Effect of Sprinklers on HHR

[Ken]

Hi, I am trying to model how the application of water through
sprinklers affect the fire growth rate and heat release rate.

Currently I have a 10m x 15m x 5m enclosed room with a 1 MW fire
(medium t2 fire) in the middle. The sprinkler is on the top of the
room above the fire.

The goal of my simulation is to generate a curve that displays the
heat release rate over time. Sprinklers were activated at around 200
seconds into the simulation, but looking at the output files, the heat
release rate did not go down until 600 seconds after the application
of the sprinklers. The heat release rate should go down with the
application of sprinklers and it should happen fairly quickly. Also,
when I checked the temperatures from the output files, I can see that
the temperatures are going down.

Could you point me in the direction of how the water is used to reduce
the heat release rate? As enthalpy is conserved, is the heat from the
fire just equated to the sensible heat increase in the water up to its
boiling point. Also I believe there is sufficient oxygen to sustain
combustion.

Any help is appreciated, thanks.


The following is the code that I used:

&TIME T_END=1500 /

&MESH IJK= 20,20,20, XB=0.0 , 10.0 , 0.0 , 15.0 , 0.0 , 5.0 /

&MISC SURF_DEFAULT='WALL', POROUS_FLOOR=.FALSE. /

&SURF ID = 'WALL'
COLOR = 'BLACK'
Transparency = .3 /

&SURF ID= 'BURNER' , HRRPUA=1000, TAU_Q=-288 /

&OBST XB= 0.00 , 10.0 , 0.00 , 0.00 , 0.00, 5.00 , /
&OBST XB= 0.00 , 0.00 , 0.00 , 15.0 , 0.00 , 5.00 , /
&OBST XB= 0.00 , 10.0 , 15.0 , 15.0 , 0.00 , 5.00 , /
&OBST XB= 10.0 , 10.0 , 0.00 , 15.0 , 0.00 , 5.00, /
&OBST XB= 0.00 , 10.0 , 0.00 , 15.0 , 5.00 , 5.00, /
&OBST XB = 0.00 , 10.0 , 0.00 , 15.0 , 0.00 , 0.00 /


&VENT XB= 5.00 , 6.00 , 7.00 , 8.00 , 0.00 , 0.00, SURF_ID='BURNER' /

&PART ID='water drops',
WATER=.TRUE.,
DIAMETER=750,
SAMPLING_FACTOR=1, /

&PROP ID='K-1',
QUANTITY='SPRINKLER LINK TEMPERATURE',
ACTIVATION_TEMPERATURE= 75,
OFFSET=0.10,
FLOW_RATE=200,
SPRAY_ANGLE=30,80,
DROPLET_VELOCITY=1.2,
PART_ID='water drops',
SMOKEVIEW_ID='sprinkler_upright', /

&DEVC ID='Spr_15', XYZ=5.5,7.5,5.0, PROP_ID='K-1', /

&DEVC XB= 0,10,0,15,0,5 , ID='HRR', QUANTITY='HRR', SETPOINT=0 /
&DEVC XB= 5,6,7,8,0,0.20 , ID='HRR2', QUANTITY='HRR', SETPOINT=0 /
&DEVC XYZ= 5.5,7.5,0 , ID='my timer', QUANTITY='TIME', SETPOINT=0 /

&DEVC XYZ= 5.5,7.5,5 , ID='Sprinkler Temp', QUANTITY='TEMPERATURE'

&DEVC XYZ= 5.7,7.7,5 , ID='Temp Fire 1', QUANTITY='TEMPERATURE' /
&DEVC XYZ= 5.3,7.3,5 , ID='Temp Fire 2', QUANTITY='TEMPERATURE' /

&SLCF PBX=5.5, QUANTITY='TEMPERATURE' /
&SLCF PBY=7.5, QUANTITY='TEMPERATURE' /

&BNDF QUANTITY='AMPUA', PART_ID='water drops' / accumulated mass per
unit area

&BNDF QUANTITY='NET HEAT FLUX' /

&DEVC QUANTITY='AMPUA',
PART_ID='water drops',
ID='Water mass',
STATISTICS='SURFACE INTEGRAL',
XB= 0,10, 0, 15, 0,0 /



&TAIL /

[Dave McGill]

Hi Ken,

There are two problems. First, your fire is a burner and the sprinklers
will have no effect on it. Second, you have a sealed box and the decline
in HRR you are observing is due the decline in oxygen levels. If you run
your case without the sprinklers, I think you will see the same decline.

Regards

Dave

[Kevin]

Ask yourself this very important question -- what information are you
providing to the model with which it can predict the suppression of
this fire? The answer is -- very little. You essentially have a cold
gas burner releasing gas (which is not specified -- what is it?) at a
rate to give you a t-squared HRR. You spray water and then what should
happen? What would happen if you did this experimentally? What physics
are you trying to replicate? On what basis should FDS change the HRR?

Please read the Technical Reference Guide, Volume 1, to learn more
about the sprinkler model and its assumptions. Then ask questions if
certain aspects are not clear.

[Ken]

Thanks for your inputs. I went back and refined my model a bit. Now
my ignition source are two nozzles that spray out liquid
fuel(heptane), similar to the spray burner example. Also I have a
combustible load just above the liquid fuel. I set the nozzles to
stop spraying fuel after 20 seconds into the simulation. I ran the
simulation and the fuel ignites just after the simulation and moved on
to ignite the combustible load as well, which is what I expected.

Now my question is 'What does the FDS do to ignite the heptane fuel
after it is sprayed into the room?' From what I found, the ignition
temperature for heptane is 215 degrees Celsius, however, when I
introduced the heptane into the room, it is at ambient temperature (20
degrees Celsius). So, what mechanism in FDS causes the fuel to ignite
immediately?

Thanks.

[Dave McGill]

Hi Ken,

The quick answer to your question is "nothing". In FDS, if fuel and air
mix, they will ignite. (I will qualify that by saying that I think they
have to mix at the stoichiometric ratio.)

Dave

[Kevin]

Dave is right. Even ambient temperature droplets evaporate into fuel
gas. The fuel gas mixes with oxygen and burns, regardless of the
temperature. This is the model, of course. In reality, if you spray
heptane nothing would happen until you use some sort of ignition
source. So assume in FDS that there is always some sort of virtual
ignition source to start the reaction.

> >>> &TAIL /- Hide quoted text -
>
> - Show quoted text -

[Ken]

I wanted to verify that my model was mesh-independent, so I ran it
with 4 different meshes. The scenario is the following: A ignition
source starts in the beginning of the simulation and goes away after
20 seconds. On top of the ignition source is a combustible load. The
whole simulation was 50 seconds. The 4 different mesh sizes were 64k
cells, 125k cells, 200k cells, and 500k cells.

For the first 3 meshes, the Heat Release Rate (HRR) curves are pretty
much the same, the HRR raises for 20 seconds and decays to zero after
the ignition source goes away. However, for the finest mesh, 500k
cells, the HRR curve is completely different from the previous 3
meshes. For the 500k cell curve, the HRR keep going up even after the
ignition source goes away.

For each simulation, all I've changed was the mesh, everything else
was the same for each simulation. So my question is does making the
mesh finer change my results?

[dr_jfloyd]

Without knowing exactly what you did some of this is only conjecture:

For ignition to be sustained on a combustible material in FDS the heat
being conducted from the surface of the material into the material
cannot be enough to drop the surface temperature below the point where
the material will pyrolyze. It sounds like your coarse grids were on
the wrong side of that threshold (that the object did not heat quite
enough to sustain ignition) and with the fine grid that you crossed to
the other side of that threshold (that the object did heat enough to
sustain ignition). To illustrate with completely made up numbers,
let's say your object needs 25 kW/m^2 for 20 s to reach sustained
ignition and the predictions of your grids (from coarse to fine) were
23.0,24.5,24.9, and 25.1 kW/M^2. This would be a case where your grid
is clearly converging but happens to cross a threshold that some other
phenomena occurs.

What you need to do as the FDS user is add additional output
diagnostics to see what is happening (heat flux, temperature, burning
rate, etc.). You should also ask yourself how appropriate the
properties are of the combustible object for your simulation. Where
did you obtain them from? Have you verified that exposing that
material in a manner similar to how the properties were obtained
yields similar results?