Hello everyone,
I used turbo-machinery module for the wind turbine with version 5.2. First I run it in the steady mode. Then I added the turbomachinery module. I tried full transition , transition-couple, frozen and frozen-coupled mode. However the calculation cannot converge. I choose a very small time step less than 1e-7 and very low rotation speed. I attached the result and share the link of the mesh below . Could you please help me about that? Thank you
MESH STATOR : https://www.dropbox.com/s/i6k0vkpcuhpux ... .cgns?dl=0
MESH ROTOR: https://www.dropbox.com/s/jmwu1vzdbogwb ... .cgns?dl=0
All the best, Yan.
divergence in turbomachinery module
Forum rules
Please read the forum usage recommendations before posting.
Please read the forum usage recommendations before posting.
divergence in turbomachinery module
- Attachments
-
- listing.dat
- (9.94 MiB) Downloaded 345 times
-
- setup.xml
- (7.68 KiB) Downloaded 317 times
-
- compile.log
- (10.12 KiB) Downloaded 308 times
-
Yvan Fournier
- Posts: 4209
- Joined: Mon Feb 20, 2012 3:25 pm
Re: divergence in turbomachinery module
Hello,
The listing does not show how the computation diverges, but your time step seems much too small (the max CFL is less than 0.1, while it should be around 1. Ideally, to have a mean CFL closer 1, a max CFL around 5 or even 10 is acceptable, and possibly better if it improves the mean CFL).
As your post does not include some of the recommended info, I might be missing other things.
Regards,
Yvan
The listing does not show how the computation diverges, but your time step seems much too small (the max CFL is less than 0.1, while it should be around 1. Ideally, to have a mean CFL closer 1, a max CFL around 5 or even 10 is acceptable, and possibly better if it improves the mean CFL).
As your post does not include some of the recommended info, I might be missing other things.
Regards,
Yvan
Re: divergence in turbomachinery module
Hello Yvan ,
Thank you. I will try to improve the time step. There is no diverge shown on the listing file, but the maxmum velocity is so large. I check the result in Paraview when I choose transition_coupled. and found the problem on the interface of rotor and stator. The figure shows the pressure of flow field. I also attached other info I can provide. Could you please find any problem about that?
When I chose full transition module , the error output is "SIGSEGV signal (forbidden memory area access) intercepted!" before the calculation. I found this because I choose the periodic boundary condition. When I use symmetry boundary condition, the case can run. So I think the periodic boundary condition can not go with the transition module???
All the best,
Yany.
Thank you. I will try to improve the time step. There is no diverge shown on the listing file, but the maxmum velocity is so large. I check the result in Paraview when I choose transition_coupled. and found the problem on the interface of rotor and stator. The figure shows the pressure of flow field. I also attached other info I can provide. Could you please find any problem about that?
When I chose full transition module , the error output is "SIGSEGV signal (forbidden memory area access) intercepted!" before the calculation. I found this because I choose the periodic boundary condition. When I use symmetry boundary condition, the case can run. So I think the periodic boundary condition can not go with the transition module???
All the best,
Yany.
- Attachments
-
- preprocessor_01.log
- (6.66 KiB) Downloaded 333 times
-
- setup.log
- (22.37 KiB) Downloaded 303 times
-
Re: divergence in turbomachinery module
Hello，Yvan，
If you have time could you please have a look at my case? I'm really appreciate that. I 'm working on that nearly one year. The mesh is simple with good quality. I choose turbomachinery module of transition but after some time steps, the error file output:
Error in turbomachinery mesh update:
Number of boundary faces has changed from 154422 to 154452.
There are probably unjoined faces, due to an insufficiently regular mesh;
adjusting mesh joining parameters might help.
I modified the joining parameters and tried different turbulence modules,but it didn't work.
The module of coupled didn't work too. What confused me is that I cannot even find the problem. Everything looks OK .But it still didn't run.
Thank you very much for your help.
Regards. Yany.
If you have time could you please have a look at my case? I'm really appreciate that. I 'm working on that nearly one year. The mesh is simple with good quality. I choose turbomachinery module of transition but after some time steps, the error file output:
Error in turbomachinery mesh update:
Number of boundary faces has changed from 154422 to 154452.
There are probably unjoined faces, due to an insufficiently regular mesh;
adjusting mesh joining parameters might help.
I modified the joining parameters and tried different turbulence modules,but it didn't work.
The module of coupled didn't work too. What confused me is that I cannot even find the problem. Everything looks OK .But it still didn't run.
Thank you very much for your help.
Regards. Yany.
Yvan Fournier wrote: ↑Mon May 07, 2018 9:12 pm Hello,
The listing does not show how the computation diverges, but your time step seems much too small (the max CFL is less than 0.1, while it should be around 1. Ideally, to have a mean CFL closer 1, a max CFL around 5 or even 10 is acceptable, and possibly better if it improves the mean CFL).
As your post does not include some of the recommended info, I might be missing other things.
Regards,
Yvan
-
Yvan Fournier
- Posts: 4209
- Joined: Mon Feb 20, 2012 3:25 pm
Re: divergence in turbomachinery module
Hello,
The meshes you posted and those in your setup file do not have the same names though ? Are they the same meshes ? Also after how many iterations do you have the issue ?
You mesh looks quite good, but checking more closely, the rotor does not seem completely centered (i.e. its center si not quite at 0,0,0 but slightly shifted in the y direction. So in the rotor definition, you need to slightly shift the invariant point to account for this (or recenter your mesh). Loading rotor.cgns alone in ParaView, I see an X rage from -1.34994 to 1.34998, and a Y range from -1.35 to 1.34997, so using (2e-5, -1.5e-5, 0) as an invariant point should help. The rotor is still not perfectly round (2.69992 X span, 2.69997 Y span, so 5e-5 difference between x and y spans.
The shift may be due to the CAD tool. What tool did you generate the CAD with ? The error can be either user error or CAD precision issues. Use error is easier to fix, but I would not be surprised is the issue is simply due to CAD precision issues. Shifting the mesh a bit and stretching it by the correct ratio in the x and y directions in a preprocessing step should help...
Regards,
Yvan
The meshes you posted and those in your setup file do not have the same names though ? Are they the same meshes ? Also after how many iterations do you have the issue ?
You mesh looks quite good, but checking more closely, the rotor does not seem completely centered (i.e. its center si not quite at 0,0,0 but slightly shifted in the y direction. So in the rotor definition, you need to slightly shift the invariant point to account for this (or recenter your mesh). Loading rotor.cgns alone in ParaView, I see an X rage from -1.34994 to 1.34998, and a Y range from -1.35 to 1.34997, so using (2e-5, -1.5e-5, 0) as an invariant point should help. The rotor is still not perfectly round (2.69992 X span, 2.69997 Y span, so 5e-5 difference between x and y spans.
The shift may be due to the CAD tool. What tool did you generate the CAD with ? The error can be either user error or CAD precision issues. Use error is easier to fix, but I would not be surprised is the issue is simply due to CAD precision issues. Shifting the mesh a bit and stretching it by the correct ratio in the x and y directions in a preprocessing step should help...
Regards,
Yvan
Re: divergence in turbomachinery module
Hello Yvan,
Thank you for the reply.
Yes， you are right . The rotation zone is not exactly a circle. I modified the mesh today. However, the issue still exist after about 10 iterations.
I uploaded the new mesh and setup file.
ROTOR: https://www.dropbox.com/s/jmwu1vzdbogwb ... .cgns?dl=0
STATOR: https://www.dropbox.com/s/i6k0vkpcuhpux ... .cgns?dl=0
setup.xml : https://www.dropbox.com/s/iq53a5u5nf4m2 ... p.xml?dl=0
I'm wondering is there any bugs for the turbomachinery module? Thank you very much.
All the best, Yan.
Thank you for the reply.
Yes， you are right . The rotation zone is not exactly a circle. I modified the mesh today. However, the issue still exist after about 10 iterations.
I uploaded the new mesh and setup file.
ROTOR: https://www.dropbox.com/s/jmwu1vzdbogwb ... .cgns?dl=0
STATOR: https://www.dropbox.com/s/i6k0vkpcuhpux ... .cgns?dl=0
setup.xml : https://www.dropbox.com/s/iq53a5u5nf4m2 ... p.xml?dl=0
I'm wondering is there any bugs for the turbomachinery module? Thank you very much.
All the best, Yan.
-
Yvan Fournier
- Posts: 4209
- Joined: Mon Feb 20, 2012 3:25 pm
Re: divergence in turbomachinery module
Hello,
There are no known bugs in the turbomachinery module which is used quite often for major studies).
Is the crash the same type you had before ? Please post the listing and error logs (and usier subroutines if you have some), because running the computation is a bit long.
Regards,
Yvan
There are no known bugs in the turbomachinery module which is used quite often for major studies).
Is the crash the same type you had before ? Please post the listing and error logs (and usier subroutines if you have some), because running the computation is a bit long.
Regards,
Yvan