Bayesian damage localization and identification based on a transient wave propagation model for composite beam structures

Abstract

This paper proposes the use of a physics-based Bayesian framework for the localization and identification of damage in composite beam structures using ultrasonic guided-waves. The methodology relies on a transient wave propagation model based on wave and finite element scheme that efficiently provides time-domain signals that are compared with the ultrasonic measurements within a multilevel Bayesian framework. As a key contribution, the proposed methodology enables the localization and identification of damage using just the signals without any baseline comparison or further transformation, hence reducing additional sources of uncertainty. The proposed Bayesian approach allows (1) the localization of the defect, and (2) the identification of different candidate damage hypotheses and their ranking based on probabilities that measure their relative degree of belief. The methodology is illustrated in carbon fiber reinforced polymer beams with different layups. An investigation into how the measurement noise can impact the identified damage properties is also provided. The results show the effectiveness and efficiency of the proposed approach in reconstructing and identifying different types of damage in long and complex composite beams at a relatively low computational cost.