The Operating Window and Design of Experiments for Robustness

Introduction

In DOE we traditionally look for the combination of input parameters that result in the lowest signal to noise ratio. Essentially we want the standard deviation of the output to be as small as possible relative to the mean of the output.?

This is done while keeping the standard deviation of each input at a value that is indicative of what can be expected in actual production.?

This in turn presents an issue with the practical application of DOE. In modern manufacture the standard deviation of each input and the output is very small. This can mean that the measurement of the noise is difficult (the differences between the maxima and minima are small relative to the resolution of measuring equipment used), and the overall process of DOE for robustness becomes dubious.

The operating window

A solution to this problem is the operating window.

The operating window is the range of values that the input can take while still maintaining a tolerable value for the output. This could mean a value that the end user will tolerate or value that corresponds to functionality of the system.

For those interested in history - the operating window was developed at Xerox in the late 70s primarily by an engineer called Don Clausing. This could have been the beginning of a dominant American influence upon quality as opposed to Japanese, but this is not how things turned out.

When the operating window is used in experimental optimisation of a system the approach is opposite to the traditional approach. Instead of looking for the combination of input parameters that reduces the range of values that the output will take, the range in the output is kept constant and the allowable range in the inputs (the operating window) is increased. This means that the relative improvement in the robustness of the system as it is optimised will be the same, but the range in each of the inputs and the output is easier to measure.?

Therefore, by using the operating window the same outcome of traditional DOE can be had, but less effort is required with the precise measurement of input parameters and the output.

Summary and recommendations

The operating window allows for a more accurate measurement of the relative improvement in robustness of the system being considered with less effort. Therefore, if you are ever performing a DOE for robustness, or something like Six Sigma even, and the differences between the maxima and minima of the parameters are small relative to the resolution of the measuring equipment being used, then consider using the operating window approach.

Source: http://ezinearticles.com/6289851

lupe fiasco car game schwarzenegger celtics amy lee glock alliteration