having problem to run model with two steps 1

[Tina a]

Hi,

I am running my analysis in two steps. The model is running in step 1 but not converging in step 2. The model had same time incrementation in both steps. I did change the increment size in step 2, the model tries to converge taking a long time but eventually dies. Could it be because of the boundary conditions that the model is dying? I have two objects and I have to fix one of the objects in each step and I did the same so I don't know what should be done in step 2 so that it would run. Any suggestion would be helpful.

 

[Dave442]

Convergence difficulties can arise for a bunch of reasons. It's hard to say what your issue is with the info you provided. I would check the .dat and .msg files for any warning/error messages to try to clarify the problem.

 

[Tina a]

Hi Dave442,

I looked at the warnings but I could not figure out the exact cause. Could you please have a look at the .dat file and help me know how can I fix this? Thank you.

 

[Dave442]

Sorry I'm travelling and can't look at it. Look through the .dat and .msg files and copy/paste any warning/error messages here. Someone might help you.

Also provide more details of your analysis. Give more details of your aim, material properties, boundary conditions and the errors you encounter. It's hard to provide useful input with the info provided.

 

[Tina a]

I have a spine model where screws are placed posteriorly (proximally and distally). And two rods are placed into the screws on right and left side. I want to distract one rod at a time and then fix it and distract the other rod. This is to compare differences in stress on the rods when distracting both rods at once and distraction of one rod at a time.

In step 1, I fixed the right rod to distract the left rod and in step 2, I deactivated the boundary condition for the right rod so that it could distract in next step. In step 2, the left rod was fixed (because at this time I want to distract only the right rod) in similar way how right rod was fixed.

The problem in my model is that the left rod could fully distract in step 1 but the model is not converging in step 2. For boundary condition, I used displacement/rotation and fixed the rods in all degrees of freedom along the entire length of rod (I don't know if this is correct or not to apply boundary condition along the entire length, please suggest if there are other ways to fix the rod entirely).

Warning messages:

In step 1:
OVERCONSTRAINT CHECKS: THERE ARE 125718 BOUNDARY CONDITIONS
SPECIFIED IN THIS MODEL. OVERCONSTRAINT CHECKS FOR BOUNDARY
CONDITIONS SPECIFIED ON SLAVE NODES OF RIGID BODIES, OF *TIE
OPTIONS, OR OF *COUPLING OPTIONS REQUIRE 47 Mb OF MEMORY. IF THIS
IS A PROBLEM, PLEASE TURN OFF OVERCONSTRAINT CHECKS USING
*CONSTRAINT CONTROLS, NO CHECKS or INCREASE THE MEMORY USED BY THE
PRE-PROCESSOR.
In step 2:
WARNING: OVERCONSTRAINT CHECKS: THERE ARE 133336 BOUNDARY CONDITIONS
SPECIFIED IN THIS MODEL. OVERCONSTRAINT CHECKS FOR BOUNDARY
CONDITIONS SPECIFIED ON SLAVE NODES OF RIGID BODIES, OF *TIE
OPTIONS, OR OF *COUPLING OPTIONS REQUIRE 49 Mb OF MEMORY. IF THIS
IS A PROBLEM, PLEASE TURN OFF OVERCONSTRAINT CHECKS USING
*CONSTRAINT CONTROLS, NO CHECKS or INCREASE THE MEMORY USED BY THE
PRE-PROCESSOR.

Now, I am trying to simulate exact same procedure but this time distracting right rod first. The rod could distract for few increments but now the model is taking long time to converge (seems like it will take forever like in previous model).

I don't know if overconstraint could be the reason since in my previous model the rod distracted fully in step 1. The similarity I see in both models is that for the distraction of right rod its taking time to converge. In new model, warning message says that it requires 50 Mb of memory in both steps, may be because of this, this model is taking time to converge in step 1.

What could be the issue? Any help would be highly appreciated.

Thanks again.

 

[Dave442]

If I was to guess I'd say your problem is your boundary conditions. You mention that in the second step you apply a zero displacement constraint to the left rod. Displacement constraints are applied relative to the initial configuration. So in the second step you are distracting the right rod while displacing the left rod back to it's starting configuration.

Try a velocity constraint on the left rod in step 2. Also, I don't know what you mean by "distracting".

 

[IceBreakerSours]

An imprecise way of conveying what the warning says is that you are asking a node to do move in more than one way at the same time and the solver can not resolve this conflict (except in a few limited situations). Then you need to put your detective hat on and check how you specified your BCs/constraints/contact and how those might cause such conflicts. The .MSG or .DAT file might contain more information on which nodes are complaining. You can then use the Viewer and the Create Display Group menu to visualize where those nodes are located.

*********************************************************
Are you new to this forum? If so, please read these FAQs:

http://www.eng-tips.com/faqs.cfm?fid=376

http://www.eng-tips.com/faqs.cfm?fid=1083

 

[Tina a]

Hi Dave442,

Thank you very much. I applied velocity constraint on left rod in step 2, the rod could distract for few more increments but eventually the model died. There are two rods (upper and lower) on each side- two rods on left side and two rods on right side. The rods on each side are connected by an axial connector and I have used connector displacement to displace rods along U1. This is what I meant by distracting the rods. I too think the issue is boundary condition. Like I described about my model, if you were to displace the rods on the left side without displacing the rods on the right side, what boundary condition would you use to fix the rods on the right side? Maybe the way I am applying boundary conditions to the rods is incorrect.

When I was fixing the rods on each side, I did not deactivate or activate the axial connector accordingly, I think this also could be the issue (but again if this could be the issue then step 1 could not have been completed may be). As I applied velocity constraint on the left rod, the rod could further distract so applying proper boundary condition in step 2 may solve this issue. But I am not sure what kind of boundary condition can be applied. I am trying to find the main issue and solve it.

One question, if the local coordinate system is incorrect in some places, could this also cause convergence issues?

Thank you so much again for your help.

 

[Tina a]

Hi IceBreakerSours,

Thank you very much. I did remove the boundary conditions on the nodes where tie constraints were already applied. But my model is still not converging in step 2. Step 1 is complete. Could there be any other method to fix the rod fully rather than applying boundary condition (displacement/rotation: U1,U2,U3,UR1,UR2,UR3=0) across the entire length of rod? The rod element type is C3D4. As I have mentioned in above reply, I am trying to figure out the issue and solve it. Thanks again.

 

[Dave442]

it sounds like there is a lot going on in your model. is it possible to simplify the model into smaller parts and verify that you can model each part as you intend. Then you can start to build up the assembly in steps and have confidence that everything is interacting as expected?

 

[Tina a]

I could finally run the model. Thanks to both of you.