It is often desired to determine whether two different products which perform the same functions differ in reliability, whether a redesign has indeed improved the reliability of a product, or whether two different stress levels change the reliability of the same product. Such determinations may be made using the Weibull distribution and comparing the resulting lives of the two different products, or of the former and the redesigned product, or of the same product at two different stress levels. This chapter provides tools to enable such comparisons to be made statistically, by determining the confidence level at which the two versions differ at their mean lives and at their 10% failure lives.
The test of comparison at the mean life, when the data are Weibull distributed, can be conducted for two cases [1, pp. 111 113]:
When the ?'s for the two samples are equal.
When the ?'s are not equal.
These cases are presented next.
Test the two different products, the two different designs of the same product, or the same product at two different stress levels, and obtain two sets of times to failure data, one for each product, design, or stress level.
Plot the data on Weibull probability paper, or analyze the data by other methods to determine the Weibull pdf's parameters.
Determine the Weibull slope of each set of data, ? 1 and ? 2.
Determine the...
TABLE OF CONTENTS 
