Fault detection and classification with the rebmix R package

Nagode, Marko (author), Panić, Branislav (author), Klemenc, Jernej (author), Oman, Simon (author)

Fault detection and classification is an important part of assessing the structural and system health status. The classification and detection of faults and faulty units is mostly done with statistical methods. After the data are measured and collected, the use of statistical software is necessary. Currently, many statistical software packages are being developed for the R programming language, as a result of R implementation being open source and free to use. This paper focuses on the rebmix R package, which concentrates on mixture model estimation. Mixture models, in particular Gaussian mixture models, are the main driver for many practical applications, such as clustering and classification. Hence, in this paper, we have expanded the rebmix for the estimation of the Gaussian mixtures. The results acquired on three different fault classification datasets were promising. Additionally, the process of obtaining those results is shown in detail, giving the researchers in the fault classification field useful resources for their research.

• Language: English
• Keywords: fault classification, fault detection, mixture models, REBMIX algorithm, EM algorithm
• Work type: Article
• Typology: 1.01 - Original Scientific Article
• Organization: FS - Faculty of Mechanical Engineering
• Publication status: Published
• Publication version: Version of Record
• Year: 2023
• Number of pages: 12 str.
• Numbering: Vol. 185, art. 109628
• PID: 20.500.12556/RUL-151319
• UDC: 519.876.5:004.92
• ISSN on article: 0360-8352
• DOI: 10.1016/j.cie.2023.109628
• COBISS.SI-ID: 166878723
• Publication date in RUL: 04.10.2023

Record is a part of a journal

• Title: Computers & industrial engineering
• Shortened title: Comput. ind. eng.
• Publisher: Elsevier
• ISSN: 0360-8352
• COBISS.SI-ID: 4909066

Licences

• License: CC BY-NC-ND 4.0, Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International
• Link: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Description: The most restrictive Creative Commons license. This only allows people to download and share the work for no commercial gain and for no other purposes.

Secondary language

• Language: Slovenian
• Keywords: klasifikacija napak, detekcija napak, mešani model, algoritem REBMIX, algoritem EM

Projects

• Funder: ARRS - Slovenian Research Agency
• Project number: P2-0182
• Name: Razvojna vrednotenja