Risk Analysis of Structural Engineering Systems Using Bayesian Inference

Abstract

Engineering phenomena described through models come with uncertainties arising out of the modelling assumptions and observational statistical discrepancies. A probabilistic framework helps in quantifying the uncertainties involved. Conventional statistics are not very useful for any inferring from a limited amount of data which is often the case for many real life engineering problem. Bayesian inference tackles this issue by integrating past experience with current information. This chapter introduces the basic concepts of Bayesian inference and illustrates through three problems in structural engineering, how it can be useful. The emphasis is in showing how Markov chain Monte Carlo (MCMC) simulation is used in Bayesian inference.