Detecting De-Laminations In Composite Beams Using Natural Frequencies And The Bayesian Inference

We show a mathematical model of the vibration of a laminated beam, which is made up of two elastic beams bonded together by a resilient layer. De-laminations are assumed to occur in this resilient layer. As the size and location of the de-lamination varies, the natural frequencies of the beam change. We shows how to use this variation and apply the Bayesian inference to de-lamination de-laminations. The Bayes' theorem tells us the relationship between the structural parameters and the natural frequencies with conditional distribution of those quantities. The desired distribution is numerically simulated using the Markov-Chain-Monte-Carlo (MCMC) algorithm. We further extend the method when the prior knowledge of an exact number of de-laminations is unknown.