Research Papers

A Measure of the Information Loss for Inspection Point Reduction

[+] Author and Article Information
Kristina Wärmefjord, Rikard Söderberg

Department of Product and Production Development, Chalmers University of Technology, SE-412 96 Göteborg, Sweden

Johan S. Carlson

 Fraunhofer-Chalmers Research Centre, Chalmers Science Park, SE-412 88 Göteborg, Sweden

J. Manuf. Sci. Eng 131(5), 051017 (Sep 25, 2009) (6 pages) doi:10.1115/1.4000105 History: Received April 20, 2009; Revised August 11, 2009; Published September 25, 2009

Since the vehicle program in the automotive industry gets more and more extensive, the costs related to inspection increase. Therefore, there are needs for more effective inspection preparation. In many situations, a large number of inspection points are measured, despite the fact that only a small subset of points is needed. A method, based on cluster analysis, for identifying redundant inspection points has earlier been successfully tested on industrial cases. Cluster analysis is used for grouping the variables into clusters, where the points in each cluster are highly correlated. From every cluster only one representing point is selected for inspection. In this paper the method is further developed, and multiple linear regression is used for evaluating how much of the information is lost when discarding an inspection point. The information loss can be quantified using an efficiency measure based on linear multiple regression, where the part of the variation in the discarded variables that can be explained by the remaining variables is calculated. This measure can be illustrated graphically and that helps to decide how many clusters that should be formed, i.e., how many inspection points that can be discarded.

Copyright © 2009 by American Society of Mechanical Engineers
Topics: Inspection
Your Session has timed out. Please sign back in to continue.



Grahic Jump Location
Figure 1

The different stages in the geometry assurance process

Grahic Jump Location
Figure 2

The three case studies; from above the ringframe, the side member, and the side panel. To the left with all the original inspection points. To the right with the inspection points remaining after the cluster based reduction. After the reduction the share of remaining points are 25%, 40%, and 80%, respectively.

Grahic Jump Location
Figure 3

The efficiency measure EM2a applied to the three case studies and for different number of reduction quotas, q/p. The large circles show the value of EM2a when q is determined due to the stop criterion rUV=0.45.

Grahic Jump Location
Figure 4

The two encircled points form one cluster

Grahic Jump Location
Figure 5

Points belonging to one of the clusters for the ringframe

Grahic Jump Location
Figure 6

An example of how a representative and additional points can be chosen from a cluster using information about capability index




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