Learning Forum

[jvancappellen]

Hi,

I am having some problems with a simulation in where I deform an initial flat plate into a double curved plate. My goal is to validate analytical membrane and bending strain calculations that I did. But without a working simulation this is not possible.

The problem is that my simulation won't converge for the displacements that I apply at the plate. I am using a material model with a stress-strain curve up to 100% strain. The displacements are applied by imposing a constraint on every node into x-y-z direction. These displacements are applied over 10 seconds. And should give the desired shape. However, the simulation does not converge at a certain point.

I am using the arclength method to solve the simulation and large deflection is turned on.

The nodal constraints are applied with an APDL which gives the nodal displacement in every direction, I don't have any other boundary conditions applied. Underneath you can see the mesh that I am using, the dimension of the plate are approx. 4500x500mm with 12mm thickness.

[Ashish Khemka]

nnCan you please share a snapshot of the error message or the exact error you see?.Regards,nAshish Khemkan

[jvancappellen]

nThanks for your reply! The error message that I get is the following:nThe equivalent total strain that I'm seeing is the following:nWhile my displacements are imposed linear, see the following figure;nI think that the displacement that I impose in y-direction is too large. But the material model that I'm using is going up to 150% strain and this isn't reached.. nI just looked into the element moments and these are really large at certain places of the plate, I don't really know why though. See the figure underneath with the elemental moment sum. nThanks in advance for your reply! I can also show a visualization of the nodal displacement vectors that I applied if that's helping. nKind regards.n

[Ashish Khemka]

nnYou can let your solution proceed further by increasing the number of equilibrium iterations. To do this add a command under the Analysis Settings branch:nnUse the Command Object and add the following command:nnneqit,100By default 25/ 26 iterations are used and using the above command will increase the number of iterations to 100.Regards,nAshish Khemkan

[jvancappellen]

Hi ArraynYour suggestion did not help. nHowever, I somehow came to check the element type that I am using. I've never defined to use shell181 elements because I thought that it was the default shell element. The only commands I had used before in the preprocessor were:nnKEYOPT,1,1,0nSECTYPE,1,SHELLnSECDATA,12,,,15nnNow I've added: ET,matid,181nAnd my simulation finally works. Now I am really curious about what is the default shell element? I can't find it anywhere unfortunately, maybe it is somewhere in the files on my PC?. I always thought that the default was shell181 whenever you create a flat geometry with a certain thickness.. Hopefully you can give me answer to this.nKind regards,nJoost