Learning Forum

[Rahul123]

 Hello 

I am try to simulate a case for air flow, where my inlet boundary condition is mass flow rate. But, I am confused about correct boundary conditions for outlets.  I am attaching a figure of my setup here. 

Flow from inlet should go and get stored in the tank in the figure. I also want to calculate pressure at point P1. And, outlet 1 and outlet 2 can be opened when desired. I tried by defining outlet 1 and 2 as pressure outlets. But, then the flow does not go to the tank at all. What is the correct boundary condition to achieve this?

Thanks in advance

Regards

Rahul

[DrAmine]

Please insert the screenshot/picture instead of attaching it.

[Rahul123]

 

Thanks for the reply Amine, here is the setup. I want to setup a target mass flow rate as the parameters when I define outlets. So, basically I setup inflow mass rate at inlet and target mass flow rates at the both the outlets (which I defined as pressure outlets ) as the parameters

[Rob]

Think through the physics:

what comes in (kg) = what goes out (kg) + hold-up (kg)

in steady state. Now consider the material density: how (why) will it fill the tank? 

 

[Rahul123]

Thanks for the reply. For ex: I defined my mass flow rate as 0.0004 kg/s as the input. My target mass flow rate in outlet 1 is 0.0001 and outlet 2 is 0.0001 as well. So, the rest of 0.0002 kg/s should fill in the tank. But, this not happening. Do I need to define an artificial boundary condition at the tank boundary to define the flow direction?

[Rob]

Look at the material density: where does it go if you do that?  In a transient calculation as long as the material can go somewhere you can use mass in & out. 

[Rahul123]

My simulation is steady state. If I understood you correctly, you suggest me to use one more outlet 3 at the tank outer boundary. This makes it 3 outlets in total, out of which 2 outlets  (outlet 1 and 2) are real outlets same as the physical model, while outlet 3 is an artificial outlet defined only in the fluent model?

One more problem, I have is: I also want to measure pressure at point P1 as shown in the figure. I know my mass flow rate at inlet, but when I do that pressure at P1 comes much lower than the physical model. Any suggestion regarding that too?

Should I change the gauge pressure at outlets in order to achieve that? as, of now, I am using default gauge pressure (0) at both the outlets but with the target mass flow rate.

Thanks in advance!

 

[DrAmine]

You need to account for another degrees of freedoom: Transient accumatuon in the domain and / or compressiblity. Is the gauge pressure at both outlet the same?

[Rahul123]

Yes, the gauge pressure is same at both the outlets but the target mass flow is different at both the outlets. I know my inlet mass flow rate, but it gives much lower pressure than experiment at P1. How can I resolve it? That fluid gets stored in the tank and I get the correct pressure at P1. I am using 101325 Pa as operating pressure.

 

Thanks!

regards

[DrAmine]

If mass flow and total pressure inlet are known at the inlet then you need to decide which quantity is important for you. I do not think that you know the total pressure at inlet.

Are you now running transient? Are you now running compressible?

[Rahul123]

Thanks for the reply Amine. I am running steady state simulation and I am using air flow, so yes it is compressible.

I know the mass flow at inlet from the physical experiments. The pressure at P1 as shown in the figure was measured in the experiment as well. When I use the same mass flow as experiments, the pressure at P1 is much lower. But, I guess I can fix that by defining the desired pressure at outlets? Would that be a correct approach? 

And, how can I fix the problem of air not getting stored in the tank?

Thanks!

[DrAmine]

So air is modeled as ideal gas? 

Yes you fix that by increasing the pressure at outlet: but you said that here we have ambient conditions (surrounding). Or do you have any losses at the outlet which need to be accounted for?

Can you please add a screenshot of the geometry from Fluent?

You are still not using the transient solver although we have highlighted that a couple of time.