henki
Mechanical
- Jun 27, 2011
- 22
I am currently using CalculiX as my FE solver. I already asked this question in CalculiX forum, but it does not hurt to ask it here also.
My problem is to model a steel box with blast loading inside of it. The problem includes great plastic strains, high strain rates and thermal softening. The dynamics of the problem will be solved explicitly, whereas the stress update will be implicit. I am trying to create an Abaqus UMAT routine of Johnson-Cook-material model to be used with CalculiX (the UMAT creation in CalculiX is the same as with ABAQUS, just some less important variables are not supported). The solving algorithm would be based on this article (I'm not sure if it is allowed to post the article here without the writers permission, so I will not post it for now):
-Zaera R., Fernandez-Saez J., An implicit consistent algorithm for the integration of thermoviscoplastic constitutive equations in adiabatic conditions and finite deformations, International Journal of Solids and Structures, 43 (2006), pp. 1594-1612
Lately I have been looking into continuum mechanics (still not an expert) and I've been browsing through the ABAQUS user manual. There are some things which I don't understand and I hope that someone with better knowledge and experience of the subject could help me a bit.
As I understand (correct me if I am wrong), ABAQUS gives the user the corotational Cauchy stress and corotated strain increment to be used with the UMAT routine. Apparently Green-McInnis-Naghdi rate is used in CalculiX (hat{omega} = Omega = dot{R}*R^T). Also the time increment of the dynamic analysis is provided. With these I can probably create the corotated rate-of-deformation-tensor in UMAT (D = delta_epsilon/delta_t) if I would happen to need it. From the corotated stress, I can compute deviatoric stress and von Mises stress, which should both already be corotated. With these I can compute plastic-rate-of-deformation D^p, equivalent plastic strain rate dot{hat{epsilon^p}} and equivalent plastic strain hat{epsilon^p}. And with these I can compute the adiabatic thermal effects.
The stress update algorithm itself (fully implicit backward Euler scheme) is clear to me. The corotational configuration is what creates the problem. I don't understand what to do with the incremental rotation tensor drot = R_(n+1)*R_
^T, which is provided for the user. Should I use it in my UMAT routine to update some variables or tensors to corotational configuration? But aren't all the variables already in corotational configuration, since ABAQUS automatically corotates the strains and stresses which are used as a basis for other variables and tensors? Also apparently ABAQUS automatically rotates the updated stresses and the tangent modulus to the global configuration at the end of the code, so drot is not needed there either.
I would be grateful for any help. The answer probably lies in ABAQUS user manual, but I have not been able to find or understand it.
-Henki
My problem is to model a steel box with blast loading inside of it. The problem includes great plastic strains, high strain rates and thermal softening. The dynamics of the problem will be solved explicitly, whereas the stress update will be implicit. I am trying to create an Abaqus UMAT routine of Johnson-Cook-material model to be used with CalculiX (the UMAT creation in CalculiX is the same as with ABAQUS, just some less important variables are not supported). The solving algorithm would be based on this article (I'm not sure if it is allowed to post the article here without the writers permission, so I will not post it for now):
-Zaera R., Fernandez-Saez J., An implicit consistent algorithm for the integration of thermoviscoplastic constitutive equations in adiabatic conditions and finite deformations, International Journal of Solids and Structures, 43 (2006), pp. 1594-1612
Lately I have been looking into continuum mechanics (still not an expert) and I've been browsing through the ABAQUS user manual. There are some things which I don't understand and I hope that someone with better knowledge and experience of the subject could help me a bit.
As I understand (correct me if I am wrong), ABAQUS gives the user the corotational Cauchy stress and corotated strain increment to be used with the UMAT routine. Apparently Green-McInnis-Naghdi rate is used in CalculiX (hat{omega} = Omega = dot{R}*R^T). Also the time increment of the dynamic analysis is provided. With these I can probably create the corotated rate-of-deformation-tensor in UMAT (D = delta_epsilon/delta_t) if I would happen to need it. From the corotated stress, I can compute deviatoric stress and von Mises stress, which should both already be corotated. With these I can compute plastic-rate-of-deformation D^p, equivalent plastic strain rate dot{hat{epsilon^p}} and equivalent plastic strain hat{epsilon^p}. And with these I can compute the adiabatic thermal effects.
The stress update algorithm itself (fully implicit backward Euler scheme) is clear to me. The corotational configuration is what creates the problem. I don't understand what to do with the incremental rotation tensor drot = R_(n+1)*R_
^T, which is provided for the user. Should I use it in my UMAT routine to update some variables or tensors to corotational configuration? But aren't all the variables already in corotational configuration, since ABAQUS automatically corotates the strains and stresses which are used as a basis for other variables and tensors? Also apparently ABAQUS automatically rotates the updated stresses and the tangent modulus to the global configuration at the end of the code, so drot is not needed there either.I would be grateful for any help
. The answer probably lies in ABAQUS user manual, but I have not been able to find or understand it.-Henki