Urgent: Problem with Map Solution/adaptive meshing

[biofriend]

Hi
Dear Dr. Bergstrom and my other friends

I am trying to use *map solution option in Abaqus 6.8 to remesh the object in which the meshes are highly distorted, map the solution variables from previous failed job and continue the analysis, However I think that I simply just do not get a part of this procedure and therefore have not been successful in it so far:

Let me briefly explain my approach and I would be very grateful if you had any comment on the part in which I am making mistake/mistakes :
1- I run job 1 in the model with restart request ON, frequency=1.
2- The job fails due to excessive elements distortion.
3- I extract the deformed shape from the ODB and make a solid geometry out of it.
4- I remesh the deformed geometry and reset the BCs , load and interactions boundaries like they were in the old model.
5- I edit the model attribute to restart from the desired frame(increment) and step of the job 1.
6- I set the predefined fields for the new model ( initial states) to be those of job 1 (initial conditions).
7- I create job 2, write an input file and introduce *map solution , step=,inc=,unbalanced stress=, into it.
8- Then I type abaqus job=2 oldjob=1 in abaqus command.
It fails!

I have tried some other ways with few changes in the procedures mentioned above they all failed showing different error messages.

Where do you think I am going wrong? This is an important step of my thesis which should have been finished long before, I’m out of time.

Your any comments are highly appreciated.

Thank You very much for your consideration,
Regards,
Biofriend

[Jorgen]

I have to admit that I don't use "*Map Solution" frequently. Your approach looks good to me though. What error message do you get?

You mentioned that your initial simulation failed due to element distortion. Are you using Standard or Explicit? Have you tried to use ALE adaptive remeshing?

Sometimes it helps to activate element distortion, or slightly increase the strain hardening at large strains (in order to distribute the deformation).

-Jorgen

[biofriend]

Hi
Dr. Bergstrom
Thanks for your ever-so-kind replies,
I am using Visco/Standard and it seems that I have not a better way to use explicit so ALE is not among my options(it doesn't cater for visco).

Error messages can be different depending on how I am changing the subtle points in my procedure so that I am a little confused. There should be a fine mistake somewhere in my method and I really do not find the Abaqus manual very clear.It could not be compared with MatLab for an ordinary user.

However, I am still working on this case and if I could break it I will let you know.
Any comments,any,are highly appreciated,
Yours
Biofriend

[biofriend]

Hi
Dear Dr. Bergstrom

I have succeeded in executing the map solution and I found out the problem was in using predefined fields in the remeshed model which caused error.

Now I have two other questions.
do I have to calculate the magnitude of load in the beginning of running each remeshed model and put it instead of original initial load values? cause it seems that it starts from zero and do not continue it from the value it should be at that time increment which is not right.

the other is that if I want to extract the time-displacement diagram, is there anyway to do it at once like I could do in an ordinary job or I have to acquire nodal displacements from different split jobs and put them together in MATLAB or Excel?

Thank You very much for you help in advance
Yours Sincerely
Biofriend

[Jorgen]

(1) I suspect that you either have to calculate the correct load at the beginning of the run. You can perhaps do this by simply referring to the same load amplitude definition since the time is continuous.

(2) Again I suspect that you need to extract the time-displacement response in two steps and then combine in an external program.

-Jorgen