Analysis of Staircase Reliability Tests with Least Square Method

Abstract

The design of stair-case tests is straightforward in mathematical terms, but it may be timeconsuming and cost-sensitive in actual engineering applications. The strict requirement for constant step intervals between each trial and assumption of step interval value often repre-sents a considerable engineering challenge. The flexibility of the method presented in the current study allows its application for staircase tests with variable step intervals between trials. The method is based on classical Least Square Method and its final outputs are approximations of mean (X50) value and its standard deviation, similar to the Bruceton approach. The algorithm presented overcomes uncertainty about its data and distribution type, and leads to applicable engineering results.