basic question about force/torque-calculation

Flyinglosi · November 2, 2016, 14:44

Hallo guys,

today I had a discussion with a colleague of mine about the principles of calculation force and torque on a solid structure with cfd and the following question appeared:

Let's assume we are estimating the air-flow about a solid structures (like for example a wing) with a 2d-simulation-model. Now we want to calculate force and torque acting on the structure. As I understood from the discussion, programs like openFoam are integrating the pressure on the outer surface of the structure.

Would it be also possible to define a control-volume (containing the structure) and use linear and angular-momentum of the air-flow to calculate force and torque? If yes, does it make any difference and what are the benefits of each method?

Thanks for your answers

Greetings from Austria

Stephan

FMDenaro · November 2, 2016, 14:59

Note that, for viscous flows, the pressure is only a part (the isotropic one) of the stress tensor... the complete set of forces requires the action of the remaining stresses

Flyinglosi · November 2, 2016, 16:05

So you are talking about considering shear-stress, when evaluating the surface-forces?

But how about the other proposed method (with a control-volume and evaluation via momentum). Is that a total wrong idea?

agd · November 3, 2016, 12:17

Are you referring to something like Trefftz plane analysis?

Flyinglosi · November 3, 2016, 14:36

Yes, but with cfd, you know the velocity on the surface of a Chosen Controller volume, so calculating the momentums would be easy.

agd · November 3, 2016, 15:11

It can be done, because as you note, you have the quantities necessary to perform the calculation. The reason you don't see more of it (based on my experience/recollection/opinion) is that in CFD we also have what we need on the body to do a direct calculation. Additionally, when multiple bodies are involved in a simulation a Trefftz-type analysis cannot separate out the individual contributions. Thus, it is impractical for applied CFD used in simulations for store separation, component analysis, and the like. So while it can be a useful technique in certain testing situations (where it may be difficult to measure fluid quantities on the body surface), it doesn't really offer any advantage in CFD where all the information is known on both the body surface and the farfield.

Flyinglosi · November 3, 2016, 16:39

Thanks for your answer. The reason why I asked is based on the following "problem". We are doing calculations on a special kind of rotor (a cyclogyro) and the predicition-quality of torque/power seems to be pretty poor (after comparing cfd with measurement-data).

So I thought it is maybe more accurate to use this farfield method.

November 2, 2016, 14:44	basic question about force/torque-calculation	#1
Flyinglosi New Member Stephan Lanser Join Date: Nov 2016 Posts: 4 Rep Power: 9	Hallo guys, today I had a discussion with a colleague of mine about the principles of calculation force and torque on a solid structure with cfd and the following question appeared: Let's assume we are estimating the air-flow about a solid structures (like for example a wing) with a 2d-simulation-model. Now we want to calculate force and torque acting on the structure. As I understood from the discussion, programs like openFoam are integrating the pressure on the outer surface of the structure. Would it be also possible to define a control-volume (containing the structure) and use linear and angular-momentum of the air-flow to calculate force and torque? If yes, does it make any difference and what are the benefits of each method? Thanks for your answers Greetings from Austria Stephan

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Question about number of processors shall I use to speed up the calculation	lnk	Main CFD Forum	1	October 3, 2012 08:55
[Other] very basic question about createPhi.H	feldy77	OpenFOAM Meshing & Mesh Conversion	2	November 22, 2011 08:16
Basic question on motorBike tutorial	newToOpenFoam	OpenFOAM Running, Solving & CFD	2	July 14, 2011 09:50
[Gmsh] A basic Gmsh question: 1D transfinite	the_phew	OpenFOAM Meshing & Mesh Conversion	0	May 10, 2011 10:58
Basic question to buoyantBoussinesqPisoFoam	kriz	OpenFOAM	2	June 12, 2010 03:35

November 2, 2016, 14:59		#2
FMDenaro Senior Member Filippo Maria Denaro Join Date: Jul 2010 Posts: 6,772 Rep Power: 71	Note that, for viscous flows, the pressure is only a part (the isotropic one) of the stress tensor... the complete set of forces requires the action of the remaining stresses

November 2, 2016, 16:05		#3
Flyinglosi New Member Stephan Lanser Join Date: Nov 2016 Posts: 4 Rep Power: 9	So you are talking about considering shear-stress, when evaluating the surface-forces? But how about the other proposed method (with a control-volume and evaluation via momentum). Is that a total wrong idea?

November 3, 2016, 12:17		#4
agd Senior Member Join Date: Jul 2009 Posts: 357 Rep Power: 18	Are you referring to something like Trefftz plane analysis?

November 3, 2016, 14:36		#5
Flyinglosi New Member Stephan Lanser Join Date: Nov 2016 Posts: 4 Rep Power: 9	Yes, but with cfd, you know the velocity on the surface of a Chosen Controller volume, so calculating the momentums would be easy.

November 3, 2016, 15:11		#6
agd Senior Member Join Date: Jul 2009 Posts: 357 Rep Power: 18	It can be done, because as you note, you have the quantities necessary to perform the calculation. The reason you don't see more of it (based on my experience/recollection/opinion) is that in CFD we also have what we need on the body to do a direct calculation. Additionally, when multiple bodies are involved in a simulation a Trefftz-type analysis cannot separate out the individual contributions. Thus, it is impractical for applied CFD used in simulations for store separation, component analysis, and the like. So while it can be a useful technique in certain testing situations (where it may be difficult to measure fluid quantities on the body surface), it doesn't really offer any advantage in CFD where all the information is known on both the body surface and the farfield.

November 3, 2016, 16:39		#7
Flyinglosi New Member Stephan Lanser Join Date: Nov 2016 Posts: 4 Rep Power: 9	Thanks for your answer. The reason why I asked is based on the following "problem". We are doing calculations on a special kind of rotor (a cyclogyro) and the predicition-quality of torque/power seems to be pretty poor (after comparing cfd with measurement-data). So I thought it is maybe more accurate to use this farfield method.