How Do I Repair a Broken Experimental Design?


You’ve just started using Design of Experiments and, as a new to DOE practitioner, are following the advice to be as “bold” as you can with factor ranges. But, there’s a problem. You have made your ranges so wide that certain factor combinations cause the process to “break.” You are unable to measure a useful response value for some trials, and now your model won’t run! What now? What’s needed are sufficient new trials – subject to constraints that prevent them from “breaking” – be added to the existing good trials so that the mended design can fit the originally proposed model. In this session, Dr. DOE (AKA Tom Donnelly) will show how to perform this type of repair using the Augment Design with Constraints option in the DOE platform. Follow along with Dr. DOE as he shows you how do to this in JMP.


Article Labels
Article Tags