[behrouz]

hi to all
i try to calculate skin drag for submerged cylindrical body as picture.

domain:
i define enclosure with L for upstream, 2L for downstream and L for sides.( L= length of cylinder)

1. is that enough L for side in domain?

mesh: domain divide into 3 section so i can control mesh sizing in different par better.
min value for Yplus=0.5 , max value for Yplus= 54 and average value for Yplus=11

Turbulance model: k-epsilopn

steady state

bubble diameter= 1 mm and volume fraction= 0.1 - 0.6

now i have some issue and will happy any body can help me on this.

2. in cfd-post there is no expression for skin drag and drag coefficient and i read other post, and they suggest expression, but im not sure about this expressions.
when simulate mono phase with water the skin drag coefficient will be

ave(wall shear)@wall/0.5* ave(Density)@inlet * ave(Velocity)@inlet^2

is that correct?

3. when simulate bubbly flow (water and air) for calculate Cf there is 2 wall shear ( for water and air)
witch one should be used??? should be use average value for water and air?

4. convergence is slowly and not quite well, any idea on better convergence on this kind of multiphase flow simulation? ( i read https://www.cfd-online.com/Wiki/Ansys_FAQ already)

5. for wall on sides use of wall boundary with no slip flow correct?

tnx for guide.

[ghorrocks]

1: Is your domain big enough? Do a sensitivity study and find out. The domain size you need depends on the conditions you are modelling, so you really need to work it out for yourself.

2: No idea, you can define it anyway you like. If you are trying to match other people's work then define it the same way they did.

3: Use the force() function and you will get the total force.

4: With no details of what you are doing we cannot help. But in general, convergence is helped by improved mesh quality, smaller time steps and double precision numerics.

5: Which wall are you talking about?

[behrouz]

Originally Posted by ghorrocks

1: Is your domain big enough? Do a sensitivity study and find out. The domain size you need depends on the conditions you are modelling, so you really need to work it out for yourself.

2: No idea, you can define it anyway you like. If you are trying to match other people's work then define it the same way they did.

3: Use the force() function and you will get the total force.

4: With no details of what you are doing we cannot help. But in general, convergence is helped by improved mesh quality, smaller time steps and double precision numerics.

5: Which wall are you talking about?

Dear ghorrocks
thanks for your answer, as alwayse you are best
1. you right, i should check domain study more carefully.
2. ok
3. force() function give total force include skin drag force and normal drag?
4. how can i share more detail of my work? convergence diagram and output result file will be enough?
5. i mean wall that named 'far field' in picture, witch boundary condition should be use?

[ghorrocks]

3. The force function gives the total force acting on a surface. It has both normal/pressure and tangential/friction components.

4. Post some images of what you are modelling, some images of your mesh and your output file. Put the images directly in the post and attach the output file as an attachment, compressed if you need to.

5. If you make it a no slip wall it will then generate a wall boundary layer - you probably don't want that. A slip wall is better. If you intend to run this at different angles of attack then a slip wall won't work either - in this case the usual approach is to include the sides of the domain in the inlet boundary. (This assumes your inlet is prescribed velocity of some description and your outlet is a pressure boundary of some description). Then you can change the angle of attack.

[behrouz]

Originally Posted by ghorrocks

3. The force function gives the total force acting on a surface. It has both normal/pressure and tangential/friction components.

4. Post some images of what you are modelling, some images of your mesh and your output file. Put the images directly in the post and attach the output file as an attachment, compressed if you need to.

5. If you make it a no slip wall it will then generate a wall boundary layer - you probably don't want that. A slip wall is better. If you intend to run this at different angles of attack then a slip wall won't work either - in this case the usual approach is to include the sides of the domain in the inlet boundary. (This assumes your inlet is prescribed velocity of some description and your outlet is a pressure boundary of some description). Then you can change the angle of attack.

Dear ghorrocks
thank you for your reply
3. force() function calculate total force on surface, but i there any function that just calculate skin fraction drag ??
4. of course, i will use your advise and run again and post different pic and out put file on this post.
5. very intersting advise, but use of 'opening' boundary condition in normal simulation ( that there is no attack angle) is correct?

[ghorrocks]

3. areaInt(Wall Shear)@Wall
5. There are many ways you can set up a simulation where you are trying to model a body travelling in an infinite fluid. The boundaries can be handled in many different ways.

[behrouz]

Originally Posted by ghorrocks

3. areaInt(Wall Shear)@Wall
5. There are many ways you can set up a simulation where you are trying to model a body travelling in an infinite fluid. The boundaries can be handled in many different ways.

thank you so much.