System Reliability Estimation with Failure Modes Analysis Applying Bayesian Approach - Case Study

This paper describes an approach for estimating system's reliability using the information and data gathered and analyzed by FMEA of the system and its components during design. This approach suggests estimating failure mode probability distribution, based on their Prior estimation, following the Bayesian methodology. In order to avoid a massive and time consuming procedure, only critical, dominant and uncertain failure modes should be considered. The prediction of failure mode probability is based on estimating the lower and upper probability limits at defined percentiles of failure mechanisms which cause this failure mode. In reality, often there is very limited understanding and ability to estimate those limits or “confidence interval” for the failure probability. In order to eliminate the effect of this limited ability, very strict rules and procedures were defined.