Markov Chain Simulation for Estimating Accelerated Life Model Parameters

In this research, Markov chain Monte Carlo (MCMC) method was used to derive posterior knowledge of accelerated life model parameters in a Bayesian inference framework. The concept is discussed through a case study considering the fatigue life of a mechanical component. In the first step a comprehensive model including the relationship among parameters, design variables, material properties and available prior information is constructed. The accelerated life test data are then linked to this representation using a proper likelihood function. At the final stage evolution of model parameters during the Bayesian sequential updating are studied and the convergence of whole process is verified. For further validation, the approach is illustrated with examples using traditional MLE approach