Problem with the 2D rotating blade-CFX-12.1

Araz · September 3, 2011, 04:54

Hi,

I'm trying to do a steady state analysis of the 2D (one element extruded) Rotor blade of a specific compressor. It is located in the mid-span, hub and shroud are symmetry planes.
In the Turbo mode of the CFX-pre, we can only define the angular velocity of the blade and there is no place to introduce the radius. the only way that I found was to define the radius when creating the mesh.
I think that for a 2D simulation we need linear velocity of the blade not the rotational speed. please correct me if I'm wrong.
I tried different ways to setup the simulation but it is not working.
I defined the rpm for rotor, and used the rotating domain option, while using the Stationary Frame for the INLET and OUTLET boundaries. it didn't work.
I used rotating INLET and OUTLET but the solver didn't even start.
I calculated the relative total pressure, total temp. and relative velocity angles of the inlet flow and used them as the INLET boundary conditions and set the domain and the blade as Stationary. it didn't work again.
In the blade (wall) boundary details, I tried to introduce the angular (and the next time linear) velocity of the blade but each time the solver exited with error message.
I also tried using the general mode instead of Turbo mode, but didn't make much difference.
I first used the Translational periodicity and then Rotational periodicity but again bad result.
In all cases I'm starting with a low outlet static pressure and increase it slightly to the actual value in order to prevent the outlet boundary to act as a wall, but this problem happens for the working cases and when i reach the actual static pressure, 100% of the both inlet and outlet act as a wall which prevent the fluid from flowing inside or outside of the domain.
May you please help me to solve the problem? how can I setup the simulation?

Thanks in advance,
Regards,
Araz

ghorrocks · September 3, 2011, 06:54

2D simulations can be either tranlational or rotational. In your case it sounds like a rotational one might be more suitable. Have you done the axial rotor/stator example which comes with CFX? That will show you the basics of how to set this up. And read the best practises guide while you are at it.

If you want more specific help please post an image of your geometry.

Araz · September 3, 2011, 07:10

Dear Glenn,

Thanks for your reply, I have done that tutorial before and gone through the best practices guide. After doing all these I posted my question.

Here is my blade from a front view.

Thanks again

ghorrocks · September 3, 2011, 07:23

Have a look at the geometry modelled in the tutorial again. Hopefully you can see the difference between it and what you have modelled. You have modelled a very strange rotating machine. I think you will find you cannot model this 2D, but need to model a single blade passage in 3D.

Araz · September 3, 2011, 07:50

Thanks a lot Glenn,

As I said, it is a thin slice of the blade in the mid-span. It is working properly in Fluent, So I hope I can find a way to do so in CFX.
what I don't understand is that, using the velocity triangles, I calculated all the relative flow variables, so by using them as the boundary conditions, we can set the blade and the domain as Stationary and the setup would be exactly like a Stator. I have no problem with stator but I don't know why this is not working properly.
Anyway, Thanks a lot for your time and help

ghorrocks · September 3, 2011, 08:54

I see, OK then the 2D approach makes sense, providing rotation effects (coriolis, centripetal) are not significant.

Then there should be no need for rotating frames of reference at all. Just get the AOA right and it should all work from there. Do not use turbo mode to set this up as it is a simple stationary model.

September 3, 2011, 04:54	Problem with the 2D rotating blade-CFX-12.1	#1
Araz Member Araz Join Date: Feb 2011 Location: Canada Posts: 32 Rep Power: 15	Hi, I'm trying to do a steady state analysis of the 2D (one element extruded) Rotor blade of a specific compressor. It is located in the mid-span, hub and shroud are symmetry planes. In the Turbo mode of the CFX-pre, we can only define the angular velocity of the blade and there is no place to introduce the radius. the only way that I found was to define the radius when creating the mesh. I think that for a 2D simulation we need linear velocity of the blade not the rotational speed. please correct me if I'm wrong. I tried different ways to setup the simulation but it is not working. I defined the rpm for rotor, and used the rotating domain option, while using the Stationary Frame for the INLET and OUTLET boundaries. it didn't work. I used rotating INLET and OUTLET but the solver didn't even start. I calculated the relative total pressure, total temp. and relative velocity angles of the inlet flow and used them as the INLET boundary conditions and set the domain and the blade as Stationary. it didn't work again. In the blade (wall) boundary details, I tried to introduce the angular (and the next time linear) velocity of the blade but each time the solver exited with error message. I also tried using the general mode instead of Turbo mode, but didn't make much difference. I first used the Translational periodicity and then Rotational periodicity but again bad result. In all cases I'm starting with a low outlet static pressure and increase it slightly to the actual value in order to prevent the outlet boundary to act as a wall, but this problem happens for the working cases and when i reach the actual static pressure, 100% of the both inlet and outlet act as a wall which prevent the fluid from flowing inside or outside of the domain. May you please help me to solve the problem? how can I setup the simulation? Thanks in advance, Regards, Araz

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Ansys 13 CFX problem on LInux Opensuse 11.4	serendipity	CFX	10	August 7, 2012 04:24
[TGrid] overlapping problem for stator and rotor blade	kennethaktuba	ANSYS Meshing & Geometry	0	January 1, 2010 02:31
cfx mesh problem...	mactech001	ANSYS Meshing & Geometry	0	November 5, 2009 03:19
stress analysis of a rotating beam(wind blade)	Cheyannne	ANSYS	0	October 19, 2009 18:06
Fan Modelling Problem in CFX	Jenny	CFX	8	September 11, 2007 14:59

September 3, 2011, 06:54		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	2D simulations can be either tranlational or rotational. In your case it sounds like a rotational one might be more suitable. Have you done the axial rotor/stator example which comes with CFX? That will show you the basics of how to set this up. And read the best practises guide while you are at it. If you want more specific help please post an image of your geometry.

September 3, 2011, 07:23		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	Have a look at the geometry modelled in the tutorial again. Hopefully you can see the difference between it and what you have modelled. You have modelled a very strange rotating machine. I think you will find you cannot model this 2D, but need to model a single blade passage in 3D.

September 3, 2011, 07:50		#5
Araz Member Araz Join Date: Feb 2011 Location: Canada Posts: 32 Rep Power: 15	Thanks a lot Glenn, As I said, it is a thin slice of the blade in the mid-span. It is working properly in Fluent, So I hope I can find a way to do so in CFX. what I don't understand is that, using the velocity triangles, I calculated all the relative flow variables, so by using them as the boundary conditions, we can set the blade and the domain as Stationary and the setup would be exactly like a Stator. I have no problem with stator but I don't know why this is not working properly. Anyway, Thanks a lot for your time and help

September 3, 2011, 08:54		#6
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	I see, OK then the 2D approach makes sense, providing rotation effects (coriolis, centripetal) are not significant. Then there should be no need for rotating frames of reference at all. Just get the AOA right and it should all work from there. Do not use turbo mode to set this up as it is a simple stationary model.