Abstract:Non-intrusive load monitoring (NILM) detects the energy consumption of individual appliances by monitoring the overall electricity usage in a building. By analyzing voltage and current characteristics, NILM can recognize the usage patterns of various appliances, thus facilitating energy conservation and management. To implement non-intrusive load classification and recognition more effectively, this study proposes an ensemble learning algorithm based on soft voting, which comprises a decision tree, K-nearest neighbor algorithm, and multilayer perceptron (EL-SV ). In this study, the voltage and current features in the plug-load appliance identification dataset (PLAID) and worldwide household and industry transient energy dataset (WHITED) are used as input data. The dataset is examined thoroughly and preprocessed before it is fed into the EL-SV . During preprocessing, six different normalization techniques are applied to the data to improve the accuracy and reliability of the machine-learning model, thus rendering the proposed algorithm more adept at classifying and recognizing appliances. The proposed method is validated by comparing it with other machine learning algorithms in terms of accuracy, precision, recall, and F1 score under the six different normalization methods. For the PLAID, the proposed algorithm can achieve high accuracy scores of 99.79%, 98%, 98.11%, 98.36%, 96.42%, and 98.76% under the min–max normalization, MaxAbs scaler, robust scaler, z-score normalization, L1 normalization, and Yeo–Johnson transformation, respectively. Similarly, for the WHITED, the proposed algorithm can achieve high accuracy scores of 9- .31%, 98.14%, 98.3%, 98.35%, 97.65%, and 98.02% under the abovementioned normalization methods. The results show that the proposed EL-SV algorithm outperforms the other ten machine learning algorithms examined in this study.

Single Appliance Recognition Using Statistical Features Based K-Nn Classification

A residual convolutional neural network with multi-block for appliance recognition in non-intrusive load identification

Appliance Fingerprinting Using Sound from Power Supply

Classification of Electricity Consumption Behavior Based on Improved K-Means and LSTM

A Recognition Algorithm of Household Appliances Based on Bayes Classification Model

Tracking states of massive electrical appliances by lightweight metering and sequence decoding

Household Appliance Recognition Through a Bayes Classification Model

Multi-State Household Appliance Identification Based on Convolutional Neural Networks and Clustering

Appliance Detection Using Very Low-Frequency Smart Meter Time Series

A Feature Engineering‐Based NILM Framework for Appliance Recognition Considering Data Class Imbalance

Neural Network Ensembles to Real-time Identification of Plug-level Appliance Measurements

Non-Intrusive Energy Use Monitoring for a Group of Electrical Appliances

Appliance Event Detection -- A Multivariate, Supervised Classification Approach

Appliance Identification via Combinatorial Fusion Analysis-Assisted Bayesian-Optimized Classifier

Discover The Fingerprint Of Electrical Appliance: Online Appliance Behavior Learning And Detection In Smart Homes

Non-intrusive energy estimation using random forest based multi-label classification and integer linear programming

Non-Intrusive Load Classification and Recognition Using Soft-Voting Ensemble Learning Algorithm With Decision Tree, K-Nearest Neighbor Algorithm and Multilayer Perceptron

Deep Learning Based Energy Disaggregation and On/Off Detection of Household Appliances

Mining the Big Data of Residential Appliances in the Smart Grid Environment

Toward Energy-Awareness Smart Building: Discover the Fingerprint of Your Electrical Appliances

A Fusion Framework Using Integrated Neural Network Model for Non-Intrusive Load Monitoring