CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM Running, Solving & CFD (http://www.cfd-online.com/Forums/openfoam-solving/)
-   -   buoyantBoussinesqSimpleFoam - Turbulent transient flow in a room with inlets and outl (http://www.cfd-online.com/Forums/openfoam-solving/105591-buoyantboussinesqsimplefoam-turbulent-transient-flow-room-inlets-outl.html)

Mojtaba.a August 3, 2012 04:53

buoyantBoussinesqSimpleFoam - Turbulent transient flow in a room with inlets and outl
 
Hello everyone.
I am trying to simulate a room with some inlets located on the top walls and some outlets on the side walls. Air has a velocity of 1 m/s at inlet patch. and some stuff are located inside the room. in other words i am trying to simulate an HVAC application. first I am using pisoFoam solver in order to have a reasonable pressure and velocity field. after that i will try to implement energy equation in order to solve for temperature field as well. But unfortunately i have stuck at first.
I am comparing my results with Fluent results. there is some problems with pressure field. Fluent results show that pressure varies from 0.9 to about 1.3 pa, but my values are much more less than that. they are at the range of 0 to 0.3 pa.
I have used zeroGradient for inlet and walls, and fixed value 0 for outlet.

Any suggestions would be appreciated ~
Mojtaba ~

uli August 21, 2012 18:30

For incompressible flows only differences in pressure are of interest. Do you use the same boundary conditions for pressure in Fluent and OF?

owayz August 22, 2012 05:25

Hi Mujtaba,
Uli is right. For incompressible flows you don't have to specify pressure as an absolute value since pressure difference is more important and it is more convenient to understand and interpret pressure difference than pressure itself. Also in openFOAM for incompressible flows pressure is specified as pressure/rho in a sense pressure and nu are seen as per unit density.
But for compressible flows Ideal Gas equation is needed to calculate rho and that is why pressure is important.
What I think is that you are probably using a reference pressure 0 in openFOAM or you have specified 0 pressure somewhere in some boundary condition.
The way I see your results and fluent results, they are pretty close. Because you have almost the same pressure difference. May be you can increase your confidence or improve your openFOAM results by doing some grid convergence study for parameter of interest or by increasing convergence criteria. Also apart from residuals of Momentum and pressure, keep a closer look at time step continuity errors, they should be sufficiently small (I guess this ensures mass balance in the system).

Regards,
Awais

Mojtaba.a August 22, 2012 07:56

1 Attachment(s)
Quote:

Originally Posted by uli (Post 378060)
For incompressible flows only differences in pressure are of interest. Do you use the same boundary conditions for pressure in Fluent and OF?

Thank you uli for your note. Well I am trying to validate my OF simulation with a fluent simulation which has been done in an article. There is no information about how pressure boundary conditions are defined. I set my boundary conditions as a default of all other OF cases. and that means "zeroGradient" for inlet and walls and "fixedValue uniform 0" for outlets.

Quote:

Originally Posted by owayz (Post 378120)
Hi Mujtaba,
Uli is right. For incompressible flows you don't have to specify pressure as an absolute value since pressure difference is more important and it is more convenient to understand and interpret pressure difference than pressure itself. Also in openFOAM for incompressible flows pressure is specified as pressure/rho in a sense pressure and nu are seen as per unit density.
But for compressible flows Ideal Gas equation is needed to calculate rho and that is why pressure is important.

Hi owayz, and thanks for your useful answer. well now I understand that pressure difference is more important than the pressure itself in incompressible flows. Now I am facing a new problem. I tried to improve my simulation to approach to a better solution for my case. I am now trying to use "buoyantBoussinesqSimpleFoam" for my simulation to append buoyancy effects into my case. I ran the simulation and everything seems fine except the pressure again. I tried to make a structured mesh with refinements but there was no change in results. you can see my post regarding to structured mesh refinement in this thread (no answer until now):


I wonder which one,p or p_rgh are the results of pressure, which I am searching for. As you know momentum equation with buoyancy effects is as follows:

Where rho in buoyancy term is being calculated using boussinesq approximation. As you can see, parameter p which has been used in this equation is more likely to be p_rgh in OF. The reason is that in this equation rho*g is subtracted out of p and is shown in a new term. So I think I am searching for p_rgh instead of p in OF. My values in OF seems fine but pressure contour plots are different from what I am searching for. I am looking for some layered pressure distribution among room which decreases from ceiling to the floor of the room. But my results have very complex contours and are more similar to velocity field. Your note informed me that maybe I can compute p with units of pa not pa/(m^3/kg) or pressure/rho. In this case I have to find p like this:
P=p_rhg*rho
or
P=p_rhg*rhok
Am I correct & How can I compute this p in openFOAM?


Quote:

What I think is that you are probably using a reference pressure 0 in openFOAM or you have specified 0 pressure somewhere in some boundary condition.
actually I'm using both. I am using reference pressure 0 in openFOAM and also I have specified 0 pressure in outlet BC.


Quote:

The way I see your results and fluent results, they are pretty close. Because you have almost the same pressure difference. May be you can increase your confidence or improve your openFOAM results by doing some grid convergence study for parameter of interest or by increasing convergence criteria. Also apart from residuals of Momentum and pressure, keep a closer look at time step continuity errors, they should be sufficiently small (I guess this ensures mass balance in the system).
well yes, my continuity errors are small. and there is no problem on this issue.


All times are GMT -4. The time now is 10:37.