Learning Forum

[hurch346]

Hello everyone,

I’m working on a condensation model in ANSYS Fluent where humid air condenses on a wall adjacent to a cold air stream. The condensation layer on the wall creates an obstruction that should increase the pressure drop, but the Lee and Mixture model I’m using doesn’t seem to capture this effect adequately.

In my simulation:

• I’m modeling humid air condensing on a cooled wall, which obstructs the airflow path and should increase the static pressure drop.
• Experimental data shows a pressure drop of 80 Pa on the condensate-forming side, but I’m only achieving a 30 Pa drop in the model.

I’ve tried adjusting:

1. Wall friction coefficient and friction height – While this does increase the pressure drop slightly (from 30 Pa to 50 Pa), it’s still significantly lower than my experimental results.

I also attempted the Euler-Euler model (which includes liquid momentum equations) to see if it would better account for the condensate’s impact, but the results didn’t vary much from the Mixture model.

I’m looking for suggestions on how to better capture the increased pressure drop due to wall condensation, or any strategies to get the model to account for the condensate blocking the airflow path.

Questions:

1. Is there a way to explicitly model the condensate layer’s impact on pressure drop without switching to a full Eulerian multiphase model?
2. Are there additional parameters or customizations within the Mixture or Lee models that could simulate this increased blockage effect?

Thank you in advance for any insights!

[Rob]

Assuming you're using a full multiphase model as opposed to a wall film model condensation will have a physical presence in the domain. 

A deviation from an expected result could be lack of condensation, lack of time to build up condensation, how you're adding flow into the domain or more general issues (mesh, material properties etc) are potential CFD related problems. We also need to consider whether the experimental data is correct.

I don't automatically blame either data set when starting to look into differences: I've seen too many cases where CFD models are correct but linked to a poor experiment but also too many poor CFD models with good experimental data. I've also seen many case where both the CFD and experiment are good, but the CFD didn't actually model the correct conditions. 

[hurch346]

Thank you for your response,

I am confident that the model is correct, as I have verified mesh independence, and my temperature and condensation measurements are within 5-20% of the model values over a range of conditions.

I believe my model doesn't take the interaction between the condensation film and the bulk air into account. In the experiment, the condensation is drop-wise, so there is a complex interaction between the air and condensate droplets, increasing the pressure drop relative to a duct with no condensation. 

Ideally, it would be nice for the model to account for this, but I assume it would take a lot of computational power and time, so the second best option for me would be to manually tune the pressure drop. I have tried the more complex Eulerian-Eulerian model, which solves for the momentum of both phases, however, this did not change my results at all so I am unsure how to progress the model at this stage. 

Would you have any ideas on how to either manually increase the pressure drop, or how to make the model account for the condensation-air interactions?

 

Thank you

[Rob]

How much liquid have you got on the walls? Are you accounting for volume change in the gas if cooling is significant? I'd be very wary of artificially increasing the pressure loss without knowing why: is the flow then a result of your modelling or directly down to set values?