Version of Phoenics: 3.2 Operating system: Windows 2000.

I wish to simulate a fan extracting hot air from a domain. The fan has a constant velocity (say 10 m/s), but the mass flow rate is not known before hand, because the density is a variable. The ambient conditions outside the domain are T=20 deg. C, and density=1.1892 kg/m3.

I have created the fan using an option available in the "Flair" menue of the Satellite pre-processor (Version 3.2.0). Note that I am not using the VR interface.

Oddly, the Flair menu seems to produce two totally different types of patches for identical types of fans. The first type of patch looks like this:

PATCH (FAN1 ,NORTH ,#1,#2,#2,#2,#1,#2,1,1) COVAL (FAN1 ,V1 , FIXVAL , 1.000000E+01) COVAL (FAN1 ,TEM1, 0.000000E+00, 2.000000E+01)

The form of this patch looks OK, but when Earth is executed no flow velocities results in the domain at all. (Note that there is an "opening" available on the opposite side of the domain which would allow a flow to be established).

The second type of patch generated by Flair looks like this:

INLET (FAN3 ,NORTH ,#3,#4,#2,#2,#1,#2,1,1) VALUE (FAN3 ,P1 ,-1.189200E+01) VALUE (FAN3 ,V1 , 1.000000E+01) VALUE (FAN3 ,TEM1, 2.000000E+01)

Executing Earth with this patch produces reasonable flow velocities within the domain, but I am concerned that the second line of the patch, namely:

VALUE (FAN3 ,P1 ,-1.189200E+01)

is fixing the mass flow rate based on an ambient density, rather than the variable density present inside the domain at the upstream side of the fan.

So I have several questions:

1. Why does Flair create different patches for identical fans ? 2. Which Patch is the correct one (if any) ? 3. Why does the first Patch not produce any flow velocities ? 4. Is the second patch fixing (a) the mass flow rate, or (b) the velocity ?

Any assistance, advice, or comments would be greatly appreciated. Regards,

Craig

Dear Craig

I don´t know much about "fans types" in FLAIR so I will try to help you based on my general knowledge on PHOENICS. Therefore, I can´t answer your first question.

If you want a mass outflow (or inflow) at the "fan" patch you should put a COVAL (or VALUE) function for pressure P1 into the Q1. That is why your "second type patch" works as an mass extractor (a fan) and the "first type" don´t. This, answer your second question and, with a little effort, your third question also.

For this message I will agree with your nomeclature about "PATCHes" to refer to boundary conditions, but, just to point it, the PATCH function marks a physical region into the calculation domain and are the COVALs fucntions which fix the BC. Nevertheless, INLET is a special PATCH type BC.

For it, the COVAL function: VALUE(FAN3 ,P1 ,-1.189200E+01) fix the mass flow rate through the INLET region and VALUE(FAN3 ,V1 , 1.000000E+01) indicates the momentum flux that came with this mass flux (in this case to the Y transport eq.).

The mass flow condition is used for continuity eq. and the momentum flux COVAL by momentum transport eqs. So, you need to fix BOTH to satisfy the system of equations in PHOENICS. It is not enough to give the velocity at the outlet and that is why, in the "first patch type" case, you haven´t flow through your domain. You have no outgoing mass across the "fan".

To deal with the variable density in your domain you need a reference to ground (GRND) inside the COVAL function for P1. After, you should insert some code lines in Group 13 into ground.f file. It could take a little long to explain here.

Do you have any PHOENICS included model for density dependence? Try to read info about it at CHAM web page, I am sure you will find some examples in your PHOENICS distribution.

Regards

Kike

Dear Kike,

Thank you for your helpful reply.

Although I may be corrected, there would appear to be some significant issues (errors) with the Flair menu.

The first Flair-generated patch is wrong, since it results in no extraction from the domain at all.

Regarding the second patch, most fans operate as constant volumetric flow rate devices, not constant mass flow rate devices. This difference can be vitally important in some situations for which Flair is advertised as being an appropriate tool. The menu describes the boundary condition simply as a "fan", and does not indicate that in reality it is creating a constant mass flux boundary condition.

Flair already allows for variable density flow fields, and I could probably work out how to add the appropriate coding to Ground in order to create the correct fan boundary conditions. This is not, however, the simple point-and-click solution that I think most users would prefer, and indeed what most Flair users probably believe is already present.

Dear Craig

Yes, it is probably a weak point of PHOENICS as a "user friendly code", it is not completely oriented for "click-and-go" like others commertial codes. I am sure you can contact CHAM personnel to give your comments about the FLAIR menu.

Nevetheless, it is not too difficult to introduce the varibale density outlet/inlet condition via GROUND. If you are interested ask me for help.

Regards

Kike