Reliability growth test design — Connecting math to physics

This paper addresses the Physics of Failure approach as a complement to the traditional design of reliability growth tests to determine the test duration in view of the test items´ cumulative damage model. This approach is to improve what was a pure mathematical model which does not address the life and use operational profile of the item subject to reliability growth. The paper emphasizes the Physics of Failure, PoF, rationale where the failure rates of the failure modes are a function of applied stress types, levels, and their durations. During its life, a product/item experiences cumulative damage introduced by the environmental and operational stresses in its actual use. This damage reduces its designed strength to a lower level and the comparison of that strength or the requirement to the actual stresses yields product reliability at the end of a specified time period. The goal of a reliability growth test is to, through product improvements during test, achieve this required reliability. The paper explains how the reliability growth test is de signed with regards to the product expected life, and how to best simulate the life environments in the test. Reliability data analysis is also addressed in this paper where failure occurrence times are mapped to the product life, a new approach to the data organization, which eliminates dependence of the test results on the test sequence or the initial test time. The methodology for the test duration and environments planning explained in this paper does not demand knowledge of initial product reliability. In that manner the PoF test design bridges the gap between the pure mathematics and the technical properties of the tested items and the underlying laws of the physics of failure.