Characterization of wear particles and their relations with sliding conditions

Akihiko Umeda, Joichi Sugimura, Yuji Yamamoto

Research output: Contribution to journalArticle

27 Citations (Scopus)

Abstract

Wear particles and worn surfaces generated in pin-on-disk steel sliding experiments are studied by microscope image analysis and two types of neural networks. Features of wear particles described by four particle descriptors depend strongly on sliding conditions. A multilayer neural network successfully learns the relations between wear particle features and sliding conditions. If the network is trained with data representing typical features, it also recognizes the particles having similar features. This suggests that the network can be used as a tool for condition monitoring in which the network identifies wear particles produced under unknown sliding conditions to predict that conditions. A self-organizing neural network using the competitive learning rule classifies the wear particles based on their features without any supervisor data. Particle features are expressed by the position on a two-dimensional feature map. This type of network is useful in finding typical particle features, which in turn can be used as supervisor data for the multi-layer neural network. In another application of the self-organizing network, microscopic images of both wear particles and worn surfaces are automatically classified, and characteristics of each surface are represented by the distributions of weights on the feature map.

Original languageEnglish
Pages (from-to)220-228
Number of pages9
JournalWear
Volume216
Issue number2
DOIs
Publication statusPublished - Apr 1 1998

Fingerprint

sliding
Wear of materials
Supervisory personnel
Multilayer neural networks
Neural networks
organizing
Steel
Condition monitoring
Image analysis
Microscopes
image analysis
learning
microscopes
steels
Experiments

All Science Journal Classification (ASJC) codes

  • Condensed Matter Physics
  • Mechanics of Materials
  • Surfaces and Interfaces
  • Surfaces, Coatings and Films
  • Materials Chemistry

Cite this

Characterization of wear particles and their relations with sliding conditions. / Umeda, Akihiko; Sugimura, Joichi; Yamamoto, Yuji.

In: Wear, Vol. 216, No. 2, 01.04.1998, p. 220-228.

Research output: Contribution to journalArticle

Umeda, Akihiko ; Sugimura, Joichi ; Yamamoto, Yuji. / Characterization of wear particles and their relations with sliding conditions. In: Wear. 1998 ; Vol. 216, No. 2. pp. 220-228.
@article{b2c6f827059047d6b0a07f932d0673bd,
title = "Characterization of wear particles and their relations with sliding conditions",
abstract = "Wear particles and worn surfaces generated in pin-on-disk steel sliding experiments are studied by microscope image analysis and two types of neural networks. Features of wear particles described by four particle descriptors depend strongly on sliding conditions. A multilayer neural network successfully learns the relations between wear particle features and sliding conditions. If the network is trained with data representing typical features, it also recognizes the particles having similar features. This suggests that the network can be used as a tool for condition monitoring in which the network identifies wear particles produced under unknown sliding conditions to predict that conditions. A self-organizing neural network using the competitive learning rule classifies the wear particles based on their features without any supervisor data. Particle features are expressed by the position on a two-dimensional feature map. This type of network is useful in finding typical particle features, which in turn can be used as supervisor data for the multi-layer neural network. In another application of the self-organizing network, microscopic images of both wear particles and worn surfaces are automatically classified, and characteristics of each surface are represented by the distributions of weights on the feature map.",
author = "Akihiko Umeda and Joichi Sugimura and Yuji Yamamoto",
year = "1998",
month = "4",
day = "1",
doi = "10.1016/S0043-1648(97)00260-3",
language = "English",
volume = "216",
pages = "220--228",
journal = "Wear",
issn = "0043-1648",
publisher = "Elsevier BV",
number = "2",

}

TY - JOUR

T1 - Characterization of wear particles and their relations with sliding conditions

AU - Umeda, Akihiko

AU - Sugimura, Joichi

AU - Yamamoto, Yuji

PY - 1998/4/1

Y1 - 1998/4/1

N2 - Wear particles and worn surfaces generated in pin-on-disk steel sliding experiments are studied by microscope image analysis and two types of neural networks. Features of wear particles described by four particle descriptors depend strongly on sliding conditions. A multilayer neural network successfully learns the relations between wear particle features and sliding conditions. If the network is trained with data representing typical features, it also recognizes the particles having similar features. This suggests that the network can be used as a tool for condition monitoring in which the network identifies wear particles produced under unknown sliding conditions to predict that conditions. A self-organizing neural network using the competitive learning rule classifies the wear particles based on their features without any supervisor data. Particle features are expressed by the position on a two-dimensional feature map. This type of network is useful in finding typical particle features, which in turn can be used as supervisor data for the multi-layer neural network. In another application of the self-organizing network, microscopic images of both wear particles and worn surfaces are automatically classified, and characteristics of each surface are represented by the distributions of weights on the feature map.

AB - Wear particles and worn surfaces generated in pin-on-disk steel sliding experiments are studied by microscope image analysis and two types of neural networks. Features of wear particles described by four particle descriptors depend strongly on sliding conditions. A multilayer neural network successfully learns the relations between wear particle features and sliding conditions. If the network is trained with data representing typical features, it also recognizes the particles having similar features. This suggests that the network can be used as a tool for condition monitoring in which the network identifies wear particles produced under unknown sliding conditions to predict that conditions. A self-organizing neural network using the competitive learning rule classifies the wear particles based on their features without any supervisor data. Particle features are expressed by the position on a two-dimensional feature map. This type of network is useful in finding typical particle features, which in turn can be used as supervisor data for the multi-layer neural network. In another application of the self-organizing network, microscopic images of both wear particles and worn surfaces are automatically classified, and characteristics of each surface are represented by the distributions of weights on the feature map.

UR - http://www.scopus.com/inward/record.url?scp=0032052998&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0032052998&partnerID=8YFLogxK

U2 - 10.1016/S0043-1648(97)00260-3

DO - 10.1016/S0043-1648(97)00260-3

M3 - Article

AN - SCOPUS:0032052998

VL - 216

SP - 220

EP - 228

JO - Wear

JF - Wear

SN - 0043-1648

IS - 2

ER -