This paper presents a new method for projecting the reliability growth of a complex continuously operating system. The model allows for arbitrary corrective action strategies, and it differs from other models of this type by using all available data rather than failure mode first occurrence times only. It also differs from other reliability growth projection models in that it provides a complete inference framework via the posterior distribution on the system failure intensity. A unique feature of this approach relative to other Bayesian techniques is the analytic expression for the failure intensity contribution from unobserved failure modes. Expressions for estimating the initial failure intensity, growth potential failure intensity, and the cumulative number of failure modes expected in future testing are also developed. Extensions to the basic framework are also developed. The first accounts for multiple systems under test, and the second develops the posterior distribution while allowing for uncertainty on the Fix Effectiveness Factor values that are assessed. Two separate goodness-of-fit procedures are presented for assessing the appropriateness of the underlying model assumptions.Index Terms-Bayesian reliability, fix effectiveness, reliability growth, reliability growth projection.
This paper presents a reliability growth simulation test bed that is useful in examining the reliability growth of complex systems.It has many potential applications, including cost vs. reliability growth analyses and reliability growth program planning. The most significant application of the simulation though, is that it provides a method for examining the robustness of existing reliability growth models under varying assumptions and test conditions.The simulation allows for a user to more accurately model the test event, in that corrective action strategies and corrective action delay times and can be employed. Because these events tend to violate the assumptions of some models, two existing and widely used reliability growth models are examined under a variety of test conditions and assumptions. The simulated system reliabilities from the test bed are compared to the projections obtained from both the Crow Extended Model and the AMSAA Maturity Projection Model (AMPM).The results indicate that both models perform acceptably under test conditions that adhere to the assumptions of the models themselves. The performance of the Crow-Extended Model tends to degrade rapidly as the assumptions are violated, more specifically when the classification of modes is altered in any way. The AMPM appears to be robust to varying test conditions due to its limited assumptions. Other issues do exist though involving the selection of the proper form of the estimator which could have a significant impact on the accuracy of its projections.
Reliability assessment techniques constitute an important element of the reliability growth program. This paper examines the accuracy and robustness of two widely used reliability growth assessment techniques under a number of realistic corrective action processes. These methods are also compared to a newly developed assessment approach. The new approach provides a more robust assessment across a broader spectrum of cases. These include various corrective action processes as well as cases in which the number of failure modes in the system is not large compared to the number of failure modes surfaced during testing. The results indicate that each of the techniques perform acceptably well under test conditions that adhere to the assumptions on which they are based. The performance can be severely degraded when the assumptions are violated, and this is the case with the Crow-Extended Model. The AMSAA Maturity Projection Model (AMPM) and the newly developed alternate technique have a somewhat limited number of assumptions though. They appear to be more robust to the various test conditions examined here, with the alternate being the most useful across the broadest spectrum of cases.
This paper presents a new reliability assessment model that allows for the combination of developmental and operational data from different test events for continuously operating systems. The model offers an alternative to traditional reliability assessment using a single operational test only. Reliability degradation between developmental and operational testing is explicitly modeled through the use of a nuisance scale parameter, and a complete inference framework is provided via the posterior distribution.The approach serves as a natural extension of the current approach to reliability growth and demonstration used in the Defense industry while explicitly modeling the additional uncertainty that exists in the problem. Analogous Operating Characteristic (OC) curve quantities are developed from the posterior distribution. Use of these results will generally lead to tighter uncertainty intervals and result in lower reliability design goals. This approach can help to directly reduce the programmatic risks that may exist due to reliability demonstration in a constrained environment with operational test data alone.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.