Ways to simplify a flow hrough cylinder with swirl

skarthik · July 15, 2015, 11:26

Dear All,

I am working on a project to develop a model to predict the flow distribution in a micro gas turbine in Fluent.

I have defined the boundary conditions with values for inlet mass flow rate and outlet pressure, and a no-slip condition on the walls. Please refer to the attached image. Capture.JPG

I am able to solve this 3D model as a whole, but I would like to reduce the computation time by simplifying the model. Is it possible to take only a quadrant and solve for it? I have not tried it, but have heard that CFX can do this type of simulation for turbo machines (mixing plane model). I guess this is something like the axisymmetric method but still 3D and also includes my swirl vanes, which would not be possible in 2D axisymmetric.

Please comment if this would work and where should I look for some guidance to do this type of simulation. Also, let me know if there is another way to do this.

Thanks,
Karthik

SMagee · July 15, 2015, 12:13

It seems to me that what your looking for is called rotationally periodic conditions or cyclic conditions. From the image it looks like you have 60º (hexagonal) simmetry.
So, reduce the model to a 60º sector, in the mesher sellect the boundary as symetric, then in Fluent use the TUI to apply periodic conditions this is, if I remmenber correctly in: boundary conditions->make periodic.
I find this option to be very useful when modeling cylindric objects.

skarthik · July 15, 2015, 13:45

Sebastian,

Thank you.. This looks like what I am looking for. Will try it out and let you know how it goes.

To all,
Googled rotationally periodic glows and found this: 'http://www.cfd-online.com/Forums/fluent/105041-fluent-periodic-boundary-condition-solved-tutorial.html'
Could be really useful!

SMagee · July 16, 2015, 04:02

There is also a page or two in ANSY's help that may be helpful

skarthik · July 16, 2015, 07:35

Hey,

I get an error when using match control saying that the geometry/topology do not match.

I just realized that the swirl vanes are actually not symmetric. There are 7 vanes. I think this could be the reason for the error, as I tried to cut my model into a section of 90 degrees.
So I am working on a solution for this, and was wondering if you would know of any method to remove the swirl vanes, and still provide a swirl motion. I know that this could be done at the inlet of the model using cylindrical coordinates and defining a tangential velocity/mass flow rate. But I want the inlet to be axial and the swirl should start only at the inner cylinder. Do you know how to apply this to an arbitrary plane located at the place where the swirl vanes would be?
Thanks.

SMagee · July 16, 2015, 10:43

You could design a turbine with 6 vanes and use 60º or 120º but not 90º. Or the cobustor to have 51.4º symetry, and use the apropiate settings.
Or if you want to keep the current design divide the model in two, the first from the inlet to the begining of the combustor cylinder, after the vanes; the second from there on till the outlet.
The first would have 360/7=51.4º symmetry, the second 60º. For this option you would have to export the profile of the velocities at the outlet of the first part and supply these as the inlet velocity profiles to the second. (you will have to append the profile from 0º to 8.4º (60º-51.4º) to make the full 60º profile that you need for the second part)

I don't know of any simpler way to accomplish the swirl from uniform inlet.

skarthik · July 16, 2015, 10:51

Thank you for your input. I will try them out.

skarthik · July 21, 2015, 06:25

Hey Sebastian,

I tried to implement your suggestions.. I am able to run the simulation properly when I removed the swirl vanes from the model and modified the holes (14 annular holes) so that I can use a section of 51.4degrees with cyclic region symmetry. The match control works fine without errors.

But when I included the swirler, things dont work well, and I get an error in match control saying that the two periodic faces do not have the same topology or geometry. I am not sure what goes wrong, because the two faces are similar. I have attached the images here.

Is it something wrong with the coordinate system that I have defined?

Thanks!
Karthik S

1.jpg

2.jpg

July 15, 2015, 11:26	Ways to simplify a flow hrough cylinder with swirl	#1
skarthik New Member Karthik S Join Date: Jul 2015 Posts: 18 Rep Power: 11	Dear All, I am working on a project to develop a model to predict the flow distribution in a micro gas turbine in Fluent. I have defined the boundary conditions with values for inlet mass flow rate and outlet pressure, and a no-slip condition on the walls. Please refer to the attached image. Capture.JPG I am able to solve this 3D model as a whole, but I would like to reduce the computation time by simplifying the model. Is it possible to take only a quadrant and solve for it? I have not tried it, but have heard that CFX can do this type of simulation for turbo machines (mixing plane model). I guess this is something like the axisymmetric method but still 3D and also includes my swirl vanes, which would not be possible in 2D axisymmetric. Please comment if this would work and where should I look for some guidance to do this type of simulation. Also, let me know if there is another way to do this. Thanks, Karthik

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Swirl flow recirculation	mfmohdyasin	OpenFOAM Running, Solving & CFD	4	June 12, 2015 11:20
flow around a circular cylinder with velocity inlet and outflow outlet	shuoxue	OpenFOAM	1	March 3, 2014 11:42
benchmark: flow over a circular cylinder	goodegg	Main CFD Forum	12	January 22, 2013 12:47
flow around a cylinder	pXYZ	Main CFD Forum	14	July 25, 2011 11:05
flow resistance of cylinder in parallel flow	Ralf Schmidt	Main CFD Forum	3	May 11, 2009 03:26

July 15, 2015, 12:13		#2
SMagee New Member Sebastian Magee Join Date: Jul 2015 Posts: 5 Rep Power: 11	It seems to me that what your looking for is called rotationally periodic conditions or cyclic conditions. From the image it looks like you have 60º (hexagonal) simmetry. So, reduce the model to a 60º sector, in the mesher sellect the boundary as symetric, then in Fluent use the TUI to apply periodic conditions this is, if I remmenber correctly in: boundary conditions->make periodic. I find this option to be very useful when modeling cylindric objects.

July 15, 2015, 13:45		#3
skarthik New Member Karthik S Join Date: Jul 2015 Posts: 18 Rep Power: 11	Sebastian, Thank you.. This looks like what I am looking for. Will try it out and let you know how it goes. To all, Googled rotationally periodic glows and found this: 'http://www.cfd-online.com/Forums/fluent/105041-fluent-periodic-boundary-condition-solved-tutorial.html' Could be really useful!

July 16, 2015, 04:02		#4
SMagee New Member Sebastian Magee Join Date: Jul 2015 Posts: 5 Rep Power: 11	There is also a page or two in ANSY's help that may be helpful

July 16, 2015, 07:35		#5
skarthik New Member Karthik S Join Date: Jul 2015 Posts: 18 Rep Power: 11	Hey, I get an error when using match control saying that the geometry/topology do not match. I just realized that the swirl vanes are actually not symmetric. There are 7 vanes. I think this could be the reason for the error, as I tried to cut my model into a section of 90 degrees. So I am working on a solution for this, and was wondering if you would know of any method to remove the swirl vanes, and still provide a swirl motion. I know that this could be done at the inlet of the model using cylindrical coordinates and defining a tangential velocity/mass flow rate. But I want the inlet to be axial and the swirl should start only at the inner cylinder. Do you know how to apply this to an arbitrary plane located at the place where the swirl vanes would be? Thanks.

July 16, 2015, 10:43		#6
SMagee New Member Sebastian Magee Join Date: Jul 2015 Posts: 5 Rep Power: 11	You could design a turbine with 6 vanes and use 60º or 120º but not 90º. Or the cobustor to have 51.4º symetry, and use the apropiate settings. Or if you want to keep the current design divide the model in two, the first from the inlet to the begining of the combustor cylinder, after the vanes; the second from there on till the outlet. The first would have 360/7=51.4º symmetry, the second 60º. For this option you would have to export the profile of the velocities at the outlet of the first part and supply these as the inlet velocity profiles to the second. (you will have to append the profile from 0º to 8.4º (60º-51.4º) to make the full 60º profile that you need for the second part) I don't know of any simpler way to accomplish the swirl from uniform inlet.

July 16, 2015, 10:51		#7
skarthik New Member Karthik S Join Date: Jul 2015 Posts: 18 Rep Power: 11	Thank you for your input. I will try them out.

July 21, 2015, 06:25		#8
skarthik New Member Karthik S Join Date: Jul 2015 Posts: 18 Rep Power: 11	Hey Sebastian, I tried to implement your suggestions.. I am able to run the simulation properly when I removed the swirl vanes from the model and modified the holes (14 annular holes) so that I can use a section of 51.4degrees with cyclic region symmetry. The match control works fine without errors. But when I included the swirler, things dont work well, and I get an error in match control saying that the two periodic faces do not have the same topology or geometry. I am not sure what goes wrong, because the two faces are similar. I have attached the images here. Is it something wrong with the coordinate system that I have defined? Thanks! Karthik S 1.jpg 2.jpg