Optimal Planning of Accelerated Reliability Demonstration Tests for the Weibull Model

Demonstrating product reliability is challenging, which is why accelerated tests are commonly used to reduce time and cost. The Probability of Test Success ( P_ts) helps assess a test's ability to demonstrate the reliability requirement under given conditions, representing the statistical power of the accelerated reliability demonstration test. Key challenges remain, particularly selecting the number of specimens, load intensities, and number of load levels to maximize P_ts while minimizing cost. Although previous work has supported risk-based planning, recent analytical advances in P_ts formulation offer more refined optimization methods. This paper introduces a new approach that provides a simple, fast algorithm for determining optimal accelerated test settings based on product performance and reliability requirements. A notable novelty is the integration of the actual load-profile shape, which influences reliability demonstration due to Type I statistical error. The method incorporates the widely used lifetime models of a Wöhler-Weibull model and includes a brief study illustrating its practical advantages.

Keywords: Reliability Testing, Accelerated Life Testing, Life Data Analysis, Reliability Modeling.