E_COEFFICIENT; use of sprinklers

[San-FDS_USER]

Question regarding an approach to modelling sprinkler effect on a specified HRR - in particular WHAT  E_coefficient to use? I would like to start by giving a general description of my simulation.

I am modelling a fast-growing 30 MW vehicle fire ( max HRR in 15mins) in a road tunnel - Fuel is Diesel with a manually input Heat of Combustion 25 MJ/KG and a soot yield of 0.05 kg/kg; the aim of the simulation is to understand the length up to which stratification is maintained. My tunnel is ~700 m and the head-room is only 5 m. I have found that the stratification is better with mesh refinement and although i cannot say that i have a fully mesh-independent solution, my best result so far has been through using 0.1 m cell resolution in the z direction and 0.25 m in the x&y. I suppose the aspect ratio is just about okay.

Overall, I am trying to ensure whether  the thick smoke layer is high enough (un-disturbed) during the early fire growth such that passengers could escape underneath it. Please feel free to comment and ask more if  my approach looks doubtful. However my actual question is below. 

- Though I am relatively happy with the current results of my simulation, i wanted to see how adding sprinklers could benefit the design and therefore present a case for sprinkler installation.Theory says that cooling the fire through water suppression is likely to reduce the total heat release rate by almost 40-50 %.Some papers I have read but then go on to say that although the total HRR can be reduced, the stratification can be disturbed so my aim of passenger escape underneath the fire during the first phase can be undermined. This makes sense.

I have used FDS for activation of sprinkler heads(several of them) with a nominal discharge of 90l/min from each nozzle. I have used other default FDS numbers for sprinkler velocity/spray angles etc etc which i can mention if necessary.  However, if a specified HRRPUA is applied (as it is in my case), I don't think the effect of the reduction of the HRR can be seen without specifying an extinction_coefficient. One simulation that i carried out without specifying anything actually did reduce the HRR and the MLR but i believe that this was more as a result of reduction of oxygen levels and not because of the actual cooling effect of the fire source. I do not know what is a practical E_COEFFICIENT value to use. Is 4 a good guess?  The FDS manual 14.6.2 tells me that this is the right approach but can someone advise on a correct value to use for this. Early results i am seeing are quite desirable in the sense that the smoke spread is limited to a zone 50 m either side of the fire but the visibility is not that great. In areas outside this this 50m zone visibility is much better? Is this result practical, Can you please advise on a good realsitic E-Coefficent value or any other source?

 Man Thanks in advance.

Sincerely, Yours

San

[Kevin]

The E_COEFFICIENT was introduced in FDS as a simple way to model sprinkler suppression. However, it requires that you actually burn cars under a calorimeter and then apply various amounts of water spray and then infer an E_COEFFICIENT from your experiments. Otherwise, there is no tabulated or recommended values.

The reason for this should be obvious -- if you are burning a vehicle, how much of the water is expected to reach the actual burning material? Half? A third? Who knows -- it depends on the car or truck. It is not practical to compile these E_COEFFICIENTs for every conceivable fuel source.

[San-FDS_USER]

Kevin,

Thanks for your prompt response. Yes okay that is understood. So essentially it seems impractical for me to use sprinkler system for on FDS for this calcualtion. 

Also, do you think my understanding that the reason for the HRR  redcution currently being observed is correct ( reduction of Oxygen). 

If specifying a E_COEFFICIENT is not possible; I was thinking of limiting the growth to a certain HRR depeding on the time of sprinkler activation to mimic the effects of a ssystem as such . So say for instance if based on my fire growth (30MW IN 15MINUTES) , the hrr is  15MW by the time the sprinklers turned on (assume 5 minutes), then i run a calculation with a 15 MW ramp up to 5 minutes and then limit the HRR growth for the remaining 10 minutes.

[Kevin]

If you use HRRPUA, then the only way for the HRR to be different from your specified time history is due to oxygen limitation.

As I understand it, and others may comment, most practitioners choose to hold the HRR steady after sprinkler activation, as a way of simulating a "control mode" sprinkler. The idea is that the fuel may keep burning, but the fire will probably not continue to spread, assuming the sprinkler system is properly designed.

[dr_jfloyd]

See 15.6.2 in the User's Guide for a technique to limit fire growth when a sprinkler activates.