyses of household electricity usage by disaggregating the total power consumption into individual appliance-level data. In this paper, we propose COLD, a transformer-based model specifically designed to address the challenges of disaggregating high-frequency data with multiple simultaneously working devices. COLD supports up to 42 devices and accurately handles scenarios with up to 11 concurrent loads, achieving 95% load identification accuracy and 82% disaggregation performance on the test data. In addition, we introduce a new fully labeled high-frequency NILM dataset for load disaggregation derived from the UK-DALE 16 kHz dataset. Finally, we analyze the decline in NILM model performance as the number of concurrent loads increases.

Authors

E. Gryazina

2 papers

Ilia Kamyshev

1 papers

Dmitrii Kriukov

1 papers

TL;DR

Abstract

Authors

References46 items

ST-NILM: A Wavelet Scattering-Based Architecture for Feature Extraction and Multilabel Classification in NILM Signals

Edframe: Open-Source Library for End-to-End Energy Disaggregation in Python

Double Fourier Integral Analysis Based Convolutional Neural Network Regression for High-Frequency Energy Disaggregation

Review on Deep Neural Networks Applied to Low-Frequency NILM

Non-intrusive load disaggregation based on composite deep long short-term memory network

UNet-NILM: A Deep Neural Network for Multi-tasks Appliances State Detection and Power Estimation in NILM

BERT4NILM: A Bidirectional Transformer Model for Non-Intrusive Load Monitoring

Semisupervised Multilabel Deep Learning Based Nonintrusive Load Monitoring in Smart Grids

Near-real-time plug load identification using low-frequency power data in office spaces: Experiments and applications

Multi-Label Learning for Appliance Recognition in NILM Using Fryze-Current Decomposition and Convolutional Neural Network

An Event-Driven Convolutional Neural Architecture for Non-Intrusive Load Monitoring of Residential Appliance

Household Appliance Classification Using Lower Odd-Numbered Harmonics and the Bagging Decision Tree

Non-Intrusive Load Disaggregation by Convolutional Neural Network and Multilabel Classification

Context Aware Energy Disaggregation Using Adaptive Bidirectional LSTM Models

Energy Disaggregation of Appliances Consumptions Using HAM Approach

An Overview of Non-Intrusive Load Monitoring: Approaches, Business Applications, and Challenges

Shipboard Fault Detection Through Nonintrusive Load Monitoring: A Case Study

A System for Massively Parallel Hyperparameter Tuning

Approximating Continuous Functions by ReLU Nets of Minimal Width

Attention is All you Need

A Comprehensive Feature Study for Appliance Recognition on High Frequency Energy Data

SGDR: Stochastic Gradient Descent with Warm Restarts

Layer Normalization

Deep Residual Learning for Image Recognition

A new approach for supervised power disaggregation by using a deep recurrent LSTM network

Neural NILM: Deep Neural Networks Applied to Energy Disaggregation

Adam: A Method for Stochastic Optimization

PLAID: a public dataset of high-resoultion electrical appliance measurements for load identification research: demo abstract

Nonintrusive load monitoring (NILM) performance evaluation

The UK-DALE dataset, domestic appliance-level electricity demand and whole-house demand from five UK homes

Event detection for Non Intrusive load monitoring

Non-Intrusive Load Monitoring Approaches for Disaggregated Energy Sensing: A Survey

Random Search for Hyper-Parameter Optimization

An Approach of Household Power Appliance Monitoring Based on Machine Learning

Unsupervised Disaggregation of Low Frequency Power Measurements

A Tutorial on Support Vector Machines for Pattern Recognition

Nonintrusive appliance load monitoring

k-nearest neighbour classifiers: 2nd edition (with python examples)

WHITED-A Worldwide Household and Industry Transient Energy Data Set

librosa: Audio and Music Signal Analysis in Python

Dropout: a simple way to prevent neural networks from overfitting

REDD : A Public Data Set for Energy Disaggregation Research

“World energy outlook 2024 – analysis.”

Frequency normalization - both voltage and current waveforms are transformed to reference frequency f ref with time-stretch method proposed by librosa library [27]

Signal duration control - signature of duration less than t sec seconds is dropped, otherwise duration is rounded to an integer value to avoid spectral-leakage for further procedures

applicable license agreement with IEEE. Restrictions apply

Field of Study

Journal Information

Name

Page