Learning Forum

[MickMack]

Hi Guys,

I tried to implement store results at Last Time Point in my analysis but the output file remains 23.8GB so i don't believe it is working correctly.

I will outline the process i followed below and i would appreciate any assistance,

Thanks,

Michael

After reviewing previous posts particularly the one linked below i thought i understood that by using 'Last Time Point' i would only have results for the analysis at the end of the last step (2 steps in analysis). This i thought would reduce the size of my results file, 'file.rst', which i retrieve from the HPC i am using.

This is not the case and 'file.rst' is 23.8Gb as shown in the 2nd image below.

The model is shown in the last image below. It is a concrete t-beam with reinforcement. It is neshed with a global size of 0.015m and has 73,752 nodes and 62,256 elements . In a previous model i used a global mesh size of 0.025m, which created 18,849 nodes and 13,612 elements , producing a results file of 2.6Gb in size.

While i expected some increases in run time and size i did not expect it to get this big this fast. The purpose of the analysis is to check mesh refinement.

I hope i have made this clear and look forward to any help, cheers.

(/forum/discussion/10815/storing-results-in-ansys-mechanical#latest)

[MickMack]

Hi Can anyone help me with this?

[Govindan Nagappan]

Last Time point option writes result at the end of each load step. So, if you have 5 load steps, you will see results saved at the end of all 5 load steps
How many steps do you have?
YOu can use command under solution to check how many result sets are saved in the result file:
Example commands to use:

This should produce an output like this in the solution information (solver output or post output)
:


YOu can use command to create a new result file from existing result file and save only the last step results
Check the commands in the command reference manual in Mechanical APDL documentation
Sample commands
resume,file,db ! resume database file
/post1
file, file,rst
set,last
reswrite,my_new_result_file,1,1,1 !RESWRITE,filename,load step, substep, time

YOu can then import the new result file into Mechanical Just clear the existing results. Use read result file in Mechanical. (Make a copy of the files before you try this or test it on a smaller model to understand the process)

[mrife]

was the first analysis a single or two step analysis (1 or 2 load steps?)? The second FEM is about 4.5x larger than the first, so I would expect the rst file to be about 4.5x larger as well. The analysis settings are per load step, so the second model is saving the last time point for each load step. If the first analysis used only one load step then the behavior would be expected. Otherwise we will need to look deeper at the differences between the two models.
Mike

[MickMack]

The only difference between the two models is the change to the mesh size, i had thought but i just realised the first model was set to 'Store Results at Equally Spaced Points = 20'. I have included images below showing this
It is a two step static analysis. The finer mesh is producing a results file approximately 10 times larger in size.
Michael

[MickMack]

Will i check how many result sets are saved in the result file? I haven't downloaded the larger one yet but i can it will take a while.
Regarding ussing Command, i have no experience with using APDL, other than insering a few commands into Mechanical, how would i get started or where doi insert the commands

Thanks forgetting back to me Michael

[mrife]

the result file compression amount can vary quite a bit. I've not done any testing, nor seen any done, to see how well the "sparse" compression technique compresses files of the same model with differing mesh sizes. So I have nothing to go on but a hunch that the smaller models file compressed more. Try turning off the file compression and solving each again. What happens?
Mike
p.s. which version are you using?