Issues in modelling flow around airfoil

khor_albert · January 30, 2010, 23:36

I know this type of simulation is commonplace, so maybe some of you can help me out?

I am currently using a C-grid on NACA 4424 and trying to generate a flow around it to investigate the lift and drag. I have about 145000 nodes in the mesh. For low angle of attack, the values seem to be fine but weird enough, the Cl/Cd INCREASES as the angle of attack increases from 20 to 25 degrees. Stalling would have occurred by then. I think that might have been the problem of boundary conditions (I am using 2 velocity inlets and 2 pressure outlets). I checked the contours and vectors and they seem to agree that stalling has taken place; it's just not been reflected in the values! Chord length is 0.2m and velocity is 34 m/s. I am using Spalart-Allmaras model and second order upwind momentum and modified flow turbulence equations. Any ideas?

How do I actually determine turbulent length scale? Can I determine it from the chord length? I read the definition in the CFD-Wiki but I do not understand.

What is the difference between backflow conditions (as shown in pressure outlet) and normal inflow conditions (as shown in velocity inlet)? Do I use the same value (modified turbulent viscosity, turbulent length scale etc) for both of them?

I know that's a lot of questions but being relatively new to CFD, all these are still giving me a headache; hopefully someone can help me out. Cheers

jack1980 · February 2, 2010, 06:04

- Stalling will be hard to model. Try using another turbulence model, like realizable k-epsilon.

- The turbulence intensity and length scale depend on how good your physical wind tunnel is. You could start with intensity 1% and length equal to your chord.

- Backflow conditions are the conditions used when there is air flowing back into your domain at your pressure outlet. You can use the same conditions. All in all backflow is unlikely in your setup.

khor_albert · February 2, 2010, 18:43

I tried all the recommendations, but I still don't get the results I expected and I faced convergence problems as well as continuity residual never fell below 0.1.

I still get Cl and Cl/Cd at 25 degs to be HIGHER than Cl and Cl/Cd at 20 degs, which don't match the physics of the problem. Other recommendations? I have additional materials on the airfoil (that's what my activity is all about), so that will make the flow more messy and separated.

Your idea's probably good, Jack but I just need something stronger in this occasion.

January 30, 2010, 23:36	Issues in modelling flow around airfoil	#1
khor_albert New Member Khor Yee Shin Join Date: Jan 2010 Location: Glasgow UK Posts: 6 Rep Power: 16	I know this type of simulation is commonplace, so maybe some of you can help me out? I am currently using a C-grid on NACA 4424 and trying to generate a flow around it to investigate the lift and drag. I have about 145000 nodes in the mesh. For low angle of attack, the values seem to be fine but weird enough, the Cl/Cd INCREASES as the angle of attack increases from 20 to 25 degrees. Stalling would have occurred by then. I think that might have been the problem of boundary conditions (I am using 2 velocity inlets and 2 pressure outlets). I checked the contours and vectors and they seem to agree that stalling has taken place; it's just not been reflected in the values! Chord length is 0.2m and velocity is 34 m/s. I am using Spalart-Allmaras model and second order upwind momentum and modified flow turbulence equations. Any ideas? How do I actually determine turbulent length scale? Can I determine it from the chord length? I read the definition in the CFD-Wiki but I do not understand. What is the difference between backflow conditions (as shown in pressure outlet) and normal inflow conditions (as shown in velocity inlet)? Do I use the same value (modified turbulent viscosity, turbulent length scale etc) for both of them? I know that's a lot of questions but being relatively new to CFD, all these are still giving me a headache; hopefully someone can help me out. Cheers

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
modelling airfoil flow near stall	nico	FLUENT	0	October 17, 2007 06:16
modelling blood flow in arteries with starcd	sara	Siemens	5	April 10, 2007 10:17
Simulation of transonic flow over NACA0012 airfoil	MSc Student	Siemens	2	August 9, 2006 14:49
simulating separated flow past the airfoil	Yusoff	CFX	7	January 16, 2001 20:19
Troubles modelling flow through a grid	Hans Klaufus	CFX	1	June 28, 2000 17:43

February 2, 2010, 06:04		#2
jack1980 Senior Member Jouke de Baar Join Date: Oct 2009 Posts: 127 Rep Power: 17	- Stalling will be hard to model. Try using another turbulence model, like realizable k-epsilon. - The turbulence intensity and length scale depend on how good your physical wind tunnel is. You could start with intensity 1% and length equal to your chord. - Backflow conditions are the conditions used when there is air flowing back into your domain at your pressure outlet. You can use the same conditions. All in all backflow is unlikely in your setup.

February 2, 2010, 18:43		#3
khor_albert New Member Khor Yee Shin Join Date: Jan 2010 Location: Glasgow UK Posts: 6 Rep Power: 16	I tried all the recommendations, but I still don't get the results I expected and I faced convergence problems as well as continuity residual never fell below 0.1. I still get Cl and Cl/Cd at 25 degs to be HIGHER than Cl and Cl/Cd at 20 degs, which don't match the physics of the problem. Other recommendations? I have additional materials on the airfoil (that's what my activity is all about), so that will make the flow more messy and separated. Your idea's probably good, Jack but I just need something stronger in this occasion.