Bayesian approach of reliability parameter estimation using WinBUGS

Uncertainties are integral part of Probabilistic Safety Assessment (PSA) and arising from incomplete knowledge, simplified assumptions/idealization in modeling complex process/phenomena and unpredictable variation in performance of the system under study. Uncertainties influence the decision making process and help to determine whether robust decision can be made or more information is needed first. Poorly informed decision based on point estimate usually is not optimized and leads to unnecessarily excess resource allocation. The distinction and quantification of uncertainties is vital for Risk Informed Decision Making. The unpredictable variations results in aleatory uncertainty and are embedded in Basic Event Mathematical Models of PSA. The parameters of these Mathematical model if not known with certainty gives rise to epistemic uncertainty and can be presented in the form of appropriate probabilistic distributions. As the part of effective implementation of Risk Informed Decision Making (RIDM) Process, systematic plant failure data collection, pre-processing and statistical analysis is proposed. This paper discusses the statistical analysis of typical failure data from Electrical Power supply at Waste Management Plant. Bayesian approach of statistics are used for the analysis.