Design of Experiments represents a family of techniques.
Experimental designs provide the ability to:
- Investigate multiple process variables at the same time.
- Identify which variables have significant effects on the process output.
- Study the relationships between variables to identify interactions.
The DOE approach is selected depending on the objective of the experimentation, the type of process, and the number of variables that will be studied. Experimental strategies often start with screening experiments.
DOE techniques include:
- Screening Experiments: A special extreme type of Fractional Factorial. Often used at the start of an experimental sequence; few experimental runs but yields important information about key variables.
- Fractional Factorials: Less runs (than Full Factorials) but less information, too. Studies a predetermined fraction of a Full Factorial.
- Full Factorials: Generates lots of information but requires many runs. Usually used to study variables at 2 or 3 levels (settings).
- Response Surface Analysis (RSA): An optimizing design in which the main independent variables are already known. Limited runs, highly selective information.
- EVOP: An iterative optimizing design; experiments are run within the existing range of process parameters. Relatively high number of runs, selective information.