This course introduces statistically based quality improvement methods useful in industrial settings, including observational methods and factorial design of experiments to discover influential factors and to analyze sources of variation. Many experiments involve the study of the effects of two or more factors leading to the development of factorial designs.
In a factorial design, experimental “runs” are conducted at all possible combinations of the low and high levels of each factor. Factorial designs not only allow us to determine whether a particular factor of interest has a significant effect on the response, but it allows us to quantify this effect and also determine whether there are significant interactions between factors. An interaction occurs when the magnitude of a factor’s effect depends upon the levels of the other factors.
The number of factors that can be considered in factorial design of experiments is limitless. However, the number of experimental runs increases exponentially as the number of factors increases. Fortunately, we can run carefully planned “fractions” of the full experimental design, provided we understand what information we can and cannot get from such fractions.
Generally, fractional factorial designs are helpful when there are many possible factors of interest that are not very well understood, with the “hope” that some of the factors will be insignificant, while others would be good candidates for future study. This course expands on the core principles of full factorial designs by also introducing fractional factorial designs.
Course Goals:
- Design, conduct, and analyze two-level factorial experiments for various phases of engineering work, including new product development, process development, and process improvement
- Calculate and interpret the effects of process variables on process performance
- Validate the assumptions of the assumed mathematical models utilized in factorial experiments