I have written some UMAT subroutines before for Abaqus, but in recent special case, I do not understand why the code does not work. I got NAN. value for DSTRAN in second step and it is really strange, because as far as I know abaqus generates itself these kind of values, such as STRESS and STRAN at the begining of an increment. However, I put a run with this code and the job is completed, but when I track the compilation procedure using WRITE(*,*) statement, in the text file, I recieve values changed to NAN. for sure, I also printed inputs, such as E11,E22, Nu_12 , .... they all pass into code with non zero values. I also update STRESS and Jacobean matrix within UMAT code which is mandatory. My question is that, if I am sure that my system does not have any problem with Abaqus subroutine compilation, because with a 20*20 mesh i got answers but if I change the mesh to the 40*40 or somthing more than 20*20 , DSTRAN gets NAN.
clearly mesh size can affect your calculation. I do not know Abaqus, but you should be able to change element type and check how much your problem is nonlinear. Nonlinearity can have huge impact on the problem too. if your problem is nonlinear try to increase number of iteration.
You ever could see these kind of problem in nonlinear analyzes. I had a similar problem with my software to FEM simulation of cold flat rolling. My solution was to point program to use only direct iteration in the step with convergence problems on Newton-Raphson.
A small subroutine like "if achieve X number of N-R iterations doesn't converge and direct iteration converged, jump N-R" should work.
Thanks
but in my case abaqus show that it had converged in the first iteration but when i want the stress result it 's not there. So i write my output and inputs of code and i saw DSTRAN hots NaN and also in msg file Average force and moment got NaN.
My question is while there is no iteration and attempts and abaqus shows convergence there are NaN data in were I said and also why in mesh more than 20*20 for this case and why of the meah is 20* 20 or less i have right answers.
I think there is a problem about excessive refinement (h-method). Maybe the level of the element shape function is not enough (p refinement). If the shape function is very simple, could be not enough to represent the phenom. I don't know If Abaqus qive you option to p refinement (to charge the shape function level). You could try It.
Hi Sapanta, I recommend you to write your UMAT file using python and enter coordinates using Patron C++ language if possible. It will probably assist you in finding why you are not getting results. https://www.python.org/doc/ may help. I also had the same problem where I can't see my impactor as my impactor is relatively very small compared to the impacting body. impactor=0.9*impacting body. I also prefer doing a small scale model due to numerical efficiency issue. This may not solve the issue but hopefully it will help you to identify where the issue is. I am using abaqus so the mesh cant be generated automatically to see the result of your model unless you reduce the scale of it. And check the allowable wave speed for each material you are attempting. It relies on density and elastic modulus.
- Badges
- Science topic