Modeling of a warehouse wood pallet

[Chiara Crosti]

Hello everyone,

I would like to ask you help about the model of a warehouse wood pallet I’ve been dealing with.

I placed the pallet stacks uniformly and to each ones I gave the proper RHR curve. I started the fire in one pallet stack and I defined that if the close wood stacks are reaching the temperature of 275 C then they are involved by fire as well.

 

The compartment is 20mx75mx6m, with windows of 66 mq and combustible of 387 500 kg, what looks weird to me is that the fire is extinguished in only 15 min and that when the FDS analysis ends I got a final RHR value of 550 MW while the Temperature is only of 100 C. I am attaching the RHR curve and the Temperature-time curve of the right columns.

 

Could you please give me any suggestions?

 

Thank you so much.

[Dave McGill]

Hi Chiara,

We'll need to see the input file.

Dave

[Chiara Crosti]

Sure.

I attached the input file. There few comments in italian I hope you don't mind.

Thank you very much.

Chiara

[Dave McGill]

Hi Chiara,

I ran your simulation overnight; it's run for a bit over 300 s.

At that point, the fire appears to be spreading to the second row of obstructions. 

I suspect the fire is going out prematurely because there are two competing forces controlling the burning of the wood pallets.

Here is the SURF.

&SURF ID                = 'PALLET_3.6'

      COLOR             = 'YELLOW'

      THICKNESS         = 0.01

      EMISSIVITY        = 0.9

      MATL_ID           = 'WOOD'

      HRRPUA            = 7531.0

      RAMP_Q            ='fire_ramp_3.6'

      IGNITION_TEMPERATURE = 275 /

&RAMP ID='fire_ramp_3.6', T = 0 , F = 0 /

&RAMP ID='fire_ramp_3.6', T = 50 , F = 0.02 /

&RAMP ID='fire_ramp_3.6', T = 100 , F = 0.06 /

&RAMP ID='fire_ramp_3.6', T = 150 , F = 0.14 /

&RAMP ID='fire_ramp_3.6', T = 200 , F = 0.25 /

&RAMP ID='fire_ramp_3.6', T = 250 , F = 0.38 /

&RAMP ID='fire_ramp_3.6', T = 300 , F = 0.55 /

&RAMP ID='fire_ramp_3.6', T = 350 , F = 0.75 /

&RAMP ID='fire_ramp_3.6', T = 400 , F = 0.98 /

&RAMP ID='fire_ramp_3.6', T = 403.3 , F = 1 /

&RAMP ID='fire_ramp_3.6', T = 1116.1 , F = 1 /

&RAMP ID='fire_ramp_3.6', T = 1842.2 , F = 0 /

The ramp should control the fire, and have it burn for ~1800 s. However, a thickness has also been provided, with a value of only  0.01 m. FDS may be calculating that the fuel has been depleted before the ramp reaches its endpoint. (Perhaps one of the developers will pipe in as to whether or not that makes sense.) I suggest you run a couple of small test cases, with and without a value for thickness, and see if that is what is happening.

I'm not sure why your column temperature is only 100 C. I'm seeing values of ~280 for the columns near the fire in the AST boundary file (see attached image).

As an aside, if you run the complete model again, then I suggest you increase the value of RESTART=20. Saving the restart files every 20 s is ultra-conservative, and will really slow down the simulation. (I would change it to at least 200 s)

Dave

 

[Chiara Crosti]

Thank you so much Dave for you help.

According on what you wrote I just tried to run the simulation deleting the thickness but FDS came out with a warning saying that a thickness has to be provided.

What I am doing now it is to try to run the analysis with a bigger thickness to see if it is there the problem.

I will let you know. If you have any other suggestions please let me know.

I thank you again for your precious help.

Chiara

[Dave McGill]

I'd forgotten about the need for a thickness value. Trying with a much bigger value makes sense.

[Chiara Crosti]

Hello Dave,

unfortunately using a bigger thickness does not make any difference. Do you have other suggestions?

I really cannot see the mistake I am making!!

Thank you so much for your help.

Chiara

[Dave McGill]

Hi Chiara,

I just checked the oxygen levels in your original simulation and they drop significantly at ~ 500 s. That is the major reason why the HRR is much less than expected. See the attached graph. If you take the roof off, I suspect you will get the values you are expecting.

See the attached file. I've simplified things to see if the material will ramp up to a requested value, and it does. I did remove the Heat of Combustion for the material WOOD, as it is not necessary. 

I noticed that the peak HRRPUA for the wood pallets is 10,864 kW. I suggest you check that value, it is very high for wood. 

Dave

[Dave McGill]

Chiara,

On Wednesday, 16 December 2015 13:11:10 UTC-5, Dave McGill wrote:

...

I noticed that the peak HRRPUA for the wood pallets is 10,864 kW. I suggest you check that value, it is very high for wood. 

I had second thoughts about my comment on the peak HRR and checked the SFPE Handbook. The value you have is on the high side, but credible.

Dave 

[Chiara Crosti]

Thank you again Dave for the time you spent on my simulation. I agree with you the problem on the fire duration  is the lack of oxygen.

Nevertheless what it is still not clear to me, it is why the RHR curve decreases after 10 minutes but then it increases again. It's this last increase that it is confusing to me, I attached the curve.

What do you think it could be?

Chiara

[Dave McGill]

Hi Chiara,

That is when the compartment becomes air-limited. See the two attached images of the HRRPUV, which show the combustion regions before and after the transition.

The second image shows something else that is probably affecting the overall HRR. The flame front is flush with the  Xmin  and Zmax domain boundaries. This indicates that unburned fuel is probably exiting the domain. If  the domain was extended in both directions I suspect the overall HRR would be considerably higher.

Dave