0
Research Papers

Evaluating the Use of Artificial Neural Networks and Graph Complexity to Predict Automotive Assembly Quality Defects

[+] Author and Article Information
Apurva Patel

Mechanical Engineering,
Clemson University,
Clemson, SC 29634-0921
e-mail: apurvap@g.clemson.edu

Patrick Andrews

Mechanical Engineering,
Clemson University,
Clemson, SC 29634-0921
e-mail: pcandre@g.clemson.edu

Joshua D. Summers

Professor Mechanical Engineering,
Clemson University,
Clemson, SC 29634-0921
e-mail: jsummer@clemson.edu

Erin Harrison

Assembly Planning,
BMW Manufacturing Co., LLC,
Greer, SC 29651
e-mail: Erin.Harrison@bmwmc.com

Joerg Schulte

Liaison Office,
BMW Manufacturing Co., LLC,
Greer, SC 29651
e-mail: Joerg.Schulte@bmw.de

M. Laine Mears

Professor
Automotive Engineering,
Clemson University,
Clemson, SC 29634
e-mail: mears@clemson.edu

Contributed by the Computers and Information Division of ASME for publication in the JOURNAL OF COMPUTING AND INFORMATION SCIENCE IN ENGINEERING. Manuscript received September 3, 2016; final manuscript received June 2, 2017; published online July 26, 2017. Assoc. Editor: Yong Chen.

J. Comput. Inf. Sci. Eng 17(3), 031017 (Jul 26, 2017) (10 pages) Paper No: JCISE-16-2069; doi: 10.1115/1.4037179 History: Received September 03, 2016; Revised June 02, 2017

This paper presents the use of subassembly models instead of the entire assembly model to predict assembly quality defects at an automotive original equipment manufacturer (OEM). Specifically, artificial neural networks (ANNs) were used to predict assembly time and market value from assembly models. These models were converted into bipartite graphs from which 29 graph complexity metrics were extracted to train 18,900 ANN prediction models. The size of the training set, order of the bipartite graph, selection of training set, and defect type were experimentally studied. With a training size of 28 parts, an interpolation focused training set selection with a second-order graph seeding ensured that 70% of all predictions were within 100% of the target value. The study shows that with an increase in training size and careful selection of training sets, assembly defects can be predicted reliably from subassemblies' complexity data.

FIGURES IN THIS ARTICLE
<>
Copyright © 2017 by ASME
Your Session has timed out. Please sign back in to continue.

References

Figures

Grahic Jump Location
Fig. 1

Bipartite graph of assembly model

Grahic Jump Location
Fig. 2

Basic procedure for ANN prediction

Grahic Jump Location
Fig. 3

Graph order growth of assembly model

Grahic Jump Location
Fig. 4

Algorithm used for generating bipartite graphs for individual parts

Grahic Jump Location
Fig. 5

Extrapolating test set

Grahic Jump Location
Fig. 6

Interpolating test set

Grahic Jump Location
Fig. 7

Building and testing the ANN structure

Grahic Jump Location
Fig. 8

Error distribution with respect to training size

Grahic Jump Location
Fig. 9

Error distribution based on graph order

Grahic Jump Location
Fig. 10

Error distribution based on defect type

Tables

Errata

Discussions

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In