Search
Close this search box.
Notifications
Clear all

problem with rate independent problem

4 Posts
2 Users
0 Likes
519 Views
Posts: 6
Topic starter
(@velvia1000)
Active Member
Joined: 15 years ago

Dear all

I am using crystal plasticity umat written by Huang. but when I set the rate sensivity component n=200(or m=0.005), the ABAQUS can not solve it and give me some error like this.

SOLVER PROBLEM. NUMERICAL SINGULARITY WHEN PROCESSING NODE

PART-1-1.320 D.O.F. 2 RATIO = 1.07695E+13.

and

ELEMENTS ARE DISTORTING EXCESSIVELY. CONVERGENCE IS JUDGED UNLIKELY.

I think it is because we have small number with high value power and it cause singularity.

I should say that it is work fine for small n(10 or 20)

Please help me.

Topic Tags
3 Replies
Posts: 3998
(@jorgen)
Member
Joined: 4 years ago

First, can you specify some more details of umat written by Huang. I am afraid that I am not sure I know that umat.

Also, do you get that error right at t=0, or later when the elements are highly distorted?

-Jorgen

Topic Tags
3 Replies
Posts: 6
Topic starter
(@velvia1000)
Active Member
Joined: 15 years ago

-It is the single crystal plasticity based on Asaro paper.you can find it here:

[url] http://www.columbia.edu/~jk2079/fem/umat_documentation.pdf [/url]

-it abropted after some incremenets and I just do it for simle tension test and I think we shouldnt have high distorsions.

-in this umat, we use power law for crystalline creep that cause this problem for high value of n (rate sensitivity exponent)

f(x)=x|x|^(n-1) which x=tau(shear stress on slip plane)/g(current strengh)

-for small value of n it is converged well but for larger value around 100, the problem occured.i think, it is happen because at the beginning, we have small value of x that power by large n.

thanks

Ramin

Reply
Posts: 3998
(@jorgen)
Member
Joined: 4 years ago

Oh I see. I think the problem is that the rate equation becomes a stiff ODE for high n values. If the UMAT is written impropertly that will lead to numerical instabilities when run.

-Jorgen

Reply
Share: