CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   Propeller calculation (http://www.cfd-online.com/Forums/cfx/73501-propeller-calculation.html)

toil March 10, 2010 05:31

Propeller calculation
 
5 Attachment(s)
:confused: Need Help!
Hi guys:
Recently I am working on a simple propeller ( rotor ) to calculate its aerodynamic characters, such as its moment and thrust. For convenience I neglected the propeller's hub so the propeller has two blades only. I pasted the picture below.(figure 1)

Sine there isn't stator, I thought a single rotating domain can work it out. So I created a single domain using RFR to set its rotating speed as propeller rotating speed. I pasted a domain picture below as well. (figure 2)

The boundaries were indicated on the 2nd picture, I specified the Inlet Velocity on boundary INLET ( such as 5 m /s ); Zero relative pressure on boundary OPEN; Mass Flow Rate on boundary OUTLET. Those boundary were all stationary. The working material is Idea Air, and using SST model for turbulence.

Here come problems.
1. The U V W and P-Mass RMS converge well, but the thrust and moment of propeller do not converge ( they decrease ). see the picture I pasted.(figure 3)

2. There is a big vortex ring just below the blade disk, and it moves down as the calculation carry on. I pasted a pressure contour on iteration 302 and 502. (figure 4)

3. The blade wake seems strange too, its geometry behaviors like above UNKOWN vortex ring. Bases on the propeller theory, the propeller wake should contract just blow the blade disk and maintain a certain radius screw downstream. But the calculated wake shows a diffused part . The diffused wake moves down as calculation carry on, I thought it may indicates that the flow filed is not converged yet? I don't know am I right?(figure 5)

4. By my dubious thought, I continued the calculation. As I forecasted, the reverse flow appears on the outlet boundary face as the big Unknown vortex ring passes by. The solver added wall automatically, when the vortex ring over passed, the calculation becomes normal again. Because this is my 1st propeller simulation. I wonder you guys have met those phenomena before? or I just didn't simulate it in a correct way.

I really need help.

Tell me if I am wrong. Thanks!

By Toil

Attesz March 10, 2010 07:07

Read this:
http://www.cfd-online.com/Forums/cfx...imulation.html

regards,
Attesz

ghorrocks March 10, 2010 07:31

Q1 - You probably need to run for longer. It has not cleared the start up transient yet.
Q2 - Looks like a startup vortex to me. Keep running the simulation to let it convect out of the domain.
Q3 - Don't jump to conclusions yet. Run the simulation for longer until it has stabilised.
Q4 - The warning just means there is some reverse flow at the outlet. As long as it clears it should be fine.

toil March 10, 2010 08:25

Quote:

Originally Posted by Attesz (Post 249316)

To Attesz:
Thank you very much for helping me with this issue.
I have seen the Thread that you linked to me. I think we face different problems.
Firstly I don't even know whether my method is right? Can I solve a propeller problem using a single domain with RFR? Are those boundary conditions are setted appropriate? Are you using a stationary domain which contains a rotating domain to calculate your rotor? Can you explain the method you use?
Secondly, because this is my fist time to calculate a propeller, I am not quite sure about the calculated phenomena, I want to know did any other have experienced those phenomena?
By the way, Because I thought it is a steady flow in rotating frame, so I set this simulation steady.
Thank you again!
By Toil

toil March 10, 2010 08:39

Quote:

Originally Posted by ghorrocks (Post 249323)
Q1 - You probably need to run for longer. It has not cleared the start up transient yet.
Q2 - Looks like a startup vortex to me. Keep running the simulation to let it convect out of the domain.
Q3 - Don't jump to conclusions yet. Run the simulation for longer until it has stabilised.
Q4 - The warning just means there is some reverse flow at the outlet. As long as it clears it should be fine.

To ghorrocks:
Thanks a lot for your answers and advices.
According to your answers, a propeller can be calculated using a single domain with RFR if there is no stator, right?
I am quite agree with you that the vortex ring is a star vortex, and I'll follow your adivce to continue the calculation. Maybe when the start vortex ring convect out of the calulating domain, the propeller wake and its thrust will be converged.
I'll paste my newest progress.

By Toil

Attesz March 10, 2010 08:40

Hi,
Firstly, You use Fluent, I'm not so experienced in that. I think RFR meand Rotating Frame...
Anyway, I used three domains:
1. stationary, in front of the rotor
2. rotating, containing the blades,
3.stationary behind the rotor
The rotating domain has to be as "little" as possible, it should contain only the real rotating fluid near the blades, otherwise you get wrong results.
So if you models a piece in the air, and the propeller in it, so you can use a stationary domain which contains the rotating one.

toil March 10, 2010 09:08

To Attesz:
Thanks for your reply!
I am using CFX not Fluent. The RFR is short for Rotating Frame of Reference, you are right. The way I use also called SRF which is short for "Single Rotating Frame" in Fluent I think (Let me know if I am wrong). The method you using I think is so called MFR which contains a stationary frame for your stationary domain and a rotating frame for your rotating domain. I used this method before, but I just can't get continuous propeller wake stream line. I think this is mostly caused by the different relative velocity in those two domains.
Can I have a look at your rotor wake? By the way how did you set you boundary conditions?
I'll also follow your method to calculate this propeller once you let me know your method detail such as the boundary settings, I'll share you the latest calculation progress!
Thanks for your help.
By Toil

ghorrocks March 10, 2010 17:38

I do not agree with Attesz. The only reason you would have both a rotating part and a stationary part is if you needed that to get the geometry or boundary conditions correct. In your case it looks like you can make the entire thing a RFR, so you have no need for a stationary component. This has the advantage that you do not need a GGI and visualising the results will be easier as they are all in the same frame of reference.

Attesz March 11, 2010 08:09

Hi,
I've now time to read the posts detailed at least.
I've never used Rotating Frame. In your case It's a single rotating domain, with a stationary wall, right? So it's the exact opposite when you rotates the blades in a stationary fluid?
If I'm right, and if you really haven't any stator parts, this way is better than mine. But usually the stator has an important effect to the blades, so to the forces, drag etc.
I can send you images at the weekend, please, write a message otherwise i will forget it:rolleyes:
My boundaries: opening everywhere:D total pressure at the inlet and static pressure at the outlet faces. I have not so good experiences using mass flow outlet, but if you have measures about it, then use that of course. If this rotor operates in the air, it is easier to set the environment parameters, as I did it in my simulation.
The calculation of the forces is very sensitive to the fineness of the mesh around the blades. If you get wrong results, this is the first point to examine. I've measured the thrust on a plane behind the rotor, but here told me cleverer people than me, that measuring on the surfaces of the blades is the correct way, so use this.

By the way, I've used three domains as I wrote with interfaces, because I have stators and wanted to simulate with them. The boundaries was set as I write above, with SST turbulence model. But I get wrong results by measuring the lift forces, due the coarse mesh.

Good luck,
Attesz

Attesz March 11, 2010 11:24

http://smartech.gatech.edu/bitstream...200412_phd.pdf

same story :)

toil March 15, 2010 21:48

I am sorry, I am absent from this Forum for quite a long time because of my other calculation tasks. Now let's get back to my thread, I did carry on my propeller calculation, unfortunately once the START VORTEX RING moving through my outlet boundary, my calculation got disconverged. I don't know why and I am still trying.
I think changing my outlet boundary condition to average static pressure will lead better result.
I will continue update this thread for my leaset progress.
By the way, if you guys have any ideas, plesae tell me!
Thanks!

By Toil

Shafiul December 8, 2010 16:45

Hello Toil,

I'm still waiting for your updates!!:confused:
Specially to know your boundary conditions and stationary/rotatting frame issue to make sure I'm following the right way.

Please reply asap.

Thanks,
Shafi

cfd_analysis December 18, 2012 13:45

Hi all,
I am going to 3D flow analysis on 8 bladed aircraft propeller. I came to know that there two ways to do it in fluent.

Multiple Rotating Reference frame (MRF)
This is done by creating three domains.
1 stationary, in front of the propeller
2. rotating, containing the blades,
3.stationary behind the propeller


Single Rotating Reference Frame (SRF)
Propeller is stationary and the flow rotates around it (domain is rotating)

I would like to know in which method one is correct method and which method is easy to carry out CFD analysis.

ghorrocks December 18, 2012 18:12

This is the CFX forum. Try the Fluent forum for your question.

And please do not post the same question multiple times. I have removed the duplicate posts.

cfd_analysis December 19, 2012 05:12

Hi ghorrocks,
Thanks for your information.


All times are GMT -4. The time now is 10:32.