Computational Modeling of Progressive Damage and Failure in Composites

  
         
 
here has been a tremendous resurgence in interest in advanced composites as primary structural materials in recent years. This is fuelled in part by major aircraft manufacturers, whose next-generation airliners will have airframes made with unprecedented amounts of

lightweight composites. This trend, which results in considerable savings in fuel costs, is set to continue well into the future . The extremely high demand placed upon the structural performance of composite materials means that the design envelopes for composite structures will be stretched ever further. Hence, there is now an urgent need for reliable failure theories and damage propagation methodologies that will accurately mirror the complex failure mechanisms found in composite structures. These multiple failure modes include fiber rupture, fiber-matrix debonding, matrix micro-cracking and delamination. They invariably occur at different length scales and it is important to include pertinent microstructural details into the failure prediction methodology. This rather ambitious goal is unlikely to be achieved without extensive computational modeling. Currently, the most common computational scheme for modeling damage and failure progression involves some form of material property degradation at the ply level. However, it suffers from several disadvantages, including possible computational instability.

Figure 1

Figure 2

A new computational technique, known as the element-failure method (EFM), has been developed in the Department of Mechanical Engineering, to model and predict the complex progressive damage in fiber-reinforced composite materials and structures. In this technique, the effect of damage is incorporated into modification of nodal forces in the finite element method (FEM). A micromechanics-based failure criterion, known as the strain invariant failure theory (SIFT), has been incorporated to determine the necessary directions and magnitudes of nodal force modification. Using the SIFT-EFM computational technique developed, damage evolution and propagation in composite laminates under different test conditions have been successfully investigated.

Figure 1 shows an example of an application of SIFT-EFM to failure of pin-loaded composite laminates. The laminates are quasi-isotropic (QI), with the lay-up [0/90/45/-45] 2s . Depending on the relative dimensions of the laminate and bolt, three possible overall failure modes have been identified. Firstly, when the laminate width W is relatively small compared to distance E of the bolt from the free edge , net tension failure of the laminate occurs. When W is relatively large compared to E , however, net shear failure results. Lastly, when both W and E are approximately equal, net bearing failure becomes the dominant failure mode. The failure pattern just before the ultimate load drop for the net tension failure case and the corresponding predicted and experimental load-displacement traces are shown in Figures 1 (b) and (c). Similarly, the detailed damage processes in the net bearing failure and the shear out cases have been investigated and analysed.

Damage patterns for OHC and OHT specimens may be similarly analysed. The analysis correctly predicts the effects of stacking sequence and hole size on ultimate failure loads. An example of failure patterns and predicted ultimate loads for OHC is shown in Figure 2. The comparison with experimental data shows good agreement.

Figure 3 shows an example of cascading delamination patterns typical of low-velocity impact damage of composite laminates. The prediction by SIFT-EFM shown on the left bears a striking resemblance to the experimental pulse-echo scan of delamination damage on the right.

Figure 3

Funding support from Boeing Co., Asian Office of Aerospace R&D ( AOARD) and an NUS Academic Research Fund (ARF) Grant is gratefully acknowledged.

 

 
     
 

Contact person

Assoc Prof TE Tay
Tel: 6516 2887,
Fax: 6779 1459
E-mail: mpetayte@nus.edu.sg
     
 


8