Design of Type-I censored reliability qualification test based on subsystem test data in a parallel system

Reliability qualification testing (RQT) is conducted during the product development phase to verify whether a product meets established reliability standards. The Type-I censored experiment is commonly employed in RQT due to its specific advantages. However, the conventional RQT framework demonstrates notable limitations when applied to small-sample scenarios in complex systems. Specifically, extended testing durations increase overall costs, while short-term test schemes often involve excessive producer and consumer risks. These limitations primarily stem from the traditional approach’s reliance solely o n system-level test data, while available reliability information from subsystem testing is typically disregarded. To address these issues, a novel RQT design method is proposed in this study, under the widely accepted assumption that both subsystems and s ystems follow exponential distributions. The proposed approach accounts for parallel system configurations, incorporates available subsyste m test data, and applies the fractional-random-weight Bootstrap method, which effectively mitigates the limitations associated with small sample sizes. A case study demonstrates that this method reduces both risk and test duration.