Derivation of fire sizes

[thebuildinginspector]

Hi all,

I'm a learner at the dark arts of fire engineering, so be gentle and please excuse any ignorance on my part! Here goes, whenever I’ve read through any fire engineering reports in relation to SHEV’s and fire sizes the fire always seems to be given a perimeter, an area and a peak heat release rate.  The perimeter and the area being essential to use in the smoke production calcs further down the line and I've found good examples in BRE 368 and Table 1 of BS7346-4 , but, for example, a restaurant building I’m looking at doesn’t fall into any of these categories in the table for either fire size or the size of the building .
 
In that case I thought it best to derive my own, so I figured out that there might be two approaches for an unsprinklered fire, either
 
1.In lieu of convective heat release data, to calculate the fire size at the critical time I could work out the RSET and then substitute the seconds needed into the t squared fire growth equation which would give a fire size, but it's pretty unrealistic as the fire is assumed to continue to grow and not reach a steady state and doesn't take into account the fuel is likely to be consumed and change to ventilation controlled. However, I could use the perimeter and area of the room and assume that the whole room is involved.  This leads to a MASSIVE fire size and is unrealistic.
 
2.This leaves me with the option of using MJ/m2 or kW/m2 density information from CIB W14 or 7974 -Part 1  , but then I need to define the area of the fire to work out the total output - which is difficult without assuming a fire that consumes all the fuel in the room (100m2), leaving me with an unrealistic fire size as it assumes that all the room is involved at once!


My brain hurts.  Any thoughts?
 

[Phoenix]

The approach depends upon what you are aiming to demonstrate. 

If you're trying to get a rough idea of the time to reach some stipulated tenability limit (such as, for example, base of smoke layer down to 2.5m from the floor) then the t squared fire growth formula should give a rough but reasonable indication of this time.  This is not usually "unrealistic" when making comparisons of ASET v RSET because we are only examining the early stages of a fire before it becomes ventilation controlled or before it reaches a fully involved steady state.

If you're looking at later development of the fire, after everyone has escaped from the immediate vicinity, then why would it be unrealistic to assume that the fire would involve the whole room or compartment?  That seems reasonable to me for an unsprinklered building, especially if there is the possibility of ventilation occurring by window breakage or similar.

You can look at the heat content of the materials in the room but you will have to make some simplifying assumptions that may or may not be realistic.  This is a route that ties in with some specific aims.  What are your aims?

Stu

[thebuildinginspector]

Stu - Thanks for your response

I've got a situation with a conversion of a listed building to a mixed use development that has a restaurant area at ground floor level and sleeping accommodation above, with the occupants to some of the bedrooms having to pass through the atrium on open balcony/walkways to enter the protected stair at 2nd floor level.

I've looked through the exemplars in BS 9999 (Exemplar 22) and it suggests that the accommodation at all levels must be separated from the walkways, which it is, except for ground floor level where it can't be separated because it is heavily listed - BS 9999 also suggests that sprinklers are provided, but then because of the listing of the ceiling it's very difficult to do this.  The fire load in the base of the atrium is controlled and meets the description in 9999, but the fire load in the rooms that face off the atrium isn't controlled.

In order to demonstrate acceptability the only way I can think to do this is to prove that a fire in the rooms at groundfloor level off the atrium wil ensure tenable conditions (temp, visibility or FED) long enough for the sozzled hotel guests to escape. The RSET for the occupants to traverse the walkway, crossing the atrium, and then enter the protected stair I have calculated to be around 1400 seconds, therefore the fire has the potential, using t2 fire growth, to be enormous at this point.

You're right, an unsprinklered fire would spread to the extent of the room or compartment and I thought I would need a steady state fire to do the smoke calcs, but I guess using the heat release rate of a t2 fire at that point at which an occupant enter the atrium would be fine; although the fire is assumed to still be growing at that point and not fully developed and a steady state.  I think this might just have to be a limitation the study - if anything it provides a degree of conservatism in the design.

[Phoenix]

The t squared approximation for fire growth is only appropriate up to flashover, as you're aware - typically a few minutes.  After that it's wrong.

The rooms that adjoin the atrium may be assumed not to be capable of spilling smoke into the atrium if they are separated from the atrium by fire resisting construction.  If this is the case then you have a quick fix and you only have to deal with the smoke from the controlled fire load(s) in the base of the atrium.

If the rooms that adjoin the atrium are not separated by fr then there are models available for smoke spilling from doorways or similar openings.  How, specifically, the fire is behaving in the room of origin is not a dominating factor - probably more important is the size of the opening.  Have a look in BR368 and 7974-2 for models but bear in mind that these are not designed to be used for fill time calculations.  Some may be incorporated into fill time analyses but may not be valid if the atrium is large (greater in area than, say, 500 - 1000m²).

1400 seconds sound about right for sleeping guests in a hotel but it is possible to vastly improve this in a number of ways.  If sufficient staff are available at night they can do a round of the occupied rooms to rouse all the people.  Having bedroom lights come on with the alarm helps, as do vibrating pillows or headboards for everyone - makes it hard to stay in bed (or more fun for some maybe).

Stu

[thebuildinginspector]

Stu - Thank you for your reply

The rooms facing out on to the base of the atrium aren't protected from the base and unfortunately can't be because of the heritage of the building.  I intend to have a stab at the calcs for a fire in the base of the atrium, with a controlled fire load and also with a car in at as it's a function/exhibition space and also to run one with a fire in rooms off the base of the atrium to look at the affect of spill plume, a kind of sensitivity analysis I suppose, for the location.

I'd agree with your point about the opening, that it's most likely to affect the volume of smoke along with the depth of the balcony and thinking it through logically the size of the fire can't be all that important. I suppose the fire may well have started to decay looking at the RSEST when occupants enter the atrium and by using a t2 fire for calcs at this point it will be a worst case scenario and adds a good safety margin.

[kurnal]

Can I make a couple of points- neither of which are intended as criticisms. Just  observations from a non engineer who often sees fire engineered solutions some of which are excellent and some of which are based on questionable principles. If I need fire engineering solutions I use one of two engineers to advise me - one of whom is Phoenix. So I am not having a go. I respect the science and the competence of practicioners.

The first point is summed up in Building Inspectors comment:

"In order to demonstrate acceptability the only way I can think to do this is to prove that a fire in the rooms at groundfloor level off the atrium wil ensure tenable conditions (temp, visibility or FED) long enough for the sozzled hotel guests to escape."

I would suggest that using fire engineering you can prove anything you want to and pull the wool over the eyes of  many of the  enforcement agencies. As we used to say "Bull**** baffles brains". I would argue that the process should start not by setting out the goal and choosing calculations and methodology that will support that goal, but instead the engineering analysis should seek to identify all potential problems and from there move on to seek and analyse alternative  solutions. I very often see solutions that have been engineered from the goal backwards which then will only work for a very narrow set of parameters and circumstances.

Secondly the issue of pre movement time. The studies that have been carried out are useful but in practical terms they are probably mathematical expediency and means rather than useful data. In your case- 1400 secs RSET. I would suggest that someone who hasnt started to make their escape 20 minutes after the fire alarm has sounded is not likely to leave under their own volition so it is pointless trying to engineer an ASET as long as this. I feel sure this data is simply based on mathematical averages and as Phoenix suggests this is the area in which practical measures need to be taken to minimise pre movement time.

 

[SamFIRT]

Human behaviour is the issue here not engineering; both the guests and the staff…. including the temporary staff for the exhibitions.

I believe this scenario is a disaster waiting to happen!

[thebuildinginspector]

Kurnal - I've only been involved with Fire Engineering in any great detail for the last couple of years and I'd agree with your sentiments over the use of it - proceed with caution.  You're right, bulls**t does baffle brains and I've been subjected to a few presentations and received a few strategies that have relied on "prettiness" and style over substance.  I might be being a bit controversial, but I also get the impression that some practices also trade on their reputation and that I'm met with an attitude of, "what could you possibly know" when occassionally querying some aspect of the report.  The VAST majority are great and I love working with fellow professionals, but there are a few bad apples - I guess the same can be said of some Building Control Body's!

I also agree with your comments on the derivation of the goal of the study and as a matter of course I usually look at carrying out a desktop risk assessment to determine the most realistic worst case scenario of fire and also occupant affected - meaning that the design will cover a range of fire loads, ignition risks and occupant mobility or immobility.  This has got to be the starting point, otherwise, as you say, the strategy is for such a narrow set of parameters it's unlikely to work.

I share your concerns over the pre-movement time, but I've got to attribute a number to this and it's the only figure I can really use.  It's back to the "management" chestnut I suppose, as in sleeping risk buildings where people are unfamiliar, this seems to be the variable that changes the pre-movement time value.  It doesn't sit too comfy with me either!

[thebuildinginspector]

SamFIRT - I'm sure all reasonable steps will be taken to ensure that it's not, but as you say it's difficult to account for human behaviour.

[CivvyFSO]

Using the full room as a perimeter is also wrong for another reason. Where the perimeter of the fire is used in calculations its impact within the calculations is one regarding the entrainment of air, and the effect that this entrainment of air has on the smoke plume. To assume that the perimeter of the fire is at the walls negates the possiblity of entrainment at these points. This is the reason why a corner plume is calculated differently, leading to less smoke produced, but much more heat.

BS7346-5 might be of some help, there is an explanation of how to employ the t² fire growth to calculate room filling times and temperatures. Basically, get a spreadsheet program, start your fire off on the t² basis, over the next 'x' number of seconds this fire will grow to its new fire size, take these changing fire sizes and calc the mass of smoke, then the temp of the smoke, then the volume of smoke created over those 'x' seconds. This volume of smoke is then assumed mathematically to lie evenly at roof level, thus subtracting from the height of rise for the next iteration. Fire grows again, more smoke is created, height of rise drops again. And so on. (Check the BS for the actual method though, as it is not 'quite' that simple)

There are a few methods to calculate the time to flashover, but there are certain methods that look at the heat release rate required for flashover to occur in a room of given dimensions and given materials. This could be factored in to the above, basically giving you a specific output limit for flashover. i.e. The maths will be wrong from that point onwards.

Also, IMHO BS9999 is a potential liability when it comes to atria guidance.

[thebuildinginspector]

CivvyFSO - Thanks for your advice, it's greatly appreciated!

Drew