Weibayes testing: What is the impact if assumed beta is incorrect?

Abstract

This paper evaluates Weibayes small sample, zero and sudden death failure test and data analysis techniques in order to mathematically and graphically quantify the risk associated with assuming an incorrect value of the Weibull shape parameter, beta (β). Specifically, the risk associated with assuming a value of β that is higher than the “true” value of β is investigated, as this is the scenario that can cause users of the Weibayes approach to (1) understate the test time needed to correctly define a reliability requirement and (2) interpret the test results incorrectly and overstate the reliability actually demonstrated. The work product of this effort is a MS Excel® spreadsheet that automates the process of quantifying the risk associated with the Weibayes approach.