[Adrian Portillo-Juan]

Hi:

I am currently working with the self-propulsion tutorial at Star-CCM+ and I was wondering why one of the inputs for the virtual disk is the rotation rate instead of the ship velocity (which is obtained), while in self-propulsion experiments it is done the opposite way (the input is the ship/towing carriage velocity and the output is the propeller rpm.


Thank you so much in advance

[Aeronautics El. K.]

The virtual disk model uses the propeller's open water curves. In order to do that, it has to somehow know the advance ratio. To do that, it takes the speed of advance from the plane you have defined as the inflow to the virtual disk.
Now the only thing missing is the rotation rate that you give it.
I'm guessing that's the sequence but I may be wrong.

[a.gmira]

Good afternoon Lefteris,

I am having the same doubt. I need to know which rotation rate of the virtual disk gives me the velocity of the ship that I want to study (same concept as in experiments). However I dont find out how to impose the velocity of the ship as an input and extract the rotation rate of the propeller as an output. (I am using the same set up as in the tutorial, with Body Force Propeller Method)

From what I have read, the Operation Point Input (in the parameters of the Virtual Disk) has only three options: Rotation Rate, Thrust or Torque. Do you know if it is possible to impose the velocity of the ship and extract the rotation rate as an output? or should I just follow an iterative process in which I impose different rotation rates until I reach the velocity that I want?

Thank you very much in advance.

A.gmira

[Aeronautics El. K.]

This is a typical self-propulsion simulation, if I understand you correctly.
What I do is to set the virtual disk to work on thrust instead of rotation rate. Then, I set the thrust equal to the resistance.
You can have a report which gives you the rotation rate of the virtual disk.

[a.gmira]

Good morning Lefteris,

Exactly. That is what I thought. However, since there is a thrust deduction fraction, the thrust is no longer equal to the resistance, it is higher, correct?

So here the question is if the software is using the thrust that we impose as the operating point of the propeller in the performance curves or, on the other hand, if it is already considering the thrust deduction fraction so the thrust that we impose is actually equal to the resistance.

Thank you again for your replies.

A.gmira

[Aeronautics El. K.]

Originally Posted by a.gmira

Good morning Lefteris,

Exactly. That is what I thought. However, since there is a thrust deduction fraction, the thrust is no longer equal to the resistance, it is higher, correct?

So here the question is if the software is using the thrust that we impose as the operating point of the propeller in the performance curves or, on the other hand, if it is already considering the thrust deduction fraction so the thrust that we impose is actually equal to the resistance.

Thank you again for your replies.

A.gmira

If you do it right, it sorts itself out.

[a.gmira]

Hi Lefteris,

I am doing it by iteration, imposing some revolutions and taking as output the velocity of the ship. This proceedure seems to be the only one possible in Starccm+ since the resistance of the ship and the resistance of the ship+propeller is not the same (and we do not know the second one), so we can not put thrust equal to the resistance of the ship.

Are you talking about selfpropulsion in ships?

A.gmira

[Aeronautics El. K.]

Originally Posted by a.gmira

Hi Lefteris,

I am doing it by iteration, imposing some revolutions and taking as output the velocity of the ship. This proceedure seems to be the only one possible in Starccm+ since the resistance of the ship and the resistance of the ship+propeller is not the same (and we do not know the second one), so we can not put thrust equal to the resistance of the ship.

Are you talking about selfpropulsion in ships?

A.gmira

Hello!

Yes, I'm talking about self-propulsion.

When I use a rigid body prop, then yes, I apply the guessed rpm (although there are controllers that can be implemented).

When I use the virtual disk, I use the thrust option.
Thrust and resistance don't have fixed values, they are linked so, the thrust is updated on every time step based on the resistance until the values of both converge.
I have done the validation like that multiple times, so there definitely more than one ways to skin the cat.

[a.gmira]

Thank for your reply!

But which value do you impose in the thrust? You impose a fixed value, dont you?

[a.gmira]

Hi again Lefteris!

Do you have an static VOF Wave as in the tutorial? or you are imposing your velocity to that wave? I think here is the issue.

Thank you in advance!

A.gmira

[Aeronautics El. K.]

Originally Posted by a.gmira

Thank for your reply!

But which value do you impose in the thrust? You impose a fixed value, dont you?

Nope
That's what reports are for. You can use them in various places, including to define the thrust of the VD.

Originally Posted by a.gmira

Hi again Lefteris!

Do you have an static VOF Wave as in the tutorial? or you are imposing your velocity to that wave? I think here is the issue.

Thank you in advance!

A.gmira

I use the flat vof wave 99% of the time, with the speed set to the speed of the vessel and no current/wind

[a.gmira]

Thank you very much Lefteris! I understand. I have run some simulations with your advice and the results seems to be reasonable!

On the other hand, I am simulating the case with different Inflow Plane Offsets (in the help of Starccm+ they advise to put an Offset between 1% and 10% of the propeller diameter). However, the results change from one Offset to another in a notable way. Could you tell me some advice related with this issue?

Also, do you recommend to implement the Induced Velocity Correction? In the tutorial they do not do it, but I have done it and again, the results are very different.

Finally, related to the VOF wave model, How do you set the velocity of the flat wave if it is not in the current node? I set the velocity in that node. And what do you use the 1% of the time when you are not using the flat VOF wave?

Feel free to answer any of these questions and thank you again for your time!

A.gmira

[Aeronautics El. K.]

Originally Posted by a.gmira

Thank you very much Lefteris! I understand. I have run some simulations with your advice and the results seems to be reasonable!

On the other hand, I am simulating the case with different Inflow Plane Offsets (in the help of Starccm+ they advise to put an Offset between 1% and 10% of the propeller diameter). However, the results change from one Offset to another in a notable way. Could you tell me some advice related with this issue?

Also, do you recommend to implement the Induced Velocity Correction? In the tutorial they do not do it, but I have done it and again, the results are very different.

Finally, related to the VOF wave model, How do you set the velocity of the flat wave if it is not in the current node? I set the velocity in that node. And what do you use the 1% of the time when you are not using the flat VOF wave?

Feel free to answer any of these questions and thank you again for your time!

A.gmira

Good luck with the inflow plane.
I had a study some time ago, it does change of course because the inflow sees different velocity distribution.
I don't know what the right answer is. I usually go for 10% of the diameter.

I stopped using the induced velocity correction based on my validations.

If you have a current you add or subtract the current velocity. Otherwise it's just the velocity of the ship I guess.

That 1% of the time is when I've made a mistake
Or just the requirement is different so, whatever the project demands really.

[melvinardan]

In the context of self-propulsion simulations, it is common to use the propeller rpm as the input because it allows for a more direct control of the propeller's behavior, which can be more complex than simply adjusting the ship velocity.

The propeller rpm determines the rotational speed of the propeller blades, which in turn affects the thrust generated by the propeller. By controlling the propeller rpm directly, it is possible to study the effects of different propeller designs and operating conditions on the performance of the propulsion system.

On the other hand, in the case of towing carriage experiments, the ship velocity is typically used as the input because it is more directly controlled by the towing carriage system. The propeller rpm is then measured as an output to evaluate the performance of the propulsion system.

In the virtual disk model used in Star-CCM+, the rotation rate is used as an input because it represents the speed of the disk that is generating the fluid flow around the ship. The disk is used to model the effect of a free stream on the ship's motion, and the rotation rate is used to control the magnitude of the flow velocity. This allows for a more direct control of the fluid flow around the ship, which can be important for accurately simulating the self-propulsion behavior of the ship.

[Aeronautics El. K.]

Originally Posted by melvinardan

In the context of self-propulsion simulations, it is common to use the propeller rpm as the input

When using a sliding mesh and modelling the actual propeller or when using a virtual disk?

Originally Posted by melvinardan

By controlling the propeller rpm directly, it is possible to study the effects of different propeller designs and operating conditions on the performance of the propulsion system.

Again, assuming you're talking about a rigid body propeller, right?

Originally Posted by melvinardan

In the virtual disk model used in Star-CCM+, the rotation rate is used as an input because it represents the speed of the disk that is generating the fluid flow around the ship. The disk is used to model the effect of a free stream on the ship's motion, and the rotation rate is used to control the magnitude of the flow velocity.

Do you mean in a simulation where the ship is allowed to surge (1-DoF)? Because if the ship is fixed (0-DoF) the virtual disk behind the ship would do virtually nothing to the freestreem ahead of the ship.

Originally Posted by melvinardan

This allows for a more direct control of the fluid flow around the ship, which can be important for accurately simulating the self-propulsion behavior of the ship.

More direct control as opposed to what? How does the virtual disk affect the flow around the ship when the ship is not allowed to surge?
"Accurately simulating the self-propulsion", I'm assuming you can show us a reference or a validation result of that (with the virtual disk model), showing the differences between using the rpm and the thrust as input?

[a.gmira]

Hi Lefteris,

Thank you for your advice, it has been very useful! The Induced velocity correction seems to give worse results.

However, I have run the case with the thrust as an input equal to the resistamce, taking into account your recommendations, and with a flat VOF wave with velocity the one I had as an output from the previous simulation (in which the revolutions where the input) but it does not give me excatly the same revolutions that I put as an input in the first one. Let me know if this is clear to you... The difference is little (the input revolutions are 2300 rpm while the output revolutions are 2308 rpm). Do you know why could be this or if it makes sense?

Also, could you give me any recommenadation for Realease and Ramp Time for the DFBI? The results also change a bit depending of this factor.

Finally, ¿do you fix the motion in X? I have been doing it by letting free the motion in X, however I dont see much differences in results between both procedures (the ship velocity in X when it has converged is almost cero, as it is reasonable). ¿Is there a better procedure?

Thank you very much!

A.gmira

[Aeronautics El. K.]

Originally Posted by a.gmira

Hi Lefteris,

Thank you for your advice, it has been very useful!

Glad to have helped! Like and subscribe Joking

Originally Posted by a.gmira

The Induced velocity correction seems to give worse results.

I'm not surprised.

Originally Posted by a.gmira

However, I have run the case with the thrust as an input equal to the resistamce, taking into account your recommendations, and with a flat VOF wave with velocity the one I had as an output from the previous simulation (in which the revolutions where the input) but it does not give me excatly the same revolutions that I put as an input in the first one. Let me know if this is clear to you... The difference is little (the input revolutions are 2300 rpm while the output revolutions are 2308 rpm). Do you know why could be this or if it makes sense?

That's 0.35%. Probably less than the error of your mesh even.

Originally Posted by a.gmira

Also, could you give me any recommenadation for Realease and Ramp Time for the DFBI? The results also change a bit depending of this factor.

I think I set this up as multiple of the residence time but I don't think it matters a lot. As long as it converges and the motions are constant in the end.
Are you using equilibrium or free motion?

Originally Posted by a.gmira

Finally, ¿do you fix the motion in X? I have been doing it by letting free the motion in X, however I dont see much differences in results between both procedures (the ship velocity in X when it has converged is almost cero, as it is reasonable). ¿Is there a better procedure?

I only allow 2 degrees of freedom (pitch, heave) but I suppose it depends on what you're after. And I don't use the free motion, I think for a virtual disk case the equilibrium is good enough.

[a.gmira]

Hi Lefteris,

I have been trying to fix some stuff of the my selfpropulsion set up (Virtual disk with Body Force Propeller method) but the wake fraction that I calculate is really small because it calculates an advance ratio of the propeller higher than the one I calculate (V/nD).

Do you know if the problem could be the Inflow Plane Offset? I have read in the guide that this offset is set in the negative-Z axis direction. Since my positive-Z axis points forward, should I put a negative offset? However, I have tried this and the advance ratio increments a bit...

Am I doing something wrong?

Thank you very much

A.gmira