Experimental Determination of Cohesive Zone Models for Epoxy Composites P.F. Fuchs & Z. Major # Society for Experimental Mechanics 2010 Abstract In this work, a new test set-up was applied in order to determine cohesive zone models experimentally. A high speed camera in combination with a digital image correlation system was used to record the local displace- ments enabling the detailed determination of crack opening values. The J-Integral method was used to calculate the cohesive stresses. The analyzed materials were composites made of glass fiber reinforced epoxy resin layers. Two different specimen geometries and the difference between warp and weft of the glass fiber mats were analyzed. As the specimen geometry didn’t have a significant influence, the difference between warp and weft, regarded by the loading direction, lead to considerably different cohesive zone laws. The initial part, the linear increase to a maximum stress, was very similar, while the damage evolution was either exponential or bilinear in shape. In future work, the derived cohesive zone models will be used to perform finite element simulations on laboratory specimens and on component scale. Thus, by comparison to the measurement result, the cohesive zone models can be evaluated. Keywords Fracture mechanics . Cohesive zone model . Digital image correlation . J-Integral method . Epoxy composite Introduction Introduced by Barenblatt [1] and Dugdale [2] in the 1960`s, cohesive zone models (CZM) have been of growing interest recently, since they can be used for fracture simulation. Cohesive zone elements which follow the traction- separation behaviour as described by the CZM, were implemented in various FE-codes, allowing for the simula- tion of crack initiation and propagation. Due to new methods presented in several publications e.g. Remmers et al. [3], Zhang and Paulino [4] and Yang and Deeks [5], a crack path prediction became possible too. The advantage of the CZM models is that they can be used to describe a wide range of different damage mechanisms. For example in Yang and Cox [6] a cohesive element for the simulation of three-dimensional, mode dependent process zones is presented and used for delamination and splitting cracks in laminates. For these simulations, accurate CZM are essential but their experimental determination is still challenging. Existing experimental methods were reviewed by Sorensen and Jacobsen [7]. Two reasonable procedures were pointed out: First the direct tension experiment which was used e.g. by Ting et al. [8], with the presumption of a uniform damage evolution across the ligament, which is difficult to achieve in experiment. Second the J-Integral approach, first used by Li and Ward [9], which was chosen by Sorensen and Jacobsen for their own experimental work. Other methods (e.g. Chen et al. [ 10 ]) are semi- experimental, adapting the CZM parameters by comparing measurement results with model predictions. In the present work, an advanced test set-up was used to experimentally determine the CZM. The approach was similar to the one used by Zhu et al. [11], but was applied on different specimen geometries and materials. The crack end opening was determined by the evaluation of high P.F. Fuchs (*) Polymer Competence Center Leoben GmbH, Roseggerstrasse 12, 8700 Leoben, Austria e-mail: peter.fuchs@pccl.at Z. Major Institute of Polymer Product Engineering, Johannes Kepler University, Altenbergerstrasse 69, 4040 Linz, Austria DOI 10.1007/s11340-010-9370-2 Experimental Mechanics (2011) 51:779–786 Received: 26 January 2010 / Accepted: 11 May 2010 / Published online: 29 May 2010