The design of experiments (DOE) approach can be used to figure out how to optimize an assembly line in a Six Sigma industry. A less known use for it is in Designing for Six Sigma during the reserach and development stage before a manufacturing environment is ever built. Because the cost of simulations are significantly less than rapid prototyping, during the research and development stage of a new technology, DOE can be used to limit the scope of parameters and discover correlations before spending a lot of money in the lab. One such simulation technique that I have used extensively with DOE is Computational Fluid Dynamics (CFD).
I have used DOE in research and development in the following disclosed situations:
- The Air Force Institute of Technology (PI: Prof. McMullen), I designed a scramjet nozzle that utilized magnetogasdynamics to accelerate flow and provide more efficient thrust compared to the baseline model. The CFD code used was Fluent with post processing with Tecplot. The rapid scaling of parameters would not have been possible without a test matrix.
- Worcester Polytechnic Institute (PI: Prof Sullivan), I used DOE to optimize my potential flow CFD code written in Matlab to solve a stream function over a complex 2D domain which utilized the following potential flow equations:
- Baldwin Filters (Champion), I used a DOE technique to optimize the design of a paperless sand filter utilizing vortex tubes using CFDesign. MiniTab was used for the correlations. This research and development project was then used for a construction and agriculture equipment filter in international sales.
- University of Notre Dame (PI: Prof. Jumper), I utilized a fractional test matrix to rapid scale an aerodynamic surface utilizing ANSYS Fluent, Tecplot360, and Excel for correlations.