Assembly problem identification by using cluster analysis

SUMMARY

Cluster analysis combined with the principal component analysis can provide an analytical tool for locating and identifying the major problem areas of an assembly misfit. Cluster analysis can be used to divide a significant number of variables or location measurements of an assembly into subgroups which differ based on the similarities and the attitude of the variables in each cluster. At the optimum cluster level, zones of variations of the assembly are identified in which every cluster would contain a few key variables which have caused the misfit of the assembly. Principal component analysis can be implemented for these few variables to obtain information on the components or the detail parts variation which have caused the assembly mismatch problem.

Notes

Please address correspondence to PO Box 8326, Wichita, Kansas 67208, USA.