OBST + HOLE vs HVAC

[appheadspace]

Which is better approach to simulate horizontal opening between 2 enclosures?

I have tried 2 approaches and compared volumetric flow rate with experimental value (~10^-6)

1. obst + hole -- this is giving high volumetric flow(~10^-4) rate than experiment at steady state.
2. HVAC duct -- this is giving lower volumetric flow(~10^-8) rate than experiment at steady state.

Kindly help.

[appheadspace]

PFB input details

!----------------------------
! Mesh: two 3 ft (0.9144 m) cubes stacked with 1 inch partition, total height ~1.8542 m
Upper and lower compartment walls maintained at constant cold and hot temps
!----------------------------

!----------------------------
! OBST + HOLE
!----------------------------

&MESH ID='Lower_Comp_Mesh', IJK=36,36,36, XB=0.0,0.9144,0.0,0.9144,0.0,0.9144/
&MESH ID='Partition', IJK=36,36,3, XB=0.0,0.9144,0.0,0.9144,0.9144,0.9398/
&MESH ID='Upper_Comp_Mesh', IJK=36,36,36, XB=0.0,0.9144,0.0,0.9144,0.9398,1.8542/

&OBST ID='Partition', SURF_ID='partition_surf', XB=0.0,0.9144, 0.0,0.9144, 0.9144,0.9398, OUTLINE=.FALSE./

&HOLE ID='Opening_hole', XB=0.3048,0.6096, 0.3048,0.6096, 0.9144,0.9398 /

&DEVC ID='Q_opening_z_0.9144', XB=0.3048, 0.6096, 0.3048, 0.6096, 0.9144, 0.9144, QUANTITY='W-VELOCITY', SPATIAL_STATISTIC='AREA INTEGRAL' /
&DEVC ID='Q_opening_z_0.9229', XB=0.3048, 0.6096, 0.3048, 0.6096, 0.9229, 0.9229, QUANTITY='W-VELOCITY', SPATIAL_STATISTIC='AREA INTEGRAL' /
&DEVC ID='Q_opening_z_0.9313', XB=0.3048, 0.6096, 0.3048, 0.6096, 0.9313, 0.9313, QUANTITY='W-VELOCITY', SPATIAL_STATISTIC='AREA INTEGRAL' /
&DEVC ID='Q_opening_z_0.9398', XB=0.3048, 0.6096, 0.3048, 0.6096, 0.9398, 0.9398, QUANTITY='W-VELOCITY', SPATIAL_STATISTIC='AREA INTEGRAL' /

!----------------------------
! HVAC Vents & Network
!----------------------------

&MESH IJK= 36,36,73, XB=0.0,0.9144, 0.0,0.9144, 0.0,1.8542/

&SURF ID='hot_wall_surf', COLOR = RED, TMP_FRONT=32.222 /
&SURF ID='cold_wall_surf', COLOR = BLUE, TMP_FRONT=-6.667/
&SURF ID='partition_surf', COLOR = GRAY, ADIABATIC = .TRUE./

&OBST ID='Partition', SURF_ID='partition_surf', XB=0.0,0.9144,0.0,0.9144,0.9144,0.9398, OUTLINE=.FALSE./

&VENT ID='VENT_LOWER', XB=0.3048,0.6096, 0.3048,0.6096, 0.9144,0.9144, SURF_ID='HVAC'/
&VENT ID='VENT_UPPER', XB=0.3048,0.6096, 0.3048,0.6096, 0.9398,0.9398, SURF_ID='HVAC'/

&HVAC TYPE_ID='NODE', VENT_ID = 'VENT_LOWER', ID='LOWER_NODE', DUCT_ID = 'Opening'/
&HVAC TYPE_ID='NODE', VENT_ID = 'VENT_UPPER', ID='UPPER_NODE', DUCT_ID = 'Opening'/

&HVAC TYPE_ID='DUCT', ID='Opening', NODE_ID='LOWER_NODE','UPPER_NODE', AREA=0.0929, LENGTH=0.0254, LOSS = 0.4,0.8/

&DEVC ID='Opening_Q', DUCT_ID='Opening', QUANTITY='DUCT VOLUME FLOW'/

[mcgratta]

What units are you using for flow rate? Do you mean 10^6 = 1 000 000 m^3/s ?

[appheadspace]

yes m^3/s

[drjfloyd]

Somewhere you have an error.

1 million m3/s through an 0.3 m x 0.3 m hole is Mach 30,000 or 3 % of light speed.

[appheadspace]

oh sorry they are minus powers.

Experimental value : ~10^-6 m3/s
obst + hole : ~10^-4 m3/s
HVAC duct : ~10^-8 m3/s

[drjfloyd]

Without knowing the details of the experiment cannot say which is the better option. The better option is which ever more closely represents the physical construction of the experiment.

~40 C temperature difference is not a very strong source of buoyancy. The accuracy with which you represent the actual experimental conditions are going to matter.

[drjfloyd]

I have delete your post. DO NOT post copyrighted material to the forum. This is a violation of copyright law and risks the FDS project being deleted from github. If you need to refer to published work provide a DOI and/or citation.

[appheadspace]

I was not aware of that. Sorry for that and Thank you for the intimation.
PFB reference.
DOI : https://doi.org/10.1016/0017-9310(62)90185-0

It is 3ft cubical compartments one above another with cold and hot temp maintained in upper and lower compartments. The partition separates them and heat and air exchange happens in the square opening in the center. I have taken 1 inch partition with 12x12 inch opening. Kindly help.

[drjfloyd]

When modeling an experiment, details matter. You have not provide a complete input file but just looking at your mesh lines, they do not seem adequate to represent the actual experimental apparatus which was larger in interior volume than the total volume of your mesh lines.

You are measuring volume flow; however, according to the paper the only measured values are temperature above and below the opening and power consumption for the heater. In a sealed domain with a partition between hot and cold sides the volume flow should be essentially zero. At steady-state the net mass flow at the opening is zero. The warm air will be less dense so it will require more volume flow but with a 40 C temperature difference the excess amount in the upwards direction is going to be small. The paper estimates the volume flow using a 1D momentum equation for unidirectional flow with an assumed orifice coefficient; however, the actual flow in this case is complex and may not be adequate represented by the assumption made.

[appheadspace]

hvac.txt
obst_hole.txt

I did reverse calculation to find volumetric flow rate.

1. Taking Gr from plot and given experimental conditions, I have calculated temperatures
2. Then based on Heiselberg correlation using temperatures and geometry, I have calculated volumetric flow rate.

Let me know if u need to see total calculation or any other info. Thank you.

[drjfloyd]

My initial comments stand.

Your FDS input file is not the same as the experiment. The experimental apparatus shown in Fig 2 of the paper is not two stacked boxes with all the lower walls heated and all the upper walls cooled. You have made the signifcant assumption that modeling the experimental apparatus which consisted of one wall of the upper chamber open to a cool reservoir and one wall of the lower chamber open to a reservoir of warm fluid is equivalent to your inputs.

A correlation is not measured data. Correlations involve assumptions and simplificaitons and introduce uncertainty which can be difficult if not impossible to quantitfy. The experiment measured the heat input at the heater and temperatures inside the box. Your first step should be seeing if you are getting the correct total heat transfer rate between the hot and cold sides.

The actual walls of the appartatus are not perflectly adiabatic. There will be some heat transfer across the walls. This may not be a significant effect but it will be a source of added uncertainty if not accounted for.