Calculix Solver Doesn't Display Results

apadzak

Hey everyone,

I came across a strange behavior of the post processor in Mecway 4. Running "Pipe Clip" simulation I can't get results nor model displayed in the post-processor. Automatic scale results set up shows 0 MAX and 0 MIN. Changing MAX to any positive value displays it on the scale but model doesn't display. I have another Win 7 on the same machine and all works fine, I have used the same CalculiX installation bConverged 2.8 for both OS'es. I have reinstalled both Mecway and Calculix but the behavior still exist. Any suggestion?

best regards,
apadzak

VMH

apadzak,

I'm couldn't answer your question but the results displayed okay on Mecway version 4.0 with bConverged ccx2.8p2 solver on a Windows 10 Pro 64-bit. See video below. The file used in the video is attached.

apadzak

Hi VHM, thanks for your post. As I said, I have two operating systems on one PC, one is experimental another one clean. On both of them I have Mecway 2,3,4 installed and 2.8 bConverged CalculiX. On experimental system, Mecway behaves as expected, when using CCX solver, but on the "clean system" it doesn't. I have reinstalled both, CCX and MY and used MY with and without license but nothing has changed. I suspect some system compatibility and maybe if someone had the same problem would share it here or maybe Victor could suggest something. Thanks, apadzak.

Victor

Hello Apadzak, could you please post the .frd file that Mecway is opening to display those incomplete results? You can find it in %temp%\MecwayInstance#######\Ccx after solving.

There is a problem that causes results not to be read correctly which comes from a missing environment variable that CCX requires - PRINTF_EXPONENT_DIGITS=2 and that will appear as a corrupt .frd file.

Also, could you have a look at the table in the solution. Does it show any data? Are node coordinates all zero or something else like NaN? That will determine if it's a graphics problem or the data Mecway has is corrupt.

Tim_Gowing

Hi Victor,
I have written out an input file from Mecway for Calculix and I've solved using Jeff's bConverged Launcher (bConverged_2_8_win_003_X).
I've imported the results file into Mecway and see the mesh but I don't get a stress contour displayed; the contour min/max values are evident.
I am expecting that I've missed a step in how to display the results.
Regards, Tim

Sergio

Could it be that this is a no lineal analysis? Maybe you are in the initial increment were you don't have results, so the mesh is empty. Try moving the time cursor in the lower side of the window to advance in the increments.

Is a little confuse this behavieur, in other postprocessors the initial increments is always adjusted to blue color (zero stress/strain...)

Regards

Tim_Gowing

Sergio, Thank you,
The time slider was the vital step. I'm really impressed with Mecway, it is a real treat to write out an input deck and it runs in Calculix. I only just realised that I could read the results back into Mecway; I'm more inclined now to also try and run ccx from within Mecway. I read an enhancement request with regard to ccx runtime messages and tailing the .sta file, are these implemented now?
Regards, Tim

Sergio

As far I know, not yet, but you can show the CCX output from the dedicated button on Mecway, and for tailing the .sta files I'm using Tail for Windows that made the job smoothly. Once I know that the simulation will run I just look for the Tail window to see the advance, and even adjust a tick noise for every increment in order to hear and don't have to take look. Give a try to run CCX from Mecway, I didn't find still a problem that can't be run from within, even if you don't have the right cards implemented in GUI you can use the CCX custom editions options to complete the input file and still run inside Mecway.

Would be nice to have this two windows (ccx output and .sta records) integrated in Mecway.

Regards

Victor

.sta type output is on my list but I can't promise. Also, it might end up parsing the increment and iteration counts from the main output instead.

apadzak

Hi Victor,
Thanks for your message. Sorry for late reply, I was away, now I should be quicker with answers. I couldn't find .frd file, whole folder is attached. Result tree don't display the table. There is something what I ignored before but in the solution folder "Components" there is 359 elements which don's have material assigned to it - pls see attachment. In the "Components & Materials" branch in the outline tree all looks ok with regards to materials assignment and displaying highlighted components.
Regards,
apadzak

Victor

This is kind of curious. The .frd file is in there under the Ccx folder in your attachment and I can open it OK and it seems to be a proper .frd file. Can you please try opening it and see if that works?

Also, could you have another go at the table when the problem is happening? Use the button near the right-hand end of the lower toolbar.

The error message on the component is spurious and you can ignore it. It won't be there in v5.

apadzak

Hi All!
Victor, I have tried another file which can be analyzed with MW and CCX, I have chosen CadAssembly. CCX wasn't able to solve as in the Pipe Clip example. I have selected exporting solution to CSV and only MW is able to do it, CCX didn't generate anything so I have used clipboard to past the solution manually. I did the same with both CCX solved solutions. Pipe Clip solution looks like it doesn't change time step, even when I tried to change it manually for 0 to 20 time steps. I think it may be some system variable missing, which is required by CCX. Please have a look at attachments.

apadzak

...and this are results for the PipeClip test performed on the testing system. I was able to manipulate time steps in the table and received all results... I will have a look at CCX requirements and see if I am missing anything on the other system.
Thanks,
apadzak
P.S. Sorry for the size of the file...

Victor

Hi apadzak

Could you please try opening the .frd file in the MecwayInstance64574871\Ccx folder you attached in a previous message? That seems to be a correct file and I can open it OK, so it looks like CCX is working correctly, but apparently your Mecway is ignoring all the data values when it reads it, and only reads the names of the field variables and the node numbers. This check should determine if the problem is reading .frd files in general, or only when it opens them automatically after solving.

After this, I'd like to ask you to use a special debugging version of Mecway to help narrow it down further if you don't mind.

apadzak

Hi Victor,
I did try to open .frd file and it opened in the system where Mecway works fine (there was some error message that CGX can only display displacement) but I couldn't animate results though. I think in the other system I tried to open the .frd file and it was opening without model display. I will confirm that anyway. I don't mind testing Mecway with some enhancments, it will be a pleasure actually.
Thanks,
apadzak

apadzak

Hi Victor,

I did have a look at the .fdr file in the system with problems and it displayed in the same way as in the system where all results worked fine. Please have a look at the screen videos I made and test suite summery. Videos are here: https://www.dropbox.com/sh/5jyi4t5ixd737eb/AAClwV1KUfHOswGfJM0kAMlMa?dl=0

Are results of CCX Test Suite of any help?
I did it and there wasn't any error displayed but maybe I missed something.

If you have any suggestions please let me know.
Thanks,
apadzak

Victor

Thanks. Could you please open that same .frd file that you opened in CGX but using Mecway instead (File->Open, not Solve with CCX)? This check will tell me if it's a fault in the .frd reading code or something related to running the external solver.

No need to check CCX. It's apparently generating a valid .frd file because you found that CGX can open it OK, and I can open it with Mecway myself too.

apadzak

Hi Victor,

Opening .frd file Via Mecway was successful. Good news. Is that mean Mecway can't find the right file to lad it after analysis automatically? If you need some more details please let me know.

Thanks,
apadzak

Victor

Could be. Can you please try solving the attached model only with CCX. It's a simple thermal analysis so it should produce different field variables - temperature and heat flux instead of displacement and stress. If it's still showing displacement and stress in the outline tree, that indicates it's reading some other file.

Next I'll create a modified program to help debug this. That'll take a few more days. Thanks for all your effort in helping with this.

apadzak

Hi Victor,
The same with thermal analysis. Works ok on experimental system, doesn't work on other system apart going to temp folder and opening CCX result file. By the way, is there any chance that I can change where the solution is being saved. System partition isn't best for me due to the size constraints so I would prefer to have some flexibility in that case.

I am happy to help with testing the bug...

Thanks,
Arek

P.S. Can it be some confusion related to two Win OS on the PC?

Victor

For anyone else reading this. It turned out to be a number format problem that happens with some localization settings in Windows. It's fixed in version 5. You can also work around it by changing the number format in Windows to use dots for decimal separators.

apadzak

Thanks Victor for solving it!