PSI - Issue 25

Wojciech Danek et al. / Procedia Structural Integrity 25 (2020) 19–26 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

20 2

and contains both transverse cracks and multiple delamination. Due to this fact, this type of damage is called the barely visible impact damage (BVID). Usually, such damage is not critical itself for the structure, i.e. the reinforcing fibers remain undamaged at BVID, while shear stresses initiating delamination can be a reason of a propagation of an initial delamination, which may lead to structural failure, especially when such a structure is subjected to compression stresses (Sun and Hallett, 2017; Andrew et al, 2019). In order to evaluate the structural residual life after BVID the analysis of the damage propagation and failure mechanisms is necessary. Following this, numerous attempts have been undertaken for a proper reflection of these mechanisms using finite element (FE) modeling (see e.g. Shi et al, 2012; Thorsson et al., 2018; Metoui et al., 2018; Soto et al., 2018; Baluch et al., 2019; Tuo et al., 2019). The development of effective FE models may significantly reduce the number of experimental studies on the evaluation of structural residual life, and additionally, allows for the development of reverse models for simulation of various scenarios of BVID and the prediction of damage propagation and structural failure modes. However, the development of a FE model, which properly reflects the failure mechanisms is challenging, and many parameters need to be taken into consideration and properly adjusted. Omitting the parameter adjustment may result in significant errors in the results, which in turn, may result in under- or overestimation of the true structural residual life. In the recently published studies several parameters were analyzed. In particular, Shi et al. (2012) analyzed an impact force and energy as well as a displacement, and compared them with the experimental data. Sun and Hallett (2017) investigated an influence of an offset of an impact position on the resulting damage. The authors of (Francesconi and Aymerich, 2017) analyzed the effect of bridging during delamination and its influence on the resulting BVID. A great review on parameters influencing on LVI is presented in (Andrew et al., 2019), where the authors mentioned a material characteristics (like fiber/matrix/interface systems), geometrical factors (like thickness or curvature), an impactor characteristics (like impactor shape, size, and velocity) as well as an environmental characteristics (like moisture or temperature). Besides the physical factors influencing on the results of FE modeling, the technical factors of the development of the FE model as well as performing numerical calculation also have a significant influence on the obtained result. Some of such factors were analyzed in (Gaitanelis et al., 2019; Zhou et al., 2019). The authors of these studies analyzed, in particular, the influence of properties of cohesive elements and a number of cohesive zones, mesh density and overall meshing factor, loading parameters, and contact definition. However, the specificity of BVID modeling is still an open question, which needs to be deeply investigated in order to increase its comparability to real BVID form. In this paper, the authors investigated the reflection of BVID using a numerical model based on selected parameters of FE modeling, namely, an influence of mesh density, a physical consideration of a layered structure of a laminate in FE model, and an influence of the boundary conditions applied during experimental studies. As the measure of BVID the damage extent was assumed for further comparison with other experimental results (e.g. with ultrasonic C-Scans), which is planned to be performed in next steps within the research project realized by the authors of this paper. 2. Preparation of numerical model The numerical model prepared in this study is a reflection of impact damage introduced using a test rig (described in (Katunin, 2015)) designed for impact testing of composite plates (Fig. 1). The impact on the applied composite plate was introduced using the defined parameters, i.e. a hemispherical impactor with a diameter of 10 mm (Fig. 1) and the impact energy of 20 J. To create the numerical model the Finite Element Method implemented in LS-Dyna ® software was used. The impactor considered in this study was modeled as a rigid body due to the fact that the material of which it was made (steel) has a much higher stiffness than the plate made of CFRP composite material. By modeling the impactor as a rigid body it was possible to reduce the computing time. The impactor has only one Degree of Freedom (DOF) perpendicular to the composite plate, which was connected with the allowable displacement of the impactor in the test stand.