Learning Forum

[fatih_g]

Hi,

I try to simulate a coughing scenario with inert particles using DPM and multiphase flow and its interaction of particles with the human face directly at the opposite side of the coughing mouth (inlet) as given in the figure below. I also defined another inlet (room_inlet) which is the surface where the inlet is located to simulate a constant room airflow in the room in the same direction with the cough. So, when there is no airflow from the room_inlet, it is fine. But when the room_inlet has a value larger than 0, e.g., 1 m/s, the emitted particles flow back to the inlet (mouth) as if there is a vacuum. So this is the first problem.

The second problem is similar. When I define the outlet which is the circular surface (human face) as the pressure outlet., the particles in the vicinity of the outlet are absorbed as if there is a vacuum in the outlet. All of the initial/gauge pressures are 0 Pa. Basically, I count the particles that is trapped at the outlet.

So, is the problem about the pressure values, inlet/outlet types or anything else?

Thank you for your help in advance,

Fatih

[Rob]

Why are you using Eulerian Multiphase and DPM model for a cough? I'd hope the emitted particle concentration wasn't too high for the DPM model, and I've only ever used DPM for this type of application. My other concern is the cell resolution difference between the room and manikin: if they're in the same mesh they shouldn't be that dissimilar. Can you check you have one CFD model and not two separate mesh zones on top of each other.

[fatih_g]

First of all, I need to say that I'm fairly new to CFD. I want to simulate the particles with an airflow coming out of the mouth during the emission time. I use 800 particles distributed in a range between 10 to 300 micrometer. What I can find for this multiphase flow (particles+airflow from the mouth) was to use Eulerian multiphase flow coupled with k-epsilon turbulence model. I also want to add a constant room airflow in the same direction of the cough.
Secondly, I put the manikin just for the figure. As you said, the manikin and flow domain didn't have a shared topology. The flow domain has 36688 cells. Since I have the Student version, I need to have a cell number less than 512k. So, using shared topology the cell number exceeds 512k. Now I removed the manikin. My aim is only to count the particles passing through the surface which represents the human face. Hence, the existence of the manikin is not so important.
So, what is your suggestion about the model and also the vacuum effect when I add a constant room airflow at the inlet and outlet? I would appreciate any help. Thank you.

[Rob]

If you have an open volume you can inject particles wherever you want (just add an injection). Providing you have an inlet & outlet in the domain I'm not sure how you're getting a vacuum. Remember the injection releases with velocity components: are they point in the right direction?

[fatih_g]

Thank you for the answer. As you suggested, I removed the multiphase model. As illustrated in the figure below, I already inject particles from the inlet_mouth. I also added an airflow from the same inlet_mouth. Now if there is not any ambient airflow from the inlet_room which is a velocity inlet in the same direction with the particle injection, everything is fine. But if there is an airflow from the inlet_room, the particles are absorbed back to the inlet_mouth in a few time steps.

So, how can I create a constant ambient air velocity from the inlet_room without this vacuum effect? Is it related with the gauge pressures defined at the boundaries or the inlet types or any thing else?

[Rob]

I'm not sure why you've got a refinement zone on the outlet side. Setting two velocities is OK, but run without particles and see what the flow is doing. From there we can figure out what's going on. We'd usually add the particle source into the domain, and it's rare to have a room open like this.
https://www.ansys.com/advantage-magazine/volume-xv-issue-2-2021/how-does-hvac-airflow-influence-coronavirus-transmission may be of interest.

[fatih_g]

Thank you for your time. I've found one of your comments related to vacuum here: https://forum.ansys.com/discussion/30669/fluent-boundary-conditions
But I couldn't figure out the problem yet. Whenever I define a constant ambient airflow velocity, particles go back and escape from the inlet.

[Rob]

Please plot the velocity contours and vectors then post those here.

[fatih_g]

Dear Rob,
You can find the velocity contour and vectors below. Here, 800 particles and air are injected with a velocity of 11.2 m/s for 0.12 s from the inlet_mouth. The velocity from the inlet_room is defined as 2 m/s. In this scenario, particles are absorbed back into the inlet_mouth as soon as they are injected. Also, the surface on the left just across the inlet_mouth is my outlet where I want to count the particles. If I define this surface as a pressure-outlet, the injected particles are quickly absorbed by this outlet surface.
Regards,

[Rob]

I'd expect a velocity inlet at one side (low z?) and pressure outlet at the +z side. I don't understand why there's a suction effect at +z unless you set velocity boundaries at both sides?

[fatih_g]

The inlet_room is the velocity inlet which covers the whole surface. By doing this, I wanted to create an ambient airflow in the whole room, not only in the low z.
I didn't understand setting the velocity boundaries at both sides. I have two velocity inlets on the right side and one pressure outlet in the middle of the surface on the left side. Could this problem be about the default pressure values or the resolution of mesh at the inlet and outlets?

[Rob]

Please can you check the mass flux and see how it looks? The vectors above suggest flow is hitting something.

[fatih_g]

I configured the back wall also as a pressure outlet and checked the mass flow rate values as you can see below. When I define a pressure outlet, particles are quickly sucked to the wall. When I define these outlets as wall, the particles are sucked back to the inlet_mouth again. There is nothing to hit in the flow domain. It's just a cuboid.

[Rob]

Looking at that, inlet mouth is zero flow, and most of the mass leaves through flow_end rather than the outlet. Switch to steady and release a continuous cough/sneeze and let's see what happens.

[fatih_g]

When I set the flow_end, which is at the whole surface except the 'outlet' on the left side, as the second outlet, the problem is solved. But if I set the circular 'outlet' surface as the only outlet, then there is a vacuum effect. Thank you.

[Rob]

If "outlet" is the only pressure outlet then all of the flow must leave the domain through the tiny surface.