Problem of excessive element distortion in ALE adaptive meshing 1

[biofriend]

Hi
My friends

In my model, which is suction of a gel to a pipette, the step visco fails to converge and yields the error of "too many attempts "at the beginning or middle of the procedure which is likely due to "excessive element distortion".
Tricks like altering *CONTROLS will only delay the inevitable as a friend truly said.

As much as I have perceived, Dynamic Explicit is the a good way in combination with ALE adaptive meshing to tackle element distortion problem. It is clear that it helps the process to go further but it still fails due to distortion!!! and gives the following error:

" There are a total of 2 excessively distorted elements The elements contained in element set ErrElemExcessDistortion-Step1 have distorted excessively."

How do you recommend me to change and adjust the Adaptive meshing parameters like frequency, number of sweeps and if it helps to use adaptive mesh controls to get rid of this problem? I tried various values for these frequency and number of sweeps but the problem still exists with a little temporal shift.
I am not very familiar with this topic and Abaqus manual was somehow vague in this regard.
Is ALE adaptive ineffective in solving my case and I have to try another way?

Links to photos that may help to have a schematic imagination of the model:
1-

http://files.engineering.com/getfile.aspx?folder=b701fb87-7f6e-4d16-9700-63d1623e7cec&file=1.JPG

2-

http://files.engineering.com/getfile.aspx?folder=55bd768e-91cc-40f4-9949-34b896f2b9c2&file=2.JPG

Many thanks in advance

 

[williwang]

I've had the same problem before. Waiting fot the answer.

Forever Young.....

 

[biofriend]

I will let you know the possible solution as soon as i have reached any answer,But till now the problem is still ambiguous and nobody seems to have an opinion toward this matter

 

[williwang]

What about export the lasted shape as igs or step file, remesh it, and restart the analysis at that point

Forever Young.....

 

[xerf]

For large deformations you should try to design the mesh such that the element distortion is minimized.

Also, you should try to reduce the frequency for ALE.

 

[biofriend]

Hi!
Dear williwang
The method which you mentioned(Mesh to mesh solution mapping )is,as far as I know,not a very time-friendly one,besides I am not familiar with that process.I have put it aside as the last card when all the other otions failed to answer.I am still experimenting ALE and will let you know my discoveries!

 

[biofriend]

Hi!
Dear Xerf

For the 1st point,I am of the opinion that your model should answer regardless of the mesh type you choose,
yeah;using tri elements or a coarser mesh help but why shouldn't I be able to elicit answer with quad for example?
So I guess that the fundamental problem lies somewhere else.

And for the 2nd,
I am not very sure now but as I highly increase the frequency it goes further without any error! But as the solution time was very long I am still not sure about the results. I will let you know the outcome.
Regards

 

[xerf]

What I wanted to say is that you should request the adaptive meshing procedure to occur more often. However, this is typically more computationally expensive. I think each time the adaptive meshing is called, it takes as much as solving for several explicit increments.

If I am not wrong, for the FREQUENCY parameter one specify the number of time increments after which the adaptive meshing is called. Therefore, by decreasing its value, the adaptive meshing will be applied more often.

Best.

 

[biofriend]

Dear Xerf
Your idea initiated some doubt,I will check this.
Thanks dear,
Bests

 

[biofriend]

Dear Xerf

As much as I have understood, increasing the frequency will request the adaptive meshing procedure to occur more often and therefore distortion will be procrastinated.

But in my case I think that distortion is inevitable since the loading is introduced to be instantaneous which means a huge load applied in a very short time interval and therefore an impact which makes the elements to distort.
As I have to introduce the load instantaneously and not as a ramp , I have to find a cure for this case.

Regards

 

[biofriend]

It seems that I have finally understood what my problem is.My load is instantaneous,therefore excessive distortion is inevitable.I have to substitute this with a ramp load.Increasing frequency also helps to prevent distortion as far as I have understood.Another interesting matter is mass scaling.It helps you to reduce computational costs.You can increase it as much as the kinetic energy of the system in a transient(quasi-static)remains negligible in comparison with the internal energy.

Thank you all my friends for your ideas and contributions,
Regards