How do we measure Robustness?

Technically

Robust Design is a process for optimization. A technique to help with this is the Signal-to Noise (S/N) Ratio.

Signal-to-Noise (S/N) Ratio is the mathematical formula used to calculate the design robustness. The larger the S/N ratio, the more robust the performance.

Dr. Taguchi has developed various types of Signal-to-Noise (S/N) ratios as universal indices for the evaluation of quality and reliability of products and processes.

The Signal-to-Noise ratio gives a sense of how close the design is to the optimum performance of a product or process.

It has been used in the communications industry to compare a line signal (the desired outcome) with the line noises (the undesired outcome).

This concept was applied by Dr. Taguchi in the 1950s to other systems, including mechanical, electrical, electro-mechanical, chemical, and many others.

"Variability in product/process function is the enemy of quality."Dr. Taguchi relates this deviation from ideal to a concept called Noise. Noise, quite simply, is variability. Moreover, factors that cause variations are referred to as "Noise Factors." By definition, noise factors are uncontrolled either from a practical or cost standpoint.

Economically

Quality Loss Function is the standard by which quality-related design factors are studied and weighed.

This notion differs from the conventional concept of quality. Instead of defining quality as a positive attribute of a product, it is defined as the financial loss or cost to society caused by undesired variance in the shipped product.

It includes costs such as warranty, liability, and lost customer goodwill. The important consequence of this concept is that it brings engineering choices into the realm of economics, something scientists and engineers often find difficult.

The other main result of the Quality Loss Function is that the farther a product's performance varies from the "target" value, the greater the cost, which implies that it is a continuous function.

It can serve as a powerful language that permits experts on multidisciplinary product development teams to communicate more easily. This is because all engineering choices are related to a common currency-money.

The Quality Loss Function is the most rational tool for making cost-performance trade-off decisions in product and process development, decisions engineers and managers face every day.