Different order of import mesh in IGG

dijiaqiu · December 29, 2016, 08:36

Dear All

When i calculate the centrifugal compressor, i found different order of import volute and impller mesh in IGG, the result of efficiency is different. why?

Thank you

DarylMusashi · December 29, 2016, 17:22

Dear dijiaqiu,
Sorry, I don't understand the question. May you explain what you mean with "different order"?

Kind regards
Holger

dijiaqiu · December 29, 2016, 20:05

Quote:

Originally Posted by DarylMusashi

Dear dijiaqiu,
Sorry, I don't understand the question. May you explain what you mean with "different order"?

Kind regards
Holger

Sorry, I imported volute and impeller mesh in various different orders. For example, first I imported impeller mesh in IGG,then imported volute mesh. Or the reverse sequences. The result of efficiency is different.

DarylMusashi · December 29, 2016, 21:02

Dear dijiaqiu,
may I ask how much the efficiency differs in both cases? At a first glance I cannot explain this behavior. I suppose you are going to use FINE/Turbo for your project? Or are you going to proceed with an unstructured mesh (using FINE/Open)?

From my experience it is recommended to merge different meshes in AutoGrid (by creating a 3D effect) instead of using IGG if you want to use FINE/Turbo.

dijiaqiu · December 30, 2016, 08:29

Dear DaryMusashi
I merged the meshes in AutoGrid (using FINE/Turbo).The efficiency differs is 5% ~ 10%. the interface between volute and impeller show different total pressure distribution. It's something wrong with the interface, How to check the interface mesh or compute setting is right?

DarylMusashi · December 30, 2016, 09:03

Dear dijiaqiu,
for further analysis I would recommend to set the integer expert parameter IWRIT = 1. The expert parameter settings are available under "Computation Steering -> Control Variables" and are only visible when you switch to "expert mode" in the top right corner of the GUI.

IWRIT=1 provides a more detailed report in the .mf file, which is located in your computation directory. You can open it with a text editor and analyse the flow values on both sides of your interfaces now. If everything is correct all flow variables are (nearly) equal on both sides. Maybe it will be helpful to share your insights of the analysis of both cases with us, if you want.

May I ask which type of interface you use for your computation and which relative Mach number your flow has near the interface?

dijiaqiu · January 1, 2017, 08:21

Frist thank your help. the type of interface is Conservative Coupling by Pitchwise Row, near the interface relative Mach numeber is about 0.8.

May I ask what's the differences between merging different meshes in AutoGrid(by creating a 3D effect) and merging different meshes in IGG(by importing the project).

kind regards

DarylMusashi · January 1, 2017, 16:58

May I ask if you have been able to compare the flow variables on both sides of the interface in your both projects? Theoretically the values should be the same for both projects.

Secondly, have you duplicated your FINE/Turbo project and just linked the other mesh (with different loading order)? This should work without problems, as the resulting mesh is the same in both cases. I would recommend it here, because if you set up the second project manually it might be the case that you used different settings, which result in a different solution.

Have you compared the convergence of the flow variables? In which way do they differ? You can easily do it with the "Monitor" (start it in the top bar of the FINE/Turbo GUI) and load different solution files for comparison.

To answer your question: I prefer to combine projects in AutoGrid5 with 3D effects, as you have the ability to change the meshing parameter of the rows afterwards. Secondly exchanging data is faster, as you only need to exchange the .trb template file for the rows. So only the volute .cgns file (large) is needed for merging.

If you encounter Mach numbers near 1.0 in the region of your interface the "1D non reflecting" interface type might be better (with expert parameter LOCCOR=0). But as you use the same type of interface in both projects this is surely not the reason for the differing efficiency here.

December 29, 2016, 08:36	Different order of import mesh in IGG	#1
dijiaqiu New Member Join Date: Sep 2011 Posts: 7 Rep Power: 14	Dear All When i calculate the centrifugal compressor, i found different order of import volute and impller mesh in IGG, the result of efficiency is different. why? Thank you

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Gambit problems	Althea	FLUENT	22	January 4, 2017 03:19
[mesh manipulation] Importing Multiple Meshes	thomasnwalshiii	OpenFOAM Meshing & Mesh Conversion	18	December 19, 2015 18:57
[snappyHexMesh] No layers in a small gap	bobburnquist	OpenFOAM Meshing & Mesh Conversion	6	August 26, 2015 09:38
2nd order boundary conditions for 2nd order discretization?	quarkz	Main CFD Forum	30	December 26, 2011 07:12
mesh missing after import in fluent	morteza08	FLUENT	0	July 23, 2010 02:22

December 29, 2016, 17:22		#2
DarylMusashi Senior Member Holger Dietrich Join Date: Apr 2011 Location: Germany Posts: 174 Rep Power: 14	Dear dijiaqiu, Sorry, I don't understand the question. May you explain what you mean with "different order"? Kind regards Holger

December 29, 2016, 21:02		#4
DarylMusashi Senior Member Holger Dietrich Join Date: Apr 2011 Location: Germany Posts: 174 Rep Power: 14	Dear dijiaqiu, may I ask how much the efficiency differs in both cases? At a first glance I cannot explain this behavior. I suppose you are going to use FINE/Turbo for your project? Or are you going to proceed with an unstructured mesh (using FINE/Open)? From my experience it is recommended to merge different meshes in AutoGrid (by creating a 3D effect) instead of using IGG if you want to use FINE/Turbo.

December 30, 2016, 08:29		#5
dijiaqiu New Member Join Date: Sep 2011 Posts: 7 Rep Power: 14	Dear DaryMusashi I merged the meshes in AutoGrid (using FINE/Turbo).The efficiency differs is 5% ~ 10%. the interface between volute and impeller show different total pressure distribution. It's something wrong with the interface, How to check the interface mesh or compute setting is right?

December 30, 2016, 09:03		#6
DarylMusashi Senior Member Holger Dietrich Join Date: Apr 2011 Location: Germany Posts: 174 Rep Power: 14	Dear dijiaqiu, for further analysis I would recommend to set the integer expert parameter IWRIT = 1. The expert parameter settings are available under "Computation Steering -> Control Variables" and are only visible when you switch to "expert mode" in the top right corner of the GUI. IWRIT=1 provides a more detailed report in the .mf file, which is located in your computation directory. You can open it with a text editor and analyse the flow values on both sides of your interfaces now. If everything is correct all flow variables are (nearly) equal on both sides. Maybe it will be helpful to share your insights of the analysis of both cases with us, if you want. May I ask which type of interface you use for your computation and which relative Mach number your flow has near the interface?

January 1, 2017, 08:21		#7
dijiaqiu New Member Join Date: Sep 2011 Posts: 7 Rep Power: 14	Frist thank your help. the type of interface is Conservative Coupling by Pitchwise Row, near the interface relative Mach numeber is about 0.8. May I ask what's the differences between merging different meshes in AutoGrid(by creating a 3D effect) and merging different meshes in IGG(by importing the project). kind regards

January 1, 2017, 16:58		#8
DarylMusashi Senior Member Holger Dietrich Join Date: Apr 2011 Location: Germany Posts: 174 Rep Power: 14	May I ask if you have been able to compare the flow variables on both sides of the interface in your both projects? Theoretically the values should be the same for both projects. Secondly, have you duplicated your FINE/Turbo project and just linked the other mesh (with different loading order)? This should work without problems, as the resulting mesh is the same in both cases. I would recommend it here, because if you set up the second project manually it might be the case that you used different settings, which result in a different solution. Have you compared the convergence of the flow variables? In which way do they differ? You can easily do it with the "Monitor" (start it in the top bar of the FINE/Turbo GUI) and load different solution files for comparison. To answer your question: I prefer to combine projects in AutoGrid5 with 3D effects, as you have the ability to change the meshing parameter of the rows afterwards. Secondly exchanging data is faster, as you only need to exchange the .trb template file for the rows. So only the volute .cgns file (large) is needed for merging. If you encounter Mach numbers near 1.0 in the region of your interface the "1D non reflecting" interface type might be better (with expert parameter LOCCOR=0). But as you use the same type of interface in both projects this is surely not the reason for the differing efficiency here.