System reliability analysis with in-house and outsourced components

Abstract

Outsourcing is a common practice for product development, but it also poses a challenge in system reliability prediction. Design details of outsourced components may not be available for system designers, making the system reliability prediction difficult. This paper discusses how to use a physics-based reliability approach to accurately predict system reliability for systems with both outsourced and in-house components. The accuracy is achieved by removing the assumption of independent component states. The idea is demonstrated with systems whose failures are caused by excessive loading. For system designers, physics-based limit-state functions are available for in-house components, and they also have reliability testing data provided by suppliers of outsourced components. Then system designers construct limit-state functions using the testing data for the outsourced components. With all the component limit-state functions available, the joint probability density function of the states of all the components in the system becomes available, resulting in accurate system reliability prediction. An engineering example is provided to demonstrate the proposed method.