In Comsol it’s not that easy because it doesn’t save results for each increment, like other programs do. Most likely you will have to use auxiliary sweep with load as a parameter to obtain the stress-strain curve.

"I employed a displacement control strategy. The average load on the boundary is considered as a global degree of freedom. I defined a global equation in the Solid Mechanics>Global Equations 1 node in terms of the average displacement at the boundary to get the applied load value. To calculate the average displacements on the upper boundary I defined an average nonlocal coupling under the Component 1>Average 1 node."





He creates an average displacement at the boundary to get load value? And he uses a nonlocal coupling to calculate the average displacement?

This is related to the postbuckling modeling approach we discussed in one of the previous threads here. Its details are described in the article "Buckling, When Structures Suddenly Collapse" on Comsol’s website but there are also some examples in Application Gallery that you can open in Comsol and review. They feature pdf files with instructions as well.

Hi FEA way, for the auxiliary sweep, when I stop my range at 0.6 it runs fine, but when I increase it to for example 0.66 to make sure the model buckles, I get the following error message:

Regular Newton-Raphson method without any support usually won’t be able to complete the whole equilibrium path, including the postbuckling regime. Comsol has very limited capabilities when it comes to such analyses and it doesn’t have an arc length (Riks) solver which could be very helpful to overcome convergence issues. However, you can try replacing it with the approach described in previous threads and on Comsol’s website (the approach that uses the continuation parameter).

Hi FEA way, I have been trying to follow the article and set up my model the following way. It seems to run fine when the equilibrium path is linear, but does not seem to give me the full equilibrium path when I turn up the load/displacement - it does not run and gives an error message. Do you know what might be wrong?

You may have to add imperfections in form of scaled linear buckling mode shapes to this model in order to get the complete post-buckling response. In Comsol this can be done using Deformed Mesh —> Deformed Geometry module.

Do you know why it doesnt run when I dont have imperfections? Does it have to do with COMSOL itself?

It might be Comsol’s fault since it has limited capabilities for such analyses. However, in many cases imperfections are necessary to obtain the whole equilibrium path regardless of the software.