Method and software platform for electronic COTS parts reliability estimation in space applications. (October 2021)

• Record Type: Journal Article
• Authors: Rabiei, Elaheh; Huang, Lixian; Chien, Hao-Yu; Earthperson, Arjun; Diaconeasa, Mihai A; Woo, Jason; Iyer, Subramanian; White, Mark; Mosleh, Ali
• Is Part Of: Proceedings of the Institution of Mechanical Engineers. Volume 235:Number 5(2021)
• Page Start: 744
• Page End: 760
• Publication Date: 2021-10
• Subjects: COTS electronics -- non-homogeneous Bayesian estimation -- reliability analysis -- Bayesian network -- hybrid causal logic analysis -- physics of failure; Reliability (Engineering) -- Mathematical models -- Periodiclals; Risk assessment -- Mathematical models -- Periodicals; Engineering design -- Mathematical models -- Periodicals; 620.00452
• DOI: 10.1177/1748006X21998231
• Languages: English
• ISSNs: 1748-006X

Abstract

Adoption of electronic Commercial-Off-The-Shelf (COTS) parts in various industrial products is rapidly increasing due to the accessibility and appealing lower cost of these commodities. Depending on the type of application, having an accurate understanding of the COTS failure information can be crucial to ensure the reliability and safety of the final products. On the other hand, frequent large-scale testing is often cost prohibitive and time consuming for emerging technologies, especially in the consumer electronics sector where minimizing time-to-market and cost is critical. This paper presents a comprehensive Bayesian approach and software platform (named COTS Reliability Expert System), that integrates multiple pieces of heterogeneous information about the failure rate of COTS parts. The ultimate goal is to reduce dependency on testing for reliability analysis and yet to obtain a more accurate "order of magnitude" estimate of the failure rate through an efficient process. The method provides a foundation for incorporating manufacturers reliability data, estimates based on underlying physics-of-failure mechanisms and circuit simulations, partially relevant life test data of similar (but not necessarily identical) parts, and expert opinions on the manufacturing process of the COTS part of interest. The developed expert system uses Bayesian estimation to integrate all these types of evidence. The methodology is demonstrated in estimating the failure rate of a static random-access memory (SRAM) part.