Quantitative Risk Assessment with Bayesian Reliability Improvement  
Author Mingxiao Jiang




Abstract Reliability improvement and risk reduction are integrated in product development for design deficiency related hazards. For Quantitative Risk Assessment (QRA), the Risk Analysis Network (RAN) as proposed in this paper is an analytical method to link failures, hazards, hazardous situations and end user impacts (harms/damages to users). After reliability target setting and allocation for a highly reliable and complex system being developed, the reliability of each subsystem is going to be higher and higher in the reliability improvement process. The reliability estimation in each step of the reliability improvement process is getting more and more challenging, since it gets harder to obtain failure information for highly reliable design; hence, Bayesian reliability estimation approach needs to be structured and used. In this paper, a practical Bayesian reliability improvement algorithm is constructed and illustrated for product development. It is an iterative approach, including activities of reliability baseline, reliability testing, reliability update with the testing results, reliability improvement, reliability estimation for improved design as the new baseline, another reliability testing, another reliability update with the testing results, etc. Reliability results of each step in this Bayesian reliability improvement process provide inputs to the RAN in QRA for design deficiency related hazards, whose risk is reduced as far as possible.


Keywords quantitative risk analysis, Bayesian method, reliability improvement
    Article #:  2487
Proceedings ISSAT International Conference on Reliability and Quality in Design 2018
August 2-4, 2018 - Toronto, Ontario, Canada