Damage visualization based on local dynamic perturbation: Theory and application to characterization of multi-damage in a plane structure

Previously, an inverse damage characterization framework was proposed by quantifying the perturbation to local dynamic equilibrium of a beam-like structure, showing advantages in some aspects over the traditional global vibration-based and local guided-wave-based methods. Residing on the plate theory, this framework was expanded to a two-dimensional domain. Inheriting the attributes of localized canvassing using high-order spatial derivatives this approach has proven effectiveness in quantitatively characterizing damage of small dimension, regardless of its number and type. In addition, the approach requires no benchmarks, baseline signals, global models, additional excitation sources, pre-modal analysis nor prior knowledge on structural boundary. A damage imaging algorithm using the quantified dynamic perturbation was further established, enabling presentation of damage characterization results in an intuitive and prompt manner. Integrating the detection capacities in one- and two-dimensional domains, a hybrid damage visualization strategy was developed, for systems comprising structural components of different types, various geometries and diverse boundary conditions. Two independent de-noising techniques (low-pass wavenumber filtering and adjustment of measurement density), together with a hybrid data fusion algorithm, were proposed as auxiliary means to enhance the robustness of the strategy in noisy measurement conditions. The strategy was applied experimentally to the evaluation of multi-damage in a plane structure comprising beam and plate components, showing satisfactory results.

► A damage detection approach for plates based on locally perturbed equilibrium.
► Independence of baselines, global models, or prior knowledge on boundary.
► Applicability to multi-damage in a system of various structural components.
► A damage imaging algorithm able to present results intuitively.