fire protection , planning report [BUILDING CONSTRUCTION INFORMATION FROM THE CONCRETE AND MASONRY INDUSTRIES NO. 10 OF A SERIES Confirmed: Fire Losses in Multifamily Buildings Depend on Type of Construction Fg, 1. Avapiay spreading apartment fro. Mutttamily buidings bul of combustitee construction port ke lo quekly spreed and fendange the ives and gropary of every occupant Multiamily bultings should ba designed and constrycteg so that each unit can Sustain « Complate burnout winou allacing adjacent aroas Highlights of This issue © As the tire resistwity of malitamily residences increases, fre losses—measured by extent of flame soread, average dollar loss per five, and ‘umber of injuries per fire—all decrease. © Asthe number of living units in mutitamily residences increases, fire losses increase + In residences with over 20 ving units and buit ‘of wood-trame constuction, injury losses and property losses are significantly greater than in {ny other combination of constuction type and buiding size * Recommendation: The use of wood- frame Construction for muttlamily residences with Cover 20 living urits should be prohibited. ‘The second phase of a study of the relationship between construction type andifire losses inmuli- family residences has been completed The new data strongly reinforce conclusions from the fist phase that the fire-resistive qualities of construction play an important part in limiting foases from fire. Based on three measures of fre losses, the report shows a close relationship be- tween construction type and extent of fire losses, As the fire-resistive qualties of @ construction increase, fire losses decrease. The report o! the second phaso (Rot. 1) was pro pared by the University of Maryland's Department ‘of Civil Engineering under a grant from the United States Fire Acministration (USFA). The informa- tion was developed by analyzing data oblained from the USFA Fire Data Center using the National Fire Incident Reporting System (NFIRS). The sec- ‘ond phase combines new USFA data with the dala Used in the first report U.S. FIRE LOSSES Approximately 6500 fire deaths occurrad in the United States in 1980. In adition, 30,000 civilians‘and 98,000 firefighters were injured and over $6 bilion worth of property was destroyed. Res- ences have the worst experience for loss of ite ard property ftom fire in the United States, They are responsible for over 80% of all fire fatalities, 70% of all injuries, and 50% of all property losses. (Ret, 2) ‘Approximately 20% of allresidential fres occur in mutitamily buildings. Very olten these struc- tures are built to the same standards and with the same type of combustible building materials as single-family homes. This multiplies the fire dan- ger in multifamily buildings, bocauso residents and their property are much more vulnerable to fire exposure resulting from the negligent actions ‘of their neighbors than are people livingin single- family homes. DATA BASE Fire data are collected by NFIRS from states vol- tuntatily participating in the program. In the first phase of this frre-loss study, dala wore reosived {rom five states for 1975 through 1977. In the sec- ‘ond phase, data for 1978 included 10 new states {see Fig, 2). Up to and including 1978, NFIRS has data on 59.495 fires in multiple family residences, of which 35,908 are fires with a known extent of flame damage. The study was intended to investigate only the eflect of construction type. Therefore, itwasnec- essary to exclude fires that were confined fo the ‘area oF origin since these small fires would nat be significantly affected by the construction type of the structure. The NFIFS proviced a dala base of ‘8023 fires that met the required characteristics ‘and were analyzed in the study. CONSTRUCTION TYPE AND BUILDING SIZE In addition to the relationship between construc tion type end fire losses, the size of these muit- family residential buildings was also investigated to determine the relationship between bulding size and fire losses. The NFIRS coding system classilies construc- tion into eight types, described in Tabie | Alilstates except California follow this coding system. The NFIRS classification of construction types found in the model building codes is given in Table 4 California classifies construction into fourtypes, depending on the combustibiliy of major struc- tural elements. Table 2 describes these four types, Because ofthis citference the data areanalyzed separetely for California and for all states except California. The NFIRS system describes the size ot ine building by the number of living units in each building. These building categories are listed in Table 3. ‘Table 1. NFIRS Construction Types 1. ite Resistive Atatally noncombustibiebuiding in which no structural Steel's exposed and all vertical openings ere protected By approved doors. The fire-resistant covering of the steel ie typically very heavy: poured concrete, brick, ‘concrete block. or simlar material. 2 Heavy Timber ‘A.ypical millconstrucied building in which the load Dearing walls or colures are masonry oF heayy timber _andall exposed wood members have aminimumdimen- Sion of two (2) inches. I ste! or ion colurnns are used, they should be protected by a fire-resistant enclosure, 3. Prolected Nonconbustible ‘totally noncombustible building in wich no structural Soc! Is exposed. All vertical openings are protected by approved doors. The fire-resistant covering of the steel {s typically light: gypsum board, sprayed Hre-resistve Covering, rated celings, and similar materials, 4 Unprotected Noncombustible ‘Atolaly noncombustible building in which the struc- {ural sto! is exposed to the effects ofa fire, 5. Protected Ordinary “The load-bearing walls are masonry. Columns are protected by a fireresistive covering. The underside of Ail wood floor and roof dects is protected by @ fre- {osistive covering {6 Unprotected Ordinary ‘The load-bearing walls are masenry. Columns, wood floor and root decks are exposed and unprotected from fire. 7. Protected Weod Frame ‘Wal, roofs, and root structure ere wood framing. The interior wall and ceiling surfaces ot habitable spacesare prolected by a fire-resitive covering. brick-vaneer building alsin this category because the wall siructure ‘is wood framed. But for any wood-framo building i he basement does net have a fre-esistive ceiling pro- footing the underside of the fra floor the Bulding ‘should be clssiied in the unprotected-wood-trame category. 18 Unprotected Wood Frame Walls, Moors, and roof structure are wood framing. ‘There is no fre-resistive covering protecting the weod frame. A typical residential garage would fall inthis category. ‘Table 2. California Construction Types ‘Exterior | ieror | Floor and root] ‘type | wall | wall. | construction a N N N 5 N N © ec] oN c ¢ Oe. c 6 Ws Noreombusibie Table 3. NFIRS Building Categories [category | Number ot units) L 24 1 7.20, i over 0 —]tm Siates Reporting in 1879 Study 1B Adcitional sates Reporang in 1880 Study Fg 2 Inthe 1900 study ten aditonalsiates wore adsodtethe FIRS program, FINDINGS The method of evaluating the contribution of con- struction type in minimizing fire losses is based primarily on the following three measures or pa~ rameters available from the fire cata’ 1. Extent of fame damage 2. Property loss in dollars 3 Injuries ard fatalities The reported findings of the study are as follows: Fire losses (meastred by flame damage, prop- erty loss, and injuries) in multifamily residences re dependent on type of construction In particu- lar, the following are construction types in de- creasing order of abiity to mnimize fre losses Type 5 Protected Ordinary Type 8 Unprotected Ordinary Type 7 Protected Wood Frame Type 8 Unprotected Wood Frame For California the similar ranking is Type A Type B Type C Type D Because data avellabie for siudy are highly concentrated in construction types 5,6, 7, and 8, the ranking is based on a comparison of these types only. However, data from all eight NFIRS types can be very useful in assessing the influ- ‘ence of building construction on fre losses. From this study the generaltrerd of he data reveals that as ine fire resistivity o! theconstructiondecreases, tre losses increase. The investigation also determined that tire losses aro dependent on the size of the structure, Losses increase as the number of units in rutti- family residences increases, Of the building cete- gories shown in Table 3, the createst relalive losses occur in Category Ill {over 20 units), (Of patticular importance, the data revealed that injury losses and property losses in residences with over 20 units and of wood-frame construction (types 7 and 8) ere signiicantly orealer than any other combination of construction type and build: ‘ng size. This important finding is examined more closely in the following sections EXTENT OF FLAME DAMAGE The extent ol lame damage describes essentially the extent of the burned or charred area in the structure. According to the National Fire Incident Reporting Systern Handbook the extent of flame. spread “provides one means of descriving the magnitude or seriousness of the fre” and "can be used for evaluating the effectiveness of built-in fire protection features designed to limit fire spread." “The confinement and extingushment of a fire is influenced by many factors, inclucing structural comperimentation.” However, by ana. lyzing “the extent of flame spread for many fires, the effect of individual factors [in this case, con- struction type] can be determined.” (Ret. 3) The effect of construction type on the extent of flame damage is best determined by analyzing fires that extend beyond the area of origin, called extended fires. Since the tirst barriers tothe spread, of fire in a residential buliding are the walls floor, and ceiling assembly, the efficiency of these struc- tural elements in confining the fite to the space of origin is meesured In terms of the percentage of extended fires penetrating this firs ine of defense, Values are presented in Fig. 3. Table 4. NFIRS Classification of Construction Types Found in Model Building Codes Consiiction ype ae iiss Grama Fane Frenroot oncombusta | tunter | Bromcted | Unproieced | Protedied | unsrotesiog wool «| =e [a lost[o| «| «6 36 aA rT A a A = 10 wet veo usc ioe —— [a feem [ee | var |e TN vet VN Freestate —[ _Limied combustibe c teen | — ina = Wea N8C:1976 [Type A [Type 8 [ Potecied | Unpotecid | timtor Ordinary ton Na say zt i a + 2 5 6 7 8 tote: Abul nestig avo! bldg cata clmicton uches BBS typ 98, wo resin he NFS cinicaono tyes Fowever, a 9uing sified as an NFIRS type 5 would not neces iy meat ie spectations to be csaiied a8 BBC type 28fo. 34 Extent of Fame Seread a 98 bent of Pie Spa Al Sttos excep! Calor Caltoia Porcintnge of Evended Fee 90} Eslend Bevond Fre Rated Comeartnent of Oro 90} That exer Bajond ire 80 ‘20| Anes Comparten 0! Onan 60) 60 30} 20| 40) 40] 30] 30| 10) 19 20 Eicna Sryonaiiocw of Grain 60 20 40 ea danetat tere Tine constuction Type Fo 4 Average Losses Por Fira it Dotars AN StIe8 Except Catiornia Borcontage of Exton Fae a Eston Boron Aon o Onin 99] 10] 69| * 30 10) EO CenD Fo 48 ‘Average Lossas Po Fire in Dolars alone 90 Fire Ratos Corpartmant of Origin $1000 50 °C incl ‘90 | Fre Extning Beyors Aes of nig $1000 50 40 Cconstction Typ ‘Note: Ineuticent data for construction type ? (heavy timber) {or statistical meshing vatues, 4 20 beyond Fi Feo solonan $1000 20) 40 sou 90} Room at ongn co] ‘$1000 50 40) 30) 20) 90] 70} ot eo} $1,000 50) “ateal tl Cconsiucton Type ] Buiding Cotegory toa @The data indicate that for construction types 5, 6, 7, and 8 (protected ordinary, unprotected ordi- nary, protected wood-frame, and unprotected wood-frame) and for al buiding categories com- bined (three units and over), the relative proba billy offlames extending beyond the room oforigin and beyond the tire-rated compartment of origin increases as the lire resistivity of construction decreases. This is particularly evident in building category | (3 to 6 units) From the California cata it was determined that for all building categories combined (three units ‘and over), the type of interior wail construction has ‘a significant etfact on reducing the extent offlame damage. Also, for buildings of 3 to 20 units both interior and exterior wall constructions havea sig- nificant effect on reducing the extent of flame damage PROPERTY LOSS ‘The NFIRS data provide an estimate of the total dolar loss for contents and structure in each fire. Fig. 4 indicates the average dollar loss per fire measured in terms of the extent of flame damage. Values were computed by dividing the total dollar loss by the corresponding number of fires. This information can be used to compare the relative performance of the type of construction and the offect of building size on fire losses, Those com- binations of building size and construction type in which large dolar losses occur are ready identities The data reveal a large increase in fire damage measured by average dollar loss as the ire resis- tivity of the construction is teduced. There is also a marked increase in losses per fire as the size of ‘multifamily residences increases. The largest losses occurred in construction types 7 and 8 (wood-frame) in buidings with more than 20units, This, particular combination. produced losses much greater than any other comparable loss values. Even in buildings with 7 to 20 units, aver- age loss per fire in constructiontypes 7 and8 are ‘greater than the average values for other types. Observing the California data for al bulding Fg 54 Al Sales Excop! Catlornia ‘Nurber of june Per 100 Free 90 | Exteasng Boyond Are of Grist 13 4 8 6 7 8 Construction Type | Sevens Area at Oran e88ssss categories combined (three units and over), aver- ‘age losses in construction types C and D are near- ly twice as large as comparable values for types A and B CASUALTIES The greatest concern about the effects of fire re- {ates to life safely. Thus, it's important to evaluate the effects of construction type and bulding cato- gory on the number of casualties (injuries and Geaths) sustained by analyzingpast performances in actual fres. Since it was felt that casuaties from fires con- fined to the area of origin are not alfected by con- struction type, only casualties occurring in ex- tended fires were considered, Also, because ot the limited number of fatalities recorded, a statis- tically meaningful analysis wasnot possible for the fatality data, and only injury data were presented. The injury cata include persons injured at the scene as a result either of the tie or o!theaction of handling the incident Firefighter injuries were included with these data Fig. 5 presents the number of injuries per 100 fires, The data indicate for construction types 5,6, 7, and 8 that an increase in injuries per 100 fies occurs as the tire resistivity of the construction is reduced, This is particularly evident in structures with more than 20.units. Also, all the data indicate ‘@ consistent increase in injuries as the number of units in a multifamily residence increases. CONCLUSIONS AND RECOMMENDATIONS ‘The information used in the study of the relation ships between fire losses and construction type and building size was based on actual fre losses, land the results are indicative of the real-world fire performance of the different construction types. This information, if applied to the planning and construction of new multifamily buildings, can be ‘most valuable in increasing lite safety and reduc~ ing property damage due to fire ns satin oy @ Caltornia ere oe ont oo Saeeeas CU Constracton TypoTableS. Number of Living Units Issued Pormits in United States (Thousands of Units) 7076 | 1979 | 1960 | 1981 ‘One and two family | 600 | 1643 | ase | 714 eure aco err [ae [207 % mulitamiy trom 2a] ar] S| ‘Muttaniy housing units represent an incressing potion of thehouning marta, Source"? W.Dodge-—Conivact Construion Awards ‘The report recommends that "serious consider- ation should be given to prohibiting the use of \wood-frame construction [types 7 and 8] in multi- family residences with more than 20 Iving unis ‘This Is based on tne high average property losses ‘and injuries experienced in this combination of building size and construction type. If the same fires had occurred in buildings of ordinary con- struction (types 5 and 6), the report estimates there would have been @ 12% savings in the total dollar loss from fires extending beyond the area of origin and a 60% reduction in injuries, ‘As the mix of residential construction shifts to- wards an increasing proportion of mutifamily resi- dontial units where residents are supject to the action of their neighbors, the potential increases for greater danger of loss of lite and property from fire. See Table 5. By using concrete and masonryin the construc~ tion of new multifamily buildings, the added safety of fire-resistive, noncombustible construction is provided. As idenlifiedin this study, the use of non- ‘combustible construction is moat effective in mini- ‘mizing fire damage measured by extent of flame spread, dollar loss, and casualties. By analyzing a large body oi firadata, theimpor- tence of consiruciion iypein the firesalety of multi- family buidings has been evaluated. This informa: tion is important if rational decisions regarcing firesafety are to be made based on reliable infor mation of past performance. Since litie informa tion has been published aboul the performance of multifamily residential buildings in fires, this study represents a significant step forward in under- standing fire behavior in such occupancies. REFERENCES 1A Study of Fire Losses in Multi-Family Resi ences by J, Colville and B Behanami, Depart- ment of Civil Engineering, University of Mary- land, 1982, prepared forthe FederalEmergercy Fig. 6 Many communities are shifting some of the burden for {ie protection tothe private sector. OFtentheyrequie allt ‘amy Buldings to be Gult with fire resistive noncombustible onsvuction sepraing each dweting uni in oro vt the Spread of te and contain H a! 8 managoable size Management Agency, United Statos Fire Ad ministration. Available from National Technical Information Service—order No. PB62214701 2. Fire Facts, 1982 Ediion, National Fire Protec- lion Association 3, National Fire Incident Reporting System Hana book, Federal Emergency Management Agen- oy, US. Fire Administration, National Fire Data Center, August 1980. Organizations represented on the CONCRETE AND MASONRY INDUSTRY FIRESAFETY COMMITTEE BIA Brick Institute of America CRS Concrete Reintorcing Stee! institute ESCSI Expanded Shale Gay and Slate Institute NCMA National Concrete Masonry Association NRMCA Natlonal Ready Mixed Concrete Association PCA Portland Cement Assosiation PCI Prestressed Concrete institute “This publicaton is intended for the use of professional personnel competent to evaluate the significance anc limitations af its contents ane who will accept respon: sibilty fo theappication ol thematerial t contains. The Concrete and Masonty Industry Firesalety Committee disclaims any and all responsibilty for application ofthe stated inclples or for the accuracy of the sources ‘thor than work performed or information developed by the Committe, Concrete and Masonry Industry Firesafety Committee 5420 Old Orchard Road, Skokie, Illinois 60077-4321 Printed in USA. SR243.018.