Specifies the flow rate of the fluid domain in the room

Jeremy167270 · March 12, 2023, 23:40

Hi everyone, I'm simulating a data center room, and I got a problem, to simplify my problem, I'll illustrate it with attached pictures. The first picture shows the design of the whole room. The blue boxes are the servers, and the red surface has been set as the Radiator as the heat source.

My question is how do I specify the mass flow rate of each Server?

The second picture shows my goal.

If there is any part of my expression that is not clear enough, please let me know, and I will explain it.

LoGaL · March 13, 2023, 12:19

So you want to specify the mass flow exiting from each outlet? There is mass flow outlet BC for that. But then at inlet you specify pressure, not mass flow. Otherwise your problem is ill-posed

Jeremy167270 · March 14, 2023, 23:05

Quote:

Originally Posted by LoGaL

So you want to specify the mass flow exiting from each outlet? There is mass flow outlet BC for that. But then at inlet you specify pressure, not mass flow. Otherwise your problem is ill-posed

First of all, thank you for your reply.
I tried to change the inlet of the room to pressure and set the mass flow outlet on each outlet, but my calculation will diverge. Is there any other way?

LoGaL · March 15, 2023, 11:11

I find very weird that it doesn't converge. I must be honest, I think that you made a mistake somewhere.

NickFL · March 16, 2023, 02:52

In your first image you are not showing an inlet for the servers, only an outlet (in red). Both servers need to have an inlet to account for the fluid that is coming out. The inlet & outlet into the server should be the same amount with only a temperature difference. This is probably accounting for your divergence when you have more or less entering than physics allows.

A more advanced model would be to couple the flow rate through the server by knowing the resistance. Then you could treat this pair with a fan boundary condition.

Jeremy167270 · March 16, 2023, 04:56

Quote:

Originally Posted by NickFL

In your first image you are not showing an inlet for the servers, only an outlet (in red). Both servers need to have an inlet to account for the fluid that is coming out. The inlet & outlet into the server should be the same amount with only a temperature difference. This is probably accounting for your divergence when you have more or less entering than physics allows.

A more advanced model would be to couple the flow rate through the server by knowing the resistance. Then you could treat this pair with a fan boundary condition.

First of all, thanks for your reply!

The reason I don't have inlets on the servers (red faces) is that I have to set there faces as the heat source so I can't add inlet BCs on there faces.

I have tried fan's BC before, but I think that is not what I need.

NickFL · March 16, 2023, 05:35

OK, that makes some sense. That means your hand drawing above showing the air going through the servers is not correct. What you are modeling is natural convection from a vertical wall on each server. This might not be the most accurate way to model it, because the servers will likely have air flowing through them. But it might produce a good first approximation.

Two mass-flow outlets and a pressure inlet should converge quite well. Does it work if you switch them (both outlets as pressure and inlet a mass flow rate)? What equation(s) is(are) not converging? Does it converge with just flow (i.e. energy off)? What properties are you using for air (boussinesq, ideal gas)? There are many things that could be the reason.

Jeremy167270 · March 20, 2023, 05:55

Quote:

Originally Posted by NickFL

OK, that makes some sense. That means your hand drawing above showing the air going through the servers is not correct. What you are modeling is natural convection from a vertical wall on each server. This might not be the most accurate way to model it, because the servers will likely have air flowing through them. But it might produce a good first approximation.

Two mass-flow outlets and a pressure inlet should converge quite well. Does it work if you switch them (both outlets as pressure and inlet a mass flow rate)? What equation(s) is(are) not converging? Does it converge with just flow (i.e. energy off)? What properties are you using for air (boussinesq, ideal gas)? There are many things that could be the reason.

I just tried to set the inlet and outlet of the whole room as pressure inlet and outlet, and set the mass flow outlet on the air outlet side of the Servers. In this simulation, I turned off the energy and simply simulated the airflow, but the result is still divergent. The calculation result is Continuity divergence, and the property of air is constant.

The attached file is the model I actually want to simulate and the results I want to get. Is there any other way to do it?

LoGaL · March 20, 2023, 06:00

as I said, you made a distraction mistake somewhere, or perhaps the mesh is really bad. Incompressible flow with a pressure and massflow BC has to converge, no question about it.

Without you showing us mesh and settings one by one, it is impossible to understand what is going on. I have no crystal ball.

March 12, 2023, 23:40	Specifies the flow rate of the fluid domain in the room	#1
Jeremy167270 New Member Jie-Ying Zhan Join Date: Feb 2023 Posts: 10 Rep Power: 3	Hi everyone, I'm simulating a data center room, and I got a problem, to simplify my problem, I'll illustrate it with attached pictures. The first picture shows the design of the whole room. The blue boxes are the servers, and the red surface has been set as the Radiator as the heat source. My question is how do I specify the mass flow rate of each Server? The second picture shows my goal. If there is any part of my expression that is not clear enough, please let me know, and I will explain it.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
CFD analaysis of Pelton turbine	amodpanthee	CFX	31	April 19, 2018 19:02
Wrong multiphase flow at rotating interface	Sanyo	CFX	14	February 7, 2017 18:19
Out File does not show Imbalance in %	Mmaragann	CFX	5	January 20, 2017 11:20
two fluid, volumetric/mass flow rate at different volume fractions	Veronique Pe	CFX	2	November 26, 2014 06:30
An error has occurred in cfx5solve:	volo87	CFX	5	June 14, 2013 18:44

March 13, 2023, 12:19		#2
LoGaL Senior Member Lorenzo Galieti Join Date: Mar 2018 Posts: 375 Rep Power: 12	So you want to specify the mass flow exiting from each outlet? There is mass flow outlet BC for that. But then at inlet you specify pressure, not mass flow. Otherwise your problem is ill-posed

March 15, 2023, 11:11		#4
LoGaL Senior Member Lorenzo Galieti Join Date: Mar 2018 Posts: 375 Rep Power: 12	I find very weird that it doesn't converge. I must be honest, I think that you made a mistake somewhere.

March 16, 2023, 02:52		#5
NickFL Senior Member Join Date: Jun 2009 Location: Technische Universität Chemnitz Posts: 107 Rep Power: 17	In your first image you are not showing an inlet for the servers, only an outlet (in red). Both servers need to have an inlet to account for the fluid that is coming out. The inlet & outlet into the server should be the same amount with only a temperature difference. This is probably accounting for your divergence when you have more or less entering than physics allows. A more advanced model would be to couple the flow rate through the server by knowing the resistance. Then you could treat this pair with a fan boundary condition.

March 16, 2023, 05:35		#7
NickFL Senior Member Join Date: Jun 2009 Location: Technische Universität Chemnitz Posts: 107 Rep Power: 17	OK, that makes some sense. That means your hand drawing above showing the air going through the servers is not correct. What you are modeling is natural convection from a vertical wall on each server. This might not be the most accurate way to model it, because the servers will likely have air flowing through them. But it might produce a good first approximation. Two mass-flow outlets and a pressure inlet should converge quite well. Does it work if you switch them (both outlets as pressure and inlet a mass flow rate)? What equation(s) is(are) not converging? Does it converge with just flow (i.e. energy off)? What properties are you using for air (boussinesq, ideal gas)? There are many things that could be the reason.

March 20, 2023, 06:00		#9
LoGaL Senior Member Lorenzo Galieti Join Date: Mar 2018 Posts: 375 Rep Power: 12	as I said, you made a distraction mistake somewhere, or perhaps the mesh is really bad. Incompressible flow with a pressure and massflow BC has to converge, no question about it. Without you showing us mesh and settings one by one, it is impossible to understand what is going on. I have no crystal ball.