Crack progation 1

[jazzyg]

Dear all,
I am using I-deas 8 to run some crack propagation tests on various laminate layups (for example uni-directional, quasi-isotropic layups etc.) on a double cantillever beam specimen.

The problem is how do i realisticly model a crack using I-deas. I thought that i could create a partition and mesh the whole specimen. After this i thought i could create some contraints on the individual nodes that lie on the partition and get them to split (redefined into 2 nodes) after a certain value has been exceeded (i.e the discrete crack method.

Does anyone know how to enforce these contraints on the node or alternatively does anyone know any method to model crack propagation using I-deas.

Thank You
jazzyg

 

[LeonEng]

jazzyg,
I apply the following steps to model a separate-able interface(crack!):
1. create additional nodes (overlap on the existing nodes)along the crack-line.
2. delete the ELEMENTS along the crack-line (for a 2D crack model, i.e. elements on both sides of the crack-line).Note that all the nodes still remain there.
3. Manually re-create all the deleted elements. Now you have 2 coincident nodes along the crack-line (created in step 1) & you will be prompted to select the appropriate node when you create the elements. Make sure that the elements on oppoiste side of the crack line are connected to the different nodes, thus you will get an interface that is separate-able(a crack!)
The method is quite straight forward for a 2D model. For 3D model, I cut out the partition where the crack interface with, then 'add' back the cut out partition to the model. If you mesh the model, you get a model with a dis-joined partition with coincident nodes/interfaces. Merge the coincident nodes on all the interfaces except the crack interface.
If your dimension of your crack model is small(in microns?), you may want to scale up the model as well since Ideas have modeling capability limit to 0.01mm (you may need to do necessary scaling on either material properties or the simulation output parameter as well)

 

[jazzyg]

Thank you Leon

 

[jazzyg]

Oh Leon may i ask a quick question though, once i apply boudary conditions (ie. forces and restraints), will the crack begin to propagate (will the nodes in front of the crack release so to speak and move independently from each other). I ask this as these nodes are still conected to each other as they are in front of the crack.

Thank You
Jazzyg

 

[LeonEng]

jazzyg,
Only the 'crack interface' that you created will 'open up', but will not propagate through other elements. However, you can examine the tendency of further propagation of the crack by the means of strain energy release rate(G) with various technique e.g. virtual crack closure, J-intergral, and etc.

 

[brianpaul]

jazzyq,

I'm pretty confident you will still need the experimentally determined the critical and arrest fracture energies, or the fracture toughness, to know exactly "when" the crack will propogate and when it will stop.

For instance you can simulate strain energy release rate by measuring the change in elastic strain energy caused by Leon's method of "crack propogation over a certain area. But if you don't know the critical strain energy release rate, you can't determine the system load at which the crack will propogate.
And if you don't know the arrest fracture energy, you can't determine how far the crack will grow before arresting.

Atleast thats my take on the situation, and I may be wrong.