Small Heptane Pool fire FDS & Experimental Data don't match up

[Samiyah]

Hi,

I am working on modeling a small heptane pool fire and comparing the FDS results to experimental data for the same case. 

Pool area = 0.007 m^2 (0.084 m x 0.084 m)

Pool depth = 0.007 m^2

Attached is the input file, where I am using multiple meshes with the finest resolution at the pan (x and y split into 0.0084 m grid cells, and y into 0.00175 m grid cell size). Fuel property values are from multiple sources including the Technical Reference Guide and NIST data. 

The fire burns out much sooner in FDS than in the experiment. In FDS the fire is out before 150 seconds and in the experiment the fire lasts for approximately 300 seconds. I have played with the grid resolutions more by using cubic grids and having a larger fine mesh domain, placing the fuel in a steel pan (like the heptane case in the Validation model), but am stumped as to why the fire is burning out so soon. 

Your advice would be most appreciated!

-Samiyah

[Kevin]

Do you get similar results when you use larger grid cells?

[Samiyah]

Yes, but there is only a slight difference: the finer one burns longer by just a few more seconds.

[Kevin]

You are using different properties than we are in our Pool_Fires example case. What happens when you use the same properties? Try plotting the burning rate vs pool diameter. Do you see the same result as discussed in, for example, the SFPE Handbook? I have never tried this myself. I don't know if the heat conduction and edge effects of your small pool are being modeled properly.

[dr_jfloyd]

You have defined your pool thickness as 7 mm.  To achieve a burning duration of 300 s implies an average burning rate of 0.015 kg/m^2/s. This much lower than I would expect for heptane. Did you input the correct depth?

[Samiyah]

Kevin-- I'll run another case tonight using the same properties to see if there are any changes. To plot the burning rate vs pool diameter, per SFPE Handbook and Quintiere's "Enclosure Fire Dynamics" text, the mass loss rates documented are for a certain range of pool diameters for heptane. The minimum for the range is 0.2 m diameter pools. So I would need to look at the experimental data to back out the plot/trend. 

dr_jfloyd-- 50 mL was in the pan and matches the thickness of 7 mm. 

[Samiyah]

I ran the new case, keeping everything the same except for the heptane properties. The fire was out in just under 100 seconds.

[dr_jfloyd]

Looking at your inputs, you don't have any substrate beneath your heptane pool which means you don't have a steel pan and floor to act as a heat sink.  You are not using the species N-HEPTANE so the gas phase properties of the fuel vapor are not correct.  Change your REAC from ID='N-HEPTANE' to FUEL='N-HEPTANE' and get rid of the C and H (you also don't need the HEAT_OF_COMBUSTION on the MATL. There is no need to specify a different HoC than the REAC since it is the same chemical species). These will have some effect but probably not enough. My guess is that one issue is that the effective conductivity of a pool of heptane is greater than the material conductivity since there will be some convection in the pool. FDS is not modeling the second phase, and therefore, is not predicting that convection.   

[Samiyah]

I had previous cases run, with a floor and another with a steel pan and the floor, with the same results. I have BURN_AWAY=.TRUE./ is there a change in the physics when this line is included or is it only an added visual feature for Smokeview?

(I am currently running a case with the changes to the REAC line and the MATL section as suggested.)

[dr_jfloyd]

Did you have the steel pan and floor as part of the SURF with the pool (e.g. three layers on the SURF)?

[Samiyah]

No, I had set them up with the OBST lines:

For the steel pan with fuel in it --- &OBST XB=1.65, 1.75, 2.10, 2.2, 0.2, 0.25, SURF_IDS='N-HEPTANE POOL', 'STEEL SHEET', 'STEEL SHEET'/

And the floor -- &OBST XB=0.50565, 2.876, 0.9843, 3.29135, 0.1167, 0.2, SURF_ID='PLATFORM', COLOR='BRICK'/

&SURF ID        = 'STEEL SHEET'

      COLOR     = 'BLACK'

      MATL_ID   = 'STEEL'

      THICKNESS = 0.01 /

&SURF ID = 'N-HEPTANE POOL',

    COLOR = 'GREEN',

    MATL_ID = 'N-HEPTANE FUEL'

    THICKNESS = 0.04

BURN_AWAY=.TRUE. / 

[dr_jfloyd]

See the User's manual discussion on multi-layered surfaces.  Your SURF needs to have the layers not the OBST>