Reliability Growth Test and Evaluation for Multistage Systems Open Access
Downloadable ContentDownload PDF
The goal of developmental testing is to ensure that the target reliability threshold is achieved and to eliminate existing defects before the system is fielded. However, reliability prediction through testing can be a challenging process due to budget constraints and calculation complexity. Single-shot systems can only be tested once because of their nature of operations, which introduces additional issues. Due to the ever-increasing complexity and growing demands for stable field performance, reliability prediction continues to be the main focus in the development of the single-shot systems.This research describes a methodology to overcome the challenges mentioned above and enable the planning, tracking, and prediction of reliability growth of a multistage single-shot system. It proposes the incorporation of existing test data with expert judgment in a Bayesian framework that allows point and interval estimation of current and future reliability values, as well as addressing critical managerial questions regarding the optimal amount of testing and number of remaining defects. The proposed framework enables continuous updating of predicted reliability values, removing the dependence on the estimates formed at the initial stage of testing.