Hey Guys,
I'm running a coupled thermal/CFD analysis, and it's failing, giving me an error to the effect of AlgSolve.exe is not a valid windows application.
Looking in the Simulation directory, I see that AlgSolve.exe is 0 bytes.
Did something get corrupted on my system, or is this error indicative of something else?
Thanks,
Todd
Solved! Go to Solution.
Solved by AstroJohnPE. Go to Solution.
By the way, the thermal solver log file contains:
**** Fixing loads ... **** Invoking PVSS Sparse Solver ... Error. solver failure (1).
and the simulation ends with "Temperature update failed."
So I tried uncoupled, like I have successfully run before, albeit with a different model.
This time, it fails and tells me there was 1 warning and 0 errors.
But here's the log file. I can't find anything!
Autodesk (R) Simulation Steady-State Heat Transfer Version 2013.01.00.0012-W64/X64 28-Jun-2012 Copyright (c) 2012, Autodesk, Inc. All rights reserved. DATE: SEPTEMBER 4, 2012 TIME: 04:32 PM Input Model: S:\Applied Research\Mathematical Models\Absorption\Second Attempt\Absorption.ds_data\3\ds PROGRAM VERSION: 201301000012 alg-win-x64.dll VERSION: 201301000012 algconfig-win-x64.dll VERSION: 201301000012 agsdb_ar-win-x64.dll VERSION: 201301000012 solvercallback-win-x64.dll VERSION: 201301000012 amgsolve.exe VERSION: 0360000000 **** Model Unit System Settings: -------------------------------------------- Unit System : Custom Force : N Length : mm Time : s Temperature (Absolute) : deg C (K) Thermal Energy : J Voltage : V Current : A Electrical Resistance : ohm Mass : N*s^2/mm -------------------------------------------- **** CONTROL INFORMATION number of node points = 234652 number of element types = 5 analysis type code = 10 equations per block = 0 bandwidth minimization flag = 0 **** PHYSICAL CONSTANTS USED WHEN APPLICABLE The Stefan-Boltzmann Constant = 5.668E-14 Temperature Increment to Absolute = 2.731E+02 **** NON-LINEAR ITERATION CONTROL PARAMETERS Nonlinear control option = OFF Convergence criterion = 2 Maximum number of iterations = 15 Interval for monitoring = 5 Corrective tolerance = 1.0000E-03 Relative tolerance = 1.0000E-03 Relaxation parameter = 1.0000E+00 **** PRINT OF NODAL DATA SUPPRESSED **** PRINT OF EQUATION NUMBERS SUPPRESSED **** FLUID FLOW CONVECTION IS INCLUDED IN THIS MODEL. [ 1] Input fluid flow result file for part: 3 <S:\Applied Research\Mathematical Models\Absorption\Second Attempt\Absorption.ds_data\1\ds.sfv> Load Case: 1 Viscous Heating: 0 Turbulence: 0 **** HARD DISK FILE SIZE INFORMATION FOR PROCESSOR: Available hard disk space on S drive = 495577.395 megabytes **** LOAD CASE MULTIPLIERS Boundary temperature multiplier 1.000 Convection multiplier 1.000 Radiation multiplier 1.000 Heat generation multiplier 1.000 **** Invoking BCSLIB-EXT Sparse Solver ... BCSLIB-EXT solver memory status: in-core memory requirement (MB) = 1933.45 minimum memory requirement (MB) = 201.17 user specified memory (MB) = 145254.69 available physical memory (MB) = 2905.09 available virtual memory (MB) = 8208.75 memory currently allocated (MB) = 1933.45
Any ideas? Anywhere else I can look?
I guess you must run a legacy model.
PVSS solver should NOT be called any more.
I think you should switch to other available solver now.
Hi S.LI,
Thank you for the response, but I don't know what you mean.
I just built the assembly in SolidWorks and created the Simulation model. In fact, I've done it a few times, trying to solve this problem. What would make it "legacy?"
How do I choose which solver to use?
Thanks,
Todd
I thought your model had been created in a old version.
As shown in your log file, PVSS solver is called, but it shouldn't.
PVSS is an old equation solver, which is stopped in 2013 SP1.
I don't know why it's still called in your model.
What I can suggest for you is to find the UI input, and choose another equation solver, such BCS.
Hi Todd,
Your second post (2012-09-04 12:16 PM Eastern time) where you show the log file with the line "Invoking PVSS Sparse Solver" indicates that the heat transfer analysis is trying to use the PVSS sparse solver to solve for the temperatures. The PVSS sparse solver was removed from 2013 since it is an outdated solver. My guess is that you have some load in the heat transfer part of the analysis that used to cause the processor to switch to the PVSS solver, and we missed changing that logic when the PVSS was removed. You could try going to "Setup > Model Setup > Thermal Parameters", then to the "Solution" tab, set the "Type of solver" to "Sparse" and the "Type of sparse solver" to "BCSLIB-EXT". Hopefully, whatever load was causing the processor to "use" the PVSS solver can be solved with the BCSLIB-EXT solver.
Or, it could be due to the applied loads or the model size or some other combination of parameters that caused it to want to use the PVSS solver. I was not able to reproduce the problem, so I do not know what the combination is. (By the way, Algsolve.exe is the PVSS solver, so it is no wonder that it is 0 bytes. Perhaps the processor sees that the file exists and attempts to use it. I would suggest deleting the file Algsolve.exe and see if the analysis runs.)
In your third post where you get 1 warning, 0 errors during the steady state heat transfer analysis, the warnings and errors can appear in either the log file or the summary file. (Sometimes the pop-up message that indicates there is a warning or error doesn't report the correct file, so you need to look at both files. These can be viewed from the "Report" tab, under the last branch of the browser.)
Let us know what happens.
Hi John,
I see! Yes, I think that was the problem. Here's what I was doing with the independent simulation. Perhaps this will help fix the bug:
1. I ran the CFD to get fluid velocities.
2. In the thermal solver, I set my boundary conditions and ran the model to get the zero flow results. This worked fine.
3. Then I loaded the fluid velocities and ran it again.
Here the simulation would stop with no errors or warnings. I couldn't find evidence of warnings in either files, even though the dialog said one warning. I didn't notice the PVSS notice in the files, either, but I may have just missed it.
A few minutes before I got your response, I was continuing with my trial and error, and:
4. I right-clicked on the Analysis Type and selected Edit Analysis Parameters. (Note: This is the first time I edited the analysis parameters AFTER specifying the fluid velocity file.)
5. A dialog popped up that told me something about having to use the Sparse Solver, that some other solver wouldn't work with velocities, etc.
6. I didn't change anything, but clicked OK to close the Analysis Parameters dialog.
7. I re-ran the model and it worked fine.
So it looks like the problem occurs when velocities are loaded. Either it doesn't change the solver type to sparse and the solver barfs, or it does change it, but it changes it to PVSS.
Hope that helps!
Thanks again for the help. I haven't tried the coupled model again, but I'll let you know if I do.
Todd
Hi John,
Everything works now, both in the coupled and uncoupled solutions. I just have to go into the thermal solver and select Sparse instead of leaving it at automatic.
Thanks again!
Todd