Topology Optimization Failed - Not enough RAM?
Altair EmployeeHi,
I'm getting the attached error log when running topology optimization with optistruct on a formula student car part I'm designing. The machine I'm using has 8GB of RAM. Isn't that sufficient? Or is it a different issue?
Thanks!
Josh
Answers
-
Hi Josh,
Looks like the problem is with the disk space. Can you change the working directory for the analysis and try again?
** WARNING # 1971
Current run is likely to fail due to insufficient disk space.
Estimated disk space requirement = 83903 MB
Estimated disk space available = 20777 MB
Suggestions to reduce disk space requirements:
For static analysis - reduce the number of loadcases.
For eigenvalue/buckling analysis - reduce frequency range in
EIGRL card or number of modes requested.
For FRF analysis - reduce the number of loading frequencies.
For transient analysis - reduce the number of output time steps.0 -
Hi,
thanks for your reply! Can you tell me how I do that? Before running the analysis, do I just save the FEM file to the larger allocated space?
Cheers!
Josh
0 -
Hi,
I tried saving it to a larger external HDD but the simulation still failed. This time I got the new attached error log. From the error messages I'm still suspecting inadequate RAM in the machine. Can you confirm that for me? Also, if I put a coarser mesh on my model, would it then be able to run?
Thanks!
0 -
According to your log, I think you need more RAM for your box.
You're NOT University user, so upgrade to 32G ram is not problem :-)
Good luck,
0 -
Sorry, I'm not entirely sure what you mean. Are you saying I have to physically upgrade the PC with more RAM? It currently has 8GB and I'm not sure whether it is upgradeable to even 16GB. Is there a way to get it to run with less RAM? A coarser mesh maybe?
(Also, I am in fact a student and I am also part of the Altair University program, but I don't actively use it yet. Does that technically make me a university user?)
Cheers!
0 -
Your forum profile tells me:
'Are you University user? : No'
If you can not add more ram, try a coarser mesh.
0 -
Thanks, I'll try that!
Also, I am not sure whether university user relates to me actually being a student at the University of Glasgow, or whether it means I'm part of the Altair University.
0 -
Hi again,
so now I'm done with exams I've had time to continue with this project. I've deleted the previous mesh and remeshed the model with a larger element size.Then I tried re-running the optimization, but got the attached .out file with a new error message. Not sure what to make of it. It does list a few possible reasons for the failure, but I don't know which it could be. Any suggestions?
Also, looking at the new mesh, the original rigids do not seem to match it, even though I created the new mesh with 'match existing mesh' ticked. Not sure whether this is an issue, but thought I'd mention that too.
Lastly, if uploading the HM file itself would help in resolving the issue, I'd be more than happy to share it!
Thanks again!
0 -
*** WARNING: The amount of allocated memory exceeds available RAM - this may cause significant slowing down of solution progress due to the swapping out memory to the swap file. It is recommended to use out-of-core solution to reduce memory requirements.
Again, your RAM is too small for your analysis :-)
MEMORY ESTIMATION INFORMATION :
-------------------------------Solver Type is: Sparse-Matrix Solver
Direct Method
Current Memory (RAM) : 3849 MB
Estimated Minimum Memory (RAM) for Out of Core Solution : 3849 MB
Recommended Memory (RAM) for Out of Core Solution : 3849 MB
Recommended Memory (RAM) for In-Core Solution : 10151 MB
(Note: The Minimum Memory Requirement is limited by Optimization Module.)*** ERROR # 153 ***
Exactly zero pivoting encountered during Numerical Factorization;
the model may have rigid body mode.
Solver error no. = -503
spc set id = 6
index = 1
Possible reasons are:
1) insufficiently constrained model,
2) having rigid body mechanisms within the model,
3) extremely ill-conditioned rigid element sets,
4) extremely thin shells (as used for skinning) that have MID2/MID3,
5) gap elements with extremely high stiffness (KA, especially KT or MU).
Check the model and rerun the problem.
(MECHCHECK may be used to find the rigid body modes. To do so, change the input to be an eigenvalue analysis and add MECHCHECK.)
(WARNING: results obtained with MECHCHECK cannot be used because the model is changed internally.)
This error was detected in subroutine bcsmtxfct.
This error occurs in module 'lstatdrv'.I think the log file is very understandable.
0 -
Right, I'll try a coarser mesh again then. Although, can you tell me what an out of core solution is and why it doesn't run? The PC has 8GB of RAM, so out of core *should* run according to the error.
Also, I am asking these questions to get an explanation for what is going wrong. I do not understand the log file. I am not sure what zero pivoting during numerical factorization means. Can you explain please? I have also encountered the rigid body mode error before, but managed to solve it without actually knowing what rigid body mode is. Can you please explain that too? I do not know how I solved it last time, because I do not understand what it means.
Solver error nr, SPC set ID and index also do not mean anything to me, so if you could clarify these, I might be able to get it to work myself.
Thanks!
Josh
0 -
Hi,
In general, this error could be because of insufficient constraints in the model which results in ill-conditioned stiffness matrix and Eigen solver fails to extract the modes.
The normal procedure in such cases is to check the model and provide sufficient BC's.
It is recommended to run the model with MECHCHECK and (or) GROUNDCHECK option in normal mode loadstep to detect the massless mechanism where the massless mechanism will show up as rigid body modes ( This run should be considered as a step to debug the problem as the results from this solver is inaccurate)
You can find the GROUNDCHECK and MECHCHECK in the Analysis > Control CardsThe GROUNDCHECK option will perform a grounding analysis on stiffness matrix to expose unintentional constraints by moving the model rigidly and prints the information in the .out file.
Regards
Rahul R
0