CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   Lift on the Hull (http://www.cfd-online.com/Forums/cfx/64806-lift-hull.html)

hashem_1064 May 25, 2009 08:34

Lift on the Hull
 
hi dear users;
modeling a planing hull in CFX, I 'v encountered a problem;
I want to evaluate Lift on the high speed planing hull. Domain and modeling are as fallow:
1.domain is a well-dimensioned cubic
2. i modeled just underwater portion as i want to calculate Lift
3. the domain is just contained water with free slip wall boundary condition for top wall
4. portions of domain and hull which had been exposed to the air have been omitted since I want to modeling just underwater portion.
5. V= 5 m/s
6. mesh dependency study is done and near the hull mesh is fine enough.

but the problem is here: given lift by CFX is about 10 times as much as expected value! I don t know why?!!! I m baffled what I have to do.
I would be happy to get your helps.
tnx

ghorrocks May 25, 2009 20:28

Hi,

Have you modelled this as single phase? How have you included free surface effects?

Glenn Horrocks

hashem_1064 May 27, 2009 06:31

yes, it is. since I just want to calculate the Lift it is applicable and i saw some papers related to it. but something amazing that I find it just today is that problem is very sensitive to input and output boundary conditions. I m not sure about BCs which I have chosen. can you help me?

ghorrocks May 27, 2009 08:24

Hi,

The lift on a planing body will be very dependent on the wave shape it generates. I have seen somebody (an America's Cup Yacht designer) use a single phase approach to model a boat but only after he got a full 3D scan of the wave profile the boat generates, and that was used as a free-slip boundary on the top of the domain.

If you don't have a detailed 3D map of the wave profile you will have to do a 2-phase free surface model. This will take a long time to run, they are very slow.

Glenn Horrocks

hashem_1064 May 27, 2009 09:12

thank you for you answer Glenn
I modeled it as to-phase problem before and results were awful, there was a big deviation. I don t know why, but single phase modeling led to more reasonable result. the point is here: I got a wide range of diverse results depending on inlet and outlet BCs. For the case which both of them are Cartesian Velocity results are good. I don t know how much this BC is reliable?! In CFX help has been mentioned that the most robust BC is Velocity/Mass Flow for inlet and Static Pressure for outlet. When I implied this case, I found that for small changes in Outlet Static Pressure, lift would change dramatically. i don t know how is this Robust! Also I define an Expression for outlet hydrostatic pressure, but results showed the redirection in streamlines toward the less pressure level (near the moving wall).

Hashem

hashem_1064 May 27, 2009 09:14

thank you for you answer Glenn
I modeled it as to-phase problem before and results were awful, there was a big deviation. I don t know why, but single phase modeling led to more reasonable result. the point is here: I got a wide range of diverse results depending on inlet and outlet BCs. For the case which both of them are Cartesian Velocity results are good. I don t know how much this BC is reliable?! In CFX help has been mentioned that the most robust BC is Velocity/Mass Flow for inlet and Static Pressure for outlet. When I implied this case, I found that for small changes in Outlet Static Pressure, lift would change dramatically. i don t know how is this Robust! Also I define an Expression for outlet hydrostatic pressure, but results showed the redirection in streamlines toward the less pressure level (near the moving wall).

Hashem

ghorrocks May 27, 2009 19:45

Hi,

Two phase models are much more complicated and will take some time to get right. I think you will have to stick with it and investigate this further.

What about the pitch and trim of the boat. Do you know the exact position and height of the hull in the water, or does the simulation need to find this?

Does your single phase simulation assume the surface of the water is flat or do you have a wave pattern on it?

Glenn Horrocks

hashem_1064 May 27, 2009 22:07

Hi Glenn,
in fact the problem is to calculate lift force and moment which produces by trim tab. in this case there is no need to simulate dynamics of the boat.
Free surface is just a flat plane. As I told you I examined two phase modeling, even with structured mesh, but as velocity is somehow high and water spray is predominant, i m not sure about ansys cfx to get a precise answer for free surface modeling.

Hashem

ghorrocks May 27, 2009 23:21

Hi,

Then can't you just do a model of the trim tabs in isolation? That is no hull and no free surface - if you had explained this in the beginning this conversation would have been quite different.

To answer your original question - If you model the trim tabs in isolation then a simple velocity inlet and pressure outlet should work fine. Changing the outlet pressure in this case would make no difference to the result, as all pressures are relative to the outlet in this case anyway.

Glenn Horrocks

hashem_1064 May 28, 2009 13:38

Hi
Firstly I m sorry if i make misconception.
No Glenn, there is some theoretical formulation for trim tab which is extracted considering hull effects.
I don t know what exception is your case included rather than mine which causes results to be independence from outlet pressure? The only additional issue of my modeling rather than yours was the presence of underwater portion of the hull. How it could makes such an extraordinary effect?

Hashem

ghorrocks May 28, 2009 18:23

Hi,

There are many ways the outlet pressure can affect things. If you are using a material model with a dependance on pressure (compressible for instance) that can cause it.

But the most likely thing is a numerical error. Have you checked convergence, mesh and time step sensitivities and ensured your numerical model is accurate? If not then you just have a random number generator so comparisons are pointless.

Glenn Horrocks


All times are GMT -4. The time now is 20:09.