Phase change condenser

[ramrecon]

Hi all,

I have a question regarding the phase change capabilities, more specifically steam phase change that occurs in a direct condenser?

How would you go about modelling two inlets and one outlet?

First inlet is a saturated steam at 0.8 quality, secondary inlet is a colder liquid (water) entering at 20°C (the cooling water). As far as I know you cannot mix two fluids in the same domain, so is it possible to set the secondary inlet (water) as saturated steam at quality 0 at a specific temperature below the saturation point (subcooled)?

Can you model the condensed water, that is the condensed steam + the cooling water?

Also to get a good convergence since you need to set the solver to solve as incompressable flow + heat transfer, do you require a much finer mesh using the phase change rather then single phase? Would the turbulence model affect the convergence, that is k-e vs. SST k-omega for example? I belive ADV5 would also be the preffered choice in this problem.

I read that it is not recommended to go above 80psi, if you go above could it result in a more unstable solution convergence? And would it be possible to fix it by adding a more detailed thermodynamic table, expand the steam.dat file?

I am very curious about the phase change capabilities of Autodesk CFD, if anyone has any good information I take all comments gladly 🙂

And thank you very much for this forum and the good support Autodesk is providing with their CFD software.

PS. I do not have a model yet as I am just looking into the problem and possible modelling solution.

Kind regards,

[srhusain]

As far as I know, you should be able to use the saturated steam as a single material but use the scalar tracking to represent the inlet quality (0.8 and 0.0).

Hope this helps.

[ramrecon]

You would have to assign a static temperature to those inlets, right along with flow rate, velocity or pressure?

[srhusain]

Since saturated steam is being used, it follows that the operation is on a fixed pressure and temperature line(assuming equlibrium conditions), so you might try assigning a static temperature corresponding to the operating pressure. Also, the flow is assumed to be largely incompressible, so, the pressure inlet boundary conditions are not necessary if the velocities are prescribed.

[ramrecon]

I have attached a file that I am working on.

It is a pipe that has a certain outside wall temperature, inside there is a fluid that enters the pipe as a liquid but then gets affected by the hot pipe.

I cannot seem to get the solution to converge, it seems to be converging at one point but then jumps off again.

Have any suggestions on improvement to my model?

Kind regards,

[srhusain]

Phase change models are numerically challenging due to large thermal source terms which arise from changes in steam quality and changes in density. Some things you can try are:

1. Refine the mesh where sharp thermal gradients are expected, and
2. Run transient with Intelligent Solution Control turned on (in the solver options dialog) so that the solver can automatically adjust the time step for better stability.
3. If not already done so, switch to the advection scheme Adv5

[ramrecon]

Thank you for the advice, I will try the transient approach, so far it has been analysed as steady state.

Also add refinement regions around the expected larger thermal gradients.

I am using ADV5 + SST k-omega turbulence model with enhanced layers and blending to try to capture the thermal boundary layer better.