Modeling Fatigue Crack Growth in CFRP Adhesively Bonded Substrate Using XFEM

Abstract

Abstract Carbon fiber reinforced polymers (CFRP) are crucial for many industries due to their superior material properties. CFRPs have strength and toughness that are comparable to metals but with the advantage of possessing lighter weight and higher corrosion resistance. Typically, structural parts are joined by bolts and rivets resulting in difficulties keeping the integrity of these joints. In CFRP joints, screw holes are stress concentration sites that may develop cracks, splits and delamination. Alternatively, adhesive bonding can be used as a joining method for CFRP substrates to overcome the disadvantages bolts and rivets. Structural parts are usually subjected to cyclic loading. Therefore, fatigue is considered as a major design tool for these parts. Finite element analysis is a powerful tool for modeling damage in components. This paper aims to simulate fatigue crack growth in adhesively bonded carbon fiber reinforced polymer (CFRP) composite substrates using a double cantilever beam (DCB) specimen. ANSYS XFEM is utilized to simulate the crack path using the enrichment technique to assign elements to the crack path. The model calculates the stress intensity factor (SIF) based on the domain integral over the contour around the crack tip. Then, it is converted at each sub-step to the energy released rate (ERR) which includes a correction factor estimated from the cohesive zone model (CZM). The model is idealized as a 2D geometry with the nodes at the unloaded edge and the corner being constrained in the longitudinal and the tangential directions, respectively. Displacement was applied at the other end of the specimen separating the two beams in a mode I condition. Finally, the number of cycles is estimated from Paris law. To verify the proposed model, fatigue crack growth (FCG) tests were performed on an 8-layer unidirectional CFRP laminates (HexPly T700/M21) fabricated into DCB specimens. The substrates is joined by an aerospace grade adhesive (Araldite 420). The estimated energy release rate (ERR) using the developed finite element model is within 90% or more of that determined experimentally.