[mrji8011]

It is possible to simulate towing tank test by two approach. first setting the body fixed (stationary wall) and boundry pressure-inlet with specific velocity and second, setting inlet and outlet as far-field domain ( velocity=0) and impose the speed as moving wall for ship body. does anybody has idea which one is better by fluent?
thanks all

[jiangtao167]

Originally Posted by mrji8011

It is possible to simulate towing tank test by two approach. first setting the body fixed (stationary wall) and boundry pressure-inlet with specific velocity and second, setting inlet and outlet as far-field domain ( velocity=0) and impose the speed as moving wall for ship body. does anybody has idea which one is better by fluent?
thanks all

vel=0 at Boundaries may induce divergent problem and back flow warning, so imo the first approach is a better choice
good luck for u

[mrji8011]

Originally Posted by jiangtao167

vel=0 at Boundaries may induce divergent problem and back flow warning, so imo the first approach is a better choice
good luck for u

Thanks for your reply
I set pressure-inlet as BC for inlet domain. Alought I set fixed water level (freesurface level and bottom level), the level of the water change during simulation. Let me know if you have any suggestion for this error please

[jiangtao167]

Originally Posted by mrji8011

Thanks for your reply
I set pressure-inlet as BC for inlet domain. Alought I set fixed water level (freesurface level and bottom level), the level of the water change during simulation. Let me know if you have any suggestion for this error please

Initialize and Patch operation play an important part in vof simulation. have you followed the recommendations in users guide manual?

[mrji8011]

Originally Posted by jiangtao167

Initialize and Patch operation play an important part in vof simulation. have you followed the recommendations in users guide manual?

yes of course. at low speed and also at first time steps is currect

[sail]

Are you shure that the rise of freesurface is not caused by blockage effect ( domain too small)?

I assume your bottom boundary is exactely parallel to the flow direction (no bumps, etc)

is it rising everywhere or just at specific locations?)

Thetraehral mesh or hexa?

Can you post some pictures of the problem?

[mrji8011]

Originally Posted by sail

Are you shure that the rise of freesurface is not caused by blockage effect ( domain too small)?

I assume your bottom boundary is exactely parallel to the flow direction (no bumps, etc)

is it rising everywhere or just at specific locations?)

Thetraehral mesh or hexa?

Can you post some pictures of the problem?

Thanks Rail
yes I am sure that it's not because of blockage and the bottom exactly is parallel to flow direction. rising is just in inlet flow. there is reverse flow in some cells at air and water border. I fill the body hole with mesh and run it as channel with current (without body)but there is this reverse flow again.

[sail]

i'm puzzed, i've never encountered this problem. some reversed flow at the outlet can happen at the aerly stages of simulations, but at the inlet... never heard of.

so to recap:

pressure inlet with open channel bc, fixed height (no imposed waves), far enought to do not "feel" the effect of the hull; the mesh at the inlet is made such as the freesurface is not across a very big cell, but lies in a refined region
pressure outlet open channel
hull=no-slip wall
all the rest =symmetry.

gravity is turned on and intensity and irection is specified accrodingly

explicit vof
tranisient simulation with no moving mesh. timestep is adequate to sufficently small cfl

initialization and subsequent patching is made accordingly to inlet conditions.

how about numerical schemes? it's the only thing not considered... apart form turbulence but don't think this is the problem...

how about the convergence? what are your residuals for continuity and momenutm? how many iterations per timestep?

[mrji8011]

Originally Posted by sail

i'm puzzed, i've never encountered this problem. some reversed flow at the outlet can happen at the aerly stages of simulations, but at the inlet... never heard of.

so to recap:

pressure inlet with open channel bc, fixed height (no imposed waves), far enought to do not "feel" the effect of the hull; the mesh at the inlet is made such as the freesurface is not across a very big cell, but lies in a refined region
pressure outlet open channel
hull=no-slip wall
all the rest =symmetry.

gravity is turned on and intensity and irection is specified accrodingly

explicit vof
tranisient simulation with no moving mesh. timestep is adequate to sufficently small cfl

initialization and subsequent patching is made accordingly to inlet conditions.

how about numerical schemes? it's the only thing not considered... apart form turbulence but don't think this is the problem...

how about the convergence? what are your residuals for continuity and momenutm? how many iterations per timestep?

I have done all these setting (thanke you so much)
i put 20it/timestep and it converges less tahn 20 it
the resual of continuty and momentum are in order of e-4

[sail]

then, i'm sorry but i don't understand what might be the cause. i hope that somebody else can came up with a solution, or you can ask help to ansys support

if the geometry is not confiential try to upload the case somewhere and i might take a look at it, apart from that i don't really know how to help you.