Robust Design

Robust Design

Sensitivity of products to factors that cause variation in their performance is an undesirable situation.

A typical tendency is to eliminate such sources of variation (noise), or to try to control them. This approach is costly, not always successful, and may not be necessary. It may be possible to minimize sensitivity of products and processes to the sources of variation by carefully selecting levels of design parameters such as types of materials, dimensions of products/parts, process temperature, pressure, etc. Such an approach to product/process design is called a robust design or parameter design. The parameter settings that provide robustness are found by:

	planning a statistical design of experiments
	collecting data by conducting experiments
	analyzing the data collected to find optimal parameter settings

The cost of robust design is negligible compared to the cost of eliminating or controlling the sources of variation. If the parameter settings found after a robust design do not reduce variation as much as desired, it is always possible to consider the more costly way of controlling the noise factors (tolerancing).

Genichi Taguchi has developed the robust design methodology including continuous loss function concept. You can read more on Taguchi's Loss Function and Robust Design from the book: Phadke, M. S. (1989) Quality Engineering Using Robust Design, Prentice Hall.

Last updated: December 10, 2009