Cohesive Zone Methods (CZMs) are numerical methods used to numerically simulate damage initiation and growth at the interfaces of composite materials under external loading. These techniques provide estimates of the residual properties of a damaged specimen such as residual strengths, fatigue life or stiffness losses. To run these simulations, it is necessary to measure material properties such as the strengths of the interfaces. To describe the state of the structure at a given time instant during the loading, CZMs define a damage variable D which is a constitutive variable of the structure assessing its current state of health. The literature has, so far, been focusing on the damage initiation and growth of standardized test samples or the modelization of high speed impact for example.

This paper focuses on the measurement of the damage variable of a damaged test sample using guided waves. The measure of the damage variable of a specimen allows the creation of a numerical representation of the given specimen, leading to a customized set of residual properties estimation, thus removing the need to conduct extensive testing to characterize the residual strengths of materials. The technique developed in this paper relies on the Frequency Domain Instantaneous Wavenumber [2] technique, which is a guided wave-based technique measuring the wavenumber of propagating waves at every point of space from a single frequency steady state measurement. The measurement of the damage variable can be achieved regardless of the physical signification of defect, meaning that any defect weakening an interface of a composite can be modelled with different damage variable values, including delaminations, porous interfaces and wrinkles.

After a short introduction of CZM, the fundamental relationship between the guided wave wavenumber and the damage variable is presented. The process, from the measurement of the non- destructive propagation of the guided waves to the estimation of the residual strengths is presented. The measure of various damage variable distributions is then presented for numerical simulations.