You are here
How to obtain array constants, A(N), in MPC subroutine Abaqus
Normal
0
false
false
false
MicrosoftInternetExplorer4
/* Style Definitions */
table.MsoNormalTable
{mso-style-name:"Table Normal";
mso-tstyle-rowband-size:0;
mso-tstyle-colband-size:0;
mso-style-noshow:yes;
mso-style-parent:"";
mso-padding-alt:0cm 5.4pt 0cm 5.4pt;
mso-para-margin:0cm;
mso-para-margin-bottom:.0001pt;
mso-pagination:widow-orphan;
font-size:10.0pt;
font-family:"Times New Roman";
mso-ansi-language:#0400;
mso-fareast-language:#0400;
mso-bidi-language:#0400;}
Hi there,
I'm trying to model crack growth under special conditions using MPC
subroutine in Abaqus6.9. The procedure is simple:
In a 2D half-symmetric model, a virtual FIXED node has been defined out of
the body and all nodes in crack direction (symmetric line) in the body have
been tied to the virtual node (only one DOF tied in normal direction). Below
command line was used in *.inp file:
*MPC, USER (default=DOF)
1,1,10
2,2,10
...
where the first "1" is the identifier used in subroutine, the
second "1" is the node 1 in the body which is tied to "10",
the virtual FIXED node out of the body. In the subroutine, I found that the
simple definition of
A(1)= 1
A(2)=-1
JDOF(1)=2
JDOF(2)=2
is ended with 5 unsuccessful attempts to converge and an error; whereas
defining
A(1)= 1
A(2)=0
JDOF(1)=2
JDOF(2)=2
leads to a proper node release in the model based on the defined criterion
without problems in convergence. The constraint form in my model is U1-U10=0
where U is my desirable DOF identifier for nodes 1 and 10. So simply A(N)
constants based on Abaqus docs are expected to be 1 and -1 for a 2D problem.
I'm just wondering what is the proper way to obtain A(N) constants? Or are there some tricks
for node release in a symmetric line using a fixed virtual node that i must include? What would be
wrong in both ways of A definition mentioned above?
Thanks for your helps.
Regards,
Hamed
- Hamed YN's blog
- Log in or register to post comments
- 5481 reads
Comments
Full model
The new point is the method using A (1) = 1 and A(2) = 0 is successfully applicable on the full model (not half-symmetric) as well.