Udf for a transient analisys of a VAWT

ermes · June 25, 2010, 04:33

Hi,
I' working with a vertical axis wind turbine; i made static analisys and i've used MRF method but i'm not satisfied of my result.
Now i'm trying to write an udf and use the six dof solver to do this:
-define the axis of rotation (y) and the mass property of the model
-calculate at every substep of a transient analisys:
-moment generate by wind
-increment of velocity due to the difference between the moment generated by wind and that of the generator.

Can everybody help me?

Thanks.

sohail_27 · June 27, 2010, 19:50

Hi,
I am trying to run simulation (2D) for 3 bladed NACA 0012 airfoil in vertical motion with respect to central axis using gambit and fluent. I know how to create mesh for single airfoil and run simulation on it, but i don't know how to create mesh for airfoil and to rotate it in fluent. I would be glad if anyone could help me out.
Here is a youtube video of what i actually want to do. Please help me out guys, its part of my thesis.
http://www.youtube.com/watch?v=1eyEtzdHDEE

Regards.

sohail_27 · July 13, 2010, 17:07

Hi Davesmith, i wish i could help you but i dont have mesh file for a cylinder, i am dealing purely with airfoils

enry · July 13, 2010, 17:27

Hi all;
you have to create a sliding interface... have you created it?

sohail_27 · July 13, 2010, 17:30

Hi,
i have created the sliding meshes, and also got to see meshes rotatting, but at the interface region, i see a huge numerical noise (vortex shedding) how can i get rid of that???

enry · July 13, 2010, 17:37

what type of solver do you use? the best is K-epsilon realizable, K-W sst, or RSM.
However, near sliding interface the solution is not so good, so you have to put your interface far from the vawt. I use a box of 60xVawt Radius, and interface have a diameter of 10xVawt radius , and it work well. Try to re-mesh and use more cells near interface.

enry · July 13, 2010, 17:39

PS. How many revolution do you set to calculate your coefficient?
Keep in mind that solution will converge after 15-20 vawt revolution.

enry · July 13, 2010, 17:44

Sorry, i read that you are doing a static analisys. What type of coefficient would you like to obtain from static analysis? The static torque? To obtain the correct power coefficient you have to do a unsteady simulation...

sohail_27 · July 13, 2010, 18:09

hi, i kept my interface at a distance of about 5*chord, @ this i observed error from inteface but was getting disappear after some time, should i keep them at more distance? i ran simulation with, K-e standard, RNG; K-w standard, SST, but i dint observed vortex shedding from blade, even though they appear they disappear before blade enters downwind condition, using S-A model, there was some shedding but even that disappear quickly? i am looking for a case where i can understand where the vortex shedding is maximum and where is the maximum interaction of the blade and shed vortices, which model will work good for that?

enry · July 13, 2010, 18:27

which type of error do u observe?

Every model you mentioned should be good...maybe you have to compare some results with experimental data...however, you should observe vortex shadding at the blades...

sohail_27 · July 13, 2010, 18:51

yes i did observe some vortex shedding @ blade, but it got disappear as soon as blade moved to the next position, how can i account for blade vortex interaction then?

enry · July 14, 2010, 02:55

mmm...it's look like an update error... but, does fluent report any errors? Do you set properly b.c.?

sohail_27 · July 14, 2010, 13:35

yes, i gave proper BC, fluent dint gave any error

enry · July 14, 2010, 15:32

Hi sohail_27, what do u mean saying : "i did observe some vortex shedding @ blade, but it got disappear as soon as blade moved to the next position" . The next position is the delta-angle of rotation? Maybe i don't undestand very well your problem, because you posted first: "I am trying to run simulation (2D) for 3 bladed NACA 0012 airfoil in vertical motion with respect to central axis using gambit and fluent". What do u mean saying: in vertical motion? How can you move the vawt vertically if the model is 2d?

sohail_27 · July 14, 2010, 16:12

Hi enry,
I am doing 2-d simulation with NACA 0012 airfoil in circular motion. I have a circle( stationary wall) at center which would be the top view of vertical shaft. I started my simulation with single airfoil, after getting good results from single airfoil ill go for three airfoil system. I need to study about the blade and vortex interaction to minimize them.
i created the mesh file for it, but during simulation i observe numerical noise( in form of vortex shedding) from the interface region. How should i tackle this problem?

June 25, 2010, 04:33	Udf for a transient analisys of a VAWT	#1
ermes New Member Join Date: Feb 2010 Posts: 5 Rep Power: 16	Hi, I' working with a vertical axis wind turbine; i made static analisys and i've used MRF method but i'm not satisfied of my result. Now i'm trying to write an udf and use the six dof solver to do this: -define the axis of rotation (y) and the mass property of the model -calculate at every substep of a transient analisys: -moment generate by wind -increment of velocity due to the difference between the moment generated by wind and that of the generator. Can everybody help me? Thanks.

June 27, 2010, 19:50	need help for VAWT simulation	#2
sohail_27 New Member SOhail Join Date: Jun 2010 Posts: 16 Rep Power: 15	Hi, I am trying to run simulation (2D) for 3 bladed NACA 0012 airfoil in vertical motion with respect to central axis using gambit and fluent. I know how to create mesh for single airfoil and run simulation on it, but i don't know how to create mesh for airfoil and to rotate it in fluent. I would be glad if anyone could help me out. Here is a youtube video of what i actually want to do. Please help me out guys, its part of my thesis. http://www.youtube.com/watch?v=1eyEtzdHDEE Regards.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Parallelizing Transient UDF	eliasmontoya	FLUENT	2	October 20, 2009 09:47
Multi step transient UDF velocity profile problem	William177	FLUENT	1	February 3, 2008 06:47
spatial and transient B.C. using UDF	Park	FLUENT	0	April 28, 2005 21:32
initial conditions with transient UDF	Antoine Pages	FLUENT	4	September 19, 2004 11:51
UDF for transient boundary condition?	Shankar	FLUENT	1	November 22, 2003 08:10

July 13, 2010, 17:07		#3
sohail_27 New Member SOhail Join Date: Jun 2010 Posts: 16 Rep Power: 15	Hi Davesmith, i wish i could help you but i dont have mesh file for a cylinder, i am dealing purely with airfoils

July 13, 2010, 17:27		#4
enry Senior Member Herman Join Date: Nov 2009 Posts: 122 Rep Power: 16	Hi all; you have to create a sliding interface... have you created it?

July 13, 2010, 17:30		#5
sohail_27 New Member SOhail Join Date: Jun 2010 Posts: 16 Rep Power: 15	Hi, i have created the sliding meshes, and also got to see meshes rotatting, but at the interface region, i see a huge numerical noise (vortex shedding) how can i get rid of that???

July 13, 2010, 17:37		#6
enry Senior Member Herman Join Date: Nov 2009 Posts: 122 Rep Power: 16	what type of solver do you use? the best is K-epsilon realizable, K-W sst, or RSM. However, near sliding interface the solution is not so good, so you have to put your interface far from the vawt. I use a box of 60xVawt Radius, and interface have a diameter of 10xVawt radius , and it work well. Try to re-mesh and use more cells near interface.

July 13, 2010, 17:39		#7
enry Senior Member Herman Join Date: Nov 2009 Posts: 122 Rep Power: 16	PS. How many revolution do you set to calculate your coefficient? Keep in mind that solution will converge after 15-20 vawt revolution.

July 13, 2010, 17:44		#8
enry Senior Member Herman Join Date: Nov 2009 Posts: 122 Rep Power: 16	Sorry, i read that you are doing a static analisys. What type of coefficient would you like to obtain from static analysis? The static torque? To obtain the correct power coefficient you have to do a unsteady simulation...

July 13, 2010, 18:09		#9
sohail_27 New Member SOhail Join Date: Jun 2010 Posts: 16 Rep Power: 15	hi, i kept my interface at a distance of about 5*chord, @ this i observed error from inteface but was getting disappear after some time, should i keep them at more distance? i ran simulation with, K-e standard, RNG; K-w standard, SST, but i dint observed vortex shedding from blade, even though they appear they disappear before blade enters downwind condition, using S-A model, there was some shedding but even that disappear quickly? i am looking for a case where i can understand where the vortex shedding is maximum and where is the maximum interaction of the blade and shed vortices, which model will work good for that?

July 13, 2010, 18:27		#10
enry Senior Member Herman Join Date: Nov 2009 Posts: 122 Rep Power: 16	which type of error do u observe? Every model you mentioned should be good...maybe you have to compare some results with experimental data...however, you should observe vortex shadding at the blades...

July 13, 2010, 18:51		#11
sohail_27 New Member SOhail Join Date: Jun 2010 Posts: 16 Rep Power: 15	yes i did observe some vortex shedding @ blade, but it got disappear as soon as blade moved to the next position, how can i account for blade vortex interaction then?

July 14, 2010, 02:55		#12
enry Senior Member Herman Join Date: Nov 2009 Posts: 122 Rep Power: 16	mmm...it's look like an update error... but, does fluent report any errors? Do you set properly b.c.?

July 14, 2010, 13:35		#13
sohail_27 New Member SOhail Join Date: Jun 2010 Posts: 16 Rep Power: 15	yes, i gave proper BC, fluent dint gave any error

July 14, 2010, 15:32		#14
enry Senior Member Herman Join Date: Nov 2009 Posts: 122 Rep Power: 16	Hi sohail_27, what do u mean saying : "i did observe some vortex shedding @ blade, but it got disappear as soon as blade moved to the next position" . The next position is the delta-angle of rotation? Maybe i don't undestand very well your problem, because you posted first: "I am trying to run simulation (2D) for 3 bladed NACA 0012 airfoil in vertical motion with respect to central axis using gambit and fluent". What do u mean saying: in vertical motion? How can you move the vawt vertically if the model is 2d?

July 14, 2010, 16:12		#15
sohail_27 New Member SOhail Join Date: Jun 2010 Posts: 16 Rep Power: 15	Hi enry, I am doing 2-d simulation with NACA 0012 airfoil in circular motion. I have a circle( stationary wall) at center which would be the top view of vertical shaft. I started my simulation with single airfoil, after getting good results from single airfoil ill go for three airfoil system. I need to study about the blade and vortex interaction to minimize them. i created the mesh file for it, but during simulation i observe numerical noise( in form of vortex shedding) from the interface region. How should i tackle this problem?