Statistical Design of Experiments (DOE)
Traditional trial and error experiments, although effective at finding the primary causes of variation, tend to be inefficient and costly if you are working with several variables. Trial and error experiments are also unable to resolve the interactions between different variables in your process.
For example, let's say that the time required for heat treating a steel bracket changes for differing levels of an alloying element in the steel. A trial and error method will be able to provide the time required for a single level of the alloying element, but cannot provide the relationship between the time and the alloy concentration. For this, a more sophisticated experimental approach that allows changes in more than one variable at a time is needed. This is called Statistical Design of Experiments.
The image below shows a response surface map which resulted from one type of Statistical DOE. In this case, the response characteristics (for example, strength) are graphed versus two input variables simultaneously (for example, alloy level and heat treat time). This results in a response surface that predicts the level of the response for any combination of the two input variables of interest. This method can be expanded to include many more than 2 variables.
In addition to the training we offer, Process Champions can provide direct support in setting up and analyzing Statistical DOE's to help troubleshoot and improve your processes. Statistical DOE's are also one of the many tools taught in the Six-Sigma training series. Contact us for details and a quotation.
On-Site Training
(Please contact us at: info@processchampions.com
for special group pricing)
| Course | Group Pricing | Cost per person (min 8 ) | |
| Introduction to Design of Experiments | 1-day | $1499 per class | $199 |
| Custom Training | Variable | Contact us for a quotation | Contact us for a quotation |
