Page Last Updated: Monday, 11 January 2016 19:16 EDT, © 2016

HARTLEY CONSULTING
Solving
Complex Operational and Organizational Problems

PROJECT: Experimental Design for Manufacturing

Dr. Dean S. Hartley III


Project Metadata Keywords
Label Name Other Year DurationYrs
Client Milliken Industrial Fabrics Milliken & Co.
Dates 1978 0.25
Employer Milliken Inc
Partner N/A
Experimental design
Industrial issues
Inventory control
Productivity
Scheduling
Statistical inference
Textiles

DETAILS

A particular type of fabric required multiple steps in its manufacture.  One of the steps in the middle was experiencing inconsistencies in its processing.  After investigating the processes (including doing some statistical inferences), I determined that one set of variables in earlier processes were both controllable and likely to affect the processing problem.  I created an experimental design to investigate both main affects and interactions of these variables.

The statistical analysis of the results produced the best values of the variables and proved to be statistically significant.  Unfortunately, the reality was that these variables, while controllable, were not controllable at the precision required.  The machinery was simply incapable of producing consistent values of the fabric components. 

The OR lessons are important:  statistically significant does not equal physically significant and theoretically controllable does not equal practically controllable.


If you arrived here using a keyword shortcut, you may use your browser's "back" key to return to the keyword distribution page.

Return to Hartley's Projects Page