CFX Analysis of a wind turbine

Bournegenius · December 3, 2010, 21:33

I am carrying out analysis of a wind turbine in CFX as part of my final year project . Due to complicated geometry , i cannot analyze a single blade , so i have to analyse the whole scaled model in cfx . i have created three domains , Stationary 1 , Rotating and Stationary 2 . Inlet is specified at stationary 1 ( By total pressure 1atm) , in the middle there is the rotating domain ( Rotational speed = 60 rpm), while outlet is specified at Stationary 2 (by mass flow rate 4.65 kg/s) . The turbine model is in the rotating domain .
Other things u need to know abt my setup are :

1. Interface is created b/w Stationary 1 and Rotating (FLuid-Fluid interface , frozen rotor model used)
2. interface is created b/w rotating and stationary 2 (Fluid fluid interface frozen rotr model used)

Rotating Domain :
1. Boundary condition WALL is specified at the turbine geometry .
2. Shroud is set as counter rotating wall .

Please refer to the attached jpg to see the setup .

The problem i am having is that the streamlines of velocity obtained in post processing pass through the turbine (which was specified as a WALL in rotating domain ) , otherwise the flow seems to be pretty accurate .

Can anyone tell me what i am doing wrong ?

ghorrocks · December 4, 2010, 06:52

It is because you are using velocity in the stationary frame in the rotating section. Use the variable "Velocity in stationary frame" and it won't do that.

Bournegenius · December 6, 2010, 02:58

dear ghorocks , thans for ur reply . I also tried using Velocity in Stn frame , but still the streamlines pass through the blade specified as wall . Any other suggestions ? um really in trouble here ..
regards

Abdullah Asad

ghorrocks · December 6, 2010, 07:34

If you don't want lines through walls you should look at each section in its own frame of reference. Look at the stationary sections with velocities in the stationary frame of reference, look at the rotating sections in a rotating frame of reference.

There could also be other problems

Bournegenius · December 7, 2010, 01:07

What other problems could be there ? i am using solid model of my turbine inside the rotating domain specified as wall .. should i use a hollow model ? ( by hollow model i mean leave the geometrical space occupied by the turbine vacant inside rotating domain) .... could this be the problem ?

sanchovg2 · February 2, 2012, 09:40

Hey, i´m having the same problem, wanted to ask you if you could resolve the problem. and how? im getting a good convergence the presure and velocity plot are ok but the streamlines just go straight.you can not see the turbulence in the velocity vector plot. :s what can i do?

ghorrocks · February 2, 2012, 17:24

What do you mean by not seeing the turbulence in the velocity vector plot? Do you mean you want to see the turbulent eddies in the velocity vectors?

Can you describe your model - what turbulence model, what are you modelling?

sanchovg2 · February 3, 2012, 07:46

thank you for the quick answer

what i meant was that in the velocity vector plot i dont see the vortices that are supposed to be generetad by the fluid flow after a body.because the stream lines are a representation with lines of this vortices.isnt it? or im wrong :s ?
I design myself a wind turbine in pro e with some data and studies i fond on the internet.the wind velocity is 8 m/s and the rpm are 22.5
So what i did was to put the velocity with a normal velocity inlet and the rotation i did it over the reference frames and the gave rpm.
I extracted the turbine from a small circular volumen and then this one from a bigger fluid area. i connected the circular volumen and the fluid volumen with 2 interfaces.
the first one the circular front/back is an In place-FAN interface and the second one is the other area and i did it with an In plane-Baffle Interface. Then i said that the turbine region should have an tagential velocity specification method: none and reference frame absolut. im not sure if it works this way for the rotion and the flow from wind.
the physics models that im using: 3D,Steady,Ideal Gas,Segregated flow,turbulent,K-epsilon T. and im runing another simulation with Stadard Spalart-Allmaras (eddie). but im not to familiarized with this turbulenz modell.
i also wanted to ask something like this is possible with star ccm+
http://www.youtube.com/watch?v=nj_iL8PXOD8

ghorrocks · February 5, 2012, 18:27

A well designed wind turbine running at design conditions will have very little vorticies coming off it.

sanchovg2 · February 6, 2012, 05:57

Still, it doesn't matter how well something is design (aereodinamycs) it always comes to swirls. plus it is rotating there must be vortices.this is how the streamlines look like after eddie.1. Pic and the 2. its from a ventitalortor tutorial

ghorrocks · February 6, 2012, 20:13

The amount of swirl on something like this will be tiny. Look closely and you will find it.

sanchovg2 · February 7, 2012, 06:09

hi, can you help to interpreted this plots i got.i think i have some trouble with the MRF or with the interfaces :s
as it where complete separeted from the flow area. no effect from the wind turbine. the last pic is how it shoul look like

ghorrocks · February 7, 2012, 17:25

These simulation have been done in Star and I do not know star at all. You should post these question on the star forum.

junc · August 16, 2012, 14:09

Quote:

Originally Posted by Bournegenius

I am carrying out analysis of a wind turbine in CFX as part of my final year project . Due to complicated geometry , i cannot analyze a single blade , so i have to analyse the whole scaled model in cfx . i have created three domains , Stationary 1 , Rotating and Stationary 2 . Inlet is specified at stationary 1 ( By total pressure 1atm) , in the middle there is the rotating domain ( Rotational speed = 60 rpm), while outlet is specified at Stationary 2 (by mass flow rate 4.65 kg/s) . The turbine model is in the rotating domain .
Other things u need to know abt my setup are :

1. Interface is created b/w Stationary 1 and Rotating (FLuid-Fluid interface , frozen rotor model used)
2. interface is created b/w rotating and stationary 2 (Fluid fluid interface frozen rotr model used)

Rotating Domain :
1. Boundary condition WALL is specified at the turbine geometry .
2. Shroud is set as counter rotating wall .

Please refer to the attached jpg to see the setup .

The problem i am having is that the streamlines of velocity obtained in post processing pass through the turbine (which was specified as a WALL in rotating domain ) , otherwise the flow seems to be pretty accurate .

Can anyone tell me what i am doing wrong ?

hi,Bournegenius.have you solved the streamline problem?
i have a problem similar to yours. i am simulating a ducted fan, it seems that the streamlines pass through the fan, i can't find out the answer.
thanks in advance

ghorrocks · August 16, 2012, 19:37

Streamline do not pass through walls in a steady state, stationary frame of reference simulation unless:
* The integration of the streamline is not accurate
* The simulation is not converged adequately.

If the simulation is a rotating frame of reference simulation the streamlines may hit walls if you integrate velocity in the stationary frame in a rotating frame of reference. Alternately, if you integrate velocity in the rotating frame of reference they might hit in the stationary frame.

If the simulation is transient then instantaneous streamlines can hit the walls. If this bothers you then change to massless particle tracks, which include the time integration.

junc · August 20, 2012, 11:20

Quote:

Originally Posted by ghorrocks

Streamline do not pass through walls in a steady state, stationary frame of reference simulation unless:
* The integration of the streamline is not accurate
* The simulation is not converged adequately.

If the simulation is a rotating frame of reference simulation the streamlines may hit walls if you integrate velocity in the stationary frame in a rotating frame of reference. Alternately, if you integrate velocity in the rotating frame of reference they might hit in the stationary frame.

If the simulation is transient then instantaneous streamlines can hit the walls. If this bothers you then change to massless particle tracks, which include the time integration.

thank you for your reply, ghorrocks.
my problem is that i ntegrate velocity in the stationary frame in a rotating frame of reference.
can you tell me why instantaneous streamlines can hit the walls?

ghorrocks · August 20, 2012, 20:16

When a wall has some motion normal to the wall then streamlines can hit the wall. Just have a play with it and you will see. I do not feel like writing a mathematical proof of it as it is obvious when you look at what is going on.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Wind turbine simulation	Saturn	CFX	60	July 17, 2024 06:45
CFD analysis on wind turbine rotor	Ken (Wind Turbine CFD Super Rookie)	Main CFD Forum	45	February 9, 2016 15:07
analysis of wind turbine blades	sharuk	Main CFD Forum	7	December 9, 2014 19:59
Can I use Turbogrid to mesh a wind turbine blade?	Henry Liu	CFX	10	November 10, 2011 06:41
Wind turbine simulation	Saturn	FLUENT	1	June 16, 2006 03:12

December 4, 2010, 06:52		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	It is because you are using velocity in the stationary frame in the rotating section. Use the variable "Velocity in stationary frame" and it won't do that.

December 6, 2010, 02:58		#3
Bournegenius New Member abdullah asad Join Date: Dec 2010 Posts: 3 Rep Power: 16	dear ghorocks , thans for ur reply . I also tried using Velocity in Stn frame , but still the streamlines pass through the blade specified as wall . Any other suggestions ? um really in trouble here .. regards Abdullah Asad

December 6, 2010, 07:34		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	If you don't want lines through walls you should look at each section in its own frame of reference. Look at the stationary sections with velocities in the stationary frame of reference, look at the rotating sections in a rotating frame of reference. There could also be other problems

December 7, 2010, 01:07		#5
Bournegenius New Member abdullah asad Join Date: Dec 2010 Posts: 3 Rep Power: 16	What other problems could be there ? i am using solid model of my turbine inside the rotating domain specified as wall .. should i use a hollow model ? ( by hollow model i mean leave the geometrical space occupied by the turbine vacant inside rotating domain) .... could this be the problem ?

February 2, 2012, 09:40		#6
sanchovg2 New Member Join Date: Feb 2012 Posts: 19 Rep Power: 14	Hey, i´m having the same problem, wanted to ask you if you could resolve the problem. and how? im getting a good convergence the presure and velocity plot are ok but the streamlines just go straight.you can not see the turbulence in the velocity vector plot. :s what can i do?

February 2, 2012, 17:24		#7
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	What do you mean by not seeing the turbulence in the velocity vector plot? Do you mean you want to see the turbulent eddies in the velocity vectors? Can you describe your model - what turbulence model, what are you modelling?

February 3, 2012, 07:46		#8
sanchovg2 New Member Join Date: Feb 2012 Posts: 19 Rep Power: 14	thank you for the quick answer what i meant was that in the velocity vector plot i dont see the vortices that are supposed to be generetad by the fluid flow after a body.because the stream lines are a representation with lines of this vortices.isnt it? or im wrong :s ? I design myself a wind turbine in pro e with some data and studies i fond on the internet.the wind velocity is 8 m/s and the rpm are 22.5 So what i did was to put the velocity with a normal velocity inlet and the rotation i did it over the reference frames and the gave rpm. I extracted the turbine from a small circular volumen and then this one from a bigger fluid area. i connected the circular volumen and the fluid volumen with 2 interfaces. the first one the circular front/back is an In place-FAN interface and the second one is the other area and i did it with an In plane-Baffle Interface. Then i said that the turbine region should have an tagential velocity specification method: none and reference frame absolut. im not sure if it works this way for the rotion and the flow from wind. the physics models that im using: 3D,Steady,Ideal Gas,Segregated flow,turbulent,K-epsilon T. and im runing another simulation with Stadard Spalart-Allmaras (eddie). but im not to familiarized with this turbulenz modell. i also wanted to ask something like this is possible with star ccm+ http://www.youtube.com/watch?v=nj_iL8PXOD8

February 5, 2012, 18:27		#9
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	A well designed wind turbine running at design conditions will have very little vorticies coming off it.

February 6, 2012, 20:13		#11
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	The amount of swirl on something like this will be tiny. Look closely and you will find it.

February 7, 2012, 17:25		#13
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	These simulation have been done in Star and I do not know star at all. You should post these question on the star forum.

August 16, 2012, 19:37		#15
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	Streamline do not pass through walls in a steady state, stationary frame of reference simulation unless: * The integration of the streamline is not accurate * The simulation is not converged adequately. If the simulation is a rotating frame of reference simulation the streamlines may hit walls if you integrate velocity in the stationary frame in a rotating frame of reference. Alternately, if you integrate velocity in the rotating frame of reference they might hit in the stationary frame. If the simulation is transient then instantaneous streamlines can hit the walls. If this bothers you then change to massless particle tracks, which include the time integration.

August 20, 2012, 20:16		#17
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	When a wall has some motion normal to the wall then streamlines can hit the wall. Just have a play with it and you will see. I do not feel like writing a mathematical proof of it as it is obvious when you look at what is going on.