A framework of integrated reliability demonstration in system development

Abstract

In a design and development process, a final system reliability demonstration testing is needed before the system is released for production. According to classic reliability theory, a large sample size is required to demonstrate system reliability. The reason is that the reliability demonstration according to classic reliability theory does not make use of all information in the system development process. To utilize the information from system development to help reliability demonstration, a Bayesian approach is desirable. This paper presents a Bayesian framework of integrated reliability demonstration in system development. The proposed framework uses component design analysis as prior reliability knowledge of the components. The results of life test of the components are used to update the reliability of components. The updated reliability of components is then integrated to form the prior reliability knowledge of the system according to system structure. Based on the prior reliability of the system, a reliability demonstration test plan of the system is developed. An example is used to illustrate this process and application issues are addressed. With this framework, the importance of upfront design and development on system reliability is shown, and the sample size required for demonstrating system reliability could be reduced compared to the sample size required by the classic reliability theory due to the existence of prior knowledge.