Shunlai.Zang
2007-09-09, 07:24
Dr. Bergstrom
I use Abaqus/explicit to model sheet metal forming process, and Abaqus/standard to analyze sheet springback. If I use the built-in constitutive models in abaqus, everthings are well, no matter whether "HOURGLASS=ENHANCED" is used or not(I dont check whether there are differences in stress when "HOURGLASS=ENHANCED" is used or not). But If I use vumat constitutive model to model forming process, I have to delect "HOURGLASS=ENHANCED" in Abaqus/explicit input file, or I will obtain "***WARNING: THE SYSTEM MATRIX HAS 36 NEGATIVE EIGENVALUES" and "ELEMENT BLANK-MESH-1-1.3574 IS DISTORTING EXCESSIVELY" in Abaqus/standard using UMAT constitutive model, no matter how small time increment is setup. The vumat and UMAT codes have been verified, its work well, respectively. I searched the Abaqus manuals, and found that "It is recommended that *section controls, hourglass=enhanced be used in the original analysis for consistent treatment of hourglass forces in both original as well as import analyses", but If I follow the suggestions, I could not obtain springback at all. In addtion, It seems that the section controls can not be redefined in import analysis(I am not sure). How to define "hourglass stiffness" in Abaqus/standard?
So my questions are:
1. why I cant use "HOURGLASS=ENHANCED" in Abaqus/explicit when vumat and UMAT are used to import analysis?
2. whether "HOURGLASS=ENHANCED" has a significant influence on my springback prediction result?
3. Need I to define "hourglass stiffness" in Abaqus/standard, and how?
4. could you give me an explanation about the mechanism of hourglass control in import analysis when vumat and UMAT are used.
Thank you very much!
Shunlai
2007.09.09
I use Abaqus/explicit to model sheet metal forming process, and Abaqus/standard to analyze sheet springback. If I use the built-in constitutive models in abaqus, everthings are well, no matter whether "HOURGLASS=ENHANCED" is used or not(I dont check whether there are differences in stress when "HOURGLASS=ENHANCED" is used or not). But If I use vumat constitutive model to model forming process, I have to delect "HOURGLASS=ENHANCED" in Abaqus/explicit input file, or I will obtain "***WARNING: THE SYSTEM MATRIX HAS 36 NEGATIVE EIGENVALUES" and "ELEMENT BLANK-MESH-1-1.3574 IS DISTORTING EXCESSIVELY" in Abaqus/standard using UMAT constitutive model, no matter how small time increment is setup. The vumat and UMAT codes have been verified, its work well, respectively. I searched the Abaqus manuals, and found that "It is recommended that *section controls, hourglass=enhanced be used in the original analysis for consistent treatment of hourglass forces in both original as well as import analyses", but If I follow the suggestions, I could not obtain springback at all. In addtion, It seems that the section controls can not be redefined in import analysis(I am not sure). How to define "hourglass stiffness" in Abaqus/standard?
So my questions are:
1. why I cant use "HOURGLASS=ENHANCED" in Abaqus/explicit when vumat and UMAT are used to import analysis?
2. whether "HOURGLASS=ENHANCED" has a significant influence on my springback prediction result?
3. Need I to define "hourglass stiffness" in Abaqus/standard, and how?
4. could you give me an explanation about the mechanism of hourglass control in import analysis when vumat and UMAT are used.
Thank you very much!
Shunlai
2007.09.09