Design of Experiments

Design of experiments (DOE) is an important practical tool for understanding processes and improving their performance and control. This 3 day class is application oriented and, like all ILS courses, is designed to promote deep understanding, maintain student engagement, and to develop practical skills that can be directly applied on the job. It presents the statistical methods in a clear, intuitive manner and then applies them in table top exercises using popular software tools. These exercises are designed to illustrate a wide range of industrial applications and good practices for approaching and applying the methods. Instruction in basic statistics and related concepts such as process capability, statistical process control, and measurement system analysis can be provided if the need exists.

Students who take this class will learn the following:

• DOE as a process modeling approach; response surfaces, inputs, outputs, and classes of variables
• Steps in executing a DOE study
• The interpretation of ANOVA results
• Alternative experimental designs and their role in application: full factorial, 2k, fractional factorials, blocked designs, central composite designs
• Residual analysis to insure the integrity of data and models
• How to apply the models in industry including their use in general process understanding (screening experiment approaches), process targeting (use in process setup and feed forward control of the process based on input measurements), improving process robustness and capability, and process optimization. Case studies are presented for each application.
• The robust design approach (using the insights of Taguchi) to improving process capability