Rapid temperature drop in room during fire

[quantumfds]

Hello,

In one of the fire scenarios, there is a large heated room with an initial internal temperature of 20°C and an external temperature of 0°C (winter). This is defined using:
&INIT ID='Init Region', TEMPERATURE=20.0, XB=.../ for the room, and the outside temperature is set with &MISC TMPA=0.0/. Is this put in the code right?

As expected, the temperature around the inlet openings quickly drops from 20°C to 0°C. However, the overall cooling of the room happens much faster than expected. With a 5 MW fire and openings relatively close to the fire, the smoke temperature decreases, which is expected. But the issue is that the entire room, with a large volume of about 30,000 m³, almost reaches 0°C, and the smoke temperature drops to around 10°C after just 10 minutes of fire.

The cooling seems too rapid for such a short time. The inlet openings are about 30 m², and the outlet openings are around 40 m². Is this cooling rate typical due to the large openings, which account for roughly 1% of the room’s ground floor area? In this case, the natural smoke extraction is proving to be highly inefficient.

Thanks.

[drjfloyd]

INIT only sets gas temperatures. All the solid surfaces start at TMPA. At the start of your calculation you will immediately convect heat fom the warm room to the cold walls.

You can specify the intial wall temperature profile. See the User's Guide. You can do a quick hand calc to get the steady-state wall temperature profle with the inside 20 C and the outside 0 C.

[quantumfds]

Thank you. I guess I need to define the initial temperatures of the inner and outer walls, since they are currently cold.
I am not sure if the code is written properly but I tried to put the temperature profile in the surfaces:

&MATL ID='CONCRETE',
FYI='NBSIR 88-3752 - ATF NIST Multi-Floor Validation',
SPECIFIC_HEAT=1.04,
CONDUCTIVITY=1.8,
DENSITY=2280.0/

&SURF ID='Concrete Inside Walls',
RGB=255,203,101,
TRANSPARENCY=0.498039,
RAMP_T_I='Concrete Inside Walls_RAMP_T_I',
MATL_ID(1,1)='CONCRETE',
MATL_MASS_FRACTION(1,1)=1.0,
THICKNESS(1)=0.4/
&RAMP ID='Concrete Inside Walls_RAMP_T_I', T=0.0, F=20.0/
&RAMP ID='Concrete Inside Walls_RAMP_T_I', T=0.4, F=20.0/

&SURF ID='Concrete Outside Walls',
RGB=255,203,101,
TRANSPARENCY=0.498039,
RAMP_T_I='Concrete Outside Walls_RAMP_T_I',
MATL_ID(1,1)='CONCRETE',
MATL_MASS_FRACTION(1,1)=1.0,
THICKNESS(1)=0.4/
&RAMP ID='Concrete Outside Walls_RAMP_T_I', T=0.0, F=20.0/
&RAMP ID='Concrete Outside Walls_RAMP_T_I', T=0.1, F=15.0/
&RAMP ID='Concrete Outside Walls_RAMP_T_I', T=0.2, F=10.0/
&RAMP ID='Concrete Outside Walls_RAMP_T_I', T=0.3, F=5.0/
&RAMP ID='Concrete Outside Walls_RAMP_T_I', T=0.4, F=0.0/

Is this correct? Or could I just put TMP_INNER=20 to get some more realistic drop of inner temperature in the room?

[drjfloyd]

One of the best ways to learn FDS is to get in the habit of making simple test cases that run in a few minutes. Here all you need to do is establish that the correct wall temperature profile is set in initialization. You could have a very simple case with one temperature &PROF for an exterior wall and one for an interior wall. This would have given you the answer in a few minutes rather than waiting for someone to respond in the forum. I suggest you try that as practice and see.

TMP_INNER=20 would not get a more realistic drop. In the real world, heat is being conducted through the walls. Setting the entire wall to 20 C would stop that real world behavior, and thus, be less realistic. Depending on the specifics of what you are simulating and why, setting TMP_INNER=20 could be either a conservative or a non-conversative setting.