Multiscale finite element prediction of shear and flexural properties of porous CFRP laminates utilizing X-ray CT data

Despite the continuous developments in the manufacturing of CFRP materials, aiming to improve their quality, pores still remain a common defect, which in the form of initial damage, affect severely all matrix-dominated properties, especially the shear and flexural. Thus, the detection and quantification of pores in combination with FE-based predictive tools could potentially be a very efficient tool for the quality control of composite structural parts. In the present paper, a multiscale numerical methodology is used to predict the shear and flexural properties unidirectional (UD) CFRP laminates with different contents of pores utilizing data derived from X-ray Computed Tomography (X-ray CT) scans. The quantification of pores is conducted using the VG Studio Max software. In order to explore the limitations of the X-ray CT method and validate the parameters of pores quantification, the output of the analysis of CT scans is compared against optical microscopy images. A two-scale simulation is performed based on the size difference of the pores using the progressive damage modeling method. In the first scale simulation, the properties of the porous epoxy resin are predicted by means of FE models of representative unit cells (RUCs) constructed using the porosity data obtained from the X-ray CT scans, and in the second scale simulation, the shear and flexural properties of the CFRP specimens are predicted. The predicted values of mechanical properties are compared with results from short-beam shear and three-point bending tests.