Learning Forum

[pp2278]

This is my first ansys simulation and I need some help. I am trying to simulate the incubation of agar and water used for microbiological testing of water. The agar-water mixture is put in a petri dish and left for 6 days to incubate at around 35 degrees Celsius. There are 6 petri dishes in total, stacked on top of each other. The petri dishes are also flipped in a way, that the agair-water mixture is on the top (the mixture is solid and does not fall down). After performing real life incubation tests, we noticed that condensation occurs in the petri dishes during incubation and water droplets form and fall down. Now, we want to simulate the incubation, to try and figure out how and why this condensation forms. 

I have attached an image of the 2D geometry bellow. The gray square is used to model air, and the smaller green square is also air (the bigger square has larger elements, and the smaller square has smaller elements, which is important for the mesh).

The next image shows the petri dishes closer. The petri dishes are put in a plastic box, with small gaps (2 mm) between each petri dish and between the bottom petri dish and the box.

I want to simulate water evaporating from the agar-water mixture and condensating on the bottom of the petri dish. Our theory is that the agar-water mixture contributes to a 100% relative humidity inside the petri dishes. The upper part of one petri dish is colder than the bottom part of the petri dish above it. And it is this difference in temperature that causes a drop in temperature in the petri dish above, which results in condensation to form, since the relative humidity was at 100%. So we want to simulate this as closely as we can.

I completed the geometry and the mesh, and I have some questions about the setup:

1. I have activated Energy and chose Mixture for the Multiphase model. Are those all the models I need to activate?
2. This question is about the Multiphase model. I added the images with the setup. The Mixture Model has two phases, the first one is water-liquid (water) and the second is water-vapor (vapor). I added one Mass Transfer Interaction, which is evaporation-condensation. Is this the correct setup for the simulation I want to create?

3. I want to model the agar-mixture as a solid with the properties of water. But the Mixture model only works with fluids as the phases. The only issue I have with this is that as a liquid, the water on the top of the petri dish will fall down, since gravity is turned on. Is there some way to make the water-liquid not fall down because of gravity or to use a solid as a phase in the Mixture model?

4. How do I set the agar as water-liquid in the beginning of the simulation? When I go to Cell Zone Conditions, I can assign materials to the solid zones, but I cannot assign materials to the fluids. Also, every zone (solids and fluids) are set to Phase: mixture. I change it to water (phase-1), but when I click on it again, it changes back to mixture. Is this ok?

5. I would also like to assign a certain temperature and relative humidity to the air, which slightly changes throughout the simulation (around 6 days). Those changes are +- 1 degree for the temperature and +-10% for the relative humidty. Does anyone know how to do this?

6. If there is anything else I missed and is important for the simulation, please tell me.

Thank you and best regards

[Rob]

If this is your first model I recommend doing the tutorials first - multidomain and multiphase models aren't trivial. Combining the two even less so. 

If a domain is sealed you'll need to use ideal gas. Fluent conserves mass but if the volume changes may lose mass to satisfy the volume reguirements. 

I'd expect agar to be a fair bit more viscous than water otherwise it'll run down the insides of the dish. From setting fire to some at Uni (biochem module: tip, Chemical Engineers will follow instructions about putting glowing hot wires into alcohol as we know what will happen) I don't remember it moving much. So you may want to rethink that mixture. You'll also find it'll move if it's a fluid so things get a bit more complicated. 

You can use patch to add materials/phases to regions. That also works for temperature. So, you can initialise the whole domain as (for example) air, and then patch in the agar fluid, temperature and vapour fraction as needed. That may get a little complicated so I'd consider domain with just a single dish floating in space to work out all settings. 

 

 

[pp2278]

Thanks for the feedback. I watched videos explaining multiphase models and water evaporation-condensation. My main question is, would the evaporation-condensation model correctly simulate the dew that appears during incubation. The evaporation-condensation models uses temperature values to assess when to change phases of the liquids. But the temperature in our experiments is around 35 degrees celsius and changes very slightly. The condensation occurs because the air inside the petri dishes is at 100% relative humidity (caused by the agar and water inside), and when the temperature in the petri dish lowers (caused by conduction and convection with the petri dish beneath it which is colder), the dew point is breached and dew is formed on the bottom half of the petri dish. Can Ansys Fluent simulate that correctly, using only multiphase modeling, and evaporation-condensation between phases? If not, can I use other models or UDFs to model the relative humidity and consequent condensation? 

[Rob]

If you start with RH 100% at a certain condition with a compressible gas mixture as you alter the temperature you'll see condensation. However, balancing the conditions along with evaporation/condensation rates may be numerically stiff and/or require very small timesteps to maintain solver accuracy. 

The issue may be as much that setting RH 100% will cause condensation as the temperature falls slightly, so I'm not sure whether that proves your theory, or simply that condensation occurs when a saturated mixture cools. 

[pp2278]

 

 

The air inside the petri dishes is not at 100% RH when the incubation starts. We measured the RH during the incubation process. It does not start at 100%, but after some time, the RH in every petri dish reaches and stays around 100%. Moisture escapes from the agar medium and causes this to happen.

I have also set the agar viscosity at 8000 kg/(m s), to act more as a solid.

I would like to initialize and patch the atmosphere with 40% RH (patch every fluid zone, except the agar, as 0.025 water-vapor and 0.975 air for the Volume fraction). With temperature patched at ~35 degrees Celsius, can Fluent simulate the moisture escaping the agar?

You mentioned that I would need very small timesteps for accuracy. I will most likely need to simulate at least 4 days of incubation. What size of the timestep would you recommend?

As for the compressible gas that you mentioned, do I get that by using ideal-gas in the density settings for air and water-vapor? Are there any other specific settings I need to change?

(Edit): How can I use Species Transport alongside Multiphase Models? I would like to use Species Transport for initial RH values in the atmosphere. The species (air-vapor) consists of water-vapor and air. In that case, how many phases should I have?

Before, I used three phases: agar, water-vapor and air, with mass transfer (evaporation-condensation) from agar to water-vapor. Now, having Species Transport activated, are two phases (agar and air-vapor) enough? In this case, the evaporation-condensation is from agar to air-vapor; or from agar to water-vapour (in which case, again I would need three phases for the multiphase model)?

 

 

[Rob]

You may struggle with viscosity values like that as any motion will run into the very viscous flow and that won't help solver stability. You may also find the water evaporation is a function of diffusion in the agar as much as the flow in the gas phase. Phase change will be OK, but again it's going to be difficult to ensure that only occurs at the surface of the agar: the Fluent models look at phase change on a cell basis everywhere rather than at specific locations. 

The bigger problem may be time scale. Time steps may be down in the 0.0001s range, so for 4 days you may struggle.