Society of Fire Protection Engineers

45 South Volume 2 Issue 1 February 1995

The President's Column

FIGHTING FIRES BY SIMULATORS

With the ability to carry out fire safety design of buildings using fire models and computers, it should not be too long before fire simulation programs will appear so that firemen can train on computers. Simulators already exist to train airline pilots, why not the same for firemen?

Mathematical fire models can reproduce many of the chemical and physical processes that occur within and around a fire. ASET-B tells one how long it takes for a room to become filled with smoke, and what the average smoke temperature and fire intensity will be at any time. DETACT, enables one to determine when a heat detector or a sprinkler will activate.

A more elaborate model can be run using FIREFORM, in FPETOOL. FIREFORM incorporates ASET-B and DETACT within its software. FIREFORM can also modify the fire depending on the fuel, the heat loss through the walls and ceiling and tack oxygen, carbon monoxide and carbon dioxide levels.

Fire models can be used to quantify a hazard. Most people would consider having a 20 litre can of petrol in the lounge as a hazard. The question is how much. A fire model can provide a comparison between the 20 litres of petrol and the furniture suite in the lounge. Similarly, a 10 litre container of flammable solvent on a factory workbench if knocked over and spilled would burn away in say 2 minutes. In a 10 m x 15 m workroom with a 3 m high ceiling, smoke would fill the room down to eye level in about 1½ minutes, and the average temperature would be about 100°C. If the regulations only permitted 2 litres of solvent to be the limit permitted in that size of workroom, fire modelling can show that the resultant fire will only last 45 seconds, the smoke layer would be above eye-level when the fire ceased and the average smoke temperature would only be 70°C. Knowing the difference between the two fire scenarios, a fire inspector can make an appropriate decision as to hazard to occupants and hence restrict the amount of flammable solvents on hand. The same fire modelling process could also be used to demonstrate that a larger volume than 2 litres might be permissible in a larger sized workroom. Further evaluation of the 10 litre situation would indicate that with standard response sprinkler protection, the first sprinkler head would operate in 63 seconds, and a quick response sprinkler in 20 seconds. In a similar manner, Fire Service personnel can already do pre-fire evaluations of various hazardous occupancies. By adding the time needed to respond, travel to the scene and mount the fire attack, fire-fighters could develop realistic fire plan and training exercises.

The only missing component to running a fire simulation is the effect that hoses will have on the fire. We can already determine the amount of water that will be sprayed on the fire when a sprinkler activates and while the conservative assumption is generally made that the fire will remain at a steady state intensity from that point on, in fact the fire will be suppressed and a downward suppression curve could be introduced into the computer program to assess final knock-down of the fire by a sprinkler system. If we could also determine suppression curves for hose jets and fog nozzles, a useful step would be available to finalise fire simulation training programmes.

Imagine sitting down at a computer terminal, and a three-dimensional drawing of the chosen building is displayed. The fire origin point is selected and whether or not you want detectors, and which type. You select how much water will be available from street mains and how many hoses you will use. The screen has clock, thermometer, smoke height, oxygen and carbon monoxide monitors down one side. As you watch the computer draws a rising smoke column perhaps in colours to represent temperature, and then spreads it through the room and then out the doors and windows. The clock ticks off the seconds. You are advised of the activation of the first smoke detector. The computer indicate when the first and last persons have left the building and when the Fire Service leaves the fire station and arrives at the scene. You can implement various rescue scenarios and also firefighting scenarios, ie water jets or fog nozzles. The computer responds and you see the fire get smaller or larger on the screen. This training scenario is not 10 years away in the future. It is almost here. The necessary technology exists today. All that is needed is to agree on the effect of firemen's hoses on fire. This could be evaluated using such a program, starting with a low efficiency of water use, then carrying out back-analyses on actual fires in buildings raising the efficiency until the results match.

All the pieces for making a fire simulator training tool don't fit together yet, but they soon will. Then, instead of playing computer games for fun, firemen can ply them for real. Pre-fire planning by computer is just around the corner. Cliff Barnett

There still seems to be a large body of opinion amongst those in our field that designers using the Building Code Acceptable Solutions are in some way absolved of responsibility because they are 'approved'. Sure the Acceptable Solutions are referred to as 'the approved documents' by the BIA, but since the BIA has no liability for errors in the documents the approval isn't worth much to a designer in actual practice.

A building owner who submits a building consent application (including a fire safety report) has to meet the requirements of the Building Act and remains liable for any loss as a consequence of errors or omissions in the Acceptable Solutions regardless of their approval by the BIA. If the owner has engaged someone to produce the fire safety report, that person 'sub contracts' that part of the liability for errors or omissions in the fire safety report regardless of their origin.

There is a natural limit on the fire engineer's liability - he (or she) can only be expected to apply knowledge available to him (or her) at the time. In other words if new research published after the design shows that the method used was wrong but the method was 'the best available' at design time the engineer has met his responsibility for 'Due Care' and any losses can be considered 'accidental'.

As a case in point new evidence was presented to New Zealand fire engineers in Christchurch and Auckland during December 1994, which would seem to make part of the acceptable solutions obsolete and possibly unsafe under certain circumstances. I refer to the lectures given by Jonathan Sime on the human behaviour factors associated with evacuation. His research has shown, for example, that one cannot 'start the evacuation' as soon as a fire detector initiates the evacuation sounders, further that the type of audible warning used effects the start delay. Also, people are observed not to follow the logical escape route but are more likely to use the way they came in.

It would seem commercially unwise for fire engineers to continue blindly using table 3 in C2 / AS1 which assumes instant response of building occupants to an evacuation signal and optimum egress path selection.

P.S. What happened to all those problems with the building code, and their resolutions of course, that we were going to air in the newsletter. I haven't heard from anyone on this topic...are you telling me that there are no problems between those doing design and those reviewing designs for the Territorial Authorities ?

FIRE ENGINEERING DESIGN OF INDUSTRIAL AND OTHER MIXED USE BUILDINGS

Industrial and associated mixed use buildings present particular challenges to fire engineers with respect to compliance with the performance requirements of the New Zealand Building Code (NZBC).

The First Schedule to the Building Regulations 1992 in Clause A1 - Classified Uses, defines industrial use as applying "to a building or use where people use material and physical effort to:

(a) extract or convert natural resources

(b) produce goods or energy from natural or converted resources

(c) repair goods, or

(d) store goods (ensuing from the industrial process)

and gives as examples:

"an agricultural building agricultural processing facility, aircraft hanger, factory, power station, sewage treatment works, warehouse or utility"

Of the 35 clauses, B1 to H1 inclusive, of the Building Regulations 1992, the clauses which the fire engineering design needs to address are typically:

C1 Outbreak of Fire

C2 Means of Escape

C3 Spread of Fire

C4 Structural Stability During Fire

C1 Access Routes

F3 Hazardous Substances and Processes

F6 Lighting for Emergency

F7 Warning systems

F8 Signs

G9 Electricity, and

G11 Gas As An Energy Source

The Building Regulations permit allow for design and construction of buildings for the specific intended use.

ie A warehouse designed and constructed for the bulk storage of steel products such as structural and non-structural steel members, reinforcing steel, sheet steel etc, may be quite different to a warehouse designed and constructed for the bulk storage of say paper products.

Why? - Because the product in the steel storage warehouse is typically non-combustible and therefore presents a low fire hazard, whereas paper is combustible and presents a high fire hazard.

When designing industrial and/or associated mixed use buildings, it is important that the intended use is known and that the design reflects same.

It also follows that where alterations are being carried out on, or change of occupancy occurs to, an existing industrial or mixed use building that the fire risk classification used for the original building design be established to enable the fire rating behaviour of the building to be checked for adequacy in its intended new use.

The acceptable solutions to NZBC Clauses C2, C3 and C4, published by the Building Industry Authority classify high fire hazard occupancies as an occupancy having an average fire load energy density (FLED) greater than 1500 Mj/m².

What is the FLED and how is it calculated for a particular building ?

The total fire load stored in a firecell E (MJ) is the sum of all the energy available for release when the combustible materials are burned, given by:

E = S k_cMh_a

Where S indicates the sum of the different types of fuel

M is the mass of the fire load (kg)

h_a is the net calorific value of the fuel (MJ/kg)

k_c is the proportion of the fire load available to burn in the time under consideration

This gives rise to FLED, ie Fire Load Energy Density, which is E/A MJm² where A, is the floor area of the firecell in m²

Examples of individual FLEDS are:

Wood, at ambient moisture content has a FLED of around 13 MJ/kg. Dry paper and cardboard 16-17 MJ/kg and rubber tyres 32 MJ/kg.

The acceptable solutions however, don't provide a lot of guidance on design of buildings for FLEDs › 1500 MJ/m² and thus it is the responsibility of the designer to assess the total fire load and design the building to meet the performance requirements of Clauses C2, C3 and C4 of he NZ Building Code. The design procedure is essentially a "trial and error" process to analyze the likely effects of a fire given the worst likely location and time of ignition. Knowledge of the fuel loads, the number and location of occupants and the fire protection features is essential for assessing whether the performance criteria are met.

The first two steps are to determine the geometry, construction and use of the building and to establish performance requirements. The following steps revolve around scenario analysis, considering all possible scenarios.

Some parts of the analysis can be quantified with numbers, but much of the analysis requires subjective judgement as to the likely movement and consequences of a fire and the likely location and movement of people.

If the performance criteria are not met, then either the building geometry or the fire protection features must be modified until satisfactory performance is achieved. This process must be repeated for all possible fire scenarios.

To achieve the performance requirements of the Building Code, will require fire engineering design for the majority of industrial and associated mixed use buildings. Richard Brand

The week long SFPE training course Feb 13-17 is the focus of much attention as this issue goes to press, however I'm delighted to see the excellent quality of contributions to this quarter's issue. Please volunteer your ideas and articles, I'd rather not resort to as much arm twisting for the May issue.

There has developed a 'de-facto' branch of our chapter in Christchurch around the staff and students of the fire engineering course at Canterbury University. Andy Buchannan seems to be the prime contact there and he has contributed the following to the newsletter.

CHRISTCHURCH NEWS ITEMS

SFPE Activities

There have been two recent meetings of the SFPE in Christchurch, held in association with the NZFPA, the IFPE and the Canterbury Structural Group.

SFPE Chapter president Cliff Barnett spoke on fire separation distances, comparing the Acceptable Solutions with the results of his calculations for radiant heat transfer from building to building. Mr John Henry of the Christchurch City Council and Mr Jonh Sinclair of the Fire Service gave their views of the fire spread problem. It is apparent that submissions should be made to the BIA to clear up anomalies in the Acceptable Solutions.

Visiting UK consultant Dr Jonathon Sime gave an afternoon seminar with Mr Hamish MacLennan of Holmes Consulting Group, both speaking on human behaviour in fire. One of the clear messages was that people threatened by fire need information, the more the better. Voice assisted evacuation was promoted for large buildings where early warning of the fire and evacuation instructions can drastically reduce the time before people start to move towards exit ways. The indecision time is oftem much longer than the actual travel time which is usually calculated.

M.E. (FIRE) DEGREE

The Master of Engineering in fire engineering at the University of Canterbury is about to get into its second full year of operation. In 1994 there were 5 full time students and several part time students. Of the five students graduating in March 1995, only one is yet to find employment.

Preliminary enrolments for 1995 show five full time students, and 15 part time students, four in Christchurch, eight in Auckland, two in Wellington and one in Hamilton. Most of the part time students from outside Christchurch will be attending a week long block course in Process Safety and Reliability at Canterbury in late February, and will take a heat transfer or physics course in their home town in 1995. The specific fire engineering papers will be taught by video and remote learning in Auckland and Wellington in 1996 and 1997.

The current M.E.(Fire) graduates have been completing research projects in Christchurch. Ivan Bolliger and Darin Miller have been carrying out full scale burns in a converted shipping container at the Fire Service Training Facility at Woolston. They have been trying to simulate flashover and backdraft with both methane and wood crib fires. Tony Enright has been investigating the instrumentation needed in these and similar large scale test facilities. Faran Rahmanian has been looking at the advantages and disadvantages of residential sprinkler systems, with the help of local sprinkler contractors and others. Hans Gerlich has been testing full scale steel framed walls in the BRANZ / BTL furnaces, as part of his investigation into the fire performance of Gib Board walls on light steel frames. Dr A Buchanan

The subject of branches of the chapter was discussed at the last executive meeting, with the conclusion that all such activity should be encouraged. Should you wish to organise meetings in your area the executive will do what it can (within budgetarry restraints) to assist. ...Ed