Vents don't get captured.

[Siva D]

Hello greetings.

Long time. I am trying to simulate the flow of moist air in an empty warehouse. However, the size of the domain is too large and the size of the vents is so small. The mesh size I have used does not even capture the vent due to its small size. Is there any to capture the vents without using small grids ? My PC is not capable of running a very fine grid whose size is of the order of the size of the vent. Please help me. I have attached my .FDS file. Please look at it and let me know. 

Thank you in advance.

[dr_jfloyd]

Use localized leakage.  See the User's Guide and the NIST Helium experiments in the Validation Guide.

[Siva D]

Hi

I have used Localized leak approach to model the small diffusers. However, the velocity obtained from the model is so low and I don't think the VOLUME_FLOW I am giving on the &HVAC line is getting recognized. My code is as follows.

&MESH IJK=150,100,18,    XB= 0,150, 0,100, 0,18/

&TIME T_END=1200.0 /

&SPEC ID= 'WATER VAPOR'/

&SURF ID='Supply', SPEC_ID(1)='WATER VAPOR', MASS_FRACTION(1)=0.012, TMP_FRONT= 17.6/

&VENT ID= 'DiffuserA' XB=17,18  1,1, 3,4,  SURF_ID='Supply', COLOR='RED'/
&VENT ID= 'DiffuserB' XB=17,18, 2,2, 3,4,  SURF_ID='Supply', COLOR='RED'/

&HVAC ID='DUCT', TYPE_ID='LEAK', VENT_ID='DiffuserA', VENT2_ID='DiffuserB', AREA=0.099, VOLUME_FLOW= -0.943/

&OBST XB=0,150, 1,2, 0,18,/

&SLCF PBX=17.5,   QUANTITY='MASS FRACTION',SPEC_ID='WATER VAPOR' /
&SLCF PBX=17.5,   QUANTITY='VELOCITY', VECTOR=.TRUE. /

&TAIL/

Thank you

[dr_jfloyd]

Ignore my previous advice.  I must not have had enough coffee when reading your post and completely misread it.

Currently your mesh is 30x30x30 with a domain size of 150 m x 110 m x 18 m.  This is giving you aspect ratios of over 8:1.  We recommend in the user's guide that you avoid exceeding 2:1 to 3:1.

The default background species is AIR with 40 % HUMIDITY.  If you truly want your domain to have no moisture at the start, just set HUMIDTY=0 on MISC.  You don't need to define a DRY AIR species. All you need is on SPEC input for WATER VAPOR. For your Suppy SURF define the MASS FLUX of AIR and the MASS FLUX of WATER VAPOR.

The MASS_FLUX you give on SURF is applied to every VENT. It is not divided up over all VENTs.  So you are trying to inject 11 kg/s at dozens of locations. Since each VENT is on the order of 0.1 m^2, 11 kg/s over 0.1 m^2 is not in line with low Mach number assumption.

Ultimately your grid size is going to define what you can and cannot resolve via a CFD model. A VENT can't be smaller than one grid cell in area. FDS rounds the VENT XB to the grid size.  If a dimension gets rounded down to zero, that VENT won't appear. You will need to make sure your VENT XB are large enough to be seen by the grid. 

In the end, I think you will need a lot more grid cells than you are currently using. 

[Siva D]

Hi

My problem was nothing to do with the moisture already present in the air. I just wanted more moisture in the air that was being supplied through the vent.

I managed to achieve it by using SPEC_ID(1) = 'WATER VAPOR' and MASS_FRACTION(1) = 0.012 on the &SURF line. 

I understand that the VENT size cannot be smaller than one grid. So, is there any way to circumvent this issue, like is there any other alternative means to model this problem without using a really small grid ??

And finally, could you please tell me why you thought the LOCALIZED LEAK approach might help me initially ? Because it might lead me to some other solution.

Thank you.

 

[dr_jfloyd]

If the real size of your vent is smaller than the current grid you have two options:

1. Make the vent bigger so it occupies at least one grid cell. This option will obviously introduce some error into your calculation. It would be up to you to assess if that error was significant or not in terms of what you are trying to do with the simulation.

2. Make the grid smaller until the vent occupies at least one grid cell. This option obviously requires more time and resources for the calculation.

The first time I read it, I read it as you were trying to capture leakage from a large building. You aren't, so using the leakage model wont help.