Application of neural networks to a diagnostic problem in quality control

Yuan Guo, Kevin Dooley

Research output: Chapter in Book/Report/Conference proceedingConference contribution

4 Citations (Scopus)

Abstract

In order to properly diagnose quality problems that occur in manufacturing the diagnostician, be it human or computer, must be privy to various sources of information about the process and its behavior. This paper describes how neural networks can be used to provide knowledge of how a product characteristic changed, i.e. shift in mean or variability, when so noted by the control chart application. This is potentially useful because there might be some underlying knowledge about the physical phenomena in question that relates the behavior of the observed characteristic to its processing variables. When a change in the process is detected, a feature vector of process-related statistics is sent to the neural network. The neural network then classifies the change as a shift in mean or variance, and passes this information on to the diagnostician. This paper addresses various issues concerned with this problem, namely: process change detection, feature vector selection, neural network training patterns, and convergence error rates. Construction and corresponding simulation experiments show that the neural network approach training procedure yielded a net that correctly classified the problem 94 percent of the time, averaged across classes and magnitudes. Error rates for this configuration ranged from 1.2 to 19.3 percent for mean changes and 3.7 to 9.4 percent for variance changes. Clearly the change structure identification problem can be successfully addressed by a neural network approach, resulting in very low error rates and relatively robust performance across different magnitudes.

Original languageEnglish (US)
Title of host publicationAmerican Society of Mechanical Engineers, Production Engineering Division (Publication) PED
EditorsSteven Y. Liang, Tsu-Chin Tsao
Place of PublicationNew York, NY, United States
PublisherPubl by ASME
Pages111-122
Number of pages12
Volume44
StatePublished - 1990
Externally publishedYes
EventWinter Annual Meeting of the American Society of Mechanical Engineers - Dallas, TX, USA
Duration: Nov 25 1990Nov 30 1990

Other

OtherWinter Annual Meeting of the American Society of Mechanical Engineers
CityDallas, TX, USA
Period11/25/9011/30/90

Fingerprint

Quality control
Neural networks
Statistics
Processing
Experiments

ASJC Scopus subject areas

  • Industrial and Manufacturing Engineering
  • Mechanical Engineering

Cite this

Guo, Y., & Dooley, K. (1990). Application of neural networks to a diagnostic problem in quality control. In S. Y. Liang, & T-C. Tsao (Eds.), American Society of Mechanical Engineers, Production Engineering Division (Publication) PED (Vol. 44, pp. 111-122). New York, NY, United States: Publ by ASME.

Application of neural networks to a diagnostic problem in quality control. / Guo, Yuan; Dooley, Kevin.

American Society of Mechanical Engineers, Production Engineering Division (Publication) PED. ed. / Steven Y. Liang; Tsu-Chin Tsao. Vol. 44 New York, NY, United States : Publ by ASME, 1990. p. 111-122.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Guo, Y & Dooley, K 1990, Application of neural networks to a diagnostic problem in quality control. in SY Liang & T-C Tsao (eds), American Society of Mechanical Engineers, Production Engineering Division (Publication) PED. vol. 44, Publ by ASME, New York, NY, United States, pp. 111-122, Winter Annual Meeting of the American Society of Mechanical Engineers, Dallas, TX, USA, 11/25/90.
Guo Y, Dooley K. Application of neural networks to a diagnostic problem in quality control. In Liang SY, Tsao T-C, editors, American Society of Mechanical Engineers, Production Engineering Division (Publication) PED. Vol. 44. New York, NY, United States: Publ by ASME. 1990. p. 111-122
Guo, Yuan ; Dooley, Kevin. / Application of neural networks to a diagnostic problem in quality control. American Society of Mechanical Engineers, Production Engineering Division (Publication) PED. editor / Steven Y. Liang ; Tsu-Chin Tsao. Vol. 44 New York, NY, United States : Publ by ASME, 1990. pp. 111-122
@inproceedings{48dd898e45104d53bdd47aa177cec012,
title = "Application of neural networks to a diagnostic problem in quality control",
abstract = "In order to properly diagnose quality problems that occur in manufacturing the diagnostician, be it human or computer, must be privy to various sources of information about the process and its behavior. This paper describes how neural networks can be used to provide knowledge of how a product characteristic changed, i.e. shift in mean or variability, when so noted by the control chart application. This is potentially useful because there might be some underlying knowledge about the physical phenomena in question that relates the behavior of the observed characteristic to its processing variables. When a change in the process is detected, a feature vector of process-related statistics is sent to the neural network. The neural network then classifies the change as a shift in mean or variance, and passes this information on to the diagnostician. This paper addresses various issues concerned with this problem, namely: process change detection, feature vector selection, neural network training patterns, and convergence error rates. Construction and corresponding simulation experiments show that the neural network approach training procedure yielded a net that correctly classified the problem 94 percent of the time, averaged across classes and magnitudes. Error rates for this configuration ranged from 1.2 to 19.3 percent for mean changes and 3.7 to 9.4 percent for variance changes. Clearly the change structure identification problem can be successfully addressed by a neural network approach, resulting in very low error rates and relatively robust performance across different magnitudes.",
author = "Yuan Guo and Kevin Dooley",
year = "1990",
language = "English (US)",
volume = "44",
pages = "111--122",
editor = "Liang, {Steven Y.} and Tsu-Chin Tsao",
booktitle = "American Society of Mechanical Engineers, Production Engineering Division (Publication) PED",
publisher = "Publ by ASME",

}

TY - GEN

T1 - Application of neural networks to a diagnostic problem in quality control

AU - Guo, Yuan

AU - Dooley, Kevin

PY - 1990

Y1 - 1990

N2 - In order to properly diagnose quality problems that occur in manufacturing the diagnostician, be it human or computer, must be privy to various sources of information about the process and its behavior. This paper describes how neural networks can be used to provide knowledge of how a product characteristic changed, i.e. shift in mean or variability, when so noted by the control chart application. This is potentially useful because there might be some underlying knowledge about the physical phenomena in question that relates the behavior of the observed characteristic to its processing variables. When a change in the process is detected, a feature vector of process-related statistics is sent to the neural network. The neural network then classifies the change as a shift in mean or variance, and passes this information on to the diagnostician. This paper addresses various issues concerned with this problem, namely: process change detection, feature vector selection, neural network training patterns, and convergence error rates. Construction and corresponding simulation experiments show that the neural network approach training procedure yielded a net that correctly classified the problem 94 percent of the time, averaged across classes and magnitudes. Error rates for this configuration ranged from 1.2 to 19.3 percent for mean changes and 3.7 to 9.4 percent for variance changes. Clearly the change structure identification problem can be successfully addressed by a neural network approach, resulting in very low error rates and relatively robust performance across different magnitudes.

AB - In order to properly diagnose quality problems that occur in manufacturing the diagnostician, be it human or computer, must be privy to various sources of information about the process and its behavior. This paper describes how neural networks can be used to provide knowledge of how a product characteristic changed, i.e. shift in mean or variability, when so noted by the control chart application. This is potentially useful because there might be some underlying knowledge about the physical phenomena in question that relates the behavior of the observed characteristic to its processing variables. When a change in the process is detected, a feature vector of process-related statistics is sent to the neural network. The neural network then classifies the change as a shift in mean or variance, and passes this information on to the diagnostician. This paper addresses various issues concerned with this problem, namely: process change detection, feature vector selection, neural network training patterns, and convergence error rates. Construction and corresponding simulation experiments show that the neural network approach training procedure yielded a net that correctly classified the problem 94 percent of the time, averaged across classes and magnitudes. Error rates for this configuration ranged from 1.2 to 19.3 percent for mean changes and 3.7 to 9.4 percent for variance changes. Clearly the change structure identification problem can be successfully addressed by a neural network approach, resulting in very low error rates and relatively robust performance across different magnitudes.

UR - http://www.scopus.com/inward/record.url?scp=0025623091&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0025623091&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:0025623091

VL - 44

SP - 111

EP - 122

BT - American Society of Mechanical Engineers, Production Engineering Division (Publication) PED

A2 - Liang, Steven Y.

A2 - Tsao, Tsu-Chin

PB - Publ by ASME

CY - New York, NY, United States

ER -