Comparative Evaluation of Machine Learning Models and Input Feature Space for Non-intrusive Load Monitoring

Attique Ur Rehman,Brice Vall鑣,Shafiqur Rahman Tito,Tek Tjing Lie

doi:10.35833/mpce.2020.000741

Abstract

Recent advancement in computational capabilities has accelerated the research and development of non-intrusive load disaggregation. Non-intrusive load monitoring (NILM) offers many promising applications in the context of energy efficiency and conservation. Load classification is a key component of NILM that relies on different artificial intelligence techniques, e.g., machine learning. This study employs different machine learning models for load classification and presents a comprehensive performance evaluation of the employed models along with their comparative analysis. Moreover, this study also analyzes the role of input feature space dimensionality in the context of classification performance. For the above purposes, an event-based NILM methodology is presented and comprehensive digital simulation studies are carried out on a low sampling real-world electricity load acquired from four different households. Based on the presented analysis, it is concluded that the presented methodology yields promising results and the employed machine learning models generalize well for the invisible diverse testing data. The multi-layer perceptron learning model based on the neural network approach emerges as the most promising classifier. Furthermore, it is also noted that it significantly facilitates the classification performance by reducing the input feature space dimensionality.

Highlights

WITH the fast development pace of the electronics mar‐ ket, the energy demand has risen exponentially in the last two decades
All the employed machine learning (ML) models are independently trained with 20-day load data from a sin‐ gle household and later tested on a diverse set of testing da‐ ta that are not known in the training phase
Ta‐ ble IV presents the details of households in New Zealand GREEN Grid used for the training and testing purposes of the employed ML models along with the corresponding re‐ sults in terms of event detection and feature extraction

Summary

Introduction

WITH the fast development pace of the electronics mar‐ ket, the energy demand has risen exponentially in the last two decades. The variability and forecasting un‐ certainty of energy consumption patterns make it difficult for the utilities to maintain the equilibrium between demand and supply. In this context, effective energy monitoring is essen‐ tial for modern power systems. Energy monitoring offers many promising solutions for the grid stability, including but not limited to energy forecasting, demand-side management, Manuscript received: October 29, 2020; accepted: March 19, 2021. Date of online publication: June 25, 2021

Objectives

Methods

Results

Conclusion