I'm doing something about the aerodynamic performance of wind turbine airfoil with Fl

shlgzz · January 14, 2013, 08:58

Hello.all!
I'm doing something about the aerodynamic performance of wind turbine airfoil with Fluent.Setting the same simulating conditions with experiment,however,the simulating results is quite different from the experiment data.The simulating maximum CL is 2.5,while the exp data is 2.Could you help you find the reasons?Thank U!

cfd seeker · January 14, 2013, 09:44

Post some pics of your mesh and also give some more info about your problem like domain, boundary conditions, turbulence model, Re. No, sover settings etc

shlgzz · January 14, 2013, 22:09

There are 15 boundary layers near the airfoil and the height of first layer is 0.0003.Comapare with other papers,I choose the K-W SST turbulence model and default for sover settings,the operating Re is 100000.As to the mesh,the total mesh number is about 500000 although I choose the unstructed mesh.

cfd seeker · January 14, 2013, 22:40

500000 cells for airfoil simulation! oops its too much. I have obtained good results for airfoils with just around 100000 cells. Anyhow mesh resolution near the wall is also good but this does not effect Cl. Turbulence model is also ok. Have you used 2nd order upwind for all equations? What about the domain size? How many chord lengths upstream, downstream and in other directions?

shlgzz · January 14, 2013, 23:24

It's terrible!Just 100000 cells.You are doing dynamic stall,either?I haven't tried 2nd upwind for all equations because I have tried under the steady conditions but no good result.The domain size is about 10 lengths upstream,20 lengths downstream and 20 lengths in width direction.

cfd seeker · January 15, 2013, 01:17

Yes 100000 cells and the results were accurate with just 2-3 % error as compared toliterature results. Domain size is good enough. Use 2nd order upwind for all equations using the previuosly converged results as the initail conditions. BTW at which angle of attack you are doing your simulation?

shlgzz · January 15, 2013, 02:46

Thank you,cfd seeker.It's very kind of you to answer my question so qucikly.I am aiming at obtaining the aerodynamic performance with the aoa ranging from 10 to 30 degrees in a sine function.For the transient problem,I set the time step as 0.001 and maxsium interation mumber as 20 interations/time step.But the result is really bad.

cfd seeker · January 15, 2013, 03:41

time step is very large as compared to the smallest cell size in your mesh. Use time step of 10^-6

shlgzz · January 15, 2013, 06:25

I would have a try.Thank U very much.
Best regards.

shlgzz · January 15, 2013, 10:10

BTW,what did you set the convergence criteria as on the condition of oscillating airfoil?

cfd seeker · January 15, 2013, 14:41

Quote:

Originally Posted by shlgzz

BTW,what did you set the convergence criteria as on the condition of oscillating airfoil?

what do you mean by oscillating airfoil? do you mean flapping motion?

shlgzz · January 15, 2013, 22:29

The relevant description is that the unsteady aerodynamic charcateristics of the NREL S809 airfoil undergoing sinusoidal pitch oscillations about the quarter chord point at various reduce frequencies,mean angles of attack and pitch oscillation amplitude of about 10 degrees at Re=1000000 were investigated through solutions of two demensional Navier-Stokes equations with transition corrected k-w turbulence model.

January 15, 2013, 02:46		#7
shlgzz New Member Wheeler Join Date: Jan 2013 Posts: 24 Rep Power: 13	Thank you,cfd seeker.It's very kind of you to answer my question so qucikly.I am aiming at obtaining the aerodynamic performance with the aoa ranging from 10 to 30 degrees in a sine function.For the transient problem,I set the time step as 0.001 and maxsium interation mumber as 20 interations/time step.But the result is really bad. Last edited by shlgzz; January 15, 2013 at 02:48. Reason: bad spelling

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Ducted wind turbine (BC for the shroud)	Pepita	CFX	4	June 29, 2013 08:09
FSI - Wind Turbine	AUN	CFX	13	August 29, 2012 17:44
Negative lift in Modeling an airfoil design for wind turbine	Ram209e	Main CFD Forum	3	August 8, 2012 23:39
[ANSYS Meshing] Aerodynamic wind turbine modeling	welaa	ANSYS Meshing & Geometry	0	February 13, 2012 15:57
[snappyHexMesh] 2D wind turbine airfoil (snappyHexMesh OR blockMesh)	openfoam1	OpenFOAM Meshing & Mesh Conversion	0	January 5, 2010 03:15

January 14, 2013, 09:44		#2
cfd seeker Senior Member Join Date: Mar 2011 Location: Germany Posts: 552 Rep Power: 20	Post some pics of your mesh and also give some more info about your problem like domain, boundary conditions, turbulence model, Re. No, sover settings etc

January 14, 2013, 22:09		#3
shlgzz New Member Wheeler Join Date: Jan 2013 Posts: 24 Rep Power: 13	There are 15 boundary layers near the airfoil and the height of first layer is 0.0003.Comapare with other papers,I choose the K-W SST turbulence model and default for sover settings,the operating Re is 100000.As to the mesh,the total mesh number is about 500000 although I choose the unstructed mesh.

January 14, 2013, 22:40		#4
cfd seeker Senior Member Join Date: Mar 2011 Location: Germany Posts: 552 Rep Power: 20	500000 cells for airfoil simulation! oops its too much. I have obtained good results for airfoils with just around 100000 cells. Anyhow mesh resolution near the wall is also good but this does not effect Cl. Turbulence model is also ok. Have you used 2nd order upwind for all equations? What about the domain size? How many chord lengths upstream, downstream and in other directions?

January 14, 2013, 23:24		#5
shlgzz New Member Wheeler Join Date: Jan 2013 Posts: 24 Rep Power: 13	It's terrible!Just 100000 cells.You are doing dynamic stall,either?I haven't tried 2nd upwind for all equations because I have tried under the steady conditions but no good result.The domain size is about 10 lengths upstream,20 lengths downstream and 20 lengths in width direction.

January 15, 2013, 01:17		#6
cfd seeker Senior Member Join Date: Mar 2011 Location: Germany Posts: 552 Rep Power: 20	Yes 100000 cells and the results were accurate with just 2-3 % error as compared toliterature results. Domain size is good enough. Use 2nd order upwind for all equations using the previuosly converged results as the initail conditions. BTW at which angle of attack you are doing your simulation?

January 15, 2013, 03:41		#8
cfd seeker Senior Member Join Date: Mar 2011 Location: Germany Posts: 552 Rep Power: 20	time step is very large as compared to the smallest cell size in your mesh. Use time step of 10^-6

January 15, 2013, 06:25		#9
shlgzz New Member Wheeler Join Date: Jan 2013 Posts: 24 Rep Power: 13	I would have a try.Thank U very much. Best regards.

January 15, 2013, 10:10		#10
shlgzz New Member Wheeler Join Date: Jan 2013 Posts: 24 Rep Power: 13	BTW,what did you set the convergence criteria as on the condition of oscillating airfoil?

January 15, 2013, 22:29		#12
shlgzz New Member Wheeler Join Date: Jan 2013 Posts: 24 Rep Power: 13	The relevant description is that the unsteady aerodynamic charcateristics of the NREL S809 airfoil undergoing sinusoidal pitch oscillations about the quarter chord point at various reduce frequencies,mean angles of attack and pitch oscillation amplitude of about 10 degrees at Re=1000000 were investigated through solutions of two demensional Navier-Stokes equations with transition corrected k-w turbulence model.