Firehole News
Converging on Delamination, February 2011
Advances in Composite Simulation
Converging on Composite Delamination
Delamination is known to be one of the most common failure modes in composites, and yet a reliable solution for predicting its initiation and propagation has remained elusive. The most commonly used methods are limited by the need for pre-defined crack initiation, convergence issues and computational inefficiency.
Recent work by Firehole Composites has demonstrated the ability to accurately and efficiently model the simultaneous evolution of intra-laminar material failure and delamination for composite structural analysis. While the Helius solution proves to be quite accurate, the most significant difference is the efficiency with which the solution is obtained, enabling up to a 90% reduction in runtime compared with the existing COTS solution.
Too good to be true? To demonstrate, these technologies were used to model a common benchmark problem involving the de-bonding of a laminated composite stiffener from a laminated composite strip. Results were compared with existing commercial delamination modeling capabilities. While the Helius solution proves to be quite accurate, the most significant difference is the efficiency with which the solution is obtained, enabling up to a 90% reduction in runtime compared with the existing COTS solution.
EXAMPLE: Skin Stringer Debonding
(data taken from NASA/TM-1999-209097)
The following example employs a benchmark problem included with the Abaqus package by Simulia Corp. The problem utilizes data from a NASA study published in NASA/TM-1999-209097.
In the study, tension loading of the beam lead to the load displacement and damage pattern shown in Figure 2.
Figure 2 – Load Displacement and Damage Pattern of Example Skin Stringer. Click for Larger View
Baseline Results
To represent the present state-of-the-art COTS solution, several simulations were run using the delamination modeling capability that comes with Abaqus/Standard (COH3D8 cohesive elements). The finite element mesh and input file were taken directly from the Abaqus benchmark problem packaged with the Abaqus software.
It is noted that the Abaqus solution requires a value for viscosity to aid the process of convergence. Figure 3 shows the results of solutions run for varying amounts of viscosity.
The following observations are made about the results:
- There are 4 different solutions for 4 different viscosity values
- None of the solutions match the load/deflection response from the experiment.
- There is no matrix cracking reported; only debonding is modeled (unable to simultaneously model intra- and inter-laminar response).
- Run time = 179 minutes
Helius:MCT Cohesive Solution
To compare, a simulation was done using Helius:MCT progressive failure analysis combined with Helius:MCT Cohesive functionality. Results are shown in Figure 4. Note that the predicted load/displacement curve agrees both qualitatively and quantitatively with the typical measured response shown above. The Helius approach goes a step further than the existing Abaqus solution in accounting for both delamination and intra-laminar effects; the resulting contour output shows the delamination (red elements) and the localized matrix failure (green elements). Finally, the Helius solution took just 13 minutes to complete – a reduction of >90%.
Figure 4 – Helius:MCT with Cohesive Contour Plot and Load-Displacement Curve. Click for Larger View
A Complete Solution for In-Plane and Out-of-Plane Composite Simulation
As mentioned earlier, significant limitations of delamination/debond analysis is the need for a pre-defined crack initiation, convergence difficulties, and computational time burden. Helius:MCT with Cohesive has taken these challenges head on.
- By combining cohesive layer modeling with mulitscale progressive failure analysis, the ability to accurately predict matrix cracking enables the simulation to determine crack initiation within the solution.
- The computational efficiency of the MCT decomposition and the convergence-enabling technology of IDSM, the simulation is able to quickly arrive at an accurate solution, delivering substantial runtime reductions over existing cohesive technology.