Fire Scenarios

Abstract

The engineering approach to fire safety design requires the selection and evaluation of fire scenarios that may occur in a building. Each fire scenario represents a unique combination of events and circumstances that influence the outcome of a fire in a building, including the impact of fire safety measures. The SFPE Engineering Guide to Performance-Based Fire Protection [1] refers to fire scenarios as “a set of conditions that defines the development of fire and the spread of combustion products throughout a building or part of a building.”

Appendices

Appendix 1: Fire Scenarios in Risk Model FiRECAM

FiRECAM™ is a fire risk and cost assessment model developed by the National Research Council of Canada [19, 29]. As a result of simplifying assumptions made, the model is applicable to apartment and office buildings. FiRECAM calculates the expected risk to life and fire cost expectation based on a hazard analysis of a number of scenarios. These scenarios and their probability of occurrence are hard coded in the model.

The approach employed in FiRECAM is to consider only three generic fire types that represent the three distinct types of fires that may occur. They are (1) smoldering fires where only smoke is generated, (2) nonflashover flaming fires where a small amount of heat and smoke is generated, and (3) flashover fires where a significant amount of heat and smoke is generated with a potential for fire spread to other parts of the building. The design fires can occur on each floor of the building, and each fire could happen with the apartment door open or closed. In addition, scenarios are considered with occupants being awake or asleep, and suppression systems being effective in extinguishing the fire or not. Within each fire scenario analysis, the times of occupant response and evacuation are based on analysis of the impact of fire detection systems, alarm systems, and other possible perceptions that occupants may receive during the fire.

The probabilities of these three fire types, for both apartment and office buildings, were obtained for Australia, the United States, and Canada [30]. They were obtained based on independent analyses of fire statistics in these three countries. The definition of fire type is based on the severity of the fire when it was observed and recorded by the fire fighters on their arrival. Obviously, small fires can develop into fully developed, postflashover fires if they are given enough time and the right conditions. For risk assessment purposes, however, the fire conditions at the time of fire department arrival are the appropriate ones to use. They represent the fire conditions that the occupants are exposed to prior to fire department extinguishment and rescue operations. In the event of no fire department response, then the eventual conditions of the fire at extinguishment, either by itself or by occupant intervention, are the ones to be used.

The reason why fires can develop into different types with different probabilities is because they are governed by a number of random parameters that cannot be predicted, such as the type of ignition source, the point of ignition, and the arrangement of the combustibles. Table 38.3 shows the probabilities of the three fire types, after ignition, for apartment buildings. It is interesting to note that the probabilities are quite similar among the three countries, even though there is no reason that these numbers should be the same due to climatic and cultural differences. Table 38.3 also clearly demonstrates the importance of considering all fire types. For example, flashover fires, which can pose significant hazards to the occupants, have a relatively low probability of occurrence; whereas nonflashover and smoldering fires, which pose lower hazards to the occupants, have a higher probability of occurrence.

Table 38.3 Probabilities of fire types for apartment buildings

Design Fires

In addition to the random parameters, described in the previous section, that govern the type of fire that can develop, the condition of the door to the compartment of fire origin is another random parameter that also affects the fire growth. The fire type and the door condition can be combined to create six design fires that allow all the random parameters that govern fire growth to be easily considered. These six design fires are (1) smoldering fire with the fire compartment entrance door open, (2) smoldering fire with the fire compartment entrance door closed, (3) flaming nonflashover fire with the fire compartment entrance door open, (4) flaming nonflashover fire with the fire compartment entrance door closed, (5) flashover fire with the fire compartment entrance door open, and (6) flashover fire with the fire compartment entrance door closed. The probability of each of these design fires is the product of the probability of the fire type (Table 38.3) and the probability of the door to the compartment of fire origin being open or closed. The probability of the door being open or closed can be estimated based on experience. For example, the entrance door to an apartment unit can be assumed to be mostly closed (for security and privacy reasons), whereas the entrance door to an office room can be assumed to be mostly open (to allow work interaction).

The scenarios used in FiRECAM are shown in Fig. 38.6, which demonstrates the various parameters used that may impact fire development and smoke movement as well as occupant response. The model does not attempt to decrease the number of scenarios, although it is evident from the results that some scenarios such as the smoldering scenarios and flaming nonflashover scenarios contribute less to the overall risk to life. Important scenarios, as identified by the model, are the flashover scenarios with door open and sprinklers nonfunctioning.

Fig. 38.6

Event tree showing fire scenarios in FiRECAM

Appendix 2: Example Demonstrating Selection of Fire Scenarios

Example

The fire protection team is in the process of performing a performance-based design for a complex building with multiple occupancies, including a parking garage on the four floors below grade and shopping areas on the first four floors, which are interconnected through an atrium and a 20 story hotel tower. The complex is fully sprinklered with a central alarm and voice communication, and the atrium has a smoke exhaust system.

Solution

The ten steps identified in Table 38.1 are followed for the solution of this example.

• Step 1—Location of Fire

• A brainstorming session has identified the following fire locations to be considered in the analysis:

• • Fire in a hotel room

  • Fire in the underground parking garage

  • Fire in the atrium

  • Fire in the restaurant of the hotel adjacent to the hotel lobby

  • Fires in stores of the commercial area

• Step 2—Type of FireThe type of fire that may start at each location depends on the type of combustibles, fuel load, and ignition sources.

• Hotel Room One type of fire that may be expected in a hotel room are those that start with a cigarette thrown into a garbage container that ignites curtains and then spreads to a couch and bed. Another type of fire may start in a garbage can but then ignites the wood cabinet with a TV and clothing items in the drawers. Fire development for these two fires may be different, although after flashover both fires may have similar characteristics.

• Underground Parking Garage The type of fire expected in an underground garage is one that involves a car and then spreads to adjacent cars.

• Atrium In the atrium area, the expected fire could be a fire of a Christmas tree that is placed there during the holidays or a fire involving couches and tables located there.

• Restaurant In the case of restaurants, the fire may start in the kitchen area or it may start in the sitting area. These two types of fire are different.

• Commercial Area Fire in this area could potentially start in any store. This may result in different types of fires depending on the combustible materials and fire loads in each store as well the size and ventilation characteristics. Examples of different fires in stores are fires in clothing stores, bookstores, and shoe stores. A survey of commercial stores done in 2004 has identified a number of different types of fires that should be considered for commercial areas [5, 6].

• Due to space limitations and to avoid repetition, only the fires in the hotel room, the parking garage, and the atrium are considered in the remainder of the example.

• Step 3—Potential Fire Hazards

• For this type of occupancy, no special hazards are anticipated. However, the authorities having jurisdiction may request consideration of arson or hazards as a result of functions or events that may be held in the atrium space. This could include exhibitions and displays of goods and merchandise.

• Step 4—Systems Affecting Fire

• The building is fully sprinklered with central alarm with voice communication. In addition, the atrium and the parking garage have smoke management systems. The effects of these systems should be considered.

• Step 5—Occupant Response

• Consideration is given here to the response of the occupants to the various warnings and their likelihood to extinguish the fire. For this example, a probability of response and effectiveness in extinguishing the fire is assigned as 0.3 for all fire scenarios.

• Step 6—Event Tree

• For each fire location, an event tree is constructed so that the different fire scenarios can be identified. Figure 38.7 shows the event tree for fire starting in a hotel room. As the figure shows, four scenarios are associated with this fire type. The probability of occurrence of this fire type could be obtained from statistics. For this example, however, it is assumed that the probability of fire starting at the three locations is the same. Probabilities for each of the events shown in the tree can also be obtained from statistical data; however, because this is a qualitative analysis, expert judgment can be used.

  Fig. 38.7

  

  Event tree for hotel room fire

• Figure 38.8 shows the event tree for the fire starting in the parking garage. This tree considers the events of manual suppression, sprinkler activation and effective control of the fire, effective smoke ventilation, and barriers that are effective in containing the fire. The fire in this location results in six fire scenarios.

  Fig. 38.8

  

  Event tree for parking garage fire

• Figure 38.9 shows the event tree for the atrium fire. It considers the same events as the parking garage fire, so it results in six scenarios. (Although it is possible that sprinklers operate but venting does not, for brevity, this potential scenario is not considered here.)

  Fig. 38.9

  

  Event tree for atrium fire

• Step 7—Consideration of Probability

• The probabilities of the various events shown in the event trees produced in Step 6 can be determined from statistical data and other sources. However, because at this stage of the process the analysis is qualitative, expert judgment can be used for the initial screening of the fire scenarios. For this example, the probabilities of each of the events will be described in qualitative terms and then converted to probability values to facilitate the calculation of the scenario probabilities. For this, the descriptions and values shown in Table 38.4 are used.

  Table 38.4 Values associated with probability descriptions

• The very high value of 0.95 corresponds to the probability of effectiveness of sprinkler systems in hotel rooms, whereas the value of 0.7 is associated with the probability of smoke detector activation. Using these values, the probabilities of the events of the event trees are assigned and the scenario probabilities are calculated, as shown in Figs. 38.10, 38.11, and 38.12.

  Fig. 38.10

  

  Probabilities of scenarios for hotel room fire

  Fig. 38.11

  

  Probabilities of scenarios for parking garage fire

  Fig. 38.12

  

  Probabilities of scenarios for atrium fire

• Step 8—Consideration of Consequence

• In this step, a qualitative evaluation of the consequence of each of the scenarios is performed. This evaluation is done using engineering judgment based on the type of fire, the location of the fire, and the effectiveness of the active fire protection systems. This evaluation considers the impact of the fire on both property as well as life safety. To facilitate this assessment, Table 38.5 shows the different consequence levels that are chosen for this example. The level is determined by considering both the property losses and the occupant impact. For example, the consequence level of a scenario with $30,000 in losses and serious injuries is “high.”

  Table 38.5 Consequence levels and associated loss estimates

• Based on the levels shown in Table 38.5 and considering the fire type, fire location, and effectiveness of the active fire protection systems, the consequences of the scenarios in the three event trees are determined as shown in Figs. 38.13, 38.14, and 38.15.

  Fig. 38.13

  

  Consequence of hotel room fire scenarios

  Fig. 38.14

  

  Consequence of parking garage fire scenarios

  Fig. 38.15

  

  Consequence of atrium fire scenarios

• Step 9—Risk Ranking

• Figure 38.16 presents the risk-ranking matrix developed based on the results of Steps 7 and 8. The matrix has six levels of probabilities of occurrence, from extremely low to very high, and five levels of consequence estimates. The levels for the probabilities of scenario occurrence for this example have been set as shown in Table 38.6.

  Fig. 38.16

  

  Risk-ranking matrix of identified scenarios

  Table 38.6 Scenario probability values used for risk-ranking matrix

• The three levels of shaded areas in Fig. 38.16 represent areas of different risk levels, with the darker area representing high risk and the lighter area representing low risk. The white areas represent very low-risk scenarios.

• As shown in Fig. 38.16, no scenario falls in a high-risk area. Scenarios S23 and S24 are moderate-risk scenarios and should be considered for quantitative analysis. Scenarios S25, S33, and S34 are low-risk scenarios that can also be considered further. In addition, Scenario S14, although it falls into a very low-risk area, may be considered for further analysis, as it is a scenario in a different section of the building with different fire protection systems and different impacts, and it has an extremely high consequence. All other scenarios do not require further analysis and can be dropped.

• Step 10—Final Selection and Documentation

• The final selection of the design fire scenarios is done in this step, and the fire scenarios are documented in detail. As indicated in Step 9, scenarios S14, S23, S24, S25, S33, and S34 should be considered for further analysis. To facilitate the quantitative analysis, Table 38.7 describes the characteristics of these scenarios. The quantitative analysis of these scenarios should consider both the impact of the fires on life safety and property. The procedure outlined in the section “Development of Fire Scenarios” in this chapter can be followed for the analysis.

  Table 38.7 Description of design fire scenarios