DBSCAN-Optimized AFHMM Non-Intrusive Load Monitoring Method

Read full text Summarise with VIRAT
Authors: Hanxun Qiao, Cheng Zhang, Kai Zhang, Ruoyu Yao, Shangshang Yu, Bing Wang
Cited by: 0
References: 12

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Abstract Non-intrusive load monitoring (NILM) technology, due to its low installation and maintenance costs and enhanced user privacy protection, has a wide range of application scenarios in load monitoring. This paper proposes an efficient and highly adaptive non-intrusive load monitoring method based on DBSCAN-optimized AFHMM to address key challenges in existing techniques. First, to address the difficulty of uniformly presetting cluster numbers for appliances with varying state counts, appliance power data is clustered using the density-based spatial clustering of applications with noise (DBSCAN) algorithm to refine operational states. Secondly, to address the exponential inference complexity of Factorial Hidden Markov Models (FHMMs), an improved Additive Factorial Hidden Markov Model (AFHMM) is proposed. It decomposes joint state inference into independent marginal distribution estimations through convex optimization relaxation, thereby reducing complexity to polynomial time and enhancing algorithmic efficiency. Finally, simulation results on the IAWE dataset validate that the proposed algorithm demonstrates superior performance in both accuracy and efficiency.

Fields of Study & Subjects

Smart Grid Energy Management Energy Load and Power Forecasting Bayesian Methods and Mixture Models

Publication details

Journal / source
Journal of Physics Conference Series
Volume
3198
Issue
1
Pages
012019–012019
Publication date
1 March 2026
Language
EN