[adilsyyed]

Let say I am modeling a steady state flow in a pipe or tank with a heat source. There is only water at the start of the simulation. But I suspect that phase change might occur on a small scale.

There is no steam present at the start of the simulation and no boundary condition has steam involved.

Do I define only water as phase or both steam & water as two phases.
Or do I only define/add steam to the available materials?

[ghorrocks]

This is a multiphase simulation, so you will need some sort of multiphase model to handle it. How you define the material model depends on exactly what physics you want to model in the phase change.

[adilsyyed]

Originally Posted by ghorrocks

This is a multiphase simulation, so you will need some sort of multiphase model to handle it. How you define the material model depends on exactly what physics you want to model in the phase change.

Ok I understand that now.

Say there is double pipe heat exchanger with no mixing of fluids. The two fluids are water and steam. That is a multiphase problem again but no mass transfer occurs. How do I model that.

[ghorrocks]

In default mode CFX models this as a multiphase flow. You are likely to get the "isolated volumes" warning, but should be OK to proceed as long as you understand the warning. But this a good time to activate the "check isolated volume =f" expert parameters option and use a water single phase flow in one domain and single phase steam in the other. Then you do not need a multiphase model.

[adilsyyed]

Originally Posted by ghorrocks

In default mode CFX models this as a multiphase flow. You are likely to get the "isolated volumes" warning, but should be OK to proceed as long as you understand the warning. But this a good time to activate the "check isolated volume =f" expert parameters option and use a water single phase flow in one domain and single phase steam in the other. Then you do not need a multiphase model.

Thanks for the fast reply.
In order to model two individual single phase flows, I have to enable beta physics and disable constant domain physics. Is that correct? Or is there another way of doing that.

[ghorrocks]

That is correct (I think, it has been a long time since I did it).

[Opaque]

Originally Posted by adilsyyed

Ok I understand that now.

Say there is double pipe heat exchanger with no mixing of fluids. The two fluids are water and steam. That is a multiphase problem again but no mass transfer occurs. How do I model that.

For clarification, using ANSYS CFX terminology, your problem is a single phase flow in separate passages.

In ANSYS CFX terminology, multiphase flow is when two (or more) distinct non-miscible phases interact in connected regions of space. Note, I do not use the word "material". A ANSYS CFX "phase" is made of variety of materials type: pure, variable composition mixture, reacting mixture, homogeneous binary mixture, etc.

ANSYS CFX can model such flow configuration; however, you must use some beta features that allows you to use different phases in the fluid domains. You can search this forum for further information (search for "constant domain physics")

Hope the above helps,

[adilsyyed]

Originally Posted by Opaque

For clarification, using ANSYS CFX terminology, your problem is a single phase flow in separate passages.

In ANSYS CFX terminology, multiphase flow is when two (or more) distinct non-miscible phases interact in connected regions of space. Note, I do not use the word "material". A ANSYS CFX "phase" is made of variety of materials type: pure, variable composition mixture, reacting mixture, homogeneous binary mixture, etc.

ANSYS CFX can model such flow configuration; however, you must use some beta features that allows you to use different phases in the fluid domains. You can search this forum for further information (search for "constant domain physics")

Hope the above helps,

Thank you for the detailed explanation. However, a colleague of mine suggested that I stay clear of "beta feature" since it means something that is still in testing mode. Should I be worried about that?

[ghorrocks]

Your colleague is correct - beta features should be used with caution as they are not fully tested and unexpected things can happen. But in your case the simulation you propose with two unconnected domains is far better modelled using the expert parameter compared to doing a full multiphase model with the default setting.

So in your case the extra care you will need in using a beta feature is a small price to pay to be able to run a far better simulation. So I recommend you go for it, but pay extra attention to the results and check for unexpected weirdness.

[adilsyyed]

Originally Posted by ghorrocks

Your colleague is correct - beta features should be used with caution as they are not fully tested and unexpected things can happen. But in your case the simulation you propose with two unconnected domains is far better modelled using the expert parameter compared to doing a full multiphase model with the default setting.

So in your case the extra care you will need in using a beta feature is a small price to pay to be able to run a far better simulation. So I recommend you go for it, but pay extra attention to the results and check for unexpected weirdness.

That makes more sense. Thank you Glenn