Your SlideShare is downloading. ×
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Nfpa 557
Upcoming SlideShare
Loading in...5
×

Thanks for flagging this SlideShare!

Oops! An error has occurred.

×
Saving this for later? Get the SlideShare app to save on your phone or tablet. Read anywhere, anytime – even offline.
Text the download link to your phone
Standard text messaging rates apply

Nfpa 557

913

Published on

Nfpa 557

Nfpa 557

Published in: Education
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total Views
913
On Slideshare
0
From Embeds
0
Number of Embeds
2
Actions
Shares
0
Downloads
51
Comments
0
Likes
0
Embeds 0
No embeds

Report content
Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
No notes for slide

Transcript

  • 1. NFPA 557 Standard for Determination of Fire Load for Use in Structural Fire ProtectionDesignChapter 1 Administration1.1 Scope. The scope of this standard is the determination of fire loading to be used as the basisfor the evaluation and design of the structural fire performance of a building.1.1.1* This document is not intended to address facilities for storage of hazardous materials.1.2 Purpose. The purpose of this standard is to provide standard methods and values for use inthe determination of design basis fire loads. This is done using a risk framework.1.3 Retroactivity. om1.4 Equivalency. .c1.5 Limitations. seChapter 2 Referenced Publications a gbChapter 3 Definitions in3.1 Distributed fire load. The overall fire load (in MJ) of the compartment. in3.2* Fire load density*. The heat energy that could be released per unit floor area of a tracompartment MJ/m2) by the combustion of the contents of the compartment and any combustiblepart(s) of the building itself. re .fi3.3 Localized fire load. The fire load at a location within the compartment that is outside thescope of normal variations in the distributed fire load w w3.4 Structurally significant fire. A fire that grows to a size that poses a threat to the structural welementsChapter 4. Design Fundamentals4.1 The methodology developed in this standard provides a risk based design fire load for use indesign and evaluation of structural fire performance.4.2 The design basis fire load is determined by the statistical distribution of fire loads inbuildings, the fire initiation frequency, and the effectiveness and reliability of the fire protectionfeatures that contribute to fire control in the early stages of the fire. Page 1 of 29
  • 2. 4.3 The fire load determined by application of this standard shall be utilized as the basis for thedetermination of the design basis fire exposure to the structure. The actual determination of thedesign basis fire exposure is outside the scope of this standard.4.4 The statistical distribution of the fire load of the building shall be determined by 1) statisticalsampling and analysis in the subject or similar buildings, or 2) through the use of suitableoccupancy based fire load statistics (mean and standard deviation) provided in Chapter 6.4.5 The frequency of fire initiations in the building shall be determined from national statisticalstudies of fire incident data.4.6 The effectiveness of fire protection features in controlling fires before the fire becomesstructurally significant shall be assessed through both functional and national statistical analysis. omChapter 5. Development of Fire Loads5.1 Types of fire loads .c se5.1.1 Fire loads shall be calculated as both localized fire loads and distributed fire loads.5.1.2 Determination of fire loads. a gb5.1.2.1 Fire loads shall be determined as an input to design of structural fire resistance. in5.1.2.2* If the values contained in chapter 6 are used, then the designer shall confirm that the inanticipated fire loads will not exceed the values contained in chapter 6. tra5.1.2.3 Prior to If there is a change in occupancy, then the building owner shall evaluate the fire reload in the new occupancy. .fi5.1.2.4 If there is a change in occupancy, and the fire load in the new occupancy exceeds the fire wload that was originally developed, then the fire resistance of the building shall be analyzed toevaluate if the existing passive fire protection meets the design objectives for the new woccupancy. If the objectives are no longer met, then modifications shall be made as necessary so wthat the building meets its fire resistance objectives. [ 5.1.2.3 and 5.1.2.4 will probably move toanother chapter at some point]5.1.3 Distributed fire loads.5.1.3.1 Distributed fire loads shall be determined to reflect the total fire load throughout acompartment.5.1.3.2 Distributed fire loads shall be determined in accordance with section 5.3 or chapter 6.5.1.4 Localized fire loads.5.1.4.1* Localized fire loads shall be determined to reflect concentrations of combustiblematerial that may pose a more severe thermal exposure than the thermal exposure that would Page 2 of 29
  • 3. result from the uniform fire load. Combustible materials shall be considered as beingconcentrated whenever the mass per unit area of one or more items is a factor of 2.5 greater thanthat established distributed fire load.5.1.4.2 Localized fire loads shall be determined in accordance with section 5.4.5.2 Structurally significant fire frequency.5.2.1 Methodology.5.2.1.1 The structurally significant fire frequency shall be developed by estimating the rate offires per year relative to numbers of buildings and area of floor space, for buildings of similaroccupancy to the building being designed.5.2.2.2* The structurally significant fire frequency shall be determined by multiplying the rate of omreportable fires per year per floor area by the fraction of fires that are structurally significant inbuildings with similar construction and fire protection systems as are proposed for the building. .c5.2.2.3 The structurally significant fire frequency shall be developed as stated in section 6.2. se5.2.2 Limitations. The limitations of the applicability of structurally significant fire frequencyestimate shall be addressed. a gb5.2.3 AHJ Approvals. Structurally significant fire frequency estimates shall be subject to the inapproval of the AHJ. in5.3 Fire load density. tra5.3.1 The fire load density shall consist of the sum of the fixed fire load and the contents fire reload divided by the floor area of the compartment. .fi5.3.2* Fixed fire load. The fixed fire load shall consist of all combustible material in or on the wwalls, floor and ceiling. This shall include power and telephone cables, plastic light fittings, wtelephones, doors, frames, trim, and carpeting. w5.3.3* Contents fire load. The contents fire load shall consist of all moveable contents andoccupant possessions within a compartment. This shall include desks, cabinets, papers, andbooks.5.3.4 The total fire load in a compartment shall be computed using the following equation:Q = ∑ k i mi hciWhere, Q = total fire load in a compartment (MJ), ki = proportion of content or building component i that can burn, mi = mass of item i (kg), and hci = calorific value of item i (MJ/kg). Page 3 of 29
  • 4. 5.3.5* The fire load in a compartment is expressed as fire load density, that is, the total fire loadper square meter of the floor area, Q′′ (MJ/m2), given by:Q ′′ = Q / A f Where A f = floor area of the fire compartment (m2).5.3.6* For items that are made of materials derived from wood, energy potential shall bedetermined be multiplying the mass by a published calorific value. In the absence of publisheddata, it shall be acceptable to use a value of 15 MJ/kg for all products that are derived fromwood. For all other materials, it shall be acceptable to determine the energy potential bymultiplying the mass by 40 MJ/kg.5.3.7 Methodology & Limitations5.3.7.1 Fire loads shall be determined by conducting a survey of one or more buildings or by omusing the values in chapter 6. .c5.3.7.2 * The fire load survey shall be conducted by either the weighing technique or theinventory technique, or a combination of the two. se5.3.7.3* Sample Size Determination: If fire loads are determined by conducting a survey, diverse acompartments shall be surveyed and a confidence interval shall be constructed. If the results of gbthe fire load survey will be applied to multiple buildings, then surveys shall be conducted inmore than one building. For design purposes, confidence intervals of no less than 99% inconfidence interval shall be used. in5.3.7.4 Where the inventory technique is used, the value(s) used for mass densities shall be trasubject to the approval of the AHJ. re5.3.7.5 The results from a fire load survey shall only be applied to the building in which it was .ficonducted or to similar buildings. w5.3.8 The fire load density determined in accordance with this section shall be subject to the wapproval of the AHJ. w5.3.9 Results of the fire load survey reported shall include the mean, standard deviation, and acumulative probability distribution for the energy content per unit area. The fixed and contentsfuel load survey reports shall be individually reported.5.4 Localized fire loads5.4.1 Localized fire loads shall be determined based upon surveys as described in section 5.3.7or upon architectural design data.5.4.2 The localized fire load determined in accordance with this section shall be subject to theapproval of the AHJ. Page 4 of 29
  • 5. 5.4.3 Localized fuel loads shall be reported by location in the building and shall include theexpected value and a measure of the variability of the value (mean and standard deviation).Chapter 6 Occupancy Based Fire Load6.1 Defining the compartment6.1.1 The compartment shall be selected as the building, or portion of the building that isbounded by exterior surfaces of the building and within fire rated boundaries that are capable ofcontaining a fire for the entire duration through burnout.6.1.2 For areas where there are no fire rated boundaries, then the entire building shall beselected. om6.2 Structurally significant fire frequency .c6.2.1* For office buildings, the structurally significant fire frequency shall be taken as 5 fires permillion square meters per year. se6.2.2 For office buildings, the structurally significant fire frequency shall be determined by amultiplying the fire frequency in 6.2.1 by the value in Table 6.2.2 that corresponds to the gbconstruction type and fire protection systems specified for the building. inTable 6.2.2 Fraction of Fires that Are Structurally Significant in Office Occupancies in No detectors Detectors present No detectors Detectors present tra Type of Construction No sprinklers No sprinklers Sprinklers present Sprinklers presentFire resistive 0.13 0.07 0.04 0.03 reProtected 0.15 0.06 0.05 0.03noncombustible .fiUnprotected 0.19 0.10 0.07 0.05noncombustible wProtected ordinary 0.21 0.10 0.03 0.04 wUnprotected ordinary 0.30 0.17 0.11 0.07Protected wood frame 0.30 0.18 0.13 0.08 wUnprotected wood 0.37 0.20 0.12 0.07frame6.2.3 For other types of buildings, the structurally significant fire frequency shall be developedfrom the data in Annex D.6.2.4 As an alternative to the above procedures, other published data can be used subject toapproval of the applicability of the data by the AHJ.6.3* Fire load density6.3.1 The average fire load shall be the sum of the average fixed fire load and the averagecontents fire load, calculated as follows: Page 5 of 29
  • 6. Q f = Q f , f + Q f ,cWhere Q f = average fire load (MJ/m2)Q f , f = average fixed fire load (MJ/m2)Q f ,c = average contents fire load6.3.2 The standard deviation of the total fire load σf shall be calculated from the standarddeviations of the fixed fire load σf,f and contents fire load σf,c as follows:σf = (σ 2 f,f + σ 2 ,c f ) omWhere σf = standard deviation of fire load (MJ/m2) .cσf,f = standard deviation of fixed fire load (MJ/m2) seσf,c = standard deviation of movable contents fire load (MJ/m2) a gb6.3.3 Movable Contents fire load.6.3.3.1 For office buildings, the average contents fire load shall be 600 MJ/m2 floor area, and the instandard deviation shall be 500 MJ/m2 floor area. in tra6.3.3.2 For other types of buildings, the contents fire load shall be determined by conducting asurvey. re6.3.4 Fixed fire load. .fi6.3.4.1 The fixed fire load shall include all combustibles which are part of the construction and wthe architectural/interior design. w w6.3.4.2 For buildings of combustible construction, the fixed fire load shall include the quantitiesof combustible materials that are used in the building construction.6.3.4.3 For buildings of non-combustible construction, the average fixed fire load shall be 130MJ/m2 and the standard deviation shall be 40 MJ/m2.6.4 Design Fire Load6.4.1 The design fire load shall be determined to achieve the risk performance criteria stated bythe code of record, using the methods described in this section.6.4.2 Where the code of record does not provide risk performance criteria for structural fireprotection, the risk performance criteria for structural collapse, RS shall be no greater than 10-6 /yr, unless another value is approved by the AHJ. Page 6 of 29
  • 7. 6.4.3 Design fire load calculation6.4.3.1 The design fire load, Lf, shall be determined from the frequency of structurallysignificant fires fSS, and the risk performance criterion as follows: σ f (0.577 + ln(− ln F )) 6Qf = Q f − π6.4.3.2 The cumulative probability function required to achieve the risk objective F shall becalculated as follows: RSF = 1− f SS om6.4.3.3 The frequency of structurally significant fires fss shall be calculated as the product of thefire frequency ff and the floor area Af as follows: .c sef ss = f f × A f a6.4.3.4 This section is applicable to the overall and the localized fire loads. gb in in tra re .fi w w w Page 7 of 29
  • 8. Annex A Explanatory MaterialA1.1.1 Examples of hazardous materials include combustible dusts, flammable and combustibleliquids, flammable solids, oxidizers and oxidizer-containing waste. Information on suchoccupancies is contained in NFPA 400, Hazardous Materials Code [NOTE: to be approved inJune 2009].A.3.2 The higher the value of the fire load density, the greater the potential fire severity anddamage, as the duration of the burning period of the fire is proportional to the fire load.A.5.1.2.2 When selecting a fire load for a building or other structure, the building owner shouldconsider the possibility of later changes in occupancy or use which could result in greater fireloads than originally contemplated. In the event that the fire load increases beyond thatcontemplated during the design, reanalysis must be performed, and it may be necessary to ommodify the fire protection applied to the building.A.5.1.4.1 The factor of 2.5 was selected based on the “Z” value used in confidence intervals. A .c“Z” value of 2.5 approximately corresponds to a confidence interval of 99% (based upon a senormally distributed variable). aA.5.2.2.2 Fires that extend beyond the room of origin are considered to be structurally gbsignificant since they represent compartment fires which flashover and spread to additionalspaces. Sprinklers or other automatic extinguishing equipment, automatic detection equipment, inand construction will influence the likelihood of a fire extending beyond the room of origin.Fires that extend beyond the room of origin are considered to present a significant challenge to inthe structure. Therefore, statistics or data for buildings with fire protection systems or traconstruction methods similar to those proposed may be used to determine the structurallysignificant fire frequency. The structurally significant fire frequency for a given building type reand set of fire protection systems and construction methods may be greater than or less than thestructurally significant fire frequency for the building type with data for all combinations of fire .fiprotection systems and construction methods aggregated. See A.6.2.1 for more information. w wA.5.2.2 Limitations may arise from several sources. Fire frequencies are culturally influencedor determined. The fraction of fires reported can vary with country and jurisdiction, depending wupon local customs and regulations. Actual fire frequencies can also vary by country based uponcultural differences and differences in regulations concerning potential ignition sources.A.5.3.2 This category consists of items such as built-in structural elements, floor coverings,cupboards, doors and frames, and windowsills. Other fixed items such as skirting boards andwall switches are generally ignored because they provide a small contribution to the total heatrelease.A.5.3.3 This category includes items such as furniture, computers, televisions, books and papers,pictures, telephones, rubbish bins and personal effects. This category involves all the items thatcan easily be taken out or put into the fire compartment. Page 8 of 29
  • 9. A.5.3.5 Note that some fire load data sources report the fire load densities based upon thecompartment bounding surface area rather than the floor area. Care is required to understand thebasis of any values in the literature.A.5.3.6 The values of 15 MJ/kg and 40 MJ/kg were selected as bounding values for cellulosicmaterials and plastics, respectively. These values were selected based on effective (sometimesreferred to as “chemical”) heats of combustion as published in Tewarson, A. “Generation ofHeat and Chemical Compounds in Fires,” SFPE Handbook of Fire Protection Engineering,National Fire Protection Association, Quincy, MA, 2002.A.5.3.7.2 The weighing technique and the inventory technique are discussed in Annex C.A.5.3.7.3 To construct a confidence interval, the sample mean, x , shall be calculated byaveraging the results from each of the compartments surveyed. Similarly, the sample standard omdeviation can be calculated as follows: .c ∑ (x − x ) N 2 se iσ= i =1 N a gbWhere σ = standard deviationxi = fire load from ith sample inx = average of all fire load samples inN = number of fire load samples traThe confidence interval can then be calculated as follows: re σ .fix±z N w wFor a 99% confidence interval, z = 2.57. It should be noted that the size of the confidence winterval may decrease if the sample size is increased due to the presence of the square root of thesample size in the denominator. However, if the sample has significant variability, the size ofthe confidence interval may not decrease below a limit value.A.6.2 For more information, see annex D.A.6.2.1 The fire frequency of 5 fires per million square meters per year represents the frequencyof fires that are reported. A fraction of the fires that start will extend beyond the room of origin.The basis for this fire frequency is provided in Section G of Annex D.A.6.2.2 These values were determined from US National Fire Statistics using spread beyond thecompartment of origin as a surrogate for structurally significant.A.6.3 See Annex B for derivation of the values used in this section. Page 9 of 29
  • 10. A.6.4.3 The design fire load equations are based on a Gumbel distribution (Type I distribution oflargest values) for fire loads. This distribution is widely used for gravity loads and has beenverified for fire loads by Korpela, K & Keski-Rahkonen, O., Fire Loads in Office Buildings,”Proceedings – 3rd International Conference on Performance-Based Codes and Fire SafetyDesign Methods, Society of Fire Protection Engineers, Bethesda, MD, 2000. om .c a se gb in in tra re .fi w w w Page 10 of 29
  • 11. Annex B Summary of Occupancy Based Fuel Load Survey DataThe fire load densities in chapter 6 were developed by identifying and assimilating fire load datafrom a number of sources: • Ingberg, S., et al., “Combustible Contents in Buildings, NBS, 1957. • Caro, T. & Milke, J. “A Survey of Furl Loads in Contemporary Office Buildings,” NIST Report GCR-96-697, NIST, Gaithersburg, MD, 1996. • Baldwin, R., et al., “Survey of Fire Loads in Modern Office Buildings – Some Preliminary Results,” BRS, 1970. • Green, M. “A Survey of Fire Loads in Hackney Hospital,” Fire Technology, February, 1977, pp. 42 – 52. • Anon, “Building Materials and Structures – Fire Resistance Classifications of Building Constructions,” Report of Subcommittee on Fire Resistance Classifications of the Central om Housing Committee on Research, Design and Constriction, Report BMS 92, National Bureau of Standards, Washington, 1942 (this paper contains the same data as the Ingberg .c paper referenced above.) • Kumar, S. and C.V.S. K. Rao “Fire Loads in Office Buildings,” Journal of Structural se Engineering ASCE 123(3) March, 1997, pp. 365 – 368. • McDonald Barnett Partners, “Pilot Fire Load Survey,” Project #3580 CRB, Auckland, NZ, 1984. a gb • Lee, B. & Parker, J. “Fire Buildup in Shipboard Compartments – Characterization of in some Vulnerable Spaces and the Status of Prediction Analysis,” NBSIR 79-1714, National Bureau of Standards, Gaithersburg, MD 1979. (Data from this survey was not in used since it was based on shipboard compartments.) tra • Korpela, K & Keski-Rahkonen, O., Fire Loads in Office Buildings,” Proceedings – 3rd International Conference on Performance-Based Codes and Fire Safety Design Methods, re Society of Fire Protection Engineers, Bethesda, MD, 2000. • Culver, C. & Kushner, J. “A Program for Survey or Fire Loads and Live Loads in Office .fi Buildings,” NBS Technical Note 858, National Bureau of Standards, Gaithersburg, MD w 1975. • w Culver, C.G. (1976), “Survey results for fire loads ad live loads in office buildings,” Building Science Series No. 85, National Bureau of Standards, Gaithersburg, MD. w • Thauvoye, C., Zhao, B., Klein, J., Fontana, M. (2008), “Fire Load Survey and Statistical Analysis,” Fire Safety Science, 9, pp.There was a tremendous amount of variability among the fire loads published in the surveyscited. The reason for this variability appears to be that within a typical occupancy classification,e.g., business, there are a number of different types of usage among spaces, e.g., general office,storage, files, etc.Culver explored the effect of a number of factors affecting the fire load in office buildings,including room size, room use, building location (geographic), building age, building height, andgovernment vs. private occupancy. While all of these factors have some effect on fire loads,Culver found that the use of the room had by far the greatest influence on fire load. Page 11 of 29
  • 12. With the exception of the Ingberg paper, none of the other papers reported space usage asaccurately as the Culver report. The Caro report stated that the spaces surveyed were offices,although further investigation has revealed that what was reported as an “office” was, in at leastone instance, a cubicle (this was determined through discussions with one of the people whose“office” was surveyed.) The mass per unit area of a cubicle is not expected to be representativeof the mass per unit floor area of office space, so the Caro findings were not used in developingthe fire loads in chapter 6.Additionally, while the Ingberg report was more specific than others in terms of space usage, thesurveys that were used to generate the data were conducted from 1928 – 1940. One paper on fireloads in India (Kumar and Rao) suggests that between the 1970’s and the 1990’s, an increaseduse of steel furniture reduced office fire loads, so the Ingberg data is not likely representative ofcurrent fire loads. Therefore, it was not used in developing the fire loads in chapter 6. omIt is also noteworthy that some surveys only included the contents fire load, while others alsoincluded fixed items as well. Some surveys “derated” combustible items that were stored inmetal cabinets, while others did not. (The logic behind “derating” items stored in metal cabinets .cis that they would not be expected to burn as efficiently as items that are not stored in non- secombustible cabinets.) Both total fire loads and derated fire loads were published in the Culverreport. The total (not derated) loads were used to develop the loads in chapter 6. a gbCulver published fire load data in units of mass per unit area. Kumar found that 99% of the fireload was cellulosic. Given that the precision in this figure is likely greater than the precision in inthe fuel load values, it is reasonable to round this up to 100%. Therefore, conversion between inmass and energy is accomplished by using an effective heat of combustion for wood. A value of15 MJ/kg was used. This value represents an upper limit for reported values of effective heats of tracombustion for wood based products as published in Tewarson, A. “Generation of Heat andChemical Compounds in Fires,” SFPE Handbook of Fire Protection Engineering, National Fire reProtection Association, Quincy, MA, 2002, for instance. Again the precision in this value is .figreater than the precision in the estimates of fuel load. wLive loads in buildings are expected to vary in a similar manner as fire loads. Indeed, Culver wfound this to be the case. This is handled in ASCE 7 (the standard that specifies the structural wloads that are used to design buildings) by specifying a value that is seldom expected to beexceeded.The contents fire loads in chapter 6 for were developed by determining a mean and standarddeviation for all of the office fire load data that was published by Culver. The mean fuel loadwas 38.2 kg/m2, and the standard deviation was 32.8 kg/m2.The fixed fire loads were handled in a similar manner; however, Culver found that the fixed fireload did not vary appreciably with room use. A stronger influence was found to be whether theroom surveyed was in a government or private building, with fixed fire loads in private buildingsbeing approximately 50% higher than those in government buildings. Page 12 of 29
  • 13. Annex C – Guidance for Fuel Load Surveys (Special Facility and Occupancy Based)To simplify the fire load estimation, surveys conducted in the past have made the followingassumptions: (1) combustible materials are uniformly distributed throughout the building; (2) allcombustible material in the compartment would be involved in a fire; and (3) all combustiblematerial in the fire compartment would undergo total combustion during a fire.Determining fire loads in a building requires measuring the mass of all the different types ofcombustibles and their calorific values. The mass of an item in a compartment can bedetermined by weighing it (weighing technique), or by determining its volume and identifying itsdensity (inventory technique). The direct-weighing method should be used for items that caneasily be weighed, such as toys and books; the inventory method may be used for heavy itemsthat cannot be weighed, such as heavy furniture and built-in shelves. In most cases, acombination of the weighing and inventory methods is used, in which some common items could ombe pre-weighed, and then the surveyor notes their inventory. To ensure a high quality of thesurvey data and to avoid inconsistencies that might occur when different individuals completethe survey forms, it is preferable that the survey is conducted by trained individuals who .cappreciate the importance of the data collected. seA standard survey form is useful to facilitate the survey process and to ensure that data is acollected in a systematic and consistent fashion for all buildings. The survey can be divided into gbthe following five sections: (1) building identification and date of investigation; (2) type ofestablishment; (3) compartment dimensions; (4) fixed fire loads (this section contains ininformation regarding building construction, weight, and type of lining materials); and in(5) moveable fire loads. traTo facilitate the survey process, it is recommended that the surveyor follow a similar procedurefor all buildings. First, the building name and address are recorded, as well as the type of reestablishment and date of the investigation. Second, the dimensions of the room(s) are measured .fiand the types of wall, floor, and ceiling lining materials are determined and noted in the fixedfire load section of the survey form. The third step identifies and classifies all contents in each wcompartment. Items that could be weighed are weighed to determine their mass; the materials wthat the item is made of are determined and recorded. For items consisting of more than one wmaterial type, the percentage of each type is determined and quantified. The mass of items thatcannot be weighed, such as heavy furniture and built-in shelving units, is determined bymeasuring their volume and using the density of the material to calculate their mass.The data collected is analyzed to determine the total fire load in each building compartment, thefire load densities (MJ/m2), and the contribution of different materials (wood, plastics, textiles,food, etc.) to the total fire load and to the fire load densities.If the results of the fire load survey are to be applied to multiple buildings, it is important tocollect data for a number of similar buildings to ensure that the survey results are valid. Samplesizes (number of compartments surveyed) will vary depending on the variation of values. Insome cases, where variations are large, it may be necessary to identify parameters that may affectfire load densities. For example, it was found in some earlier surveys that fire load densities Page 13 of 29
  • 14. decreased with increasing the area of a building. In such cases, it is preferable to group buildingsinto categories based on area and to determine fire load densities for each group.An important component of the survey is to determine the target population and the samplerequired. The first is deciding the type of buildings that will be surveyed, such as residentialbuildings, commercial buildings, shopping centers, or industrial buildings. In determining thetarget population, it is critical to identify any subgroups that may yield different results. Forexample, if dealing with residential buildings, it is important to differentiate between apartmentbuildings and houses, as the fire loads may be different.The second important decision is to determine how many rooms/buildings to use in the survey.The sample size depends on time available, budget, and necessary degree of precision. Thefollowing equation can be used to determine a sample size (number of rooms to be surveyed): om  Z ×σ  2n=   x  .c seWhere: Z = Z-value (e.g. 2.57 for 99% confidence level) σ = standard deviation x = sample mean a gbThe standard deviation could be evaluated from a small sample and then used to find the innecessary sample size. The larger the sample, the surer one can be that their answers truly reflect inthe population. traIn selecting a sample for the survey, care should be taken to choose a sample that isrepresentative of the population. For example, if one is interested in surveying houses, they reshould ensure that their sample includes houses of all sizes and price range. If one chooses .fihouses in affluent neighborhoods, they may not have the same fire load as houses in poorerareas. w w w Page 14 of 29
  • 15. Annex D - Analyses of “Structurally Significant” Fires in Buildings with SelectedCharacteristicsThese analyses first provide estimates of the rate of fires (per year) relative to numbers ofbuildings and square feet of floor space, for each of eight property use groups.Floor space survey data include only buildings with at least 1,000 square feet (93 m2) and useproperty use groupings that may differ from those used for fire data. Details on inclusion andexclusion are provided where available.Next in each section is the percentage of fires with extent of flame beyond, respectively, theroom of origin and the floor of origin. The latter is more likely to be a structurally significantfire than the former. Many properties in every category are not high-rise and may be only one omstory tall.Percentages are provided for all fires, for fires in buildings with sprinklers or other automatic .cextinguishing equipment, for fires in buildings with automatic detection equipment, and for seseven types of construction, excluding only heavy timber, for which fires are few and mis-codings appear to be a high proportion of the total. a gbFour technical papers from Finland and one from Sweden dealing with the same technical issuesalso have been reviewed. Three papers limited themselves to derivations and model-building infor mathematical methods of estimating useful parameters on these subjects. Two papers inincluded actual data from Finland, one for 1996-1999 and one for 1996-2001. The following is acomparison of the categories used in those studies and the categories used in this analysis: tra USA Finland re Public assembly Analysis divided into Analysis provided for .fi religious properties, eating “assembly buildings”. and drinking establishments, Passenger terminals may be w and other public assembly, in a second category, whose w including passenger name includes the word w terminals. “transport”. Educational Analysis provided for all Analysis provided for educational properties “educational buildings” Health care Analysis provided for Analysis provided for properties facilities that care for the “buildings for institutional sick. Facilities that care for care”; these could include the aged are grouped with either or both parts of health lodging properties in floor care and/or prisons and space data. jails, though the word “care” suggests only health care is included. Stores Analysis provided for all Analysis provided for store and mercantile “commercial buildings”. Page 15 of 29
  • 16. properties; floor space data may exclude some properties such as gasoline service stations Offices Analysis provided for office Analysis provided for properties, including fire “office buildings”; fire stations stations are included in a separate category called “transport and firefighting and rescue service buildings”. Residential Analysis provided for Analysis provided for residential other than home “residential buildings” and plus facilities that care for the separately for “buildings for om aged, because that is how institutional care”. floor space data is grouped. .cIn each section, Finnish data is provided and discussed. seThe Finnish data include figures for industrial buildings (where there is no floor space data in the aU.S.A.) and warehouses (where there is some floor space data from US sources, but isolating the gbcorresponding storage properties was deemed too speculative and sensitive for this analysis). in in tra re .fi w w w Page 16 of 29
  • 17. A. Religious PropertiesSpecific property use 130-139 includes churches, synagogues, mosques, religious educationfacilities, and funeral parlors. There is no Finnish data broken down to this level.Fires per year (to the nearest hundred) 2,100Thousands of buildings with at least 1,000 square feet 342.6Millions of square feet in buildings with at least 1,000 square feet 3,552Fires per thousand buildings per year 6.0Fires per million square feet per year 0.58 Percentage of Fires With Flame Spread Beyond Room of Origin and Estimated Number of Fires Used as Basis for Percentages No detectors Detectors present No detectors Detectors present om Type of Construction No sprinklers No sprinklers Sprinklers present Sprinklers presentFire resistive 22% 6% 0% 0% 1,982 558 33 93 .cProtected 16% 7% 7% 3% senoncombustible 776 338 29 76Unprotected 23% 15% a 0% 43% gbnoncombustible 819 239 2 14Protected 24% 12% 14% 0% inordinary 3,739 1,095 29 145 inUnprotected 29% 18% 22% 5% traordinary 4,637 1,215 27 80 reProtected 33% 17% 6% 3%wood frame 3,223 885 31 60 .fiUnprotected 39% 20% 8% 18% wwood frame 5,290 918 26 39 w wNote: These are 1989-1998 fires reported to U.S. municipal fire departments and so exclude fires reported only toFederal or state agencies or industrial fire brigades. These years are used because they are the latest for which typeof construction is included in the coded elements. All estimates are based on at least 200 reported fires (raw, notprojected estimates) in the 10 years with the indicated data known. Buildings and floor space are estimated from1992, 1995, and 1999 surveys, using linear interpolation and extrapolation for years before or between the threeyears when surveys were taken, resulting in a final formula of{(7 x 1992 estimate) + [1.5 x (1995 estimate + 1999 estimate)]}/10.Sources: NFPA analysis of NFIRS; NFPA survey; Energy Information Administration Commercial BuildingsEnergy Consumption Surveys, building characteristics tables. Page 17 of 29
  • 18. B. Eating and Drinking EstablishmentsSpecific property use 160-169 includes restaurants, cafeterias, nightclubs and taverns. Floorspace survey data are for food service establishments. There is no Finnish data broken down tothis level.Fires per year (to the nearest hundred) 11,400Thousands of buildings with at least 1,000 square feet 277.1Millions of square feet in buildings with at least 1,000 square feet 1,524Fires per thousand buildings per year 41.2Fires per million square feet per year 7.5 Percentage of Fires With Flame Spread Beyond Room of Origin and Estimated Number of Fires Used as Basis for Percentages om No detectors Detectors present No detectors Detectors present Type of Construction No sprinklers No sprinklers Sprinklers present Sprinklers presentFire resistive 16% 10% 5% 3% .c 8,566 2,090 1,879 2,893 seProtected 16% 6% 4% 4%noncombustible 4,690 1,482 1,446 2,003 a gbUnprotected 20% 10% 8% 5%noncombustible 4,991 1,193 896 836 inProtected 19% 11% 6% 4% inordinary 19,096 5,034 3,837 4,623 traUnprotected 24% 14% 8% 5%ordinary 24,670 5,325 2,917 2,469 reProtected 22% 12% 8% 5% .fiwood frame 13,513 3,499 2,180 2,210 wUnprotected 29% 19% 11% 7% wwood frame 23,985 3,901 1,902 1,303 wNote: These are 1989-1998 fires reported to U.S. municipal fire departments and so exclude fires reported only toFederal or state agencies or industrial fire brigades. These years are used because they are the latest for which typeof construction is included in the coded elements. All estimates are based on at least 200 reported fires (raw, notprojected estimates) in the 10 years with the indicated data known. Buildings and floor space are estimated from1992, 1995, and 1999 surveys, using linear interpolation and extrapolation for years before or between the threeyears when surveys were taken, resulting in a final formula of{(7 x 1992 estimate) + [1.5 x (1995 estimate + 1999 estimate)]}/10.Sources: NFPA analysis of NFIRS; NFPA survey; Energy Information Administration Commercial BuildingsEnergy Consumption Surveys, building characteristics tables. Page 18 of 29
  • 19. C. Other Public AssemblySpecific property use 100-199 excluding 130-139 and 160-169 includes exhibition halls, arenas,stadiums, ballrooms, gymnasiums, bowling alleys, ice and roller rinks, swimming facilities, cityand country clubs, libraries, museums, court rooms, passenger terminals, and theaters. Floorspace survey data are from a category called public assembly which excludes the separatecategories of religious properties and food service facilities.The Finnish data could exclude passenger terminals and could include religious properties and/oreating and drinking establishments. Their data on fires need to be converted from total fires for amulti-year period to average fires per year. Having done so, their rates of fires per millionsquare feet were 0.35 for 1996-1999 and 0.52 for 1996-2001. However, one out of sevenbuildings had unknown square feet, so it is possible these figures should be reduced by one-seventh. Either way, they are lower than the figures related to other public assembly for the U.S. omIf the three public-assembly categories are combined, the U.S. figure for all public assemblywould be 1.9, even higher than the Finnish figures. Their data on fires per thousand buildingsshowed 3.3 for 1996-2001 (no such data shown for 1996-1999). This is much lower than any .ccomparable U.S. figures. seFires per year 4,200Thousands of buildings with at least 1,000 square feet a 289.3 gbMillions of square feet in buildings with at least 1,000 square feet 4,440Fires per thousand buildings per year 14.5 inFires per million square feet per year 0.94 in Percentage of Fires With Flame Spread Beyond Room of Origin and tra Estimated Number of Fires Used as Basis for Percentages No detectors Detectors present No detectors Detectors present re Type of Construction No sprinklers No sprinklers Sprinklers present Sprinklers presentFire resistive 13% 5% 4% 2% .fi 5,087 1,757 675 2,163 wProtected 16% 6% 3% 5% wnoncombustible 2,168 815 419 1,077 wUnprotected 20% 13% 4% 6%noncombustible 2,869 727 306 343Protected 21% 11% 4% 3%ordinary 5,593 1,557 580 1,231Unprotected 31% 15% 5% 3%ordinary 8,295 1,604 416 511Protected 33% 18% 12% 5%wood frame 3,248 853 316 356Unprotected 43% 22% 10% 8%wood frame 10,823 1,282 236 250 Page 19 of 29
  • 20. Note: These are 1989-1998 fires reported to U.S. municipal fire departments and so exclude fires reported only toFederal or state agencies or industrial fire brigades. These years are used because they are the latest for which typeof construction is included in the coded elements. All estimates are based on at least 200 reported fires (raw, notprojected estimates) in the 10 years with the indicated data known. Buildings and floor space are estimated from1992, 1995, and 1999 surveys, using linear interpolation and extrapolation for years before or between the threeyears when surveys were taken, resulting in a final formula of{(7 x 1992 estimate) + [1.5 x (1995 estimate + 1999 estimate)]}/10.Sources: NFPA analysis of NFIRS; NFPA survey; Energy Information Administration Commercial BuildingsEnergy Consumption Surveys, building characteristics tables. om .c a se gb in in tra re .fi w w w Page 20 of 29
  • 21. D. EducationalSpecific property use 200-299 includes grades K-12 and college classrooms but does not includedorms or other properties common to educational complexes.The Finnish data on fires need to be converted from total fires for a multi-year period to averagefires per year. Having done so, their rates of fires per million square feet were 0.18 for 1996-1999 and 0.28 for 1996-2001. However, one out of 20 buildings had unknown square feet, so itis possible these figures should be reduced by 5%. Either way, they are lower than the figuresrelated to other educational properties for the U.S. Their data on fires per thousand buildingsshowed 5.2 for 1996-2001 (no such data shown for 1996-1999). This is much lower than anycomparable U.S. figures.Fires per year 7,700 omThousands of buildings with at least 1,000 square feet 306.1Millions of square feet in buildings with at least 1,000 square feet 8,388Fires per thousand buildings per year 25.0 .cFires per million square feet per year 0.91 se Percentage of Fires With Flame Spread Beyond Room of Origin and Estimated Number of Fires Used as Basis for Percentages a gb No detectors Detectors present No detectors Detectors present Type of Construction No sprinklers No sprinklers Sprinklers present Sprinklers present inFire resistive 7% 3% 4% 2% 12,140 9,878 1,017 4,293 in traProtected 7% 4% 2% 3%noncombustible 5,544 4,753 689 2,826 reUnprotected 9% 4% 1% 2%noncombustible 4,040 3,071 251 652 .fi wProtected 8% 4% 5% 3%ordinary 8,215 6,025 737 2,786 w wUnprotected 16% 8% 4% 5%ordinary 6,169 3,962 308 858Protected 18% 7% 5% 2%wood frame 2,794 1,595 263 647Unprotected 30% 13% 11% 3%wood frame 5,108 1,692 179 313Note: These are 1989-1998 fires reported to U.S. municipal fire departments and so exclude fires reported only toFederal or state agencies or industrial fire brigades. These years are used because they are the latest for which typeof construction is included in the coded elements. All estimates are based on at least 200 reported fires (raw, notprojected estimates) in the 10 years with the indicated data known. Buildings and floor space are estimated from1992, 1995, and 1999 surveys, using linear interpolation and extrapolation for years before or between the threeyears when surveys were taken, resulting in a final formula of{(7 x 1992 estimate) + [1.5 x (1995 estimate + 1999 estimate)]}/10. Page 21 of 29
  • 22. Sources: NFPA analysis of NFIRS; NFPA survey; Energy Information Administration Commercial BuildingsEnergy Consumption Surveys, building characteristics tables. om .c a se gb in in tra re .fi w w w Page 22 of 29
  • 23. E. Facilities That Care for the SickSpecific property use 330-339 includes hospitals and clinics. Floor space survey data includeinpatient and outpatient facilities; nursing homes are included with lodging.None of the Finnish categories seem to correspond well to this U.S. category.Fires per year 3,000Thousands of buildings with at least 1,000 square feet 78.9Millions of square feet in buildings with at least 1,000 square feet 2,022Fires per thousand buildings per year 37.8Fires per million square feet per year 1.48 Percentage of Fires With Flame Spread Beyond Room of Origin and Estimated Number of Fires Used as Basis for Percentages om No detectors Detectors present No detectors Detectors present Type of Construction No sprinklers No sprinklers Sprinklers present Sprinklers present .cFire resistive 3% 2% 2% 1% 3,894 7,660 934 13,624 seProtected 3% 2% 2% 1%noncombustible 1,198 2,157 a 363 5,704 gbUnprotected 8% 4% 0% 1% innoncombustible 279 448 38 590 inProtected 10% 3% 3% 2% traordinary 952 1,554 325 3,777Unprotected 17% 5% 0% 1% reordinary 586 594 74 659 .fiProtected 19% 7% 35% 2% wwood frame 236 299 23 464 wUnprotected 14% 14% 0% 1% wwood frame 519 306 26 223Note: These are 1989-1998 fires reported to U.S. municipal fire departments and so exclude fires reported only toFederal or state agencies or industrial fire brigades. These years are used because they are the latest for which typeof construction is included in the coded elements. All estimates are based on at least 200 reported fires (raw, notprojected estimates) in the 10 years with the indicated data known. Buildings and floor space are estimated from1992, 1995, and 1999 surveys, using linear interpolation and extrapolation for years before or between the threeyears when surveys were taken, resulting in a final formula of{(7 x 1992 estimate) + [1.5 x (1995 estimate + 1999 estimate)]}/10.Sources: NFPA analysis of NFIRS; NFPA survey; Energy Information Administration Commercial BuildingsEnergy Consumption Surveys, building characteristics tables. Page 23 of 29
  • 24. F. Stores/MercantileSpecific property use 500-589 includes department stores and other multi-line stores; facilitiesoffering sales of food, beverages, textiles, clothing, specialty items, and household goods; andfacilities offering repairs and personal or professional services. Gas stations and motor vehiclerepair and paint shops are also included. Floor space survey data include food sales, mercantile(in or out of malls), and service.The Finnish data on commercial building fires need to be converted from total fires for a multi-year period to average fires per year. Having done so, their rates of fires per million square feetwere 0.44 for 1996-1999 and 0.61 for 1996-2001. However, one out of five commercialbuildings had unknown square feet, so it is possible these figures should be reduced by 19%.Either way, they are far lower than the figures related to mercantile/store properties for the U.S.Their data on fires per thousand buildings showed 3.2 for 1996-2001 (no such data shown for om1996-1999). This is much lower than comparable U.S. figures.Fires per year 19,900 .cThousands of buildings with at least 1,000 square feet 1,393.2 seMillions of square feet in buildings with at least 1,000 square feet 13,434Fires per thousand buildings per year 14.3Fires per million square feet per year a 1.48 gb Percentage of Fires With Flame Spread Beyond Room of Origin and Estimated Number of Fires Used as Basis for Percentages in in No detectors Detectors present No detectors Detectors present Type of Construction No sprinklers No sprinklers Sprinklers present Sprinklers present traFire resistive 18% 13% 5% 4% 20,579 2,901 4,074 5,157 reProtected 17% 10% 3% 3% .finoncombustible 10,729 1,886 3,831 4,496 wUnprotected 25% 16% 5% 5% wnoncombustible 21,172 2,829 3,326 2,557 wProtected 24% 16% 7% 5%ordinary 33,577 5,038 4,623 4,730Unprotected 31% 21% 9% 9%ordinary 51,512 6,230 3,102 2,100Protected 30% 19% 10% 11%wood frame 17,184 2,627 946 782Unprotected 41% 28% 20% 6%wood frame 42,371 3,531 825 469Note: These are 1989-1998 fires reported to U.S. municipal fire departments and so exclude fires reported only toFederal or state agencies or industrial fire brigades. These years are used because they are the latest for which typeof construction is included in the coded elements. All estimates are based on at least 200 reported fires (raw, not Page 24 of 29
  • 25. projected estimates) in the 10 years with the indicated data known. Buildings and floor space are estimated from1992, 1995, and 1999 surveys, using linear interpolation and extrapolation for years before or between the threeyears when surveys were taken, resulting in a final formula of{(7 x 1992 estimate) + [1.5 x (1995 estimate + 1999 estimate)]}/10.Sources: NFPA analysis of NFIRS; NFPA survey; Energy Information Administration Commercial BuildingsEnergy Consumption Surveys, building characteristics tables. om .c a se gb in in tra re .fi w w w Page 25 of 29
  • 26. G. OfficesSpecific property use 590-599 includes general office buildings, bank buildings, fire stations andmedical, engineering, or other professional offices.The Finnish data on office fires need to be converted from total fires for a multi-year period toaverage fires per year. Having done so, their rates of fires per million square feet were 0.20 for1996-1999 and 0.23 for 1996-2001. However, one out of 14 office buildings had unknownsquare feet, so it is possible these figures should be reduced by 7%. Either way, they are lowerthan the figures related to office properties for the U.S. Their data on fires per thousandbuildings showed 3.8 for 1996-2001 (no such data shown for 1996-1999). This is much lowerthan comparable U.S. figures.Fires per year 6,500 omThousands of buildings with at least 1,000 square feet 740.9Millions of square feet in buildings with at least 1,000 square feet 12,002Fires per thousand buildings per year 8.8 .cFires per million square feet per year 0.54 se Percentage of Fires With Flame Spread Beyond Room of Origin and Estimated Number of Fires Used as Basis for Percentages a gb No detectors Detectors present No detectors Detectors present Type of Construction No sprinklers No sprinklers Sprinklers present Sprinklers present inFire resistive 13% 7% 4% 3% 7,032 3,876 1,010 4,501 in traProtected 15% 6% 5% 3%noncombustible 3,218 2,106 741 2,715 reUnprotected 19% 10% 7% 5%noncombustible 3,285 1,234 292 620 .fi wProtected 21% 10% 3% 4%ordinary 8,040 2,983 685 1,780 w wUnprotected 30% 17% 11% 7%ordinary 9,399 2,819 434 609Protected 30% 18% 13% 8%wood frame 5,380 1,681 181 339Unprotected 37% 20% 12% 7%wood frame 8,762 1,924 174 196Note: These are 1989-1998 fires reported to U.S. municipal fire departments and so exclude fires reported only toFederal or state agencies or industrial fire brigades. These years are used because they are the latest for which typeof construction is included in the coded elements. All estimates are based on at least 200 reported fires (raw, notprojected estimates) in the 10 years with the indicated data known. Buildings and floor space are estimated from1992, 1995, and 1999 surveys, using linear interpolation and extrapolation for years before or between the threeyears when surveys were taken, resulting in a final formula of{(7 x 1992 estimate) + [1.5 x (1995 estimate + 1999 estimate)]}/10. Page 26 of 29
  • 27. Sources: NFPA analysis of NFIRS; NFPA survey; Energy Information Administration Commercial BuildingsEnergy Consumption Surveys, building characteristics tables. om .c a se gb in in tra re .fi w w w Page 27 of 29
  • 28. H. Places Where People Sleep Other Than HomesSpecific property use 310-319 includes nursing homes and other facilities that care for the aged.Specific property use 430-489 includes hotels and motels, dormitories and barracks, boardinghomes, and home hotels. Board and care homes may be included in some part of this codinggroup. Floor space survey data are labeled as Lodging but are known to include nursing homes.With more than a million buildings reported, the Finnish data seems clearly to include homes(dwellings and apartments), and so would not be expected to be comparable to the U.S.A. data.In fact, the 0.44 fires per million square feet (whether institutional buildings are included or not)in Finland is lower than its U.S. counterpart by a larger ratio than is true for any other propertyclass studied. Fires per thousand buildings in Finland are 1.4 – 1.5 (depending on whetherinstitutional buildings are included), and this is much lower than the U.S.A. figures forresidential other than home plus facilities that care for the aged. omFires per year 13,000Thousands of buildings with at least 1,000 square feet 154.5 .cMillions of square feet in buildings with at least 1,000 square feet 3,245 seFires per thousand buildings per year 84.4Fires per million square feet per year 4.02 a Percentage of Fires With Flame Spread Beyond Room of Origin and gb Estimated Number of Fires Used as Basis for Percentages in No detectors Detectors present No detectors Detectors present in Type of Construction No sprinklers No sprinklers Sprinklers present Sprinklers presentFire resistive 9% 4% 4% 2% tra 6,980 14,776 1,729 14,646 reProtected 11% 4% 5% 2%noncombustible 2,564 6,154 824 8,988 .fiUnprotected 13% 5% 3% 3% wnoncombustible 1,474 2,792 166 1,414 wProtected 16% 9% 4% 2% wordinary 8,600 12,605 1,210 11,558Unprotected 23% 12% 5% 3%ordinary 7,685 9,016 383 3,149Protected 21% 13% 3% 3%wood frame 6,371 9,457 614 5,833Unprotected 32% 18% 9% 3%wood frame 11,155 10,532 369 2,627Note: These are 1989-1998 fires reported to U.S. municipal fire departments and so exclude fires reported only toFederal or state agencies or industrial fire brigades. These years are used because they are the latest for which typeof construction is included in the coded elements. All estimates are based on at least 200 reported fires (raw, notprojected estimates) in the 10 years with the indicated data known. Buildings and floor space are estimated from Page 28 of 29
  • 29. 1992, 1995, and 1999 surveys, using linear interpolation and extrapolation for years before or between the threeyears when surveys were taken, resulting in a final formula of{(7 x 1992 estimate) + [1.5 x (1995 estimate + 1999 estimate)]}/10.Sources: NFPA analysis of NFIRS; NFPA survey; Energy Information Administration Commercial BuildingsEnergy Consumption Surveys, building characteristics tables. om .c a se gb in in tra re .fi w w w Page 29 of 29

×