Condition classification of heating systems valves based on acoustic features and machine learning

Potočnik, Primož (author), Olmos Lopez-Roso, Borja (author), Vodopivec, Lučka (author), Susič, Egon (author), Govekar, Edvard (author)

The quality and condition of valves installed in district heating systems can be reflected by the soundsemitted. In this paper, a framework for a systematic approach towards the classification of valve soundsis proposed, based on acoustic features and machine learning models. The methods include the extractionof spectral and psychoacoustic features, alongside the application of a wrapper-based feature selectionmethod which, when combined with machine learning models, simultaneously selects the most informa-tive features and builds optimal classification models. The maximal balanced classification rate (BCR) wasused as the optimisation criterion in this study. Results demonstrate that the specific valve conditions canbe correctly classified with a high BCR as follows: cavitation BCR = 1, whistling BCR = 0.978, and rattlingBCR = 1. The proposed framework for a wrapper-based selection of informative features and correspond-ing machine learning models confirms the usefulness of psychoacoustic features and machine learningmodels for the classification of valve conditions. The proposed framework is, however, general and canbe applied to various acoustic-based industrial condition monitoring challenges.

• Language: English
• Keywords: valves, district heating, acoustic features, feature selection, classification, machine learning
• Work type: Article
• Typology: 1.01 - Original Scientific Article
• Organization: FS - Faculty of Mechanical Engineering
• Publication status: Published
• Publication version: Author Accepted Manuscript
• Year: 2020
• Number of pages: Str. 1-9
• Numbering: Vol. 174
• PID: 20.500.12556/RUL-121854
• UDC: 628.8:534(045)
• ISSN on article: 0003-682X
• DOI: 10.1016/j.apacoust.2020.107736
• COBISS.SI-ID: 35370243
• Publication date in RUL: 03.11.2020

Record is a part of a journal

• Title: Applied acoustics
• Shortened title: Appl. Acoust.
• Publisher: Elsevier Science
• ISSN: 0003-682X
• COBISS.SI-ID: 24982016

Secondary language

• Language: Slovenian
• Keywords: ventili, daljinsko ogrevanje, akustične značilke, izbira značilk, razvrščanje, strojno učenje

Projects

• Funder: ARRS - Slovenian Research Agency
• Project number: P2-0241
• Name: Sinergetika kompleksnih sistemov in procesov