MO-NILM: A multi-objective evolutionary algorithm for NILM classification

Ram Machlev*, Juri Belikov, Y. Beck, Y. Levron

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

35 Scopus citations

Abstract

Non-intrusive load monitoring (NILM) techniques estimate the consumption of individual appliances in a household or facility, based on readings of a centralized meter. In this work a new method for multi-dimensional NILM signals is proposed—the Multi-objective NILM (MO-NILM). While classical NILM algorithms are based on a single objective function, MO-NILM classifies NILM events by solving a multi-objective optimization problem. The main idea is to model each NILM feature as an objective function, and to mutually minimize these objectives based on the Non-dominated Sorting Genetic Algorithm II (NSGA-II). The presented algorithms can operate in real time using low sampling rates (0.25 Hz and lower) without training the system. In addition, the proposed algorithm is simple, and requires information on the average power signatures of each appliance. The method shows good performance in terms of standard measures when tested on the popular REDD and AMPds datasets.

Original languageEnglish
Pages (from-to)134-144
Number of pages11
JournalEnergy and Buildings
Volume199
DOIs
StatePublished - 15 Sep 2019

Funding

FundersFunder number
Israeli Innovation Authority60689
Israel Science Foundation2//7221

    Keywords

    • Evolutionary algorithm
    • Multi-objective optimization (MOO)
    • Non sorting genetic algorithm II (NSGA-II)
    • Non-intrusive load monitoring (NILM)
    • Power signature

    Fingerprint

    Dive into the research topics of 'MO-NILM: A multi-objective evolutionary algorithm for NILM classification'. Together they form a unique fingerprint.

    Cite this