# Pressure Jump Correction. Actuator Disc Model and numerical wiggles

 Register Blogs Members List Search Today's Posts Mark Forums Read

June 4, 2014, 05:50
Pressure Jump Correction. Actuator Disc Model and numerical wiggles
#1
Senior Member

M. Montero
Join Date: Mar 2009
Posts: 138
Rep Power: 15
Hi all,

I am working over the rotorDisk model implemented over OpenFoam and I have found a critical issue related to the SIMPLE algorithm.
The issue is described in the next paper:

http://orbit.dtu.dk/fedora/objects/o...190785/content

Quote:
 Discrete body forces are used in the present context to model the influence of wind turbines on the flow. In order to overcome the pressure wiggles introduced by discre te body forces, one approach is to smooth out the body forces by using a Gaussi an distri- bution instead of a Dirac delta distribution ]
I found the problem when simulating the same configuration with TI=10% and TI=1% keeping turbulent length scale and wind speed. The TI=10% simulation matches theoretical quite well but TI=1% underpredict the wind speed in the rotor plane and power extraction is even 30 % lower than expected. Here it a picture of what happens.

In the report, he modifies the Rhie-Chow/SIMPLE pressure correction algorithm, but it is possible to overcome this issue by distributing the force around the cell.

I think that the amplitude of the oscillation near the rotor plane is aleatory but could depend on TI, mesh discretization, schemes.... With a TI = 10%, that wiggle is still there but the amplitude of the oscillation is not so high so the error when calculating the angle of attack is not high.

Best Regards
Marcelino
Attached Images
 rotor_detail.jpg (24.7 KB, 151 views)

 June 4, 2014, 07:16 #2 Senior Member     Niels Nielsen Join Date: Mar 2009 Location: NJ - Denmark Posts: 544 Rep Power: 26 Hi Marcelino regarding the rotorDisk, do you have a simple test case? I have made a simple case, but I cannot for the life of me find you why I do not get any radial/tangential vectors from the disk. There is only axial thrust and everything looks to be fine, but just no radial/tangential vectors. I can post my case here if you want to have a look. __________________ Linnemann PS. I do not do personal support, so please post in the forums.

 June 4, 2014, 08:39 #3 Senior Member   M. Montero Join Date: Mar 2009 Location: Madrid Posts: 138 Rep Power: 15 Hi Linnemann, What OF version do you use? There were some bugs in OF2.3.0 and before about rotorDiskSource. http://www.openfoam.org/mantisbt/vie...d=708#bugnotes Be sure to use last OF2.3.x where this issue was solved and another one about the forces proyection. I do not have a simple test case that I could share, but feel free to post your case and if I can, I will help you. About the main topic of the post, it seems it is in relation to this other topic: http://www.cfd-online.com/Forums/ope...-pressure.html and http://www.cfd-online.com/Forums/ope...rce-model.html. In this latter topic, you posted also. The same problem was seen by tstovall But I do not know how to modify simpleFoam to manage correctly body forces. It would be great if someone could give some support Best Regards

 June 4, 2014, 08:46 #4 Senior Member     Niels Nielsen Join Date: Mar 2009 Location: NJ - Denmark Posts: 544 Rep Power: 26 Ahh Well I am using 2.3.x but not nearly that new a commit. I will update and see where that leaves me. I will also post a simple case for you/others to investigate. __________________ Linnemann PS. I do not do personal support, so please post in the forums.

June 4, 2014, 17:20
#5
Senior Member

Niels Nielsen
Join Date: Mar 2009
Location: NJ - Denmark
Posts: 544
Rep Power: 26
Hi

After the update i get some tangential velocities.

But everything looks very strange, on the old code i got some ok axial thrust.

Now its just a mess.

The propeller rotates around the y axis with the flow direction in the positive y-direction.

Here is a Drive link to the case.
Attached Images
 image.jpg (48.0 KB, 207 views)
__________________
Linnemann

PS. I do not do personal support, so please post in the forums.

June 5, 2014, 06:59
#6
Senior Member

M. Montero
Join Date: Mar 2009
Posts: 138
Rep Power: 15
Hi.

I have been using a "personal" updated version of OF2.3.0 where velocity and forces are correctly proyected ( vectors from cartesian CS to cylindrical CS)

In this post it is included the solution proposed by GDTech that works for me.

I attach you a picture where the flow seems ok.

Using OF2.3.x, I thought that the bug fixed proposed by the official site would solve the problem but.....I have used it and I obtain strange AoA distribution and forces like you so .... I will reopen the bug report.
Attached Images
 image1.jpg (29.9 KB, 217 views)

 June 6, 2014, 02:33 #7 Senior Member     Niels Nielsen Join Date: Mar 2009 Location: NJ - Denmark Posts: 544 Rep Power: 26 Hi was that picture made with my small case? __________________ Linnemann PS. I do not do personal support, so please post in the forums.

June 6, 2014, 04:55
#8
Senior Member

M. Montero
Join Date: Mar 2009
Posts: 138
Rep Power: 15
yes, it was but with my own OF2.3.0, not OF2.3.x.
I have modified a little bit the fvOptions file to rotate it as a propeller.
Attached Files
 fvOptions.txt (2.4 KB, 206 views) log.rotor.zip (48.5 KB, 99 views)

Last edited by be_inspired; June 9, 2014 at 06:01.

 October 21, 2014, 15:11 #9 New Member   Join Date: Apr 2013 Posts: 11 Rep Power: 11 Dear be_inspired, I am somewhat new to openFoam and I am trying to set up a case with a main rotor similar to what you have. Would it be posible to ask for some guidance on the set up as there are no examples on how to use this function. I started by modifying heatExchanger, changing fvOption from MFRSource to rotorDisk with its coefficients, and removing the excess. I appreciate your time and I think I can work it out from an example of the ./Allrun and any additional dict file if that is possible. Thanks

April 19, 2017, 16:47
#10
Senior Member

CFD
Join Date: Nov 2010
Location: United States
Posts: 243
Rep Power: 15
Quote:
 Originally Posted by be_inspired Hi all, I am working over the rotorDisk model implemented over OpenFoam and I have found a critical issue related to the SIMPLE algorithm. The issue is described in the next paper: http://orbit.dtu.dk/fedora/objects/o...190785/content I found the problem when simulating the same configuration with TI=10% and TI=1% keeping turbulent length scale and wind speed. The TI=10% simulation matches theoretical quite well but TI=1% underpredict the wind speed in the rotor plane and power extraction is even 30 % lower than expected. Here it a picture of what happens. In the report, he modifies the Rhie-Chow/SIMPLE pressure correction algorithm, but it is possible to overcome this issue by distributing the force around the cell. I think that the amplitude of the oscillation near the rotor plane is aleatory but could depend on TI, mesh discretization, schemes.... With a TI = 10%, that wiggle is still there but the amplitude of the oscillation is not so high so the error when calculating the angle of attack is not high. Best Regards Marcelino
Did you happen to implement the Guassin distribution to get rid of the wiggle?

Look forward to hearing from you.