A reliability allocation method based on Bayesian networks

With the increasing emphasis on reliability in recent years for high-tech industrial processes comprise most complex systems today. Allocating proper reliability value to each subsystem or component of a complex system as an optimization process to rise productivity is very crucial. Exists approaches like ARINC and AGREE are too simple and inadaptable for ever increasing levels of industrial systems. Other techniques like allocation method based on FTA which the allocation values are more convincing but the calculation procedure is based on minimum cut which makes it complicated and inefficient in the complex system. Finding a more practical, effective and accurate method to solve this dilemma is the objective of this paper. This paper presents a reliability allocation method based on Bayesian networks (BNs). First, the modeling aspect is defined by introducing the mapping rules between FTA and BNs, then a reliability allocating algorithm based on BNs is proposed, eventually, a case study of passive residual heat removal system is presented to demonstrate the analysis procedures.