Details
| Original language | English |
|---|---|
| Pages (from-to) | 599-607 |
| Number of pages | 9 |
| Journal | Industrial Lubrication and Tribology |
| Volume | 76 |
| Issue number | 5 |
| Early online date | 21 May 2024 |
| Publication status | Published - 26 Jun 2024 |
Abstract
Purpose: This study aims to explore the integration of machine learning (ML) in tribology to optimize lubrication interval decisions, aiming to enhance equipment lifespan and operational efficiency through wear image analysis. Design/methodology/approach: Using a data set of scanning electron microscopy images from an internal combustion engine, the authors used AlexNet as the feature extraction algorithm and the J48 decision tree algorithm for feature selection and compared 15 ML classifiers from the lazy-, Bayes and tree-based families. Findings: From the analyzed ML classifiers, instance-based k-nearest neighbor emerged as the optimal algorithm with a 95% classification accuracy against testing data. This surpassed individually trained convolutional neural networks’ (CNNs) and closely approached ensemble deep learning (DL) techniques’ accuracy. Originality/value: The proposed approach simplifies the process, enhances efficiency and improves interpretability compared to more complex CNNs and ensemble DL techniques.
Keywords
- Artificial intelligence, Feature classification, Feature extraction, Machine learning, Wear
ASJC Scopus subject areas
- Engineering(all)
- Mechanical Engineering
- Energy(all)
- Materials Science(all)
- Surfaces, Coatings and Films
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
In: Industrial Lubrication and Tribology, Vol. 76, No. 5, 26.06.2024, p. 599-607.
Research output: Contribution to journal › Article › Research › peer review
}
TY - JOUR
T1 - Wear particle image analysis
T2 - feature extraction, selection and classification by deep and machine learning
AU - Vivek, Joseph
AU - Venkatesh S, Naveen
AU - Mahanta, Tapan K.
AU - V, Sugumaran
AU - Amarnath, M.
AU - Ramteke, Sangharatna M.
AU - Marian, Max
N1 - Publisher Copyright: © 2024, Emerald Publishing Limited.
PY - 2024/6/26
Y1 - 2024/6/26
N2 - Purpose: This study aims to explore the integration of machine learning (ML) in tribology to optimize lubrication interval decisions, aiming to enhance equipment lifespan and operational efficiency through wear image analysis. Design/methodology/approach: Using a data set of scanning electron microscopy images from an internal combustion engine, the authors used AlexNet as the feature extraction algorithm and the J48 decision tree algorithm for feature selection and compared 15 ML classifiers from the lazy-, Bayes and tree-based families. Findings: From the analyzed ML classifiers, instance-based k-nearest neighbor emerged as the optimal algorithm with a 95% classification accuracy against testing data. This surpassed individually trained convolutional neural networks’ (CNNs) and closely approached ensemble deep learning (DL) techniques’ accuracy. Originality/value: The proposed approach simplifies the process, enhances efficiency and improves interpretability compared to more complex CNNs and ensemble DL techniques.
AB - Purpose: This study aims to explore the integration of machine learning (ML) in tribology to optimize lubrication interval decisions, aiming to enhance equipment lifespan and operational efficiency through wear image analysis. Design/methodology/approach: Using a data set of scanning electron microscopy images from an internal combustion engine, the authors used AlexNet as the feature extraction algorithm and the J48 decision tree algorithm for feature selection and compared 15 ML classifiers from the lazy-, Bayes and tree-based families. Findings: From the analyzed ML classifiers, instance-based k-nearest neighbor emerged as the optimal algorithm with a 95% classification accuracy against testing data. This surpassed individually trained convolutional neural networks’ (CNNs) and closely approached ensemble deep learning (DL) techniques’ accuracy. Originality/value: The proposed approach simplifies the process, enhances efficiency and improves interpretability compared to more complex CNNs and ensemble DL techniques.
KW - Artificial intelligence
KW - Feature classification
KW - Feature extraction
KW - Machine learning
KW - Wear
UR - http://www.scopus.com/inward/record.url?scp=85193524385&partnerID=8YFLogxK
U2 - 10.1108/ILT-12-2023-0414
DO - 10.1108/ILT-12-2023-0414
M3 - Article
AN - SCOPUS:85193524385
VL - 76
SP - 599
EP - 607
JO - Industrial Lubrication and Tribology
JF - Industrial Lubrication and Tribology
SN - 0036-8792
IS - 5
ER -