Forum

Important Notice for New User Registrations

To combat an increasing number of spam and bot registrations, we now manually approve all new user registrations. While this may cause a delay until your account is approved, this step is essential to ensure the quality and security of this forum.

To help us verify your registration as legitimate, please use a clear name as user name or an official email address (such as a work, university, or similar address). If you’re concerned that we may not recognize your registration as non-spam, feel free to email us at with a request to approve your username.

QBlade ForumQBlade: Simulation Setup and Evaluationsimulation step about update the …

Please or Register to create posts and topics.

simulation step about update the blade and wake

#1 · 22. April 2024, 09:53

Quote from robytian on 22. April 2024, 09:53
Hi david
I’ve recently been working on my lifting line free vortex wake code with reference to your PhD thesis, and I have a couple of queries for you:
1.bound circulation
(a)The initial blade bound circulation is 0, then the angle of attack is calculated for the blade without induced velocity, as well as the lift coefficient lookup table, and then the bound circulation is recalculated using the combined velocity without induced velocity, all the way up to convergence, and this completes the steps before the blade has even begun to rotate right?
(b)When the bound circulation converges, is it better to let the blades rotate to update the geometric information before releasing the first shed and trailing vortices? The trailing vortex at this point is the subtraction of the adjacent bound vortex, so what is the circulation of the shed vortex at this point? The next step only iterated to converge to a value that was subtracted from the initial setting of 0?
(c)Am I to understand that the subsequent steps are all, first the blades rotate, the vortex nodes move with first order Euler and all the vortex filaments contributing to the formation of the shed vortex and trailing vortex, and then the blade control points are solved for.
2.vortex core radius
(a)Are the chord lengths and twist angles at the control points averaged over the corresponding airfoils at the boundary points?
(b)Is the initial vortex core radius of an bound vortex the average of the corresponding chord lengths at neighboring foliation boundary points multiplied by 10% or some other factor?
(c)And how is the initial vortex core radius defined for the shed vortex and trailing vortex at the first time step again? Is the shed vortex the length of the chord corresponding to the control point? Is the trailing vortex the length of the string corresponding to the blade element boundary point?
3.There should be no shed vortices under linear or steady state incoming flow, and then simulations in qblade can see that the shed vortices exist because of the presence of relaxation factors?
I’m very sorry I have so many questions and thank you for answering them.
BR,
Roby

Hi david

I’ve recently been working on my lifting line free vortex wake code with reference to your PhD thesis, and I have a couple of queries for you:

1.bound circulation

(a)The initial blade bound circulation is 0, then the angle of attack is calculated for the blade without induced velocity, as well as the lift coefficient lookup table, and then the bound circulation is recalculated using the combined velocity without induced velocity, all the way up to convergence, and this completes the steps before the blade has even begun to rotate right?

(b)When the bound circulation converges, is it better to let the blades rotate to update the geometric information before releasing the first shed and trailing vortices? The trailing vortex at this point is the subtraction of the adjacent bound vortex, so what is the circulation of the shed vortex at this point? The next step only iterated to converge to a value that was subtracted from the initial setting of 0?

(c)Am I to understand that the subsequent steps are all, first the blades rotate, the vortex nodes move with first order Euler and all the vortex filaments contributing to the formation of the shed vortex and trailing vortex, and then the blade control points are solved for.

2.vortex core radius

(a)Are the chord lengths and twist angles at the control points averaged over the corresponding airfoils at the boundary points?

(b)Is the initial vortex core radius of an bound vortex the average of the corresponding chord lengths at neighboring foliation boundary points multiplied by 10% or some other factor?

(c)And how is the initial vortex core radius defined for the shed vortex and trailing vortex at the first time step again? Is the shed vortex the length of the chord corresponding to the control point? Is the trailing vortex the length of the string corresponding to the blade element boundary point?

3.There should be no shed vortices under linear or steady state incoming flow, and then simulations in qblade can see that the shed vortices exist because of the presence of relaxation factors?

I’m very sorry I have so many questions and thank you for answering them.

BR,

Roby

#2 · 22. April 2024, 21:45

Quote from David on 22. April 2024, 21:45
Hello Roby,
in general, many of the questions you are having dont have a unique answer. I would say that every implementation of the LLFVW uses different approaches to handle these things. At the same time, such detail regarding the implementation is almost never mentioned in publications or reports.
Now regarding your questions:
The initial blade bound circulation is 0, then the angle of attack is calculated for the blade without induced velocity, as well as the lift coefficient lookup table, and then the bound circulation is recalculated using the combined velocity without induced velocity, all the way up to convergence, and this completes the steps before the blade has even begun to rotate right?
At every timestep, the iteration that you describe is performed. At the beginning some circulation is assumed, based on this (and the wake filaments) induced velocities are calculated. From these the AoA results which is used (with airfoil data) to check if the initial circulation guess was correct. At the first timestep the initial guess can be zero, for successive timestep it could be the circulation from the previous timestep.
When the bound circulation converges, is it better to let the blades rotate to update the geometric information before releasing the first shed and trailing vortices? The trailing vortex at this point is the subtraction of the adjacent bound vortex, so what is the circulation of the shed vortex at this point? The next step only iterated to converge to a value that was subtracted from the initial setting of 0?
The shed vortex is the difference of the bound circulation from the current and the previous timestep. For the first timestep the “previous” circulation is zero. This in fact created the strong starting vortex in the LLFVW.
Am I to understand that the subsequent steps are all, first the blades rotate, the vortex nodes move with first order Euler and all the vortex filaments contributing to the formation of the shed vortex and trailing vortex, and then the blade control points are solved for.
See the image below. Circulation is solved for -> blade rotates -> (old) wake is advected -> (new) wake is generated to fill the gap between “old) wake and rotor.
Is the initial vortex core radius of an bound vortex the average of the corresponding chord lengths at neighboring foliation boundary points multiplied by 10% or some other factor?
The local mean chord from the neighboring aerodynamic panels is used for the trailing vortices.
And how is the initial vortex core radius defined for the shed vortex and trailing vortex at the first time step again? Is the shed vortex the length of the chord corresponding to the control point? Is the trailing vortex the length of the string corresponding to the blade element boundary point?
For the shed vortices the local chord of the shedding blade panel is used.
There should be no shed vortices under linear or steady state incoming flow, and then simulations in qblade can see that the shed vortices exist because of the presence of relaxation factors?
The shed vortices will vanish when the circulation over the blade doesnt change anymore over time. Since at the beginning of a simulation now wake exists, and the wake is then slowly created, during this transient operation shed vortices do exists. As soon as the wake is fully developed and the bound circulation doesnt change anymore over time the shed vortices will disappear.
Good luck with your implementation. And dont be afraid to try things out in the way that they make sense to you. There is no single answer to most of these questions. A lot depends on personal flavour.
Best,
David

Hello Roby,

in general, many of the questions you are having dont have a unique answer. I would say that every implementation of the LLFVW uses different approaches to handle these things. At the same time, such detail regarding the implementation is almost never mentioned in publications or reports.

Now regarding your questions:

The initial blade bound circulation is 0, then the angle of attack is calculated for the blade without induced velocity, as well as the lift coefficient lookup table, and then the bound circulation is recalculated using the combined velocity without induced velocity, all the way up to convergence, and this completes the steps before the blade has even begun to rotate right?

At every timestep, the iteration that you describe is performed. At the beginning some circulation is assumed, based on this (and the wake filaments) induced velocities are calculated. From these the AoA results which is used (with airfoil data) to check if the initial circulation guess was correct. At the first timestep the initial guess can be zero, for successive timestep it could be the circulation from the previous timestep.

When the bound circulation converges, is it better to let the blades rotate to update the geometric information before releasing the first shed and trailing vortices? The trailing vortex at this point is the subtraction of the adjacent bound vortex, so what is the circulation of the shed vortex at this point? The next step only iterated to converge to a value that was subtracted from the initial setting of 0?

The shed vortex is the difference of the bound circulation from the current and the previous timestep. For the first timestep the “previous” circulation is zero. This in fact created the strong starting vortex in the LLFVW.

Am I to understand that the subsequent steps are all, first the blades rotate, the vortex nodes move with first order Euler and all the vortex filaments contributing to the formation of the shed vortex and trailing vortex, and then the blade control points are solved for.

See the image below. Circulation is solved for -> blade rotates -> (old) wake is advected -> (new) wake is generated to fill the gap between “old) wake and rotor.

Is the initial vortex core radius of an bound vortex the average of the corresponding chord lengths at neighboring foliation boundary points multiplied by 10% or some other factor?

The local mean chord from the neighboring aerodynamic panels is used for the trailing vortices.

And how is the initial vortex core radius defined for the shed vortex and trailing vortex at the first time step again? Is the shed vortex the length of the chord corresponding to the control point? Is the trailing vortex the length of the string corresponding to the blade element boundary point?

For the shed vortices the local chord of the shedding blade panel is used.

There should be no shed vortices under linear or steady state incoming flow, and then simulations in qblade can see that the shed vortices exist because of the presence of relaxation factors?

The shed vortices will vanish when the circulation over the blade doesnt change anymore over time. Since at the beginning of a simulation now wake exists, and the wake is then slowly created, during this transient operation shed vortices do exists. As soon as the wake is fully developed and the bound circulation doesnt change anymore over time the shed vortices will disappear.

Good luck with your implementation. And dont be afraid to try things out in the way that they make sense to you. There is no single answer to most of these questions. A lot depends on personal flavour.

Best,

David

#3 · 23. April 2024, 09:12

Hi David

You are really nice, thank you very much for your help. There are very few people using the FVW method in China, and I am very interested in it, and I hope I can make some contribution in the future.

BR,

Roby

#4 · 10. May 2024, 14:12

Hi David

I would like to ask you again how you define vortex nodes, and bound vortex filaments, shed vortex filaments, and trailing vortex filaments.
The only vortex nodes involved in convection ? And then you form the shed vortex filaments and trailing vortex filaments by attaching them, and then giving them the circulation and vortex core radius, is that correct?

BR,

Roby

#5 · 10. May 2024, 18:18

Hello Roby,

the image below shows a good overview of these terms:

At every time step, the wake is convected away from the blade by updating the positions of the free wake nodes, during which a new row of wake nodes is created as the wake “detaches” from the blade.
Subsequently, the rotor is advanced by updating the positions of the bound vortex nodes.
A new trailing vortex filament is then created, filling the “gap”, between the bound vortex nodes and the newly formed row of free wake nodes.
Finally, shed vortex filaments are created by connecting this new row of wake nodes in the chordwise direction.

The circulation and core size is assigned according to the description in the documentation.

BR,

David

robytian has reacted to this post.

#6 · 15. May 2024, 08:02

Hi David

Is it sufficient to use the average of the chord lengths and twist angles corresponding to the two boundary points corresponding to the blade of the panel for the control points? Or do I need to triple spline and then interpolate if I place the control point in the middle of the blade.
Also, is it ok to use 5%-10% of the chord length corresponding to the control point for the initial bound vortex core radius, as well as for the shed vortex and trailing vortex core radius at the first time step? Thanks.

BR,

Roby

#7 · 15. May 2024, 12:05

Hello Roby,

I am taking just the average of panel chord length for the local mid-panel chord.

In the end, the % of chord length used for the intial core size is a free parameter and 5%-10% is just a suggestion that works quite well.

BR,

David

#8 · 17. May 2024, 03:40

Hi David

I had some more questions about what the criterion for the convergence of the circulation in the qblade looks like, I used one such method but some of the calculated circulation are negative, resulting in a positive induced velocity of the vortex filaments to the control point, do you have any idea what the reason for this is?

epsilon = 0.001

if abs(gamma_new – gamma_old)/(gamma_old+0.0000000001) > epsilon:

gamma_old = gamma_old + 0.05 * (gamma_new – gamma_old)

To prevent the initial circulation from being 0, I added a very small value to the denominator.

BR,

Roby

#9 · 17. May 2024, 11:57

Hi Roby,

I can’t tell why you are having such issues, but in general you get negative circulation values when you have a negative lift coefficient:

From Kutta-Joukowski: Γ = 0.5 Cl c V

BR,

David

#10 · 19. May 2024, 10:01

Hi David

Thank you very much for your answer, the previous problem was with the coordinate system and the direction of rotation of the turbine, which I have solved. Currently my problem is the convergence of the circulation at the initial time step, according to the toal velocity to solve the circulation, convergence verification, but never converge, I would like to ask you for the circulation convergence verification of this piece of how to cinsideration?

I use the method of ABEDI (Development of Vortex Filament Method for Aerodynamic Loads on Rotor Blades) ,like figure 1

BR,

Roby

Uploaded files:

You need to login to have access to uploads.