Designing Statistical Experiments in Industry
Organizer and Session Chair: Sabyasachi Basu, The Boeing CompanyDesigning Multi-Step Process Experiments
Jose G. Ramirez
W.L. Gore
Complexity, and supply chains that may expand across the globe, requires us to design experiments that must take into account: 1) the multi-step nature of manufacturing processes, 2) the large number of factors involved, 3) the different sizes of experimental units, 4) the restrictions in randomization that occur, and 5) the constraints in the number of experiments that can be run. Designing experiments for these situations present statistical and logistical challenges. We discuss these challenges, as well as design generation using JMP and SAS, using several examples including an experiment for a complex manufacturing situation involving a hard-to-vary factor, 9 process factors, and 8 mixture factors; and fractional factorial split-plots for 3, 4, and 5 step processes.
Design and Analysis of Computer Experiments with Controllable and Uncontrollable Factors
Murat Kulahci
Technical University of Denmark
Abstract: With the current advances in computing technology, computer and simulation experiments are increasingly being used to study complex systems for which physical experimentation is usually not feasible or even impossible. We introduce a split-plot type of structure for computer and simulation experiments for which we partition the factors to study into two groups: controllable and uncontrollable. For these experiments we present space filling experiments with good coverage at both factor spaces as well as an extension to Kriging to handle the particular correlation structure caused by the split plot structure of these experiments.
DoE in the Product Life Cycle
Stephen Jones
The Boeing Company
Abstract: The applications of DoE are many and varied. In a large corporation, such as The Boeing Company, there are DoE applications throughout the product life-cycle, from product conception to in-service customer support. In this presentation, we describe the application of DoE at two ends of the product life-cycle. The first application is the use of DoE to derive design allowables, which are mechanical property values used by engineers to design parts. The second application is the use of DoE to characterize process capability to support customer maintenance procedures.