In the first half of the twentieth century a set of structured methods for setting up, executing and analysing tests, have been developed. The theory is known under the general name Design Of Experiments (DOE). Unfortunately, until today very little use is made of these methods. As a consequence expensive tests lead to no results or even worse, to false results.

Because of the statistical nature of DOE many people are afraid to use it. But the principles of designed experiments are not that difficult and with a minimum of statistical knowledge a lot can be achieved. At the moment there is more than enough (cheap) software available so there is no need to go deep into the mathematics.

In our training courses a lot of attention is given to the setting up of tests and to the understanding of the practical meaning of the result of statistical analysis of tests.

With DOE we want to investigate the influence of factors (input variables) on characteristics (output variables) of a product, a process or a system. We have to make sure that noise factors, that are always present do not disturb the conclusions of our tests.

Within the overall field of DOE we can distinguish between three methodologies: classical DOE, Taguchi methods and a series of intuitive methods developed by Shainin. Each method has it's supporters and opponents and it is very popular to call one of these methods the one and only and to degrade the others as useless.

QS Consult is convinced of the power of sensible integration of methods, based on the following quote by dr. W.E. Deming:

  Each method is useful, as long as the user knows it's limitations

Services offered:

• Training course Introduction to DOE.
• Extended course on DOE with content and duration as discussed with the customer.
• Consulting on setting up tests and analysing results.

References:

Tenneco, Allanta, Volvo Cars, Arcelor Mittal, Husky Luxemburg, VDT Nederland, Eftec Belgium, …

Additional information:

• Detailed description by subject.
• Eight step plan for the application of DOE.
• Explanation of the Loss Function and Robust engineering.
• Explanation of Pre-Control.
• Some publications on DOE.
• MS Excel tool for tests with maximum 4 factors (note: Dutch excel 2003!)