[butterflymuzzy]

Hello Glenn again

I have been trying many things in the new 18.2 version but ı could not handle with convergenge problem of the drag force and mesh folding in the new version of cfx. ı want to show some picture for the trim of the ship in the ex and new versions of the ansys. in the ex version the trim converge in 1500 iterations but in the new version i could not handle.

You said that you can change the smoothing options of the mesh in the set up.

I used both of 2 (increase near boundaries and increase near small volumes) with the model exponent from 1 to 10. bu it didnt work.

can you give me any advice about mesh smoothing options ?

here are my pictures, you can see from the starting of itearion in the 16.2 after 50-60 iterations the trim started to change from -0.02 radian to -0.002 radian. but in the 18.2 version the trim started to change from -0.03 to +0.05. then after 160 itearions the itearion stopped.

Best.

note:sorry ı could not upload the photos

[ghorrocks]

Make sure you read the release notes to see what changes have been made to new versions of CFX. That will hopefully help you find what changed.

But looking back over this thread, I wonder whether your approach is the best one. Rather than model the body as a rigid body which opens up issues of moving mesh and stabilising a rigid body - how about you do a series of fixed mesh simulations at various heave, yaw, pitch and immersion depths. These should be straight forward simulations (but getting them accurate is still challenging), and will give you the forces acting on the vessel versus pitch, yaw and immersion depth. To find the steady state condition you then find where the lift force equals the mass of the object. You can also do transient simulations with a simple ODE solver.

Wouldn't this series of simple fixed mesh simulations be a better method to start with? Or is there some reason this approach is not appropriate?

[butterflymuzzy]

You are true in your side of view Glenn. But from my naval architect point of view ı did this comparision in cfd.

I had the towing tank test results of dtmb 5415 ship model with trim-sinkage and the resistance.

If I use only the fixed mesh there is no problem in every version. but for example if the experiment test resistance result is 110 Newton, I will find 90 Newton with fixed mesh.

On the other side if i use moving mesh ı will find 112 Newton. Becasue the dynamic sinkage's added resistance is coming. And the result is getting better.

And also if ı checked the trim and sinkage results with the experiment everything was ok.

And also i m using a different type of marine cfd code which is called "numeca fine marine". This program is only for marine cfd. everything is reday to use in the program. the program automaticly sets everythin (domain geometry, air water vof, water line margin line, longitudionally bouyancy point and that points gyration radiuses and mass moments etc). When i compare the results with numeca fine marine and ansys 16.2 everything is same and ok

How can i send this situation to any ansys customer service for fixing this problem in the new versions Glenn ?

Best.

Originally Posted by ghorrocks

Make sure you read the release notes to see what changes have been made to new versions of CFX. That will hopefully help you find what changed.

But looking back over this thread, I wonder whether your approach is the best one. Rather than model the body as a rigid body which opens up issues of moving mesh and stabilising a rigid body - how about you do a series of fixed mesh simulations at various heave, yaw, pitch and immersion depths. These should be straight forward simulations (but getting them accurate is still challenging), and will give you the forces acting on the vessel versus pitch, yaw and immersion depth. To find the steady state condition you then find where the lift force equals the mass of the object. You can also do transient simulations with a simple ODE solver.

Wouldn't this series of simple fixed mesh simulations be a better method to start with? Or is there some reason this approach is not appropriate?

[ghorrocks]

If you run a fixed mesh with the vessel at final sink and trim that should give you an accurate result, shouldn't it?

If you have a ANSYS TECS agreement which includes support then contact them according to that agreement. If you don't have TECS ANSYS is not going to listen to you much.

[Rizal Zul]

Hello everybody, I am Rizal,

now i am doing my thesis project by using CFD ansys 16 version. i intend to generate free surface simulation to predict ship resistance. the boundary condition in tutorial in term "free surface over bump" does not show good result. the simulation is always not finish. any body can help for explaining boundary condition used in free surface ship simulation in CFX? and may be also showing physical model that can be adopt?

Thank you

Regard*
Rizal

[nguyenvanquyetvtt]

Hi guys!
My name is Quyet, from Vietnam.
This is old post, but I have the same problem.
I am newbie in Ansys Fluent. I am calculating resistance for bulk ship 34K.
I have problem with free surface, my result is not convergent.
I attack picture in my post.Untitled.jpg
Could you explain for me about my problem? and give me some suggests to resolve this problem.
Thanks!

[urosgrivc]

For a good result, I have been using rigid body motion
with mesh deformation
With this setup near-wall region results are great and the simulation converges nicely well depends on what you are doing.
With this method, everything is taken into consideration as you need to define inertia in all directions and mass. In my case, I have fixed so the boat can turn and move up and down but not forward and backward so it stays in the domain.
It is important to be careful about initial conditions and it is better to hold the ship in place by expressions and than let it go when the flow is a bit converged, And even than it is best to use transient analysis to bring the simulation to steady state or use a physical timescale for steady state. I prefer transient one as I can see in post if there is something wrong with the simulation.

For a realy good and simple convergence use symetry if posible
and if you have problems vith fluent post them on that forum

[sriram9127]

It is a multiphase problem(air and water).
Initial velocity zero for both phase.
I want to run a simulation for 1m/s velocity of both phases and view wave pattern.

Please find the attachment. The residuals are converged, but I can't find the entire wave pattern. Give me some suggestions.

[urosgrivc]

This looks like a bullet, or is it a ship? or is it a crashing plane or...??

Damn it, write posts so people can understand, cfd is complex, or no help can be provided.

[sriram9127]

Originally Posted by urosgrivc

This looks like a bullet, or is it a ship? or is it a crashing plane or...??

Damn it, write posts so people can understand, cfd is complex, or no help can be provided.

It's a boat...if YOU can't understand...just ask.

I have attached the physics criteria below.

Feel free to ask for details.

[urosgrivc]

I actually thought it was a boat, but that is far from being enough to help you.

It is not I who needs help here, I am the one that would like to help.
I can do a simulation like this in minutes,
but am not able to provide any help based on the data provided.

And is this somehow linked to CFX or something else?

[sriram9127]

Originally Posted by urosgrivc

I actually thought it was a boat, but that is far from being enough to help you.

It is not I who needs help here, I am the one that would like to help.
I can do a simulation like this in minutes,
but am not able to provide any help based on the data provided.

And is this somehow linked to CFX or something else?

Exactly.....In that case, ask for the details rather than Unnecessary talk. Yes, I am the one who needs help...that's what I told in the first place itself. Just keep it closed to the simulation alone. Give me some suggestions on solving the problem alone. What more details are required?

I can provide.

[urosgrivc]

well, a lot more is indeed needed...

Do you already know the pitch at which the boat is angeled, or do you want to get this from the simulation (body movement reacting on fluid forces and gravitational acceleration)?
Do you know exactly how deep the boat sinks in the water?
Are there any additional jets that propel the boat further (is it necessary to include them in the simulation or not)?
and a lot more...

these are only a few data which decide if you need a simple fixed mesh system or immersed solid method with mesh deformation.
From the second method, all data (pitch, depth, drag) are obtained at once but is harder to set up.

[sriram9127]

Originally Posted by urosgrivc

well, a lot more is indeed needed...

Do you already know the pitch at which the boat is angeled, or do you want to get this from the simulation (body movement reacting on fluid forces and gravitational acceleration)?
Do you know exactly how deep the boat sinks in the water?
Are there any additional jets that propel the boat further (is it necessary to include them in the simulation or not)?
and a lot more...

these are only a few data which decide if you need a simple fixed mesh system or immersed solid method with mesh deformation.
From the second method, all data (pitch, depth, drag) are obtained at once but is harder to set up.

Thank You.
I want to get Pressure and Shear force value from this simulation. The boat is static only, with a draft of 10 cm from the waterline and remaining of the ship is above the water surface i.e 30cm.
If the water travels with a speed of 1m/s, what will be the total resistance force i.e sum of pressure and shear.
I have created an isosurface with a value of 0.5.
I couldn't visualize the wave pattern.
I am running as a transient solver.

What else I have to check in for.

[urosgrivc]

have you run the simulation for enough time for the flow to fully develop and stabilize?

[sriram9127]

Yes, the flow has stabilized, I am monitoring the force value and It shows and constant value. Also, the residuals have converged up to the order of 1e-07.

This is the problem,
1.Multiphase- air and water
2. Boundaries gave- Velocity inlet, pressure outlet, wall and symmetry plane.
3.Physics
Unsteady and K-Epsilon Turbulence model, other details are attached below.
4. The entire volume is half filled with water and the other half is filled with air.
Conditions I Doubt,
1. I have given initial velocity as 1m/s.
2. The volume fraction of Inlet and Outlet boundaries with air and water, which is the actual condition too.
3. Inlet velocity 1m/s.
4. Time step 0.0005 sec.
5. Isovalue of isosurface 0.5.

Are these parameters right???
The problem is running and the solution is getting stabilized, I am not convinced with the wave pattern created.

[ghorrocks]

Please attach your output file an image of your mesh and an image of your geometry. Also attach an image of the results you are getting.

Please attach them directly onto the forum and do not use third party upload providers.

[sriram9127]

Originally Posted by ghorrocks

Please attach your output file an image of your mesh and an image of your geometry. Also, attach an image of the results you are getting.

Please attach them directly onto the forum and do not use third party upload providers.

I have provided the images below.

[ghorrocks]

You appear to be using Fluent. This is the CFX forum. Try the Fluent forum.

[sriram9127]

Originally Posted by ghorrocks

You appear to be using Fluent. This is the CFX forum. Try the Fluent forum.

I am using STAR-CCM.
If you can suggest me any generic condition to be changed, that will also be helpful. I will look into those options in this software.