Report of the Committee on Geraldine Massey, Dillon Consulting Engr, Inc., CA [SE] Smoke Management Systems (AlL to M. E. Dillon) Jayendra S. Parikh, Underwriters Laboratories Inc., IL [RT] Harold E. Nelson, C/uz& (Air. to D.J. Kaiser) Hughes Associates Inc., MD [SE] Randolph W. Tucker, RoffJensen & Assoc., "IX [SE] (Alt. to D. L. Arnold) Paul G. Turnbull, Landis & Gyr Powers, Inc., IL [M] Daniel L. Arnold, RolfJensen & Assoc., GA [SE] (Voting Alt. to L&G. P. Rep.) Donald W. Belles, Donald W Belles & Assoc. Inc., TN [M] Peter J. Gore W'dlse, Industrial Risk Insurers, CT [I] Rep. American Architectural Mfrs. Assn. (Air. to T. E. Schumann) Jack B. Buckley, Houston, TX [SE] Michael L. Wolf, Greenheck, WI [M] Elmer F. Chapman, NewYork City Fire Dept., NY [E] (Alt. to D. Rammien) Michael Earl Dillon, Dillon Consulting Engr, Inc., CA [SE] S. E. Egesdal, Honeywell Inc., MN [M] Nonvoting Rep. Nat'l Electrical Mfrs. Agsn. CharlesJ. Green, Colt Int'l. Ltd., England [M] Bent A. Borresen, Techno Consultant, Norway GunnarHeskestad, Factory Mutuzd Research Corp., MA [I] (Alt. to C. N. Madsen) William R. Houser, U.S. Army Environmental Hygiene Agency, MD E. G. Butcher, Fire Check Consultants, England [u] (Alt. to A. G. Parnell) Winfield T. Irwin, Irwin Services, PA [M] Christian Norgaard Madsen, Techno Consultant, Norway Rel~. North American Insulation Mfrs. Agsn. Alan G. Parndl, Fire Check Consultants, England DameIJ. Kaiser, Underwriters Laboratories Inc., IL [RT] John E. Kampmeyer, Maida Engr, Inc., PA [SE] Rou Cot6, Staff Liaison John H. KIote, U.S. Nat'l. Inst. of Standards and Technology, MD [RT] This list represents the membership at the time the Committee was balloted on Gary D. Loagheed, Nat'l. Research Council of Canada, Canada [RT] the text of this edition. Since that time, changes in the membership may have FrancisJ. MeCabe, Prefco Products, PA [M] occurred. A key to classifications is found at thefront of the book. James A. Milke, University of Maryland, MD [SE] Gregory IL Miller, Code Consultants Inc., MO [SE] Committee Scope: This Committee shall have primary Erin A. M. Oneisom, l_kS. Air Force, Civil Engr Support Agency, FL responsibility for documents on the design, installation, testing, [U] operation, and maintenance of systems for the control, removal, or Lyman L. Parks, Bellcore, NJ [U] venting of heat or smoke from fires in buildings. Zenon A. Pihut, Texas Dept. of Health, TX [E] Dale Rammlen, Air Movement & Control Assn., Inc., IL [M] The Report of the Technical Committee on Smoke Management John F. Scarff, Marriott Corp., DC [U] Systems is presented for adoption. William A. Schmidt, Bowie, MD [SE] Todd E. Schumann, Industrial Risk Insurers, IL [I] This Report was prepared by the Technical Committee on Smoke J. Brooks Semple, Smoke/Fire Risk Mgmt. Inc., VA [SE] Management Systems and proposes for adoption a complete revision to NFPA 204M-1991, Guide for Smoke and Heat Venting. NFPA Alternates 204M-1991 is published in Volume 10 of the 1996 National Fire Codes and in separate pamphlet form. Eric Anderson, System Sensor, IL [M] (Alt. to S. E. Egesdal) This document when adopted will be renumbered as NFPA 204, Craig Beyler, Hughes Assoc. Inc., MD [SE] Guide for Smoke and Heat Venting. (AIt. to H. E. Nelson) RichardJ. Davis, Factory Mutual Research Corp., MA[I] This Report has been submitted to letter ballot of the Technical (Alt. to G. Heskeshad) Committee on Smoke Management Systems. which consists of 27 Victor L. Dubrowskl, Code Consultants Inc., MO [SE] voting members. The results of the balloting, after circulation of any (Ait. to G. R. Miller) negative votes, can be found in the report.

583 NFPA 204M -- A97 ROP

(Log#l) NFPA 204 204M- 1 - (.1-14 (New)): Reject SUBMITTER: Douglas E. Leihlxacher, Yonkers Fire Department, NY Guide for RECOMMENDATION: Add new text ,as follows: ~In buildings containing trnss roof construction, sufficient heat- Smoke and Heat Venting activated smoke vents shall be installed in tile roof so that heat and smoke from a fire will be automatically removed from tile building 1997 Edition by mechanical means." SUBSTANTIATION: Manyfirefighters have lost their lives when NOTICE: An asterisk (*) following the number or letter they've attempted to ventilate truss roof's that collapsed without designating aparagraph indicates explanatory material on that warning beneath them. Automatic vents would m,'xke it unnecessary paragraph m Appendix A. for firefighters to set foot on truss roofs during fire conditions. Vertical ventilation would be performed automatically. Information on referenced publicatious can be found in COMMIaTI'EE ACTION: Reject. Chapter 10 and Appendix E. Detailed information on refer- COMMrVrEE STATEMENT: To do as tile submitter requested and ences cited in brackets throughout the document can be found require the installation of vents is outside the scope of this guide. in Section B-6, Section c-g, and Section E-2. Tins guide is intended to provide guidance on how to design an effective venting system, but is not intended to serve as a code and mandate vents. Such requirements need to be addressed by a code Chapter 1 General Information (i.e., a fire prevention code or a model building code). NUMBER OF COMMITTEE MEMBERS ELIGIBLE TO VOTE: 27 1-1 Introduction. VOTE ON COMMITTEE ACTION: AFFIRMATIVE: 25 1-1.1 Previous editions of this l~guide have included tables listing vent ABSTENTION: 1 areas based on preselected design objectives. These tables were NOT RETURNED: 3 Green, Mc(;abe, Scarff based on the hot upper layer at 20 percent of the ceiling height. EXPLANATION OF ABSTENTION: Different layer depil-as were accommodated by using a multiplication KAlVIPMEYER: I am abstaining primarily because I ,am concerned factor. Curtain board and vent spacing rules were set. Minimum about the use of Sl units throughout the document. The document clear visibility times were related to fire growth rate, ceiling height, presents good information for the designer, but its usefulness is compartment size, curtain depth, and detector activation ames using engineeting equations and a set of assumptions that sometimes limited by not being presented in units familiar to building design resulted in conservative solutions. and construction personnel. ! do not feel this is sufficient to vote negatively, but should be considered in the future development of This edition has eliminated tile previous tables listing vent areas. tile document. This edition incorporates engineering equations (hand calculations) or references models. The e(tuadons or models provide the (Log #CP1 ) designer with tile necessary tools to develop vent designs based on 204M- 2 - (Entire Document): Accept sele~'ted performance objectives related to a specific building and SUBMITTER: Technical Committee on Smoke Management specific set of circumstances. Engineering equations are included Systems, for calculating vent flows, layer depths, and upper layer tempera- [ RECOMMENDATION: Replace NFPA 204M-1991, Guide for tures based on a prescribed burning rate. | Smoke and Heat Venting, with tile following complete rewrite. SUBSTANTIATION: The 1991 edition of this guide included tables An example using both hand calculations and tile LAVENT (Link- listing vent areas on the basis of preselected design objectives. The Activated VENTs) computer model is presented as an appendix. tables were based on the hot upper layer at 20 percent of the ceiling (See Appendix D.) height. Different layer depths were accommodated by a "multiplica- The majority of the information provided in this guide applies to tion factor." Curtain board and vent spacing rules were set. nonsptinklered buildings. A limited amount of guidance is provided Minimum clear visibility times were related to fire growth rate, in Chapter 8 for sprinklered buildings. ceiling height, compartment size, curtain depth and detector activatio n times using engi n eeri ng equations and a set of assu rap- 1-1.2 The following is a general description of the significant lions that sometimes led to conservative solutions. phenomena that occur d-uting a fire when a fire-ventJng strategy is This proposed complete rewrite deletes the previous tables listing tmplemented: vent areas. It incorporates engineering equations or references models. The equations or models provide the designer with the (a) Due to buoyancy, hot gases rise vertically from the combustion nece~ary tools to develop vent desigris based on selected perfor- zone and then flow horizontally below the roof until blocked by a mance objectives related to a specific building and specific set of vertical barrier (a wall or curtain board), thus initiating a layer of hot circumstances. Engineering equations are included for calculating gases below tile roof. vent flows, layer depths, and upper layer temperatures based on a prescribed burning rate. (b) The volume and temperature of gases to be vented are a For tile first time, this guide will include a computer model function of the rate of heat release of the fire and the amount of air (LAVENT) as well as engineering equations (i.e., hand calculation entrained into tile buoyant plume produced. methods). (c) As tile depth of the layer of hot gases increases, the layer This rewrite is based extensively on state-of-the-art technology temperature continues to rise and the vents open. published in the references cited in brackets throughout tile draft document and listed in Section 13-6, Section C-9, and Section E-2. In (d) The operation of vents widfin a curtained area enables some of many cases the authors of these references participated in tile task the upper la-yer of hot gases to escape and slow the thickening rate group rewrite efforts. of the layer of hot gases. With sufficient venting area, the thickening COMMITFEE ACTION: Accept. rate of the layer can be arrested and even reversed. The rate of NUMBER OF COMMITTEE MEMBERS ELIGIBLE TO VOTE: 27 discharge through a vent of a given area is primarily determined by VOTE ON COMMITTEE ACTION: tile depth of tile layer of hot gases and tile layer temperature. AFFIRMATIVE: 22 Adequate _quantities of replacement inlet air from air inlets located NEGATIVE: l below the hot upper layer are needed if the products of combustion- ABSTENTION: l laden upper gases are to be exhausted according to design. [See NOT RETURNED: 3 Green, McCabe, Scarff Figures I-7.2(a) and I-1.2(b).] EXPLANATION OF NEGATIVE: PIHUT: Use of a dimensional system that is not used currently by Curtain the design professionals in this country is my reason for the negative boards ~ t'~ Fire barrier vote on die entire NFPA 204M document. EXPLANATION OF ABSTENTION: KAMPMEYER: I am abstaining primarily because I am concerned about the use of SI units throughout tile document. Tile document presents good information for tile designer, but its usefiflness is \ Fi+e L'sr~l~i~ . LL~j? limited by not being presented in units familiar to building design and construction personnel. I do not feel this is sufficient to vote negatively, but should be considered in the filture development of the document. F'tgure 1-1.2(a) Behavior of combustionproducts under vented and curtained roof. 584 NFPA 204M ~ A97 ROP

1-3.3 It should be recognized that many large facilities have buildings or areas subject to different fire hazards. Accordingly, venting facilities should be designed specifically for each space as discussed in dais guide.

1-4 Nomenclature. The following symbols define the variables in the equations used throughout the main text of this guide. (See Appendix B for an explanat~'on of the unique nomenclature used in that appendix.)

A area (of burning surface) A. area for fresh air, below design level of smoke interface nv = actual total vent area Ava = aerodynamic vent area thermal diffusivity, k/pc O~g = fire growth coefficient C specific heat d = smoke layer depth d c depd'l of curtain board D = base diameter of the fire g = acceleration of gravity H = ceiling height above base of fire h = heat of combusdon C = heat of gasification thermal conductivity Figure l-l.9(b) Buildlng wifl~ roof vents. constant used in equation 5-3 m thermal inertia 1-1.3" The equations and procedures for hand calculations in I = thickness Section 6-I provide two different types of guidance, The first L = mean flame height addresses the ve ating of Ihnited-growth fires. These are fires that are M. = multiplier (vent size) not expected to grow beyond a predictable heat release rate. The mass burning rate secondtype of guidance is rele~mt to the venting of fires that, if ~st m uncbecked, will cominue to grow to an unpredictable size. The m t = mass burning rate per unit area engineering equations or models incorporated in flais guide ,allow an m." o mass burning rate for an infinite diameter pool estimate of bow well smoke can be confined to a curtained area and m.v = mass flow rate through vent how long the smoke interface can be maintained at a higher level T/z = mass flow rate in the plume than the de.sign elevation of the curtained area. This minimum clear- mp at mean flame height (L) visibility design time facilitates such activities as locating the fire, I appraising the fire severity and its extent, evacuating the building, incident heat flux per unit area and making an informed decision on the deployment of personnel = total heat release rate and equipment to be used for fire fighting. Q" = total heat release rate per unit plan area 1-2 Application and Scope. Qc = convective heat release rate (approx. 0.7 Q.) r radius from fire axis 1-2.1" The provisions of Chapters 2 through 7 of this guide are RTI = response time index ('~u 1/2, where 1: is the time intended to offer guidance for the design of facilities for emergency constant of the heat-responsive element for venting of products of combustion from fires in nonsprinldered, convective heating) single-story buildings. Both mamml and computer modeled solution p density methods are provided in Chapter 6 to aid in design calculation. A t = time limited amou tat of information regarding venting in s prinklered td = design interval time buildings is included in Chapter ~q. These provisions oo not attempt .growth .time. to specify under what conditions venting is to be provided; such tame to igmuon conditions are dependent upon an armlysis of the individual time to detection situation and Ioc~al building code and fire code requirements. AT gas temperature rise (from ambient) at detector 1-2.2 This gtlide does not apply to other ventilation designed for site regulation of temperature within a building, for personnel comfort AT e temperature rise (from ambient) of heat- or cooling of production equipment, or to venting provided for responsive element explosion pressu re rel ie~: See NFPA 68, (~*idefor Venting of T am.b!ent air temperature Defk~gvations. T? lgmnon temperature TIg = surface temperature S 1-2.3 This guide applies to building construction of all types. U gas velocity at detector site V = flame spread velocity 1-2.4 The concepts set forth in this guide were developed for X = radiant fraction r venting fires in large undivided floor areas with ceiling laeights z = height above base of fire sufficient to allow the design fire plume and smoke layer to develop z O = height of 'Mrtual origin" above base of fire (normally 4,6 m or greater). The application of d~ese concepts to (below base of fire, if negative). buildings of smaller area or lower ceiling heights necessitates careful engineering judgment. Chapter 2 Basic Phenomena 1-3 Determination of Occupancy Hazard. 2-1 Prlnciples of Ventlng. I-3.1 Tests and studies provide a basis for the division of occupan- cies into classesdepending upon the fnel available for contribution 2-1.1 Venting Objectives. Venting of a building is provided to slow to fire. Tlaere is awide variation in tile quantities of combustible or stop tl~e descent of a smoke layer for purposes such as: materials in the many kinds of buildings and areas of buildings. The evaluation should take into account the average or anticipated fuel (a) Providing occupants with die opportunity to travel to a safe loading and the rate of heat release anticipated from the combus- area, tible materials or flammable liquids contained therein. (b) Facilitating manual fire fighting by venting smoke and hot gases, enabling fire fighters to reach the origin or seat of the fire. 1-3.2 Cbapter 5 should be referenced to assist in quantifying types of fires in various occupancies. Cb,xracterisdc heat release rates for (c) Reducing damage to buildings and contents due to smoke and boti~ limited-growda and contimmus-growth fires in various types of hot gases. fnel arrays also are addreg~ed.

585 NFPA 204M -- A97 ROP

2-1.2 Vent Designs and Smoke Generation. Tile heat release rate of openings exist to the outside and, therefore, no pressure results a fire, the fuel geometry, the height of the clear layer above tile base from the expansion of gases. Wind effects are not taken into of the fire, and the design depth of the smoke layer are major factors account, as wind might assist or interfere with vent flows, depending affecting theproduction of smoke. Given such knowledge about a upon specific circumstances° It is also assumed that the fire fre, venting designs can be developed in accordance with this guide environment in a building space is divided into two zones -- a hot in which tile vent area is calculated to achieve a mass rate of flow upper layer and a relatively cool, clear (comparatively free of smoke) through vents that matches the mass rate of production of smoke. lower region. Where a fire grows to a size where it approaches Such a design prevents descent of the smoke layer below the design ventilation-limited burning, the building might no longer maintain a height of the clear layer. Alternate designs are possible wilere tile clear lower region, and this guide would no longer be applicable. vent flow is less than file rate of smoke production in which the Finally, caution needs to be exercised where using dais guide for descent of the smoke layer is slowed sufficiently to meet design conditions where the upper gas layer temperature approaches objectives. 600°C, as flashover might occur within the compartment. Wilere a fire develops to flashover or ventilation-limited burning, the 2-1.3 Vent Mass Flow. Vent design criteria in this g~uide assume the relationships provided in this guide are not applicable. mass flow rate through a vent is determined primarily by buoyancy pressure. Mass flow through a vent, therefore, is governed mainly by 2-3.2 Buoyancy Pressure. Buoyancy pressure is related to the depth tile free vent area and depth of the hot layer, and-its temperature. of the hot layer, the absolute temperature of the hot layer, the temperature rise above ambient of the hot layer, and the density of (a) Ventingbecomes more effective with smoke temperature differentialsbetween ambient temperature and an upper layer of the ambient air. approximately 110°C or higher. Where temperature differentials of less than 110°(3 are expected, vent flows mi~{ht be reduced signifi- 2-3.3 Vent Mass Flow. The mass rate of flow of hot gases through a candy, therefore, confideration should be gaven to using powered vent is a function of vent area, layer depth, and hot layer tempera- exhaust. NFPA 92B, Guidefor Smoke Management S3stems in Malls, ture. Atria, and Large Areas. should be consulted for guidance for power venting at these lower temperatures. 2-3.4 Temperature and Vent Flow. The temperature of the hot layer above ambient affects mass flow through a vent. Maximum (b) The vent design criteria in this guide also allow the fire to flow occurs at temperature differentials of approxi'mately 300°C reach a size where tile flame plume enters tlae upper hot layer. above ambient. Flows at other temperature differentials are Flame height may be estimated using equation 6-L diminished as shown in Figure 2-3.4o 2-2 Smoke Production. 1 2-2.1 Entrainment at the Plume Boundary. The rate of production of smoke is dependent on the rate of entrainment of air into a :) column of hot gases produced by and located above a fire. Entrain- ment is affected by tile fire diameter and rate of imat release, and is strongly ,affected by the distance between the base of the fire and the point at which the smoke plume enters the hot upper layer. I[ LL 2-2.2 Base of the Fire. The location of the base of the fire is that o level at wllich significant entrainment begins to occur. For file f, .6 purposes of the equations in this guide, this is at the bottom of the k- nurning zone. L~ < 2-2.3 Fire Size. Since smokeproduction is related to the size of a .4 fire, it logically follows that, aflfactors being equal, larger fires produce more smoke. However, entrainment ~s strongly affected by ihe distance between the base of a fire and the bottom of the hot o ...I layer. Therefore, the base of rile fire (where combustion and I.L entrainment begin) should be selected carefiflly. It is possible for a .2 smaller fire having a base near the floor to produce more smoke than a larger fire with a base at a higher elevation. Each possible fire < scenario should be considered carefully before establishing the conditions of the design fire. 0 0 200 400 ;00 800 1000 2-2.4 Entrainment and Clear Height. Entraimnent is assumed to be TEMPERATURE ABOVE AMBIENT K limited to the clear height between the base of the fire and the bottom of the hot layer. Tile buoyant plume ,associated with a fire produces a flow into tile hot upper layer. As the plume impinges on the ceiling, the plume turns andforms a ceiling jet. The ceiling jet Figure 2-3.4 Effect of temperature on mass flow through a vent. flows radiMly outward along the ceiling. 2-3.5 Inlet Air. 2-2.5 Smoke Production as a Function of Shape of l~re. Tile entrainment formulas specified in this guide predict smoke 2-3.5.1 To function as intended, a building venting system needs production ,assuming a single fire. Where the possibility of multiple sufficiently large fresh air openings at low levels. The effect of inlet fires, and, therefore,-multiple plumes exist, smoke production rates air on vent flow is addressedin 6-1.3.1. For example, where high increase beyond the rate predlcted for a single plume for a fire of equivalent output. It also should be understood that smoke upper layer temperatures of 400 K above ambient are anticipated, 80 entrainment relationships are developed primarily for file case of percent of the predicted vent flows is expected to be achieved with axisymmetric plumes. For line-like fires where a long, narrow plume an inlet area/vent area ratio of 1, whereas it is expected that 90 is created by a fuel or storage array, the smoke production relation- percent of the vent flow will result from a ratio of 2. Where relatively ships in this guide might not be valid. However, if the height of the low upper layer temperatures, such as 200 K above ambient, are smoke layer interface above the base of the fire (H-d) is large expected, a ratio of inlet air/vent area of 1 would result in about 70 compared to the largest horizonhal dimension of the fire (e.g., percent of the desired vent flows, whereas a ratio of 2 would be greater than approximately tllree times), the empirically derived expected to produce about 90 percent of the vent flow. relationships in this guide can be used to predict smoke production. 2-3.5.2 If doors and windows below the design smoke layer do not 2-2.6 Virtual Origin. Plume m,x~ flow above tile flame level is based meet file total recommended inlet air opening area, special air inlet on the concept that, except for absolute scales, die shapes of velocity provisions are necessary. and temperature profiles at the mean flame height are invariable [Heskestad 1983]. This concept leads to an expression for mass flow 2-3,5.3 It is essential that a dependable means for admitting or above the flames that involves the so-called 'Mrmal origin," a point supplying inlet air be providedpromptly after the first vent opens. source from which the plume above tile flames appears to originate. The virtual origin might be above or below the base of the fire. 2-3.5.4 Makeup Air System. The simplest method of introducing makeup air into the space is through direct openings to the outside, 2-3 Vent Flows. such as doors and louvers, which can be opened upon system 2-3.1 Buoyancy and Vent Flow. Flow through a vent in thisguide is activation. Such openings can be coordinated with the architectural calculated on tile basis of buoyancy pressure. It is ,assumed that design and be located as necessary below the design smoke layer.

586 NFPA 204M i A97 ROP

For locations where such openings are not practicable, a powered a~2Either 6 (1) The area ofa unitvent or cluster does not exceed supply system might be considered. This might be an adaptation of ~ O_ or 2d ~ where d c is the depth of the curtain board and d is the the building's HVAC system, provided capacities, outlet grille design depth of the smoke layer. These depths are measured from Io~tions, and velocities are suitable. For such systems, means the centerline of the vent. ( See Figures 4-3(a) through (d).), or (2) The should be provided to prevent supply systems from operating until width of the monitor does not exceed the depth of the curtain exhanst flow has been establishedto avoid pressurization of the fire board, d o or the design depth of the smoke layer, d, where curtains area. are not provided. Chapter 3 Vents (b) The vent spacing is such that, in plan view, the distance between any point in the plane of the roof and the nearest vent, all 3-1 Types of Vents. within the curtained area, does not exceed 2.8H (the diagonal of a s~uare whose side is 2H), where H is the ceiling height. (Also see 3-1.1 Experience has shown that any opening in a roof, over a fire, Figures 4-3(a) through (d).) relieves some heat and smoke. However, building designers and fire protection engineers cannot rely on casual inclusion of skylights, (c) The total vent area per curtained compartment under the windows, or monitors as adequate venting me,ms. Standards exist ceiling depends on the severity of the expected fire, which is (UL 793- Automati~lly Operated Roof Vents for Smoke and Heat, discussed m Chapter 5. FM 4430 - Approval Standard for Heat and Smoke Vents, UBC Standard 15-7-Automatic Smoke and Heat Vents) that include 3-5 Mechanical Vents. Where mechanical vents are considered, see design criteria and test procedures for unit vents that call for Chapter 7. simulated fire tests as well as engineering analysis. Chapter 4 Curtain Boards 3-1.2 Guidelines for the inspection and maintenance of vents are contained in Chapter 9. 4-1 General. In large, open areas, curtain boards enhance prompt activation of the vents and venting effectiveness by containing the 3-2 Vent Design Constraints. smoke in the curtained area.

3-2.1 Materials of construction and methods of installation need to 4-2 Construction. Curtain boards should be made of substantial, be used appropriately to resist expected extremes of temperature, noncombustible materials and constructed to resist the passage of wind, building movement, rain, ball, snow, ice, sunlight, corrosive smoke. environment, internal and external dust, dirt, and debris. Compat- ibility between the vent-mounting elements and the building 4-3 Location and Depth. To ensure smoke containment, curtain strncture to which they are attacbed needs to be ensured (e.g., boards, where provided, should extend down from the ceiling for a holding power, electrochemical interaction, wind lift, building StLqlcient distance to ensure that the value ofd o as shown in Figures movement). 4-3(a) through (d), is a minimum of 20 percent of ceiling height, H, where H represents the ceiling height: 3-2.2 Vents designed for multiple functions (e.g., the entrance of day-lighting, roof access, comfort ventilation) need maintenance of (a) For fiat roofs, measured from the ceiling to the floor. the fire protection time!ion that might be impaired by the other uses. Tbese impairments can include loss of spring tension, racking (b) For sloped roofs, measured from the center of the vent to the or wear of moving parts, adverse exterior cooling effects on the fire floor. Where there are differing vent heights, H, each vent should protection release mecbanism, adverse changes in performance be calculated individually. sequence such as premature heat actuation leading to opening of the vent, or reduced sensitivity to heat.

3-2.3 To avoid inadvertent operation, it is important that the actuating element be selected witla regard to the fidl range of expected ambient conditions.

3-2.4 Vents might be a single unit (entire unit opens fidly with a t H single sensor) or multiple units in rows, clusters, groups, or other arrays that satisfy the venting recommendations for the specific hazard. ~///////////////////~ 3-2.5 If the hazard is localized (e.g., dip tank, solvent storage), it is (a) Flat roof recommended that the vents be located directly above such hazard.

3-2.6 It is essential that the specific vent mecharfism and structure Figure 4-3(a) Measurement of ceiling height (H) and curtain depth be ,arranged to be inspected easily. (dc) for fiat roof. 3-3 Methods of Operation.

3-3.1 An automatic mechanism for opening the roof vents is recommended for effective release of heat, smoke, and gaseous by- products. If excessive smoke is likely to be generated prior to the release of sufficient heat to open vents, smoke detectors with appropriate linkages to open vents should be used.

3-3.2 If failure of a vent operating component occurs, it should lead to an open vent condition. Gravity should be used as the opening force, wid~ ensurance that tile opening mechanism cannot be ~///////////////////~ blocked easily by snow or roof debris or internal projections. Alternate opening mechanisms should be reliable. (b) Gabled roof

3-3.3 All mechanic~aily opened vents also should be designed to Figure 4-3(b) Measurement of ceiling height (H) and curtain depth open by manual means. (dc) for gabled roof. ~3.4 To be effective, latching mecbanisms should bejamproof, corrosion-resistant, and resishant to pressure differentials arising from windstorms, process operations, overhead doors, or traffic vibrations. 3-4 Dimensions and Spacing of Vents. The dimensions and spacing of vents ~ua be considered effective where the following criteria are m et:

587 NFPA 204M ~ A97 ROP

(d) Calculations based on tested properties and materials and expected flame flux;

(e) Mathematical models of fire spread and development.

5-3 Actual Tests of the Array Involved. Where an actual calorific test of file specific array under consideration has been conducted l j.j jjjjjj.j,j,jt and the data is in a form that can be expressed as rate of heat release, file data can then be used as input for the methods in this "//////~. guide. Since actual test data seldom produces the steady state assumed for a limited-growth fire or the square of time growth (c) Sloped roof assumed for a continuous-growth fire, engineering judgment is usually needed to derive the actual input necessary if either of these approaches are used. ffthe computer model LAVENT or other Figure 4-3(c) Measurement of ceiling height (H) and curtain depth model that is able to respond to a rate of heat release versus time (dc) for sloped roof. curve is used, the dam can be used directly. Currently there is no established catalog of tests of specific arrays. Some test data can be found in technical reports. Alternatively, individual tests can be conducted.

Many fire tests do not include a direct measurement of rate of heat release. In some cases, it can be derived based on measurement of mass loss rate using the following equation: Q= rn hc (5-1) H (Q in kW, m in kg/s, h c in kJ/kg)

In other cases it can be derived based on measurement of flame ~////////////////////~ height as follows: (d) Sawtooth roof Q= 37(L + 1.02 D) 5/2 (5-2)

Figure 4-3(d) Measurement of ceiling height (H) ~/nd curtain depth (Qin kW, L in m, D in m) (dc) for sawtooth roof. 5-4 Actual Tests of Arrays Similar to that Involved. Where an actual calorific test of the specific array under consideration cannot be NOTE: If d c exceeds 20 percent of H, H - d c should be not less found, it might be possible to find data on one or more tests that are than 3 m. similar to tlae fuel of concern in important matters such as type of For Figure 4-3(d), this concept valid where Ad/d c is much less fuel, arrangement, or ignition scenario The more the actual tests than 1. are similar to the fuel of concern, the higher the confidence that can be placed in the derived rate of heat release. Added engineering 4-4 Spacing. judgment, however, might be needed to adjust the test data to that approximating the fuel of concern. If rate of heat release has not 4-4.1 The distance between curtain hoards (or between walls been directly measured, it can be estimated using the methods without intervening curtain boards)should not exceed 8 times the provided in Section 5-3. ceiling height to ensure that vents remote from tile fire within the curtained compartmer~t are effective. 5-5 Algorithms Derived from Tests of Arrays Having Similar Fuels and Dimensional Characteristics. 4-4.2 Smaller curtained areas should be used where occupancies are particularly vnlnerable to damage. The distance between these 5-5.1 Pool Fires. In many cases, the rate of heat release of a tested curtain boards should be not less than twice tile ceiling height. This ,array has been divided by a common dimension, such as occupied spacing guidance carl be disregarded for curtain boards that extend floor area, to derive a normalized rate of heat release per unit area. down to a depth of at le:mt 40 percent of the ceiling height. The rate of heat release of pool fires is the best documented and accepted algorithm in dais class. Chapter 5 Predicting the Rate of Heat Release of Fires An equation for the mass release rate from a pool fire is as follows 5-1 Introduction. This chapter presents techniques for estimating [ Babrauskas 1995 ]: the heat release rate of warious fuel arrays likely to be present in buildings where smoke and heat ventin~ is a potential fire safety provision. It primarily addresses the esumation of fiJel concentra- fit" = rh~o l1 - e-'(kel3)D ] (5-3) tions found in storage and mamffacturing IoGttions. NFPA 92B, Guide for Smoke Managevngnt Svst~n.~ in MalL~, Atria, and Large Areas, addresses the types of fuel arrays more common to the types of building situations covered by that stancL'trd. NFPA 204 is applicable The variables rh~o and k,~ for equation 5-3 are as shown in Table 5-5.1 [Babrauskas 1995]. to situations where tile hot layer does not enhance the burning rate. [ Babrauskas 1995 ]. Tile methodsprovided in dtis chapter for estimating the rate of heat release, therefore, are bo.sed on "free burning" conditions where no ceiling or hot gas layer effects are inwdved. It is, therefore, assumed that the burning rate is relatively unaffected by the hot layer.

5-2 Sources of Data. The following sources of data appear in their approximate order of priority, given equal quality of data acquisi- tion.

(a) Actual tes~ of tile array involved;

(b) Actual tests of similar arrays;

(c) Algurithms derived from tests of arrays having similar fuels and dimensional characteristics;

588 NFPA 204M i A97 ROP

Table 5-5.1 Data for Large Pool Burning Rate Estimates

Oe.sit~ hc flU" ~ Material (kg/m o) (MJ/kg) ~l~/m2s)~ (m'l)

Cryogenics* Liquid H 2 70 120.0 0.017 6.1 LNG (mostly CH 4) 415 50.0 0.078 1. I LPG (mostly C 3H8) 585 46.0 0.099 1.4 Alcohols methanol (CH 3OH) 796 20.0 0.017 ~'t ethanol (C2H5OH) 794 26.8 0.015 ,,or Simple organic filels butan~e (C 4 H 10) 573 45.7 0.078 2.7 benzene (C 5H6) 874 40.1 0.085 2.7 hexane (C 6 H 14) 650 44.7 0.074 1.9 heptane (C 7 H 16) 675 44.6 0.101 1.1 xylene (CsHI0) 870 40.8 0.090 1.4 acetone (C 3H6 O) 791 25.8 0.041 1.9 dioxane (C 4H802) 1035 26.2 0.018** 5.4** diethyl ether (C 4 H 10 O) 714 34.2 0.085 0.7 Petroleum products benzine 740 44.7 0.048 3.6 SerOline 740 43.7 0.055 2.1 osine 820 43.2 0.039 3.5 760 43.5 0.051 3.6 810 43.0 0.054 1.6 transformer oil, hydrocarbon 760 46.4 0.039** 0.7** tirol oil, heavy 940-1000 39.7 0.035 1.7 crude oil 830-880 42.5--42.7 0.022-0.045 2.8 Solids polymethyimedlacrylate (C 5H802) n 1184 24.9 0.020 3. $ polypropyiene (C 3H6) n ' 905 43.2 O.O18 polystyrene (C 8H8) 0 1050 39.7 0.034

*For I~ools on dry land, not over water. **Esumate uncertain since only two data points available. tValue independent of diameter in turbulent regime.

The mass rates derived from equation 5-3 are converted to rates of application of external heat flux to the sample while determining heat release using equation 5-1, and tbe beat Of combustion from time to ignition, rate of mass release, and rate of heat release for the tile Table 5-5,1. The rate of heat release per unit area times tile area specific applied flux. Most prominent of the current test apparatus of the pool yields heat release data for the anticipated fire. are the cone calorimeter (ASTM E 1354, Standard Test Method for Heat and Visible Smoke Release Rates for Materials and Products Using an 5-5.2 Other Normalized Data. Odler data based on burning rate Oxygen Consumption Ca/or/meter) and the Factory Mutual calorimeter per unit area in tests have been developed. Tables 5-5.2(a) and (b) [Tewarson 1995]. In addition to these directly measured properties, list these data. it is possible to derive ignition temperature, critical ignition flux, effective thermal inertia (kpc), heat of combustion, and heat of 5-5.3 Other U.~eful Data. There are other data that are not based on results from these calorimeters. Properties not normalized that might be usefifl in developing die rate of heat ruble from these calorimeters and essential to determining flame. release curve. Examples are included in tim Tables 5-5,3(a) through spread in directions not concurrent with the flow of the flame can be (d): obtained from the LIFT (Lateral Ignition and Flame Travel) apparatus (ASTM E 1321, Standard Test Method for Determining 5-6 Calculated Fire D~scrlption Based on T~ted Properties. Material Ignition and F/ame Spread Propert/es). This section presents a concept of the use of fire property test data as the basis of an 5-6.1 Background. It is possible to make general estimates of die analytical evaluation of the rate of heat release involved in the use of rate of beat release of burning materials based on the fire properties a tested material. The approach outlined in this section is based on of that material. The fire properties involved are determined by that presented by Nelson and Forssell [ 1994]. small-scale tests. The most important of these tests are due calorim- eter tests involving bofll oxygen depletion calorimetry and file

589 NFPA 204M -- A97 ROP

Table 5-5.2(a) Unit Heat Release Rates Table 5-5.2(b) Unit Heat Release Ram for Commodities for Fuels Burning in the Open Heat release rate per unit floor area of fully involved combus- Heat Release tibles, based on negligible radiative feedback from the surround- Commodity Rate (kW) ings and I00 percent combustion efficiency. kW per Flammable liquid pool 3300/m ~ of surface m 2 of Flammable liquid spray 560/lpm of flow Commodity Floor Area Pallet stack 3500/m of height Wood pallets, stacked 0.46 m high (6-12% Wood or PMMA*(vertical) moisture) 1,420 - 0.61-m height 100/m of width Wood pallets,stacked 1.52 m high (6-12% mois- - 1.83-m height 240/m of width ture) 4,000 - 2.44-m height 620/m of width Wood pallets,stocked 3.05 m high (6-12% - 3.66-m height 1000/m of width moisture) 6,800 Wood or PMMA Wood pallets, stacked 4.88 m high (6-12% - Top of horizontal surface 720/m a of surface moisture) 10,200 Solid polystyrene (vertical) Marl bags, filled, stored 1.52 m high 400 - 0.61-m height 220/m of width Cartons, con~partmented, stacked 4.5 m high 1,700 - 1.83-m height 450/m of width PE letter trays, filled, stacked 1.5 m high on cart 8,300 - 2.44-m height 1400/m of width PE trash barrels in cartons, stacked 4.5 m high 2,000 - 3.66-m height 2400/m of width PE fiberglass shower stalls in canons, stacked 4.6 m high 1,400 (horizontal) Solid polystyrene 14001m a of surface PE bottles packed in Item 6 6,200 Solid polypropylene (vertical) PE bottles in cartons, stacked 4.5 m high 2,000 - 0.61-m height 220/m of width PU insulation board, rigid foam, stacked 4.6 m - 1.83-m height 350/m of width high 1,900 - 2.44-m height 970/m of width PS jars packed in Item 6 14,200 - 3.66-m height 1600/m of width PS tubs nested in cartons, stacked 4.2 m high 5,400 Solid polypropylene (horizontal) 800/m a of surface PS toy parts in cartons, stacked 4.5 m high 2,000 PS insulation board, rigid foam, stacked 4.2 m *PMMA, Polymethyi Methacrylate (Plexiglass. Lucite. Acrylic) high 3,300 PVC hordes packed in Item 6 3,400 PP tubs packed in Item 6 4,400 PP & PE film in rolls, stacked 4.1 m high 6,200 Methyl alcohol 740 Gasoline 3~oo Kerosene 3,300 Diesel oil 2,000

PE = Polyethylene PP = Polypropylene PS = Polystyrene PU = Polyurethane PV = Polyvinyi chloride

590 NFPA 204M m A97 ROP

Table S-&S(8) Chamcteri~'cs d Ignition Sources Table 5-5.$(©) Maximum ~ gelem~ Itat~ from Fine Det~_,x.~_'onInstitute Amdyds

Typical mum mum Approvlmmt~. Hemt Burn Flame FLmne Heat Valuw Ouqmt Tune"Helzht Width Flux 0~w) (~ (s) (ram) (nun) OkWlmt) Medium wastebasket with milk cartons 100 Cigarette 1.1 g (not Large barrel with milk cartons 140 puffed, laid on solid Upholstered chair with polyurethane foam 350 surface), bone dry, 5 1200 - - 42 conditioned to 50% !~t~z foam mattress (heat at room door) 1200 R.H. 5 1200 - - 35 Furnished living room (heat at open door) 4000-8000 Methenamine pill, 0.15 g 45 90 - - 4 Match, wooden (laid on solid surface) 80 20-30 30 14 18-90 Wood cribs, BS 5852 Part2 No. 4 crib, 8.5 g 1000 190 15 d No. 5 crib, 17 g 1900 900 17 d No. 6 crib, 60 g 2600 190 20d No. 7 cry, 126 g 6400 350 25 ~ Crumpled brown lunch bag, 6 g 1900 80 Crumpled wax paper, 4.5 g (tight) 1800 25 Crumpled wax paper, 4.5 g (loose) 5300 20 Folded double-sheet newspaper, 22 g (bottom ignition) 4000 100 Crumpled double- sheet newspaper, 22 g (top ignition) 7400 40 Crumpled double- newspaper, 22 g (bottom igni- tion) 17,000 20 Polyethylene waste- basket, 285 g, filled with 12 milk car- tons (390 g) 50,000 200b 550 200 35" Plastic u-ash bags, 120,000 filled with cellulosic to trash (1.2-14 kg) e 350,000 200b • Time duration of significant flaming b Total bum time in excess of 1800 sec c AJ measured on simulation burner • Measured from 25 mm away • ReSults vary greatly with packing density

Table 5-SJCo) ~_,,s~teristlcs of Typical Furnlshm"p u lpdtlon Sources

Malimum Thermal Total MssJmum Radiation Total Heat Rate of Heat to C~nter Mass Contest Relmuse of Floor" 0~ ~J) 0~w) 0~W/mb Waste paper baskets 0.73-1.04 0.7-7.3 4-18 0.1 Curtain& velvet, cotton 1.9 24 160-240 1.3-3.4 Curtains, acrylic/cotton 1.4 15-16 130-150 0.9-1.2 TV sets 27-33 145-150 120-290 0.3-2.6 Chair mockup 1.36 91-22 63-66 0.4-0.5 Sofa mockup 9.8 49 1.q0 0.9 Arm chair 26 18 160 1.2 Christmas trees, dry 6.5-7.4 11-41 500-650 3.4-14 • Measured at approximately 2 m away from the burning object

591 NFPA 204M I A97 ROP

Table 5-5.$(d) Ma~ Load ud Heat Release ltat~ of

Ham Comlmatible Peakm Peakq o0 s~t, Frmne Paddln~ Fabric ln~rllner (~/sec) (kW) C12 17.9 17.0 traditional easy chair wood cotton nylon - ]9.0 290" F22 31.9 traditional easy chair wood cotton (FR) cotton - 25.0 370 F23 31.2 traditional easy chair wood cotton (FR) olefin - 42.0 700 F27 29.0 traditional easy chair wood mixed cotton - 58.0 920 F28 29.2 traditional easy chair wood mixed cotton - 42.0 730 CO2 13.1 12.2 traditional easy chair wood cotton, PU olefin - 13.2 800 b CO3 13.6 12.7 traditional easy chair wood cotton, PU cotton - 17.5 460* CO1 12.6 11.7 traditional easy chair wood cotton, PU cotton - 17.5 260" CO4 12.2 11.3 traditional easy chair wood PU nylon - 75.7 1350 b C16 19.1 18.2 traditional easy chair wood PU nylon neoprene NA 180 F25 27.8 traditional easy chair wood PU olefin - 80.0 1990 T66 23.0 traditional easy chair wood PU, polyester cotton - 27.7 640 F21 28,3 traditional easy chair wood PU (FR) olefin - 83.0 1970 F24 28.3 traditional easy chair wood PU (FR) cotton - 46.0 700 C13 19.1 18.2 traditional easy chair wood PU nylon neoprene 15.0 230* C14 21.8 20.9 traditional easy chair wood PU olefin neoprene 13.7 220* C15 21.8 20.9 traditional easy chair wood PU olefin neoprene 13.1 210 b T49 15.7 easy chair wood PU cotton - 14.3 210 F26 19.2 thinner easy chair wood PU (FR) olefin - 61.0 810 F33 39.2 traditional Ioveseat wood mixed cotton - 75.0 940 F31 40.0 traditional loveseat wood PU (FR) olefin - 130.0 2890 F32 51.5 traditional sofa wood PU (FR) olefin - 145.0 3120 T57 54.6 Ioveseat wood PU, cotton PVC - 61.9 1100 T56 11.2 office chair wood latex PVC - 3.1 80 CO9gr64 16.6 16.2 foam block chair wood (part) PU, polyester PU - 19.9 460 CO7/T48 11.4 11.2 modern easy chair PS foam PU PU - 38.0 960 C10 12.1 8.6 pedestal chair rigid PU foam PU PU - 15.2 240* Cll 14.3 14.3 foam block chair - PU nylon - NA 810b F29 14.0 traditional easy chair PP foam PU olefin - 72.0 1950 F30 25.2 traditional easy chair rigid PU foam PU olefin - 41.0 1060 C08 16.3 15.4 pedestal swivel chair molded PE PU PVC - 112.0 830b CO5 7.3 7.3 bean bag chair - polDtyrene PVC - 22.2 370* CO6 20.4 20.4 frameless foam back chair - PU acrylic - 151.0 2480 b T50 16.5 waiting room chair metal cotton PVC - NA < 10 "1"53 15.5 1.9 waiting room chair metal PU PVC - 13.1 270 T54 27.3 5.8 metal frame loveseat metal PU PVC - 19.9 370 T75/F20 7.5(x4) 2.6 stacking chairs (4) metal PU PVC - 7.2 160 • Estimated from mass loss records and assumed Wl~ b Estimated from doorway gas concentrations

5-6.2 Discussion of Measured Properties. Table 5-6.2 lists the type of In Table 5-6.2, the rate of heat release (RHR), mass loss, and time fire properties obtainable from the cone or Factory Mutual to ignition are functions of the externally applied incident radiant calorimeters and similar instruments. heat flux imposed on the tested sample. The purpose of the externally applied flux is to simulate the fire environment surround- ing a burning item. In general, it can be estimated that a free- burning fuel package (i.e., one that burns in the open and is not Table 5-6.2 Relation of Calorimeter-measured Properties to Fh'e affected by energy feedback from a hot gas layer of a heat source Analysis other t~an its own fine) is impacted by a flux in the range of 25 kW/m to 50 kW/m . ffthe fire is in a space and conditions are 9 Flalne Fire Size approaching2ilashover, this can increase to the range ofS0 kW/m" Property Ignition Spread to 75 k~q/m . In fully developed, post-flashover fires, a range of 75 kW/m" to over 100 kW/m" can be expected. Thefollowing is a Rate of heat release t XXX XXX discussion of the individual properties measured or derived and the Mass loss t XXX usual form used to report the property. Time to ignitiont XXX XXX Effective thermal properties* XXX XXX (a) Rate of Heat Release. Rate of heat release is determined by Heat of combustion* XXX XXX oxygen depletion calorimetry. Each test is run at a user-specific Heat of gasification* XXX incident flux and for either a predetermined period of time or until Critical ignition flux* X~X XXX the sample is consumed. The complete results are presented in the Ignition temp.* XXX XXX form of a plot of rate of heat release against time, with the level of applied flux noted. In some cases, the rate of heat release for several t Property is a function of the externally applied incident flux. tests of the same material at different levels of applied flux is plotted * Derived properties from calorimeter measurements. on a single curve for comparison. Figure 5.6.2(a) is an example of such a plotting.

592 NFPA 204M -- A97 ROP

2000 applied flux in that test, and the effective thermal inertia of the sample. It is reported at a single temperature. If the test includes a pilot flame or spark, the reported temperature is for piloted ~ 1600 " kgnition;...... fit_here is no pilot present, the temperature is for autoigmuon. Most avatlable data ts for pdoted tgmuon. __ 1200 ! ,e'~ .... 5-6.$ Ignition. Equations for time to ignition, tito are given for both 1000 j .F z--,. thermally thin and thermally thick materials, as ~etined in 5-6.3(a) and (b). For materials of intermediate depth, estimates for ti~ 6oo ~ ; i \ necessitate considerations beyond the scope of this presentation [Drysdale 1985, Carslaw and Jaeger 1959].

~" 200 (a) Thermally Thi0J~/laterial~ Relative to ignition from a constant o ~'~ " ~ ~ ~' incident heat flux, q: , at the exposed surface and with relatively 0 200 400 600 800 1000 1200 1400 1600 small heat transfer Io~ses at the unexposed surface, a thermally thin Time (s) material is a material whose temperature is relatively uniform throughout its entire thickness,/, at t =tig. For example, at t = rig:

Texposed- Tunexposed = Tig - Tunexposed < 0.1 (Tig - To) (5-4) Equation 5-4 can be used to show that a material is thermally thin Figure 5-6-2(a) Typical graphic output of cone calorimeter test. [Ca~slaw and Jaeger 1959] if:

Often only the peak rate of heat release at a specific flux is I < 0.6(tig a) 1/2 (5-5) reported. Table 5-6.2(a) is an example. For example, for sheets ~f ngaple or oak wood (where the termal diffusivity a = 1.28 x 10- m~/s [DiNenno et al 1995]), if tie = 35 s Table 5-6.2(a) Avert I ,~ Maximum Heat Release Rat~s (kW/m2) is measured in a piloted ignition test, then, according to eqd~tion 5-5, if the sample thickness is less than approximately 0.0013 m, the 25 kW/m 2 50 kW/m 2 75 kW/m 2 Material Orientation unexposed surface of the sample can be expected to be relatively Exposing Flux Ex po~ ng Flux :ExpofingFlux close to T:_ at the time of ignition, and the sample is considered to be therm~y thin. PMMA Horizontal 65O 9O0 1300 Vertic=d 56O 720 1300 Pine Horizontal 140 240 265 The time to ignition of a thermally thin materiN subjected to Vertical 130 170 240 incident flux above a critical incident flux is: Sample A Horizontal 125 200 250 Vertical 9O 130 22O S.'m~ple B Horizontal 140 175 240 Vertical 6O 200 330 t. = 9clfTig-~ To) Sample C Horizontal 215 25O tg (5-6) Vertical 165 170 Sample D Hotizonml 7O 145 145 qi Vertical 125 125 (b) Thermally Thick Materials. Relative to the .t~e of ignition test described in 5-6.3(a), a sample of a material ofa thtckness,/, is considered to be thermally thick if the increase in temperature of (b) Mass Logs Rate. Mass loss rate is determined by a load cell. the unexposed surface is relatively small compared to that of the The method of reporting is identical to that for rate of heat release. exposed surface at t = tig. For example, at t = tig: In the typical siutation where the material has a consistent heat of combusuon, the curves for mass loss rate and rate of heat release are Tunexposed- T O < 0.1 (Texposed- To) = 0.1 (Tig - To) (5-7) similar in shape.

(c) Time to Ignition. Time to ignition is reported for each Equation 5-7 can be used to show that a material is thermally thick individual test ,and applied flux level conducted. [Carslaw andJaeger 1959] if:

(d) Effective Thermal Inertia (kpc). Effective thermal inertia is a t > 2(rig a) 1/2 (5-8) measurement of the heat rise response of the tested material to the heat flux imposed on the sample. It is derived at the time of ignition For example, according to equation 5-8, in the case of an ignition and is based on the ratio of the actual incident flux to the critical test on a sheet of maple or oak wood, if ti~ = 35 s is measured in a ignition flux and the time to ignition. A series of tests at different piloted ignition test, then, if the sample fffickness is greater than levels of applied flux is necessary to derive the effective thermal approximately 0.0042 m, the unexposed surface of the sample can inertia. Effective thermal inertia derived in this manner can differ be expected to be relatively close to T at t = t. and the sample is • . O I from and be preferable to that derived using handbook data for the considered to be thermally duck. g values of k, p, and c derived without a fire. Time to ignition of a thermally thick material subjected to incident (e) Heat of Combustion. Heat of combustion is derived by flux above a critical incident flux is:

dividing the measured rate of heat release by the measured mass loss • It 2 rate. It is normally reported as a single value, unless the sample is a tig - (x/4)k p c[ (Tig - To)/qi ] (5-9) composite material and the rates of heat release and mass loss vary significantly with time and exposure. It should be noted that a particular material is not intrinsically thermally thin or thick (i.e., the characteristic of being thermally (f) Heat of Gasification. Heat of gasification is the flux needed to thin or thick is not a material characteristic or property) but also pyrolyze a unit mass of fitel. It is derived as a heat balance and is depends on the thickness of the particular sample (i.e., a particular usually reported ,as a single value in terms of the amount of energy material can be implemented in either a thermally thick or per unit mass of material rele:Lsed (e.g., kJ/g). thermally dtin configuration).

(g) Critical Ignition Flux. Critical ignition flux is the minimum (c) Propagation Between Separate Fuel Packages. Where the level of incident flux on the sample needed to i~gnite the sample concern is forpropagation between individual separated fuel given an unlimited time of application. At incident flux levels less packages, incident flux can be calculated using traditional radiation than the critical ignition flux, ignition does not take place. heat transfer procedures [Tien et al 1995].

(h) Ignition Temperature. Ignition temperature is the surface The rate of radiation heat transfer from a flaming fuel package of temperature of a sample at which flame occurs. This is a sample total energy release rate, Q, to a facing surface element of an material value that is independent of the incident flux. It is exposed fuel package can be estimated fi'om: derivable from the calorimeter tests, the LIFT apparatns test, and other tests. It is derived from the time to ignite in a given test, the q~t = XrQ/4~r2 (5-10)

593 NFPA 204M ~ A97 ROP

5-6.4 Estimating Rate of Heat Release. As discussed in 5-6.2, t~ts have demonstrated that the energy feedback fr-pm a burning~fuel package ranges from approximately ,'25 kW/m--to 50 kW/m% For a Av re~onable conservative analysis, it is recotrug2ended that test data developed with an incident flux ofS0 kW/m be used. For a first order approximation, it should be assumed that all of the surfaces that Can be simnltaneous]y inw)lved in I)llrniJlg are releasing energy I L°°t°n° il at a t=tte equal to that determined by testing the material~n a fire propertieg calorimeter with an i0cident flux of,90 kW/m for a free- burning material and 75 kW/m" to 100 kW/m" for post-fiashover conditions.

In making this estimate, it is necessary to `a~sume that all surfaces that c:m "see" an exposing flame (or superheated gas, in the post- flashover condition) are burning and releasing energy and mass at the tested rate. If sufficient air is present, dxe rate of heat release estimate is then calculated ms the product of dye exposed m-ca and the rate of beat release per unit area ,as determined in the test calorimeter. Where there is test data taken at the incident flux of Figure 6-1.1.1 Schematic of venting system. dye exposing flame, die tested rate of heat release should be used. Where the test data is for a different incident flux, the burnin~ rate 6-1.1.2 First equilibrium conditions are assumed, with the layer should be estimated using die beat of gasification as expressedin already having formed. The smoke interface is level with the bottom equatio!a,,5-11 to calculate the mass burning rate per unit area. of the curtain boards. At equilibrium, the mass flow rate into the ., q,, smoke layer ( rhp ) equals the mass flow rate out of the vent ( rho ). m =-- (5-11) hg 6-1.1.3 The vent area calculated for equilibrium conditions would correspond to the area needed for a long-term steady fire, or the The resulting mass loss rate is then multiplied by the derived area needed at the end of a design interval for a very slow-growing effective heat of combustion and the burning area exposed to the fire. For shorter term steady fires and for faster growing fires, tile incident flux to produce the estimated rate of heat release as follows: calculated equUibrium vent area will prevent the smoke interface from descending completely to the bottom of the curtain boards. qi'= ,/*"h cA (5-12) Therefore, equilibrium calculations represent a safety factor in the design. 5-6.5 Flame Spread. If it is desired to predict the growth of fire ,as it 6-1.2 Mass Flow Rate in Plume, rhp. fipropagates over colnbusdble surfaces, it is necessm'y to estimate ame spread. The computation offl;,me spread rates is an emerging 6-1.2.1 The mass flow rate in the plume depends on whether technologystill in an embryonic stage. Predictions shotfld be locations above or below the mean flame height are considered (i.e., considered `as order of magnitude estimates. whether the flames are below the smoke interface or reach into the smoke layer). The flame height, L, is calculated from equation 6-1 Flame spread is the movement of the flame front across the surface [Heskestad 1995] as follows: of a material that is burning (or exposed to an ignition flame) where the exposed surface is not yet fidly inwdved. Pbysically, flame spread L = -1.02D + 0.235Q2/5 (6-1) can be treated `as a succession of ignitions resulting from the heat energy produced by the burning portion of a material, its flmne, ,and (L and D in m; Qin kW). any other incident heat energy imposed upon die unburned surface. Other sources of incident energy include another burning object, where: high temperature gases that can accumulate in the upper portion of D = base diameter of fire an enclosed space, and the radiant beat sources used in a test Q = total heat release rate apparatus such as One cone calorimeter or the LIFT mecbanism. For analysis purposes, flame spread can be divided into two categories, 6-1.2.2 When the mean flame height, L, is below the interface and z that which moves in the same direction as the flame (concurrent or is at or above the flame height but at or below the interface height, wind

• tt Qc = convective heat release rate (approx. 0.7 Q) qi L z = height above tile base of the fire zA = height of "virtual origin" above the base of the fire (below the V = (5d3) u . . base of the fire, tf negative) kpc(Tig-Ts) 2 6-1.2.3 When z is at or below the flame height and at or below the The values for krc and igrfition temperature are calculated from interface, the mass flow rate can be expressed as follows [Heskestad the cone calorimeter as discussed. For this equation, the flame 1995] (see 6-1.4.2 for application): length (L) is measured from the leading edge of the burning region.

5-6.6 Classification of Fires for Engineering Equations. The ~/*p = 0.0056 Qc " z~ L (6-3) engineering equations in Section 6-1 are appropriate for steady fires, limited growth fires, and t-squared forms of continuous growth fires. It should be noted that, at the mean flame height, the mass flow rate is: Chapter 6 Sizing Vents '/*pL = 0.0056 Qc (6-4) 6-1 Hand Calculations.

6-1.1 Elements of Problem. 6-1.2.4 The virttlal origin, Zo, is the effective point source of the fire plume [Heskestad 1995]: 6-1.1.1 In Figure 6-1.1.1, H is the distance between the base ofdl~ fire and the ceiling; d c is the depth of die curtain boards, and d is z o = 0.083 Q2/5 _ 1.02D (6-5) the design depth of the smoke layer; rh~ is the mass flow rate of hot gas from the fire plume into the smoke'layer; rhv is the mass flow (Qin kW, D in m) rate of hot ga.s out of fl~e vent (or vents) ; and A v is tile vent ,area (total vent area in curtained compartment, if more than one vent exists). 594 NFPA 204M -- A97 ROP

6-1.2.5 For combustibles that extend in depth, such as storages, tile geometric vent area, A v. For simple apertures, Awl can be taken as base of the fire is selected in a horizontal plane containing die worst- 0.61 times file geometric throughflow area. In other words, the o, se ignition location (i.e., the lowest point of the combustible calculated vent areas, Ava, should be increased by a factor of 1/0.61 array). Consequently, the base of the desigrl fire is often selected on to establish the geometric vent area. If the discharge coefficient is the floor of the building. different from 0:61, the calculated vent areas should be multiplied by the ratio of 0.61 to th¢ actual discharge coefficient. 6-1.73 Mass Flow Rate Through Vents, th v . 6-1.4.3 The calculated vent areas also should be increased by the 6-1.73. l The inlet area for fresh air in the building below die design multiplier, M, in equation 6-6 to account for limited inlet area for level of the smoke interface, A b can throttle the inlet flow if it is not fresh air. sufficiently large. The effective vent area with throttled inlet area is smaller than the unthrottled area, and tile calculated vent area, Av, 6-1.4.4 The required aerodynamic vent area, Ava, is calculated with should be increased by the following multiplier, M, [Hinkley 1988]: tile aid of equations 6-1, 6-2, 6-3, 6-5, 6-8, and 6-9, setting z = H - d, where H is file ceiling height above the base of the fire (usually the M = [ 1 + (Av/A/)2(To/T) ] 1/2 (6-6) foor). This vent area is distributed among individual vents within the curtained compartment_ In this case, T o is the ambient temperature and T is the smoke layer temperature. Where T = 350 K, T O = 2973 K, and the vent area 6-1.4.5 Steady Fires (Limited Growth Fires). and inlet ,area are the same (Av/A i = 1), the multiplier is 1.99. Increasing the inlet area to twice the vent ,area, so dlat Av/A i ~ 0.5, 6-1.4.5.1 For steady fires, or fires that do not develop beyond a tile multiplier is 1.08. Reducing die inlet area to 1/9 the vent area, maximum size, the required vent area per curtained compartment is so that Av/A i = 2, the multiplier is 1.90. The required vent areas, Av, calculated based on the maximum anticipated heat release rate, Q determined by 6-1.4, should be adjusted using die appropriate and Qo the associated distance from the fire base to the bottom of multiplier from eqnation 6-6, including the effects of the tempera- the curtain boards or to the design elevation of the smoke interface, ture ratio, To/T. H - d, and the estimated fire diameter, D.

6-1.3.2 Equating die buoyancyhead across tile vent tothe dynamic 6-1.4.5.2 These fires include special-hazard fires and fires in head in the vent, from Bernoulli's equation, provides die following: occupancies with concentrations of combustibles separated by sufficiently wide aisles. The minimum aisle width to prevent lateral 1/2 pu 2 = Apgd (6-7) spread by radiation, Wmi n [Alpert and Ward 1984], can be estimated from equation 6-10 for radiant heat flux from a fire and a where p is file smoke layer density, Ap = Po - P, Po is die ambient consergatively low value for file ignition flux of most materials (20.4 density, u is tile gas velocity in die vent, and g is the acceleration of kW/m'): gravity. The ma.qs flow through the vent is file product of gas density, velocity, and aerodynamic area (AVa), which, with die aid of Wmi n = 0.042 Q1/2 (6-10) equation 6-7 and the ideal gas law, becomes: (Qin kW, Wmi n in m) r 2 "~I/2[ToAT] 1/2 Tile values produced by equation 6-10 can be produced from 'hv=12P° gJ [TJ AVadl/2 (6-8, equation 5-10 ifX r is assumed to be 0.5. 6-1.4.5.3 The fire diameter, D, is taken as the diameter of a circle In this case, T is tile smoke layer temperature and AT = T- T O . having file same area as the floor area of fl~e fuel concentration.

6-1.73.3 It should be noted that the factor [ (T o AT)/T 2} 1/2 is quite 6-1.4.5.4 Tile heat release rate is taken as the heat release rate per insensitive to temperature as long as the smoke layer temperature unit area times the floor area of the fuel concentration. The rise, AT, is not small. For example, assuming T o -- 294 K (21°C), the maximum foreseeable storage height (above the fire base) and factor varies through 0.47, 0.50, and 0.47 ,as die smoke layer associated heat release rate shouldbe considered. temperature rise varies through 150 K, 320 K, and 570 K. At a temperature rise oft0 K, the factor is 0.38, and, ata temperature 6-1.4.5.5 The heat release rate per unit area might be available from rise of 20 K, it is 0.24, about 1/2 its maximum value. Consequendy, listings for a given storage height, such as Table 5-5.2(b). To roof venting by natural ventilation becomes increasingly less establish estimates for other than specified heights, it can be effective as the smoke layer temperature decreases. For low smoke assumed that the heat release rate per unit area is proportional to layer temperatures, powered ventilation as covered in NFPA 92B, die storage height, based on tests hyYu [Yu 1991] and the data in Guide for Smoke Managcnurat .Systems in Malls, Atria, and Large Areas, Table 5-5.2(b) for wood pallets. For fuel configurations that have should be considered. not been tested, the procedures of Chapter 5 should be used.

6-1.3.4 A representative smoke layer temperature rise, AT, can be 6-1.4.5.6 There is a distinctpossibility that a combustible storage estimated as a fr:tction, r, of the adial)atic temperature rise, ATa as array could collapse before die end of the design interval of the follows: venting system. (The design interval might end, for example, when manual fire fighting is expected to begin.) The fire diameter AT = rAT= r Qc/(Cplh p ) (6-9) increases, contributing to increased smoke production (via a lower flame height and virtual origin). However, the heat release rate and fire growth rate after collapse are likely to be smaller than with no where cnis the specific heat of air at consmntpressure. Tile plume collapse. Consequently, it is reasonable to assume that the net effect mass flokv, ,h~, is evahmted from equations 6-2or 6-3, with z = H - d of collapse is not significant for the calculation procedure. (where H is tile ceiling height above file base of the fire). Equation 6-2 is used if the flame height, L (equation 6-1), is smaller than (H - 6-1.4.6 Growing Fires (Continuous-Growth Fires). d) and equation 6-3 is used if tile flame height, L, is larger fllan (H - d). From experiment [Hinkley 1992], it is evident that file fraction, 6-1.4.6.1" A t-squared fire growth is assumed: r, decreases with the distance from tile fire, but a representative value, r = 0.5, can be used. Adopted values of temperature rise Q= 1000 (t/tg) 2 (6-1 l) should be limited to 1000°C. (Qin kW; tand tg in s) 6-1.4 Required Vent Area. where t is the time from effective ignition (seeFigure 6-1.4.6.1) 6-1.4.1 The required actual vent area is file minimum total area, A v, following an incubation period, and t~is the time, t, at which the of all the open vents in a curtained compartment needed to prevent fire exceeds an intermediate size of 11700 kW. The growfll time, tg, is the smoke from underspilling the curtain boards or from descend- a measure of the fire growth rate; the smaller tile growth time, the ing below tile design level of dle smoke interface. faster the fire grows. 6-1.4.2 The area, Ava, calculated according to file procedures in 6-1.4 is the aerodynamic vent area, which is always smaller titan the

595 NFPA 204M ~ A97 ROP

Table 6-1.4.6.3 Continuous-Growth Fires

Growth times of developing fires in various combustibles, assuming 100 percent combustion efficiency. (PE = polyethylene; PS = polystyrene; PVC = polyvinyl chloride; PP = polypropylene; PU = polyurethane; Continuouslygrowing FRP = Fiberglass-Reinforced Polyester)

Growth Time (s) 3000 1. Wood pallets, stacked 0.46 m high 160-320 (6-12% moisture) 2. Wood pallets, stacked 1.52 m high 90-190 (6-12% moisture) _¢ 3. Wood pallets, stacked 3.05 m high 80-120 (6-12% moisture) aooo 4. Wood pallets, stacked 4.88 m high 75-120 (6-12% moisture) 5. Mail bags, filled, stored 1.52 m high 190 6. Cartons, compartmented, stacked 60

"1" 4.57 m high 7. Paper, vertical rolls, stacked 6.10 m 17-28 high 1000 8. Cotton (also PE, PE/Cot 22-43 Acyrlic/Nylon/PE), garments in 3.66 m high rack Incubation 9. "Ordinary combustibles" rack 40-270 storage, 4.57-9.14 m high 10. Paper products, densely packed in 470 J cartons, rack storage, 6.10 m high 11. PE letter trays, filled, stacked 1.52 180 ~<~G rowlh m high on cart m time Time 12. PE trash barrels in cartons, stacked 55 ~_._Effective 4.57 m high ignition time 13. FRP shower stalls in cartons, 85 I stacked 4.57 m high 14. PE bottles packed in Item 6 85 Figure 6-1.4.6.1 Conceptual illustration o f continuous-growth fire. 15. PE bottles in cartons, stacked 4.57 75 m high 16. PE pallets, stacked 0.91 m high 150 6-1.4.6.2 Instead of growth time, ,tg, t-squared fire growth can be 17. PE pallets, stacked 1.83-2.44 m higl~ 32-57 expressed in terms of a fire growth coefficient, Ctg, ,as follows: 18. PU mattress, single, horizontal 120 19. PU insulation board, rigid foam, 8 Q= c~gt2 (6-11a) stacked 4.57 m high 20. PSjars packedin Item 6 55 (Qin kW, tins, ff.g in kWs "2) 21. PS tubs nested in cartons, stacked 110 4.27 m high Comparing equation 6-1 la and equation 6-11, the following 22. PS toy parts in cartons, stacked 4.57 120 relation exists: m high 23. PS insulation board, rigid foam, 7 (6-11b) t~g = 1000/t~ stacked 4.27 m high 24. PVC bottles packed in Item 6 9 25. PP tubspacked in Item 6 10 6-1.4.6.3 Growth times tot a number of combustible arrays have 26. PP and PE film in rolls, stacked 40 been obtained; see Table 6-1.4.6.3. These are specified for certain 4.27 m high storage heights. Actual tests have demonstrated [Yu 1991 ] that it is 27. Distilled spirits in barrels, stacked 25-40 reasonable to assume that the instantaneous heat release rate per 6.10 m high unit height of the storage array is insensitive to the storage height. Such behavior corresponds to the gro~.h time, tg, being inversely proportional to the square root of the storage he~glm Alternatively, (a) Arrival of tile emergency response team; it corresponds to the fire growth coefficient, a~, being directly proportional to the storage height. For ex,'unO]e if the storag% (b) Arrival of fire fighters from public fire department;

(c) Completion of evacuation; times the growth time from the test. For fiael configurations that (d) Other critical events. have not been tested, the procedures discussed in Chapter 5 migl!t be applicable. 6-1.4.6.5 Tile instantaneous diameter of the fire needed for calculadon of L and z o can be calculated from the instantaneous 6-1.4.6.4 A venting system needs to be able to mainuain the smoke heat release rate, O~ and data on the heat release rate per unit floor layer above the design level from the time of ignition until the end area Q~ (according to listings such as in Table 5-5.2(f~) and of the design interval, tr, where t r is measured from the time of assuming. Q n is, proporuonal, to storage height): detection, td. , ,,d/2 At die end of the design interval, the heat rel~qse rate is: v = [4Q/~Q ) (6-13) Q= 1000[(t r + td)/tg)] 2 (6-12) 6-1.4.7 Detection.

(Q in kW; tr, td, and tg in s) 6-1.4.7.1 Detection should be either by heat or smoke detectors installed at each vent, or by heat or smoke detectors installed on a The end of the design interval, t r, may be selected to correspond to regular matrix. various critical events, including: 6-1.4.7.2 The earlier the fire is detected, the earlier evacuation of the building can begin. Furthermore, for continuous-growth fires,

596 NFPA 204M ~ A97 ROP the earlier the fire is detected and vents actuated, the smaller the equivalent, for the combustible of the occupancy and the detector fire size at the end of die design interval, and the smaller the model to be installed. A temperature rise of 1O°C or less at required vent arem In the c~ase of limited-growth fires, the earlier detection is considered representative of a reasonably sensitive the fire is detected and vents actuated, the less likelyan inidal detector for a specific combustible. underspill of smoke at the curtain boards and smoke layer excursion to low heights. 6-1.4.7.2.3 The response data in NFPA 72, NationalFireAlarm Code, as well as the temperature and velocity relations in equations 6-15 6-1.4.7.2.1 For the G-alculations of the detection time, t d, of the first and 6-16 assume extensive, flat, horizontal ceilings. This assump- detector projected to operate and the detection time of the detector tion might appear optimistic for installations involving beamed controlling the actuation of the last projected vent to operate in a ceilings. However, anydelayin operation due to beams is at least curtained area prior to the end of the design interval, the design fire partially offset by opposite effects of: should be assumed f~hest possible from both detectors within die curtained area. (a) Heat banking up under the ceiling because of curtain boards or walls; and 6-1.4.7.2.1.1 Detection times for heat detectors and fusible links, the latter serving as common actuators for commercial heat and smoke (b) The nearest vent or detector usually being closer to the fire vents, can he determined with the ,aid of NFPA 7'~, NationalFire than the assumed, greatest possible distance. Alarm COd~ provided the spacing between detectors does not exceed 15m. 6-1.4.7.$ Detection Computer Programs. 6-1.4.7.2.1.2 If fl~e spacing between heat detectors (or fiasible links) 6-1.4.7.3.1 A computer program, known as DETACT-T2 [Evans and exceeds 15 m, the detection time can be determined from the Stroup 1985], is available for calculating detection times of heat following response differential equation [ Heskestad 1989(A) ]: detectors in continuous-growth, t-squared fires, equivalent to solving equation 6-14 with tile aid of equation 6-16 and effectively a d(ar,) _ u predecessor of equation 6-15. The program calculates detection --[aT-aTe] (6-14) times for smoke detectors (see 6-1.4.7.2.2) based on the effective dt RTI predecessor. The effective predecessor [Heskestad and Delichatsios 1979] assumes complete combustion of the test fuel used in the where: experiments leading to the equation, whereas equation 6-15 is based on the actual heat of combustion. However, DETACT-T2 can still be AT e = temperature rise (from ambient) of heat-responsive used, provided the projected fire growth coefficient, 0~g, is element multiplied by the factor 1.67. t = time u = gas velocity at detector site 6-1.4.7.$.2 Another computer p.rogram, known as DETACT-QS AT = gas temperature rise (from ambient) at detector sitel/2 [Evans and Stroup 1985] is avadable for calculating detection times RTI = response tlme index [ Heskestad ,and Bill 1989] (gu , of heat detectors and smoke detectors in fires of arbitrary fire where ~ is the time constant of the heat-responsive element for growth. Quasisteady gas temperatures and velocities are assumed, convective heating). i.e., instantaneously, gas temperatures and velocities under the ceiling are assumed to be related to the heat release rate as in a steady-state fire. For t-squared fires, this program would be less Tide detection time is the time, t = t d, when T e reaches the value accurate than DETACT-T2 (if the projected fire growth coefficient is associated with the rated temperature of the heat-responsive increased as described in 6-1.4.7.3.1), especially for fast growing element. fires, but DETACT-QS does provide a means of handling fires which cannot be approximated as t-squared fires. 6-1.4.7.2.1.$ In the case of contimmus-growth t-squared fires, gas temperatures for the calculation in equation 6-14 ~ua be determined 6-1.4.8 Selection of Design Basis. The vent area in a curtained from the following [Heskestad and Delichatsios 1989]: compartment should not be required to exceed the vent area calculated for the largest limited-growth fire predicted for the combustibles beneath the curtained area. Using sufficiendy small concentrations of combustibles and aisle widths at least as large as calculated from equation 6-10, it might be possible to satisfy the venting needs using smaller vent areas than required bya continu- ous-growth vent design.

6-1.5 Limitations. (6-15) 6-1.5.1 A design for a given building and its combustible contents and their distribution would comprise selecting a design basis (T in °C, tg in s, arid H in m). (limited-growth versus continuous-growth fire) and establishing the following parameters: where the interpretation A T=O is applied when the numerator of the first bracket is zero or negative and: (a) Layout of curtained areas;

H = ceiling height above the fire I~ase (b) A curtain depth; r = radius from fire axis (c) Type detector and specific characteristics;

6-1.4.7.2.1.4 Gas velocities for the calculation in equation 6-14 can (d) Detector spacing; be evaluated from a relation between gas velocity and temperature as follows [ Heskestad and Delichatsios ! 989]: (e) An appropriate design interval, tr, following detection for maintaining a clear layer (for continuous-growth fires); u/[(AT/ To )gH" 1 d/2 = 0.59(r / H) -0"63 (6-16) (f) Total vent area per curtained compartlnentg where: (g) Disu'ibution of individual vents; and

T O = ,ambient air temperature (h) An air inlet area. g = acceleration of gravity Certain limitations ~ffventing designs should be observed.

6-1.4.7.2.2 Detection times for smoke detectors can be determined 6-1.5.1.1 The distance from the fire base to the smoke interface, H - with the aid of equation 6-15 as the time to reach a certain tempera- d, is a dominant variable. Some design situations can result in ture rise, AT, at response, which is ,also the foundation for smoke smoke layer temperatures as expressed in equation 6-9 (with r = 0.5) detector spacing curves in NFPA 72, NationalFireAlarm Codg This that exceed 600°C, at which fire can flash over to all the combus- temperature rise shoukl be determined in dedicated tests, or the tibles under die curtained area, which clearly represents an

597 NFPA 204M ~,A97 ROP unacceptable design. Temperature limits exist for structural could have an important influence on the fire-generated environ- members. For example, structural steel has lost approximately half ment. A model might or might not include the effect of wind. A of its strength at a temperature of 540°C. Initiation of charting of model that does include the effect of wind is more difficult to wood members is typically assumed to occur at 280°C. The develop and validate and more complicated to use. Note that the temperature of unprotected steel and surfaces of wood will closely effect of wind is not taken into account in the LAVENT model follow the exposing smoke temperature. Practical options include discussed in 6-2.2. However, by using reasonably well-accepted enforcing limits on areas of rue/concentrations, heights of the matiaematical modeling concepts, LAVENT could be developed to combustible, or both, to limit heat release rate to levels low enough the point where it could be used to simulate this effect. to prevent the occurrence of unacceptable design temperatures. In 6-2.2, a group of phenomena described that, taken together, 6-1.5.1.2 Danger to unprotected steel directly over the fire depends represent a phys=cal basis for estimating the fire-generated environ- on local temperatures over the fire in the smoke layer, which reach ment and tile response of fusible links in well-ventilated compart- the 540°C limit earlier than the average smoke layer temperature ment fires with curtain boards and fusible link-actuated or smoke calculated using r = 0.5 in equation 6-9. In equation 6-9, r = 1 should detector-actuated ceiling vents. The phenomena include: be used to ,assess this limitation, which necessitates further restricting the storage height of the combustible. (a) Growth of the smoke layer in the curtained compartment;

6-1.5.2 The feasibility of roof venting should be questioned when (b) The flow dynamics of the buoyant fire plume; the heat release rate approaches values associated with ventilation control of die burning process (i.e., where the fire becomes (c) The flow of smoke through open ceiling vents; controlled by the makeup air replacing die vented hot gas and (d) The flow of smoke below curtain boards; smoke). Ventilation-controlled fires might be unable to support a clear layer. Tiffs limiting heat release rate is termed Qfeasible and (e) Continuation of the fire plume in the upper layer; can be estimated from the following [Heskestad 1991 ]: (f) Heat transfer to the ceiling surface and the thermal response of Qfeasible = 23,200 (H-d) 5/2 (6-17) the ceiling; (Qfeasible in kW; H ,and d in m) (g) The velocityandtemperature distribution of plume-driven, near-ceiling flows; and Venting at heat release rates greater than Qfe,~ible to maintain a clear layer necessitates larger vent areas than tinose indicated by the (h) The response of near-ceiling-deployed fusible links and smoke calculation scheme provided. detectors. 6-2 Models. All tire phenomena in (a) through (In) are taken into account in the LAVENT model, which was developed to simulate the above class 6-2.1 Mathematical Models to Simulate Fire-Generated Environ- of fire environment. Other models that could be developed for a ments and the Action of Vents. A ceiling vent design is successful to similar purpose would typically be expected to simulate these basic the extent that it controls a fire-generated environment developing phenomena also. in a space of fire origin according to any of a variety of possible specified criteria. For example, if the likely growth rate of a fire in a 6-2.2 The Physical Basis for the Fire Model LAVENT. particular burning commodity is known, a vent system with a large enough vent area, designed to provide for timely opening of the 6-2.2.1 The Basic Fire Scenario. The space to be considered is vents, can be expected to lead to rates of smoke removal that are so defined by ceiling-mounted curtain boards with a fire and with near- large that fire fighters, arriving at the fire at a specified time ceiling fusible link-actuated ceiling vents and sprinklers. The subsequent to fire detection, are able to attack the fire successfully curtained area should be considered as one of several such spaces in and protect commodities in adjacent spaces from being damaged. a large building compartment. Also, by specifying that the curtains be deep enough, they can be thought of as simulating the walls of a To evaluate the success of a particular design it is necessary to single uncurtained compartment. This subsection discusses critical predict the development of the fire environment ,as a function of any' physical phenomena that determine the overall environment in the of a number of physical characteristics dlat define and might have a curtained space up to the time of sprinkler actuation. The objective significant effect on the fire scenario. Examples ofsucin characteris- is to identify and describe the phenomena in a manner that captures tics include: the essential features of this generic class of fire scenario and allows for a complete and general, but concise and relatively simple, (a) The floor-to-ceiling height and area of the space ,and the mathematical/computer simulation. thermal properties of its-ceilihg, walls, and floor; The overall building compartment is assumed to have near-floor (b) The type of barriers that separate the space of fire origin and intake air openings that are large enough to maintain the inside adjacent spaces (e.g., full walls v~th vertical door-like vents or ceiling- environment, below any near-ceiling smoke layers that might form, mou nted cu ru'fin boards); at outside-ambient conditions. Figure 6-2.2.1 depicts the generic fire scenario considered. It is assumed that a two-layer zone-type model (c) The material type and ~ngement of the burning commodi- describes adequately the phenomena under investigation. The ties (e.g., wood pallets in plan-area ,'ua'ays of 3 m x 3 m ~/nd stacked 2 lower layer is identical to the outside ambient. The upper smoke m high); layer thickness andproperties change with time, but, at any time, the layer is assumed to be uniform in space. Conservation of energy and (d) The type, location, and method of deployment of devices that mass along with the perfect gas law is applied to the upper layer. detect the fi?e and actuate the opening of the vents (e.g., filsible links of specified RTI and distrilAuted at a specified spacing distance This leads to equations that necessitate estimates of the net rate of below the ceiling); and enthalpy flow plus heat transfer and the net rate of mass flow to the upper layer. Qualitative features of the phenomena that contribute (e) The size of the open area of the vents tlaemselves. to these flows and heat transfer are described briefly. The best way to predict the fire environment and evaluate the likely effectiveness of a vent design is to use a reliable mathematical model that simulates the various relevant physical phenomena that come Draft curtain Ceiling vents Ceiliflg jet into play during the fire scenario. Such an analytical tool should be designed to solve well-formulated mathematical problems, based on basic relevant principles of physics and fundamentally sound, well- established, empirical relationships. Even in the case of a particular class of problem, such as an engineering problem ,associated with successfitl vent design, there is " Crt a good dealof variation among applicable mathematical models that could be developed to carry out the task. Such models might differ from one another in the number and detail of the individual physical phenomena taken into account. Therefore, the list of physical characteristics that define and could have a significant effect Figure 6-2.2.1 Fire in a building space with curtain boards and on the fire scenario does not include outside wind conditions, which ceiling vents.

598 NFPA 204M -- A97 ROP

6-2.2.2 Flow through the Ceiling Vents. Flow is driven through as in the case of a point source, from the combustion zone. The ceiling vents by cross-vent hydrostatic pressure differences. The smoke layer is assumed to be relatively transparent (i.e., all radiation traditional calculation uses orifice-type flow calculations. Bernoulli's from the fire is incident on the bounding surfaces of the compart- equation is applied across a vent, and it is assumed that, away from ment). ,and on either side of tile vent, the environment is relatively lquiescent. Figure 6-2.2.2 depicts the known, instantaneous, A plume model is selected from the several available in the lydrostatic pressure distribution in die outside environment and literature, and this is used to determine the rate of mass and throughout the depth of tile curtained space. These are used to enthalpy flow in the plume at the elevation of the layer interface. It calculate the reslrlting crogs-vent pressure difference, and then the is assumed that all of dais flow penetrates the layer interface and actu,'d instantaneous mass and entbalpy flow rates through a vent. enters the upper layer.

As die plume flow enters die upper layer, the forces of buoyancy that act to drive the plume toward tile ceiling are reduced immedi- ap ately because of the temperature increase of the upper layer across environment over that of the lower ambient. AS a result, the Il ceiing Pressure in continued ascent of the plume gases is less vigorous (i.e., at reduced I vent curtained Ceiling vent velocity) than it would be in fine absence of the layer. Also, as they X'. /" space continue their ascent, the temperature of the plume gases is higher than it would be without the upper layer. Such higher temperatures are a result of the modified plume entrainment, which is now occurring in die relatively high temperature upper layer rather than in die ambient-temperature lower layer. Methods of predicting the characteristics of the modified upper plume flow are available. Cor o.,side ill Convective heating: To relatively cool I from relatively ho ,aterial Y YCEIL ceiling jet

Reradiation f Is from ceiling Pressure to relatively ~ cool floor

Figure 6-2.2.2 Flow through a ceiling vent.

6-2.2.3 Flow below the Curtain Boards. If and when the layer interface drops below the bottom of the curhain boards, the smoke heating starts to flow out of the curtained space. AS with the ceiling vents, "les this flow rate is determined by the cross-vent hydrostatic pressure difference. AS depicted in Figure 6-2.2.3, however, in this case, the pressure difference is not constant across die flow. Nonetheless, even in this configuration, file instantaneous flow rates are easily determined with well-known vertical-vent flow equations used traditionally in zone-type fire models. Figure 6.2.2.4 The fire, the fire plume, and heat transfer to the Pressure in ceiling. curtained . d space 6-2.2.5 Convective Heat Transfer to the Ceiling. Having penetrated f the interface, the plume continues to rise toward the ceiling of the curtained compartment. As it impinges on the ceiling surface, the plume flow turns and forms a relatively high temperature, high velocity, turbulent ceiling jet dlat flows radially outward along the ceiling and transfers heat to the relatively cool ceiling surface. The ceiling jet is cooled by convection and the ceiling material is heated by conduction. Eventually, the now-cooled ceiling jet reaches the I/ extremities of the curtained space and is deposited into and mixed Pre with the upper layer. The convective heat transfer rate and the next .~oa~ "~C,. YCUR ceiling surface temperature on which it depends are both strong (or outside) '~ functions of the radial distance from the point of plume/ceiling impingement, decreasing rapidly with increasing radius. I' s 6-2.2.6 Thermal Response of the Ceiling. The thermal response of Pressure the ceiling is driven by transient heat conduction. For the time period typically considered, radial gradients in ceiling surface conditions are small enough so that the conduction heat transfer is quasi-one-dimensional in space. Therefore, the thermal response of the ceiling can be obtainedfrom the solution to a set of one- dimensional conduction problems at a few discrete radial positions. Figure 6.2.2.3 Flow below a curtain board. These can be solved subject to net convection and radiation heat flux boundary conditions. Interpolation in the radial direction between the solutions leads to a sufficiendy smood~ representation 6-2.2.4 The Fire, the Fire Plume, and Radiation Heat Transfer. The of the distributions of ceiling surface temperature and convective major contributors to the upper layer flow and surface heat transfer heat transfer rate. The latter is integratedover file ceiling surface to are the fire and its plume. This is depicted in Fignre 6-2.2.4. It is obtain the net instantaneous rate of convective heat transfer losses assumed that the rate of energy release of the fire's combustion zone from tile ceiling jet. does not vary significantly from knowal free-burn values that are available and assumed to be specified (see Chapter 5). A known, 6-2.2.7 The CeillngJet and the Response of Fusible Links. Convec- fixed fraction of dais energy is assumed to be radiated isotropically, tive heating and the thermal response of a near-ceiling fusible link

599 NFPA 204M ~ A97 ROP

are detenaained from the Iocdd ceiling jet velocity and temperature. LAVENT assumes that, at all times during a simulated fire, the Velocity and temperature depend on vertical distance below the overall building space containing the curtained area of fire origin is ceiling and radial distance from the fire plume axis. If and when its vented to the outside (e.g., through open doorways). It is assumed, fuse temperature is reached, the device(s) operated by the link is furthermore, that the area of the outside vents are large relative to actuated. the area of the open ceiling vents in the curtained compartment. Therefore, if,~he total area of the outside vents is AOU T, then For specific radial distances that are relatively near to tim plume, (AOUT/AV) ~ is significantly larger than 1 (e.g., AOuT/A V > 2). If the ceiling jet is ,an inertially-dominated flow. Its velocity distribu- the outside vents are in the bounding walls of the curtained space, tion, depicted in FigaJre 6-2.2.7(a), can be estimated from the and not in adjacent spaces, they should be located entirely below the cllaracteristics of the phune, upstream of ceiling impingement. The layer interface. Paragraph B-5.5 should be referenced for the details ceiling jet temperautre distribution, depicted in Figure 6-2.2.7(b) for of od~er guidelines, ,assumptions, and limitations. a relative "hot" or "cool" ceiling surface, is dlen estimated from the velocity (which is now known), upper layer temperature, and ceiling- 6-2.3.2 A User Guide for the Computer Code. Appendix C is a user surface temperature and heat flux distributions. guide for the LAVENT computer code. The appendix includes a comprei~ensive discussion of the inputs and calculated results of a ATcj = TCj -T U = ceiling-jet teml)emture -- upper layer default siomlation involving a fire growing in a large pile of wood pallets (t~-type growth to a steady 33MW) in a 9.l-m high curtained temperature. warehouse-type space with multiple fusible-link-actuated vents and near-ceiling-deployed fusible sprinkler links. Inputs to LAVENT include those specified in 6-2.3.2(a) through (f). 0 VMAX (a) Dimensions of the Curtained Compartment of Fire Origin. Length, width, ,and height of the curtained compartment of fire l(x\\\"~ ) ~\\\\\\\\\\\\\", ~\\\\N" / " origin. (b) Dimensions of the Curtain Board. Floor-to-bottom-of-the- curtain separation distance and length of the curtain (a portion of the perimeter of the curtained space can include floor-to-ceiling ~lls).

(c) Properties of the Ceiling. Thickness, density, thermal conductivity, and heat capacity of the ceiling material. Distance below (d) Characteristics of the Fire. Elevation of the base of the fire ceiling above the floor (see 6-1.2); total energy release rate of the fire, Q, at different times during the course of the simulated fire sce.na~io (the computer code uses linear interpolation to approximate Q between these times); and the plan area of the fire, or the total energy release rate per unit area of the fire (in cases where the user supplies the latter input, the computer code estimates the changing area of the fire at any moment by using the current total energy release rate). Figure 6-2.2.7(a) Ceiling-jet velocity. (e) Characteristics of the Ceiling Vent-Actuating Fusible Links or Vent-Actuating Smoke Detectors and of the Corresponding Ceiling Vents. Horizontal distance from the fire, vertical distance below the ceiling surface, response time index (RTI), and fuse temperature of the ceiling vent-actuating fusible links; also, the clear open area, AV, of their associated ceiling vents.

NOTE 1: ff ceiling vents are actuated by smoke detectors, the ~'xXXXXXXXXX\N~ ~\\~\\\\\\\\\\\\\\\N" ( " guidelines outlined in 6-1.4.5.2.2 should be followed. LAVENT can be made to simulate this function with a very sensitive fusible link (i.e., a link with a negligibly small RTI) and an appropriate fllse temperature.

NOTE 2: As specified in B-4.1, LAVENT always assumes that the flow coefficient, CD, for ceiling vents is C D = 0.68; if the user has reason to believe that a different value, C D USER, is more Distance appropriate for a particular vent (such as ~e value 0.61 below suggested in 6-1.4.2), then the input vent area for that vent f " "Cool"ceiling, high heat transfer should be scaled up proportionately (i.e., AV, INPU T = A V ceiling CD,USER/0.68). (f) Characteristics of Fusible Sprinkler Links. Horizontal distance from the fire; vertical distance below the ceiling surface; and response time index (RTI) and fuse temperature of fusible sprinlder links. NOTE: LAVENT calculates the time that the first sprinkler link fuses and the fire environment that develops in the curtained space prior to that time. Since the model does not simulate the Figure 6-2.2.7(b) Ceiling-jet temperature. interaction of sprinkler sprays and fire environments, any LAVENT simulation results subsequent to sprinkler waterflow should be ignored. 6-2.3 The LAVENT Model Equations and Computer Code. 6-2.3.3 Computer Requirements. LAVENT is written in FORTRAN 6-2.3.1 TheModel Equations: Guidelines, Assumptlons, and 77. The executable code operates on IBM PC-compatible computers Limitations. Appendix B provides details of all equations of the and necessitates a minimum of 300 kilobytes of memory. LAVENT mathematical fire model, and its associated computer program developed to simulate all the phenomena described in 6- 6-2.4 Experimental Validation of LAVENT. LAVENT has lgtd some 2.2. LAVENT can be used to simulate and study parametrically a limited experimental validation in experiments with 3.34 rn g pool wide range of relevant fire scenarios involving these phenomena. fires in a 37 m x 40 m x 14 m high aircraft hanger [Walton and Notariannl, 1993; Notarianni, 1993]. The hanger was equipped with Included in B-5.5 is a summary of gnidelines, assumptions, and near-ceiling mounted brass disks of known RTlwhich were used to limitations to LAVENT. For example, as specified in that paragraph, simulate sprinkler links or heat detector elements. The experiments

600 t NPPA 204M -- A97 ROP

did*not Involve ceiling venm. Experimental validation of the various s~ OtherTma. Otheri.ars.e. scare .ere temp.'re, conducted mathematical rob-model equation sets that comprise the generalized [H|nkteyet ai 1992] .employing Hquld mere, smau vent spacings LAVENT simulation is also impl~cit. This b the case since the (minimum of 4.7 m), and venm open at 18ninon. Hinidey conduded ~,mb-models of L&VENT, presented ha ~ B, _ that: are based 6 n and carefully.reprOduced-c~rrelaflomof_ ace~ired in appropriate experimental studies of the isolatedphysical. (a) the prior opening of vents had little effect on the operation of phenomena that, taken together, makeup the com~. effects*of.a the first sprinkler, and LAVENT-simulated fire scenario. To learn of the experimental basis and validation ofd~ LAVENT mbmodels, the reader is referred to (b) venting suhstanthdly reduced the total number of sprinkler •L~ References for Appendix B. operations. Chapter 7 Mechanical Exhaust Systems in an independent analysis of these tests,,Gu~ noted that - sprinklers near the fire source were often delayed or did not operate 7-1 General. For mechanical veuting systetm capable offun~ionlng altogether [Gwtabson 199"2]. under the expected fire exposure, e xlmust.rates.pefcurtained . compartment arecalculated from equations 6-2 or 6-fi, with the aia 8-6 Coacitmioas, Whlle the use of _m~t_omatl~:venttngand curtain ofequadous 6-1 and 6-5. Gas temperatures are calculated from boards in spduldered _buildings is ~ under r~, the designer is equation 6-9. encouraged tome the avallaHe t~ls m~ldata ref~'~a~-dr in t~ document for soit~g problerm.~ m a.p~cular type oz 7-2 System Conversion. ffa gravity venting system ltm been hazard control [Mil[~ ~!98~:'H~ 1974,~Watemmn-I98~, designed and the projected gas tem pe/n~,re rise inthe smoke layer Trotlp t994; Hinldey et al 19~; Gus~mc~ 1992]. t exceeds. IS0°C, convenion to an eqmvalent mechanical system can be doneusiu 8 Table 7-2. Table 7-2 9-1 Imporumee, Smoke ~ heat ven~:m in the case of ether fire

Mechanical ~ Cap,eUy and maintenance t, mential _for~ eqmpmont am s~tems •that: are not subjected to-their l~endedme for manyyexn. Design l)ep/h of SmOke per Unit Area Of Gnnd~ Vent 9-2 C~N~al. 9-2.1 Various types ofappr~__ an~¢ .thermal smokeand heat 1.8 . 2.15 vents have been made a~dlable commercially. These vent, fall into two general ~*_~oories: 2.4 ~.47 $.o ~.76 (a) Mech.nicdy Opened Ven~. F.~.-o~esind.de ,i~ng4~t, pneumatic-lift, or electric motor-driven venm. ~.6 ~,0~ 4.8 8.49 (b) Grawi~ Veats,~ Examples indude PVC or acrylic drop- out pane~ 6.0 3.91 7.2 4.28 9-L2 Ge~/~ mediCally.opted ven~ ate,provided with manual releasedevicm,that allow dirt~t activation; inspection or 7-s Intake Air. Adequate intake air should be provided to make up maintenance, or buth, wwell as repbu~.mem of ac~_.~_on compo- air for mechanical exhaustsystems. Such intake might be powered nen~ (e.g.,~mive cle~cei, thermal sensors, compressed gas or noupowered. cylinden, expl~luJbO. " Chapter 8 Vemiog in Spriuldered Buildtn~ 9.2.3 Gravity-openedventsdo not ail~ctiveo[~Won, hut i~on of the ii~aUed,unit-is uece~ to ensure ibe units &l Introduction, The previous chapters represent the state of are immlled In accocdam~with triceps ~and technology of vent and curtain board desi~- in the almencesff acce~ u~te _pr~ecen~_ cm~l~mpmems~re in prate, sprinklers. A broad/y accepted equivalent design basis for using • unc~a~d.mi/'r~ee, ofsem~ deb~ma.m~e~ms e~at sprinklers, veuu~ and curtain boards together for ha:uwd control miEht interfere with.the opermion and hmetion ~e.unit. (e.g., property protection, life safety, water usage, obscuration) has not l~en universallyrecognize&-" ,, 9-2.4 The .inspection and maintenance .~ multiple&unction vents aim should enmure 4hat odntu funCdom ao not ~ the intended &2 General. For occupandes that present a high challenge to fire protection operation. sprinkler systems, concern has been raised that.~he in~ of automatic roof venting or curtain boards, of both, can be de~-imen- 9.8 Freq~ oflmlm~. ~ Imd'Ma~ tai to the perforrmu~.ce of automatic sprinlden~ Although khere ~, no universallyaccepted condmion flora fire ~ [MWer 1980], studies on a model scale [ Heskestad 1974]mggested the following: (a) Ventin 8 delays loss of visibility. All deficiencies found should be corrected immediately. (b) Venting results in increased fuel comump60" n. / 9-S.2 Mec~ Vents. (c) Depending on the location of the fire, rela~ve to the vents, the nec__es~_,y ~ater demand to achieve control is either increased or 9-3.?,1 Iris important that aa~ceptance pe~om~ce testand decreased over an unrented condition. With the fire directly under inspection ofaB ~ vents be condected immed~ the vent, water demand is decreased. Whh the fire equidistant from the vents, water den~and is increased. cfi pr I~ y and tl~t in~allatinn isit. _ _.c~rdaace~with~.he manufacturer s specili~tiom aml accepted trade practice~ 8-3 Automatic Roof Vents. A series of tests wasconducted to increase the understanding of the r01e of automatic roefvenm i • . 9-$.~ It is necessary to follow the manu~_~_urer's recommentia- simultaneously employed with autoh~ttc sprinklem [Waterman tiom re~i,'n 8 the maintenance and-rmpectiou schedule of 1082]. Thedam submitteddid uot provide a consensus on whether mechanically4~perated venu.. sprinkler control was impaired or enhanced by the presence of a~tomatic (roof) vents for the typical spacing ~m'd area. 9-~.2.S Inspection schedules should indude provisions for all units to be tested at 12-month interv'aJsor on aschedule based on a 8-4 Curtain Boards; Large scale fire tests [Troup 1994] indicated " percentage of the total units to be tested every month or every two that the presence of curtain boards ca~ cause increases in sprinkler months. Such procedures improve reHabilRy. operation, smoke production, and fire dmnage (i.e., spriniders opened wall away from the fire).

601 NFPA 204M ~ A97 ROP

9-3.2.4 Recordin~ of all pertinent characteristics of performance 9-4.1.4 Manual releases should be tested to determine that the vents and logging of thts information to allow a comparison of results with operate. those of previous inspections or acceptance tests allows a compari- son dlat provides a basis for determining the need for maintenance 9-4.1.5 All operating levers, latches, hinges, and weather-sealed or modifying die frequency of the inspection schedule to fit the surfaces should be examined to determine conditions such as any experience. indication of deterioration and accumulation of foreign material, that might warrant corrective action or suggest the need for another 9-3.2.5 A change in occupancy or in materials being used, or in inspection in advance of the normal schedule. neighboring occupancies that could introduce a significant change in the nature or severity of corrosive amaospbere exposure, debris 9-4.1.6 Following painting of the interior or exterior of vents, the accumularion, or physical encumbrance, might necessitate a change units should he opened ,and inspected to check for paint that could in the inspection schedule. "glue" surfaces together.

9-3.2.6 Special mechanisms such ,as gas cylinders, thermal sensors, 9-4.1.7 Painted heat-responsive devices should be replaced with or detectors should be checked regularly on a schedule provided by devices having an equivalent temperature and load rating. the manufacturer. 9-4.2 Gravlty-Opened Vents. 9-3.3 Gravity-Opened Vents. 9-4.2.1 All weather-sealed surfaces should be examined to deter- 9-3.3.1 The same general considerations for inspection that apply to mine conditions such as any indication of deterioration and mechanically-opened vents (see 9-3.1) also pertain to gravity-opened accumulation of foreign material that might warrant corrective vents. The dlermoplastic panels of these vents are designed to action or suggest the need for another inspection in advance of the soften and drop out from the vent opening in response to the heat normal schedule. of a fire. This makes an operational test after installation impracti- cable. Recognized fire protection testing laboratories have 9-4.2.2 Following painting of dae interior or exterior of the frame or developed standards ,and procedures for evaluating gravity-opened flashing of the vents, the units should be inspected to check for vents, including factory and field inspection schedules. paint that could "glue" surfaces together. 9-3.3.2 An acceptance inspection of all gravity-opened vents should 9-5 Air Intakes. be conducted immediately ,after installation. Compliance with tiae manufacturer's drawings and recommendations should be verified 9-5.1 Air intakes necessary for operation of smoke and heat vents bydirect examination. A suitable installation should follow accepted should be maintained clear and free of obstructions. trade practices. 9-5.2 Operating air intake louvers, doors, dampers and shutters 9-3.3.3 Changes in appearance, damage to any components, should be examined to assure movement to full-open positions. fastening security, weather tightness, and adjacent roof and flashing condition should he noted at tile time of inspection. 9-5.3 Operating equipment should be maintained and lubricated as necessary. 9-3.3.4 Prompt and careful removal of any soiling, debris, or encumbrances that could impair the operation of the vent is 9-6 Ice and Snow Removal. Removal Of ice and snow from vents is essential. an essential part of a vent maintenance program.

9-$.4 Intake Air Sources. Where necessary for the operation of vent systems, intake air sources should he inspected at the same fre- Chapter 10 Referenced Publications quency ,as vents. 10-1 The following documents or portions thereof are referenced 9-4 Conduct and Observation of Operational Tests. within this guide and should be considered part of the recommenda- tions of this document. The edition indicated for each reference is 9-4.1 Mechanically-Opened Ven t.,~. tile current edition as of tile date of the NFPA issuance of fills document. 9-4.1.1 Where feasible, release of file vent should simulate actual fire conditions. Disconnecting the restraining cable at the heat- 10-1.1 NFPA Publications. National Fire Protection Association, 1 responsive device (or other releasing device) and suddenly releasing Batterymarch Park, P.O. Box 9101, Quincy, MA 02269-9101. file restraint, allows the trigger or latching mechanism to operate normally. NFPA 68, C,uitk for Venting of Deflagrations, 1994 edition.

94.1.2 The heat-responsive device restraining cable is usually under NFPA 72, NationalFireAlarm Code, 1996 edition. considerable tension. Observation should be made of its whip and travel to determine any possibility that the vent, building construc- NFPA 92B, Guide for Smoke Manageraent Systems in Malls, Atria, and tion feature, or service pipingcould obstruct complete release. Any Large Areas, 1995 edition.. possible interference shouldbe corrected by removal of die obstruction, enclosure of cable in a suitable conduit, or other 10-1.2 Other Publications. appropriate arrangement. Following any modification, die unit shouldbe retested for evaluation of adequacy of corrective mea- 10-1.2.1 ASTM Publication. American Society for Testing and sures. Materials, 1916 Race Street, Philadelphia, PA 19105-1187.

NOTE: The whipping action of the cable upon release presents ASTM E 1321, Standard Test Method for Daennining Materiat Ignition a possibility of injury to anyone in the area. For this reason, the and F/ame Spread Properties, 1993. person conducting the test should ensure that all personnel are well clear of tile area where whipping of the cable might occur. ASTM E 1354, Standard Test Mahod for Heat and Visible Smoke Release Rates for Materials and Products Uslng an Oxygen Consumption Calorlm- 9-4.1.3 Latches should release smoothly. The vent should start to eter, 1994. open immediately and move through its design travel to tile fully- opened position without any assistance and without any problems such as undue delay indicative of a sticking weather seal, corroded or unaligned bearings, and distortion binding.

602 NFPA 204M -- A97 ROP

Appendix A Explanatory Material For manyfires involving storage arrays the time to reach 1000 kW might be much shorter than the 75 seconds depected for ultra-fast This. A. ppendix is not. part ofthe. recommendations of this NFPA document fires. but ts mclluted for reformational purposes only. The general equation is as follows: A-I-I.3 Large, undivided floor areas present extremely dil~cult fire- fighting problems, since the fire department might need to enter Q= Ctgt2 these areas in order to combat fires in central portions of the building. If the fire department is unable to enter beG~tuse of the where: accumulation of heat and smoke, fire-fighting efforts might be reduced to an application of hose streams to perimeter areas while Q= rate of heat release (kW) fire continues in the interior. Windowless buildings also present t~__= a constant describing the speed of growth (kW/s 2) similar fire-fighting problems. One fire protection tool that can be a = time (s) valuable asset for fire-fighting operations in such buildings is smoke and heat venting. Relevance of T-Squared Approximation to Real Fires.

A-l-2.1 The provisions of this guide may be permitted to be applied A t-squared fire can be viewed as a fire in which the rate of heat to the top story of multiple-story buildings. There are many features release per unit area is constant over the entire ignited surface and that would be difficnh or impracticable to incorporate into the lower the fire spreads in circular form with a steadily increasing radius. In stories of such buildings. such cases, the increase in the burning area is the square of the steadily increasing fire radius. Of course, other fires that do not have A-6-1.4.6.1 T-Squared Fires. Over the past de~tde, those interested such a conveniendy regular fuel array and consistent burning rate in developing generic descriptions of the rate of heat release of might or might not actually produce a t-squared curve. The tacit accidental open flaming fires have used a "t-squared" approximation assumption is that the t-squared approximation is close enough for for this pnrpose. A t-squared fire is a fire in which the burning rate reasonable design decisions. varies proportionally to die square of time. Frequendy, t-squared fires are classed hy their speed of growth as fast, medium, and slow Figure A-6-1.4.6.1 (a) demonstrates that most fires trove an (and occasionally ultra-f:L~t). Wilere these classes are used, they are incubation period during which the fire does not conform to a t- determined hy the time needed for the fire to grow to a rate of heat squared approximation.In some cases, this incubation period might rele:~ge of 1000 kW. The times for each of these classes are provided be a serious detriment to the use of the t-squared approximation. In in Table A-6-1.4.6.1. most instances, this is not a serious concern in large spaces covered by dais guide. It is expected that the rate of heat release during the Table A-6-1.4.6.1 Claxsiflcatlons o f incubation period would not usually be sufficient to cause activation T-Squared Fires of the smoke detection system. In any case, where such activation occurs or human observation results in earlier activation of dae Cla.gs Time to Reach 1000 kW smoke venting system, a fortuitous safeguard would result. Ultra-Fa.st 75 s Fast 150 s Medium 300 s Slow 600 s

Continuously Growing

3000

A

O

rop- 2000

== =- iv I

-i- 1000

cubation

t.~iGrowt~ ~ Time TIME .._Ef fective ~ Ignition Time

Figure A-6-1.4.6.1 (a) Conceptual illustration of continuous fire growth.

603 NFPA 204M ~ A97 ROP

Figalre A-6-1.4.6.1 (b), extracted from Nelson, Harold E., An Tile other set of dashed lines in Figure A-6-1.4.6.1 (b) shows these Engineering A nal~sis of the Ear.l~ Stages of Fire Developnwnt----The Fire at the same fire curves relocated to the origin of the graph. This is a more DuPont Plaza Hotel and Ca.¢ino, December 31, 1986, Report NBSIR 87- appropriate comparison with the generic curves. It can be seen that 3560, Nation:d Institute of Smncktrds and Technology, Galthersburg, the rate of growth in these fires is actually faster than that prescribed Maryland, 1987, compares rote of heat release curves developed by for an ultra-fast fire. This is appropriate for a test fire designed to the .'fforementioned classes of t-squared fires and two test fires challenge the fire suppression system being tested. commonly used for test purposes. The test fires are shown as dashed lines labeled as furniture and 6-ft storage. The d:Lshed curves further Figure A-6-1.4.6.1 (c) relates the classes of t-squared fire growth from the fire origin show the actual rates of heat release of the test curves to a selection of actual fuel arrays. fires used in the development of the residential sprinkler and a standard 6-ft high mTay of test cartons containing foam plastic pails that also are frequently used as a standard test fi re.

Ultr -Fast Fast Medium 6000 F j Furniture / / 5000F ''I i~ " . __~ --''6ftl" st0 rag e

_ .... "--

,$ F I //}'/// .' /~" / j/Display

~" I I I J s 0 100.__ 200 300 .400 500 600 "~u0 TIME FROM IGNITION (seconds)

Figure A-6-1.4.6.1 (b) T-squared fire, rates of energy release.

Cartons 15 ft. high. various contents, V astest if empty or containing plastic foam Thin plywood wardrobe-- r---Full mail bags, 3 ft. high / I ~5r~Wood ft. highpallets 1 ~palletstack Fastest burning CnttetrOs p~Pngl Y est e r - ~ upholstered furniture 1 Ultra-Fast Fast mattress Medium

5000 V 3000 F

0 100 200 300 400 500 600 700 TIME FROM IGNITION (seconds)

Figure A-6-1.4.6.1 (c) Relation of t-squared fires to some fire tests.

604 NFPA 204M -- A97 ROP

Appendix B The Theoretical Basis of LAVENT Conservation of Mass:

This Appendix is not a part of th~ rt~om.um.dations of this NFPA document dm U / dt = m U (13-2) but is inchded for informational pmpose.~on(~. mu = (YCEIL - Y)Pu A (B-S) B-1 Overview. In tills Appendix, the physical basis and an associ- ated mathematical model fl)r estimating the fire-generated environ- Perfect Gas Law:. ment :rod the response of sprinkler links in well-ventilated compart- ment fires with curtain boards and fiJsible-link-actuated ceiling vents PU / R ~ p / R = constant = PuTu = PAMBTAMB (B-4) is developed. Complete equations and assumptions are presented. Phenomena taken into account include the following: i.e.:

(a) The flow dynamics of the upward-driven, buoyant fire plume; T U = TAMBPAMB/Pu (B-5) (b) Growth of the elewated-temperamre smoke layer in the In the above, y is the elevation of the ceiling above the floor, curtained compartment; R = C -C is ~t//gas constant, C and C are the specific heats at a c~stan~/pressure and volume, ~spectiv~y, and p is a constant (c) The flow of smoke from the layer to the outside through open characteristic pressure (e.g., ~ ) at the floor elevation In ceiling vents; • . p' equanon B-l, (trr is the net ra~/enthalpy flow plus heat transfer (d) The flow of smoke below curtain partitions to building spaces to the upper lair and is made up of flow components as follows: adjacent to tim curtail~ed space of fire origin; , from below the curtain; = , from the plume; qCURT, from the ceiling vent; anc~f~ponent = , the total (el Continuation of the fire plume in the upper layer; h~nsfer rate. ~/HT

(f) Heat transfer to the ceiling surface and the thermzd response of qu = qCURT +qPLUME +qVENT +qnT (13-6) the ceiling :is a fnnction of radial distance from the point of plume- ceiling impingement; In equation B-2, rhty is file net rate of mass flow to tile upper layer with flow componen'(s; rhCURT, from below the curtain; ~m tlMy, (g) Tlae velocity and temperamredistribution of plume-driven from the plume; and ~hVENT, from the ceiling vent. " ~"~ near-ceiling flows and the response of near-ceiling-deployed fusible links ,as functions of distance oelow the ceiling; and m U = tnCURT + thPLUME + INVENT (B-7) (It) Distance front phmle-ceiling intpingement. Using equation B-3 in equation B-1 leads to:

The theory presented here is the basis of the LAVENT computer program, that is supported by a user guide, presented in Appendix dy / dt = qu / ( A CpP AMB T AMB ) (13-8) C, and that can be used to study parametrically a wide range of relevant fire scenarios [1, 2, 3]. if (~ = Y('~ZL_and qt, > 0 ); or (0 < Y > 'r~'rr and arbitrary qH ]- B-2 Introduction. The space under consideration is a space of a Sitl~ze b6fl~'o~ these ~onditibns are satisfied~'~ation 13-8 is always' plan area, A, defined by ceiling-mounted curtain boards with afire applicable. of time-dependent energy release rate, ~(t), and with open ceiling V ents of total tm~e-dependent' area, At/[Q.~- The curtained" ,area can The basic problem of mathematicallysimulating the growth and be considered as oneof several such s~a'c'es in a large building properties of the upper layer for the generic Figure 6-2.2.1 scenario compartment. Also, by specifying that die curtains be deep enough, necessitates the solution of the system of equations 13-2 and B-8 for they can be thonght of as simulating the walls of a single, ~_~.. and y. Where mU > O, PU can be computed from equation uncurtained compartment. This appendix presents die pltysical ] )asJs.... aa~a assocmted mathematical model for estimating the fire- generated environment ,and the response of sprinkler links in PU = (YCEIL - y)A / mu, if m U > 0 (13-9) curtained compartment fires with fitsible-link-actnated ceiling vents. .and T U can be determined from equation B-5. The overall building comparmtent is ,assumed to have near-floor wall vents that are large enough to maintain the inside environment, B-4 Mass Flow and Enthalpy Flow Plus Heat Transfer. below any near-ceiling smoke layers that could form, at assumed initial outside-ambient conditions. Figure 6-.9.2.1 depicts dae generic !?.-4.1 Flow to the Upper Layer from the Vents. Conservation of fire scenario for the space under consideration. The ,assumption of momentum across all open ceiling vents as expressed by Bemoulli's large near-floor wall vents necessitates that the modeling be equation leads to the following: restricted to conditons where y, the elevation of the smoke layer interface, is above the floor elevation (i.e., y > 0). The assumption V = C(2APCEIL/pu) 1/2 (B-10) also has important implications with regard to the cross-ceiling vent pressure differer~tial. Tltis is the pressure differential that drives elewated-temperature upper layer smoke through the ceiling vents to the outside. Therefore, below fire smoke layer (i.e., from the floor "v r = -pu v V =- vc(2pv pcF ) (B-Ill of the facility to the elew,ttion of the smoke layer interface), die inside-to-outside hydrostatic pressure differential is zero, wlfile a where V is the average velocity through all open vents, C is the positive inside-to-outside pressure differential exists at all elevations vent flow coefficient (= 0.68) , and A,OCE/L is the cross-vent the reduced-density smoke layer itself (higher pressure inside file pressure difference [4]. curtained area, lower pressn re in rite outside environment), the maximum differential occurring at the ceiling and across the open From hydrostatics: ceiling vents.

B-3 The Basic Equations. A two-layer zone-type compartment fire model is used to describe the phenomena under investigation. As is typical in such models, the upper smoke layer of total mass, mu, is (ale) ,amumed to be uniform in density, PU ' and absolute temperatnre, T U • where g is the acceleration of gravity. The following time-dependent equations describe conservation of energy, mass, ,and the perfect gas law in the upper smoke layer. Substituting equation B-12 in equation B-11 leads to the desired Conservation of Energy: INVENT result as follows: d[(YCEIL -y)PuTuACu]/ dt=qu+pAdy/dt (B-l)

605 NFPA 204M -- A97 ROP which is equivalent to die equations imed to estimate ceiling vent flow rates, Equation (6.8) and references [5 and 61. Using eqtmtion B-I 3, the desired qI,~NT result is as follows: • 0. 2+g[(,- ,,)Q]~/5/ ~a~ _,.02 ~ o; qVENT = 'hVENT CpTu (13-14) o.249[(1-~,r)Q]2/5 / DHRE -1.02 LF/_AM_E / DF/RE = B-4.2 Flow to the Layer from the Plume and Radiation from the Fire. It is assumed that tim ma~s generation rate of the fire is small if 0.249[(1- ~.r )Q]215 1 DFIRE - 1.02 _> O; compared to ~h ~r, the rate of mass of air entrained into the plume between fill'fire elevation, 3HRE, and the layer interface, or compared to other m,-Lcsflow rate components of rhU . It is ,also (Qin kW, ~ in m) assumed that all of the rh l¢#trr penetrates the layer interface and enters the upper layer. TWei'$fore: (13-18) mt'LUME = ~hENT (B-15) = 0.0054/0.071-(0.166) 5/3 = 0.02591682001.. = 0.026 (!?,-19) qm~Me = 'neNT CpTAM~ +(1- ~,)Q (e,-10) In equations B-17 through B-19, LFLAME is the fire's flame ng~l, DF/RE is the effective diamet~ of the fire source The first term on the right side of equation B-16 is die enthalpy D~TRE /4 = area of the fire source I, and a is chosen so that, .associated with ~hENT , and ~r , in the second term in equation B- ytically, the value of ~hENT is exadtly continuous at the 16, is the effective fraction of ~c. assumed to be radiated isotroplcaily elevation y = YF/RE +LFLAME • from the fire's combustion zone. B-4.3 Flow to the Layer from Below the Curtains. ffthe upper layer It is ,x~sumed fllat die smoke layer is relatively transparent and that interface, y, drops below the elevation of the bottom of th-e- it does not participate in any significant.radiation heat transfer curtains, YetrRT, mass and enthalpyflows occur from the upper exchanges. In particular, ,-all of the grQ radiation is assumed to be layer of tlle"~iffained areawhere the fire is located to adjacent incident on the bounding surfaces of the compartment. Therefore, curtained areas of the overall building compartment. The mass flow the last term of equation B-I 6 is die net amount of enthMpy added rate is the result of hydrostatic cross-curtain pressure differentials. to die upper layer from die combustion zone and its buoyancy- Provided adjacent curtained areas are not yet filled with smoke, this driven plume. Flaming fires exhibit values for ~.r of 0 < gr < 0.6 pressure difference increases linearly from zero at the layer interface (e.g., smaller values for small methane fires and higher values for to APcuRT aty = YCURT" large polystyrene fires). However, for a hazardous fire involvinga wide range of common groupings of combustibles, it is reasonable to From hydrostatics: approximate flame radiation by choosing gr ~ 0.37 [7].

A specific phime entrainment model is necessary to complete equations B-14 and 13-15 for the ~hpLtt~clv and qpLlrMl;'. The (8-20) following estimate for ~hENT [8 anH'9"ITi's adopted'ff.q'ft-llows:

Using equation B-20 together with well-known vent flow relations (e.g., equation 32 of reference [4]) mCURT and the qCURT can be estimated from the following:

0 if Y-YFIRE <- 0;

if o <(y-yFS~)/L~ <1, o.o,,b_;<,)Ql'"-'

m~n l =

(8q7)

60fi NFPA 204M -- A97 ROP

The plume above a point source of buoyancy [ 10], where the 0 if y > YCURT ; source is below the interface, is equivalent to the plume of the fire (in the sense of having identical mass and enthalp.y flow rates at the ]1/2 interface) if die point source strength is (1-)~r }Q and the elevation of the equivalent source, YEQ , satisfies the foll6wing: mCURT 0 ^1 1/2( )5/2 .~*U3 if y <- YCURT B-21) ,hpLuME = .z PAMBg [Y-YEQ ) ~EQ (13-24)

In equation 13-24, QEO ' a dimensionless measure of the strength of qCURT = ~hCURT C pTu C~-2~) the fire plume at the ~fterface, is defined as follows: • - r 1/2" ,5/2] where LCURT, is that length of the perimeter of the curtained area of fire on~m that is connected to other curtained areas of the overall building comparunenL For example, if the curtained area is in one corner of the building compartment, then the length of its two sides It should be noted that, at an arbiu-ary moment of time in the coincident with the walls of the compartment are not included in simulation of a fire scenario, ~hPLUME in equation B-24 is a known Lcrmy. Since the generic vent flow configuration under consider- value that is determined previofisT~7fkb"m equations B-15 and 13-17. au%n'h this ~ase is long and narrow, a flow coefficient for the vent flow incorporated into equation B-21 is token to be 1. Using B-24 and B-25 in order to solve for YEQ and QEQ : B-4.4 Heat Transfer to the Upper Layer. As discussed in B-4.3, where the fire is below d~e layer interface, the buoyant fire plume rises toward the ceiling and all of its mass and enthalpy flow, YEQ y-l(1 " "" ('* 1/2"~2/5 thor rrM~ and ~ or rr~*~", is assumed to be deposited into the upper la)teE~lq~ving pede'~ua~ie"~dtl~e interface, the plume continues to rise toward the ceiling of the curtained compartment. As it impinges on the ceiling surface, the plume flow turns and forms a relatively high temperature, high velocity, turbulent ceiling jet that flows radially QEQ = o.21(1- z, )Q / ( CI,T AMBmPLUME (~27) outward along the ceiling and transfers heat to the relatively cool ceiling surface. The ceiling jet is cooled by convection, and the ceiling material is heated by conduction. The convective heat As the plume crosses the interface, the fraction, rh , of transfer rate is a strong traction of the radial distance from the ~npLtr~., which is still buoyant relative to the upper layer point of plume/ceiling impingement, reducing rapidly with enwr~'ent and presumably continues to rise to the ceiling, increasing radius. It is dependent also on the characteristics of the entraining upper layer gases along the way, is predicted [ 11] to be as plume immediately u pstream of ceiling impingement. follows: The ceiling jet is blocked eventually by tl~e curtains or wall surfaces, or both. It then rams downward and forms vertical surface flows. In m* =[0;-1 < a50 the case of wall surfaces and very deep curtains, the descent of these flows is stopped eventually by t,p~wd buoyant forces and they finally mix with the upper layer. In this case it is assumed that the plume/ ceiling impingement point is relatively far from the closest curtain or wall surface (e.g., greater than a few fire-to-ceiling lengths). Under (B-28) such circumstances the ceilingjet-~r,dl flow interactions are relatively where rite dimensionless parameter o is defined as: weak and, compared to the net rate of heat transfer from the ceiling and near the plume/ceiling impingement point, the heat transfer to • "2/3 the upper layer from all vertical surfaces is relatively small. a=[1-ot+CTQEQ )/(~-1) (B-29) The symbol ~'CoNV is defined ,'is the fraction of Q, which is transferred by ctnve- ction from the upper layer gas ceiling jet to the o~ = T U / TAMB;CT = 9.115 (B-30) ceiling and wall/curtain surfaces as follows:

andwhere Q~'o is the value computedin equation B-27. The qHT =-xcoNvQ (~-~s) parameters nYUessary, to describe plume flow continuation.. in the Once the vahnes of )l.t,t~o~0 and qrrr are determined from a upper layer (Le., between 3 and 3t~rr ) are further identified (see time-dependent soluti6" ~'~'tli~e coupl~, ceiling jet/ceiling material, [11]) according to a point source ~lfi~e (see [101"). It has been convecnon/condnction problem, the hxsk of determining an determined that this plume can be modeled as being driven bya estimate for each component of rhU and qu is complete. nonradiating buoyant source of strength, Q', located a distance t B-4.4.1 Properties of the Plume in the Upper Layer Where H = YCEIL - YSOURCE > YCEIL - YHRE (B-31) Y FIRE < Y" Times when the elevation of the fire is below the interlace (i.e., when YF/RE < -~ ) shoukt be considered. below the ceiling in a downward-extended upper layer environment of temperature, Trr, and density, Pit" The relevant parameters As the plume flow moves to the center of the upper layer, the predicted [ 11 ] arenas follows: v forces of buoyancy that act to drive the plume toward the ceiling (i.e., as a result of relatively Ifigh-temperata~re, low-density plume Q = /(1+ o) B.321 gases being submerged in a relatively cool, higb-density ambient environment) are reduced immediately bec~anse of the temperature increase of the upper layer environment over that of the lower YSOURCE = Y - Y- YEQ a3/5m*2/5 [(I + a}/a]l/5 (B-3S) ambient. As a resuk, the continued ascent of the plume gases is less ( ) vigorous (i.e., ascent is at reduced velocity) and of higher tempera- The fire and the equivalent source in the lower layer and the rare dmn it would be in the absence of the layer• Indeed, some of continuation source in the upper layer are depicted in Figures 13- fl~e penetrating phnne flow will be at a lower temperature titan T U . 4.4.1 (a) through (c). Times during a fire simhlation when equation The upper layer buoyant forces on this latter portion of the flow - B-29 predicts ¢y >> 1 are related to states of the fire environment in actually retard and ~-'m possibly stop its subsequent rise to the whicli the temperature distribution above T A a~ oft_he plume flow, ceiling• at the elevation of interface penetration, is p'~'dYcted to be mostly much larger than ( Tr7 - TA-~R ). Under such circumstances, the penetrating plume flb'w is s~lli~fery strongly buoyant as it enters the The simplepoint source plume model [ 10] is used to simulate the upper layer. The plume continues to rise to the ceiling and to drive plume flow, first immediately below or upstream of the interface, ceiling jet convecuve heat transfer at rates that differ only slightly and fl~en throughout the depth of the upper layer itself. (due to the elevated temperature upper layer environment) from the heat transfer rates that could occur in the absence of an upper layer. 607 NFPA 204M ~ A97 ROP

Fire and flames in the Equivalent plume in the Continuation plume in Me lower layer lower layer extended upper layer (a) (b) (c)

Figure B-4.4.1 Conditions where equation B-29 predicts cr < 0 are related to B-4.5 Computing qlqr T and the Thermal Response of the Ceiling. times during a fire scenario when the temperature of the plume at Where the fire is beF6& the interface and the interface is below the the elevation of interface ~enetration is predicted to be uniformly ceiling, the method for calculating tbe beat transfer from the plume- less than T u. Under such cirounstances, the penetration plume driven ceiling jet to the ceiling and the thermal response of the flow is not positively (i.e•, upward) buoyant at may point ,as it enters ceiling [12] is used. This method was developed to treat generic, the upper layer. Therefore, while all of tltis flow is assumed to enter confined ceiling, room fire scenarios. As oudined in dais method. and mix with the upper layer it is predicted dlat none of it rises to [ 12] dae confined ceiling problem is solved by applying the • • ' . ,t . the ceding in a coherent plnme (i.e., Q = 0)• For this reason, unconfined ceiling heat transfer solution, [13, 14, 15] to the where to" < 0, the existence of any significant ceiling jet flow is problem of an upper-layer source in an extended upper layer precluded along with significant convective heat tl~,msfer to rite environment equivalent to equations 13-34 and B-35. Where the fire ceiling surface or to near-ceiling-deployed fllsible links. is above the interface, the unconfined ceiling methodology applies directly. The above analysis ;~ssumes that y~ < ~. However at the onset ofthefirescenarlo, YmR~

• n tim value of T U# I are dmmselves. -t undefined. • The situation at t = 0 ,assumed that the temperature distribution of the ceiling material, T, Is properly taken into account If 0 = (1 - 3,r)Q and has been computed up to this moment and is known as a function of YSOURCE = YEQ at t = 0. distance, Z, measured upward from the bottom surface of the ceiling, and radial distance, r, measured from the constant point of B-4.4.2 General Properties of the Plume in the Upper Layer. plume-ceiling impingement• The equivalent, extended upper-layer, Where the fire is below the interface, the results of equations B-32 unconfined ceiling flow and heat transfer problem is depicted in and B-33 allow the fire-driven phmte dynamics in the tipper layer to Figure B-4-4.1 (c). It involves the eq_uivalent Q' heat source from be described according to the point source phnne mode/[10]. If equation B-34 located a distance, /-/, below the ceiling surface in an the fire is at or above the interface (i:e., ~ ~-~,~ > ,~) then extended ambient environment of density, Ptl, and absolute rhPLUME = O, qPLUME = (1 - Zr)Q, a/l~i'~ point source model temperature, TII, where H is determined from equations B-31 and is contnmed in use'fo slmuktte the upper layer plume flow. All cases B-33. can be treated nsing the following final versions of original equations B-32 and B-33 ,as follows: The objective is to esdmate the instantaneous convective heat transfer flux from the upper-layer gas to the lower ceiling surface, ...... (r t) and the net heat transfer fluxes to the upper and Q,= 1-~'r)Q ty'h /(I+ty) if YFIRE < Y< Y(2EIL ; (B-34) Io~'ersL"v ~"'~ce 'of the ceiling, (lu(r,t) and q"r (r,t), respectively• With dais information, the time-depe/~dent soluu%h for the in-depth I * thermal response of the ceiling material can be advanced to I(1- A r)~). if YF/RE > Y or if y = YCEIL subsequent times. Also, qr't~-b, r can be integrated over the lower ceiling surface to obtain dVe"d~iF6d instantaneous value for qHT"

In view of the assumptions of the relatively large distance of the fire f FIBE if Y

(B-35)

m = A 4CONV,, X (,,@a, where m , ~, and 0t are calculated from equations 13-26 through (B-z6) B-30• 2 The value Awm¢ = ffDT~m~ / 4 is taken to be the actual area of the curtained~[Sii(:e, A, p~~)he portion of the vertical curtain and wall surfaces estimated to be covered by ceiling jet-driven wall flows. An estimate for this extended, effective ceiling surface area is obtained [16] where it is concluded with some generality that ceiling jet-driven wall flows penetrate for a distance of approxlmately 0.8H from the ceiling in a downward direction. Therefore:

608 NFPA 204M -- A97 ROP

2 AEFF = ;,rDEFF /4

(B-37) wbere P is tile total lengfll of the perimeter of the curtained area.

B-4.5.1 Net Heat Transfer Flux to the Ceiling's Lower Surface. The net heat transfer flux to flae ceiling's lower surface, qL, is made by means of up to tilree components: incident radiation-4 ~,4 n m-mr ; convection, qCONV,L ; and reradiation, q~dT,RAD,L ,as f'61l~s-."....

As discussed in B-4.4, tile radiant energy from the fire, ;I,rQ, is assumed to be radiated isotropically fi'om the fire widl negligible r:adiation absorption al3d emission fi'om the compartment gases.

= (e~39)

Tile convective heat transfer flux from the upper-layer gas to die ceiling's lower surface can be calculated [ 13,14] as follows:

q'(JONV,L = hL(T AD - Ts,L ) (8-40)

where Ts_L is file absolute temperature of tile ceiling's lower surface, TAn , a characteristic gas temperature, is the temperature that is measu- red adjacent to an adiabatic lower ceiling surface, and hL, is a heat transfer coefficient. Equations 41 and 42 determine hL and TAD ,as follows:

hL/i= (8-40 0.3 9/3 1/9 0.283Re H- Pr-" " (r / H)- "(r / H-O.O771)/(r / H +0.279) if O.2 < r / H

lO.22-(14.gr / H) if O< r / H <0.2; (T AD-Tu )/(Tu QH:"2/3 "~)= (B-42) 8.39f(r/ H) if 0.2 <_r/ H

where:

(B-43)

/~ _ 1/2. 1/2()*1/3 1/2 ..3/2.;*1/3 . =put, pg 1-1 ~LH ; Re H =g ta ~-H / VU; (B-44) QH"* = Q "/[ PuCpTu (gH)I/2 H21

In tlae equation B-41, Pr is tile Prandtl number (taken to be 0.7), in equation B.44, and, vtz is die kinematic viscosity of tile upper. layer gas, which is assume~d to have the properties of air. Also, QH, a dimensionless number, is a measnre of the strengtii of the plume and Re. H is a characteristic Reynolds number of tile plume at tile elevatioh" of tile ceiling.

609 NFPA 204M -- A97 ROP

Tile lollowing estimate for VU [ 17] is used where computing Re H Initially the ceiling is taken to be of uniform temperature, Tamb. from equation B-44 The upper and lower ceiling surfaces are the.n expos.~,d to the radial- and time-dependent rates of heat transfer, qfl and qL, determined from equations B-47 ,and ..B-48, resl~,ectively. For specific times in dais case, radial gradients of qv and qL are assumed to be small enough so that conduction in die c.eiling is-quasi-one-dimensional in space [i.e., T = T(Z,t;r)]. Therefore, file two-dimensional thermal Equations B-40 through B-45 use a w..due for Ttt. At t = 0, where response for die i:eiling can be obtained from the solution to a set of it is-- undefined, T U ..q hot--1 Id be set equal to TAMB . This yields the one-dimensional conduction problems for correct limiting result for the convective heat transfer to tile ceiling; T n (')Zt =TZt'r=r( '' n )' n=ltoN.. ~O' where N isthe specifically, convective heat transfer to the ceiling from an uncon- number of thscrete radial posmons necessary to obR~n a suftic=entiy fined ceiling jet in an ambient environment. smooth representation of the overall ceiling temperature distribu- tion. The r n radial positions are depicted in Figure B-4-5.3. As the fire simulation proceeds, die ceiling's lower surface tempe~tture, T¢ t, initially at TAMI~, begins to increase. At all times, die lower ceding, urface is ,lssume ! to radiate diffusely to the initially ambient tempe~.-ature floor surface and to exposed surfaces of the building contents. In response to this radiation, and to the direct radiation from d~e fire's combustion zone, the temperature of rNRAD = DEFF/2 these surfaces also increase with time. However, for specific times, it rn ~ is assnmed fllat tile effective temperature increase of these floor/ contents surfaces is relatively small compared to the chm-acteristic increases of T~ ~. Accordingly, at a given radial position of the ceiling's Iower'~h~i'face, the net radiation exchange between the . ceiling and tile floor/contents sud'aces can be approximated by the following:

where cY is d~e Stefan-Boltzmann constant and E/. and E b'?ooR are the effectwe...... emittance/ahsolptance o[ tile ceding upper suff;/ce a n d • oo • n floor/contents surfaces (assumed to he grey), respectively, both of qRERAD,U qCONV,U which are taken to be 1. Eq. (46) Eq. (45) B-4.5.2 Net Heat Transfer Flux to Ceiling's Upper Surface. It is assumed dlat the ceiling's upper surface is exposed to a relatively constant-temperature far-field environment at TAMB . Therefore, the net heat transfer flux to,,this SUl-face, qrr, is matte u}~ of two components, convection, {ICONV,U , and r'e'radiadon, qRERAD,U as follows:

qu = qCTONVLI +qRE[L4DLS (B-47)

These can be estimated from the foUowing: /

(B-48)

" 4 4 -1) , qRERAD,L qRADIJ =°(TAMB-Ts, u)/(I/~u +I/~FAR (B-49) Eq. (36) qRAD--FIRE Eq. (43) Eq. (37) where.. TsjM. is. the absolute temperature.. of the upper surface of tile ceding, hi"J Is a heat transfer coefficient, and EFAI:t and ell are the effecuve emittance/a~orptance of the far-fi~]~l and ceiling F'gure B-4.5.3 Illustration of the geometry_for boundary value upper surface (assumed to be grey), respectively, both of which are problems of the temperature distributions, T n, through the ceiling taken to be 1. at radial positions r n . Tile value tbr h U to he used [ 18] is ~Ls ffdlows:

(e,-5o) h U = I. 65( TAMB - TS,U )1/3 T~fa r¢ is assumed to be the maximum temperature of the ceiling (i.d.',~ temperature of the exposed surface at r = 0). The parametric study [15] for the thermal response of uncontined ceilings above constant and growing fires indicates generally that (h U in W/m2,TAMB arid TS] ] in K) clhanges in TJTA,fa v as a function of r / H are such dlat d(T /TMA ~¢ I/d'(~'~H) = 0(1). Therefore, it is rez~hsonable to expect B-4.5.'~ Solving for the Thermal Response of the Ceiling for qHT." ac'curateres- --hl'ts for - e equation P~-36 integral of q(z.oArtt l by The temperatnre of tile ceiling material is assumed to be governe~l interpolating between values of qCONV.L calculate-clat'" "r'a-dial by the Fourier heat conduction equation. By way of the lower positions separated by r / H inter~-Ms of-0.1 to 0.2. ceiling surface boundary condition, tile boundary value problem is coupled to, and is to be solved together with, the system of Using the above ideas, the following procedure for finding the equations B-2 and B-8. thermal response of the ceiling and solving for qHT is imple- mented:

610 NFPA 204M ~ A97 ROP

(a) Since 7t~#rr -YrrRp is a measure of H in the current is assumed that the specific link is positioned at a specified radius problem, and"D~r F /'Z~g a measllre of the 0nagimum value of g, from the impingement point, r = rl, and the distance below the N RAn is chosen as several times (Dv~ / 2) / {Y¢811. - YFIRE 1" lower ceiling surface, z = zl. TI h-as been defined as the link's In il'~ c~e, N DAr/ is ~:hosen as tl~e fi"l~t in~eger'~qu~ to or greffter assumed near-uniform tem~eratu-re. Therefore, instantaneous than [5[Dm~ 7~) //YcrsL - Yvrnr )+ 21" changes in T L are determined by the following: (b) One temperature malculation point is placed at r = 0 and the remaining Np.AD ~alculation points are distribute~ with uniform dT L / dt=(TcJ,L-k, TL ))Vc; "L1/2 / RTI (B-51) separation at ,~l between r = 0.2[yr#n - y~m# ) and r = DEby. / 2, the latter value bein'~g'tT~uppe" F~i'n~''it of~the integral of equation B-36 [i.e., r 1 = 0; ro =0.2[~r,~rr -~t'ru~ ]; where TCTL and V67..L are the walues of Vcl and TCl, respec- r~ro~n = DEFF /2; r. =r. +tAr if ~'~'~< NRA~,twhere tively, eva~'~ted near-~,.fie link position, and wISere RTI (response - n time index), a property of the link and relative flow orientation, can be measured in the "plunge test" [21,22]. The l~T,l.for ordinary (c) The boundary value problems .are solved for the NrtAn sprinkler links range from low values of 22(re.s)1/z for quick- ....temperature distributions,. Tn" At arbitrary, radius, rn, dfe"g'e are operating residenuVal sprinklers, to 375 (re.s) 1/2 for slower standard indicated m the reset portmn of Figure 13-4.5.3. sprinklers [23]. The utility of equation B-51, is shown to be valid typically through the link fusing process [24], is discussed further (d) For any moment of time during the calculation, the lower [25], and actually is used to predict link response in a parametric surface values of the Z are used to compute the corresoondinl~ study involving two-layer compartment fire scenarios. Also, the link discrete values of q...... (t~ = a...... (r = r n t'~ from eouauon response prediction methodology has been used [23], and demon- 8"40. (.A./lVV,/~,~ I *¢..{]IYV,LX ' / * strates favorable comparisons between predicted and measured link responses in a full-scale, one-room, open-doorway compartment fire (e) The ~t~/. distribution in r is approximated by interpolat- experiment. ing linearly 15~'6b~e"fithe q~'ONV,L,n" The integration indi~tedin equation 13-36 is carried dUE To compute T I from equation B-51 for a different link location necessitates estim-ates of VCj ,L and TCj,L for arbitrary link The procedure for solving for the T is the same as that used in positions, r L and z L. reference [ 15]. It requires the dnckness, thermal conductivity and thermal diffi~sivity of the ceiling material. The solution to the one- 11-5.2 The Velocity Distribution of the Ceiling Jet. Outside of the dimensional heat conduction equation involves ,an explicit finite plume/ceiling impingement stagnation zone, defined approximately difference scheme that uses an algorithm token from references by r / H > 0.-'2, find at a given r~, VCI rises rapidly frohazero at the [19,20]. For a given set of calcnlations, N < 20 equal-spaced nodes ceiling's lower surface, z = 0, to a mff.Jdmum, VA~Av, at a distance are positioned at the surfaces and through the flfickness of the z = 0.23~, c$(r) being the distance below the ~l~mg where ceiling at every radius position r Thespacing tSZ (setFigure V / VMA X" = 1'/2116]. In this region outside the stagnation zone, B-4.5.3), of these •ts selected' " to ben' large enough (Ixtsed' on a VC] Ci~tie estimated [16] as shown below: maximum time step) to ensure stability of the calculation.

11-5 Actuation of Vents and Sprinklers by Near-Ceiling-Deployed V cJ I V MAx Fusible Links. It is an objective of this guide to simulate conditions in building spaces where ceiling vents and sprinkler links can be 0 Vz I actuated by the responses of near-ceiling-deployed fusible links. The concept is that, during the course of a compartment fire, a deployed ~\\\NNN NNNNNN.'qL\...\~ link is engulfed by the near-ceiling convective flow of the elevated- f,,xxx\\xxx\ , ,; temperature products of combustion and entrained air of the fire- ) ( generated plume. As the fire continues, convective heating of the link leads to an increase in its temperature. If and when its fuse temperature is reached, the device(s) being operated by the link is actuated. 4.3 The near-ceiling flow engulfing the link is the plume-driven ceiling jet referred to previously, which transfers the flow to the lower Z / (0.2303 ceiling surface and is cooled as it traverses under the ceiling from the point of plume-ceiling impingement. In the case of relatively smooth ceiling configurations, ,'~sumed to be representative of the facilities studied in this guide, the ceiling jet flows outward radi,'dly from this point of impingement, and its g~s velocity and temperature distributions, Vet and Tc/, r~pectively, are a traction of radius from d~e imping-~ment po'fi-~t, r, distance below the ceiling, z, and time.

B-5.1 Predicting the Thermal Response of the Fusible Links. The thermal response of deployed fusible links is ~'tlculated up to their fi~se temperature, T F , by the convective heating flow model [21]. It F'~gu_re B-5.2 A plot of dimensionless ceilin~ let velocity distribution, vcj / v~t~x, as a function of • / (0.2~'~) per equaeon B-52.

where r~ H > 0.2:(B-52)

I/7 (B-52)

VMAx /V = 0.85(r/-H)-L1;6/ H = O" IO'r ~ / "H'O'9) ; v =g 1/2H1/2~,*1/3~H (B-53)

where QH is defined in equation B-44. VC] / VMAX per equation B-52 is plotted in Figure 17-5-2.

611 NFPA 204M I A97 ROP

In the vicinity of near-ceiling-detlloyed links located inside die stagnation zorie, the fire-driven flow is changing directions from an upward-directed plume flow to a outward-directed ceiling jet-type ATcj = Tcj - T u flow. There the flow velocity local to the link, the velocity that drives the link's connective heat transfer, involves generally a significant = Ceiling jet temperature-upper layer temperature vertical as well as radial component of velocity. Nevertheless, at such link locations, it is reasonable to continue to approximate die link response using equatioti 13-51 with Vf:I estimated using equations B- 0 52 and B-53 and with r/H set equa[Vto 0.2. This approximation is shown as follows: ~N'x\\\b\\\\\\\\\\\ "N\ (

Where O<_r/H <0.2:

VCj = Vcj(r l H = 0.2) (B-54)

B-5.3 The Temperature Distribution of the CeillngJet. Outside of the plume-ceiling impingement stagnation zone (i.e., where ' r / H _> 0.2 ) and at a given value of r, T m rises very rapidly from the temperature of the ceiling's lower sur~t~e, T s I., at z = 0, to a maximum, T^,A¢, somewhat below the ceiling su'fface. It is assumed that t ~itl~maximum vahie of TCI occurs at the identical distance below the ceiling ,xs does the m.'fximum of Vc.I (i.e., at z = 0.236 ). Below this elevation, T~; I drops with inc'i%asing distance from the ceiling until it reac[i'~s the upper-layer tempera- ture. T U . In dlis latter, outer re, on of the ~eiljngjet, die sb~.pe of tile norf~lalized T,, distribution, i Tc:I - TI]I/(TMA X - Tu }.has tile same characteristio,~ as that ofkV'~.!r / V~x,. A]s(), sitice tl rbl lent boundary flow exists, it is r'~ason,'il~'to expect that the characteristic thicknesses of the outer region of both the velocity Figure B-5-3 Plots Of dimensionless ceiling jet temperature and temperature distributions is tile same, dictated by the distribu- distribution, O, as a function of z / 0.233 per equation B-55 for tion of the turbulent eddies. cases where O s is < O, between 0 and I, and > O.

For these re,xsons the velocity and temperature distribution are In a manner similar to die treatment of V,-,I / VMa x,, for die approxinlated as in be identical in the oilier region of tile ceiling jet purpose of calculating T L from equation BY-'eal, OS~]'~ approximated flow, 0.236 < z. In the irmer region of the flow, between inside die stagnation zone by die description of equations B-55 and z = 0 arid 0.236, the normalized telnperaqire distribution is B-56, with r/'H set equal to 0.2 as follows: approximated by a quadratic fimctioil of z / (0.23~), necessitating the use of Tg7 = Tvt at z = 0 and TCI = TM~ ~, dTi)~ldz = 0 at z = 0.23S. Therefore, where r "fl H > "07 ~Z:

<->s ?'<-T,)= i'0-<'/0.''/'': (B-55)

V~71VMAx if 1 _<,

It should be noted that 0 s is negative when the ceiling surface Where O 0. enthalpy flow in the upper-layer portion of the plume, Q , less the integral (from the plume-ceiling impingement prior to r ) of the flux of convective heat transfer from the ceiling jet to the ceiling surface as follows:

612 NFPA 204M I A97 ROP

Vdlere 0 .._9 0 ." 2 (Y, CEJL - Y'FIRE..)',unless .,. the e q uations. , and tJle cocte are usec~ to stmutate an unconllnea ceaalng scenario wnere [YCEIL - YGURT ) = O. The ceiling of die curtained space should be relatively smooth, with protuberances° having depths, significandy less than.. 0 .1W . Except at die Iocauons of die curtain boards, below-celhng-mounted barriers to fow, such as solid beams, should be avoided. Ceiling surface protuberances near to and upstream of fusible links (i.e., ' . . "r . ~cONV ts the fractmn of Q transferred by convecuon to the between the links and die fire) should be significantly smaller than ce~Frig from the point of ceiling impingement to r ,as follows: link-to-ceiling distances.

W V is die width, i.e., the smaller dimension of a single ceiling vent (or vent cluster). Therefore, the prediction of smoke layer tlhickness' ~CEJ./,,-,Y' is reliable only after the time that ~Y'CEIL -- Yl / IWV ts greater than 1. (See also 3-4(a).) Note that In equations B-a8 and B-59, Q has been ~dculatedpreviously m diis pFaces fin additiopal limitation on.the minimum depth of the equation B-34. Also, the integral on tbe right hand sides of curtain boards [i.e., [YCEIL - YCURT ) / WV sbould exceed I I. equations B-58 and 13-59 can be calculated by approximating (I...... ~[r,l) ,as shown in equation 13-59 as a linear fimction of r At all times during a simulated fire scenario, the overall building between prevmusly calculated values of qCONV,L (r = rn,t ) . space should be vented to the outside (e.g., through opened doorways). The integral on the [eft band side of equation B-58 is calculated using Vcr of equations B-52 and B-53 and Tm of equations B-55 and B-SlY.:'From this, tile desired distribution'fbr TMA X is determined `as follows:

(TMAx_Tu)=2.6(I_2'CONV "~ 0£ ,2/3 (B-60)

ifO/2<_r/H The result of equation [',-60, together with equations B-55 and B-56 repr~ent the desired estimate for Te~. Tiffs and the equations B-52 In dais regard, compared to die open ceiling vents in the curtained through B-54 estimate for TGj are us'~d to calculate T L from compartment, the area of the outside vents must be large enough so equation I?-51. that the pressure drop across the outside vents is small compared to die pressure drop across the ceiling vents. For example, under near- B-5.4 Dependence of Open Vent Area on Fusible-Link-Actuated steady-state conditions, when the rate of mass flow into the outside Vents. As discu~ed, the influence of ceiling vent action on the fire- vents is approximately equal to the rate of mass outflow from the generated environment is dependent on the active area of the open cteiling vents, ~ outside vent agea must satisfy ceiling vents, A V . A variety of basic vent opening design strategies is IAVOUT /AV} (TuJ TAMB) z >>, 1 , or, more conservatively and possible, and a major application of the current model equations is ihdependent ot" Tu, AVO / A 2 >> 1 . The latter criteria will to evahmte these strategies within tile context of die developing fire always be reasonablytatisfieU~if ,~V)ou T/Av>2. Under flashover- environment. For example, one of the simplest strategies [9], level conditions, s3y, when T u/T,,djMB = 3, the former criterion assumes that all vents deployed in die specified curtained area are will be satisfied if [3AvouT / A V ~ >> i, say, if AVOUT = AV , or opened by whatever means at the onset of die fire. In general, A V even AVOUT is somewhat smaller than Av. will be time-dependent. To tile extent that a strategy of vent opening is dependent directly on die fusing of any one or several The simulation assumes a relatively quiescent outside environment deployed fusible links, tile Io~ation of these links ~md their charac- (i.e., without any wind) and a relatively quiescent inside environ- teristics (i.e., likely spacings from plume-ceiling impingement, ment (i.e., remote from vent flows, under-curtain flows, ceiling jets, distance below the ceiling, and the RTI) and tile fimctional and the fire plume). In real fire scenarios, such an assumption relationship between link fiasing and A v need to be specified. sbould be valid where the characteristic velocities of actualflows in These matters c~'m be examinedin the context of different solutions these quiescent environments are much less than the velocity of the to the overall problem by exercising parametrically die LAVENT fire plume near its ceiling impingement point (i.e., where the computer program [2], wltich implements all the model equations characteristic velocities are much less than VMA X of equation 13- provided in tl'ds appendix. 53). It should be noted that, for a given fire strength, Q, this latter a(ySUmption places a restriction on die maximum size of B-5.5 Concluding Remarks --A Summary of Guidelines, Assump- CEIL - YbTRE)' which is a m~asure of H, sir~cel/~3MAX is tions, and Limitations. The theory presented here is the basis of appro~imafeTy--proportional to [YCE/L - YF/RE)- ' ~" • LAVENT, a user-friendly computer program [2] that is supported by a user guide [3] and dlat can be used to study parametrically a wide In configurations where smoke flows below curtain partitions to range of relevant fire scenarios. adjacent curtained spaces, the simulation is only valid up to the time that it rakes for any one of the adjacent spaces to fill with smoke to The assumptions made in the development of tile set of model the level of the bottom of die curtain. While it is beyond the scope equations provided limit fire scenarios or ~spects of fire scenarios of diis guide to provide any general guidelines for this limiting time, that can he simulated and studied with confidence. A summary of die following rule can be useful where all curtained spaces of a guidelines and assumptions that characterize what are perhaps file building are similar and where die fire is not growing too rapidly most critic~al of these limitations follows. These are the result of die time to fill an adjacent space is of the order of the time to fill the explicit or implicit ~L~mnptions necessary for valid application of the original space. variety of submodels introduced throughout this work. The reliability of the simulation begins to degrade subsequent to L and W are tile length and widdL respectively, of the plan area of " die time dlat file top of the flame penetrates die layer elevation and the curtained space. Simulated configurations should be limited to especially if equation B-20 predicts a flame height that reaches the those with ~pect ratios, L~ W that are not much different dlan 1 ceiling.

It is assumed that the smoke is relatively transparent and that the rate of radiation absorbed by or emitted from the smoke layer is small compared to file rate of radiation transfer from the fire's combustion zone. The assumption is typically true and a simulation is valid at least up to those times that the physical features of the ceiling can be discerned visually from the floor elevation.

613 NFPA 204M ~ A97 ROP

Surfaces,"Journa/ofHeat Transfer, Vol. 104, pp. 49~-499, Aug. 1982. It should be emphasized that the above limitations are intended 19. Emmous, H.W., "The Prediction of Fire in Buildings," 17111 only as guidelines. Therefore, even when the characteristics of a Symposium (International) on Combustion, Combustion Institute, particular fire scenario satisfy tilese limitations, die results should be pp. H01-1111 (1979). regarded with caution until solutions to the overall model equations have been validated by a substantial hody of experimental dat.x 20. Mider, H.E., and Emmons, H.W., "Documentation for the Fifth Also, where a fire scenario does not satisfy die above limitations but Harvard Computer Fire Code," Home Fire Project Tech. Report 45, is close to doing so, it is possible that the model equations can still Harvard University, Cambridge, MA 1981. provide usefifl quantitative descriptions of the simulated phenom- ella. 21. Heskestad, G. and Smith, H.F., "Investigation of a New Sprinkler Sensitivity Approval Test: The Plunge Test," Technical B-6 Referenc~ for Appendix B. Report Serial No. 22485, RC 76-T-50, Factory Mutual Research Corporation, Norwood, MA, 1976. 1. Cooper, L.Y., "Estimating the Environment and the Response of Sprinkler Links in Compartment Fires with Draft Curtains and 22. Heskestad, G., "The Sprinkler Response Time Index (RTI)," Fusible Link-Actuated Ceiling Vents," Fire .~fetyJournal, Vol. 16 pp, Paper RC,-81-TP-3 presented at file Technical Conference on 137-163, 1990. Residential Sprinkler Systems, Factory Mutual Research Corpora- tion, Norwood, MA, April 28-29, 1981. 2. LAVENT software, available from National Institute of Shandards and Tecbnology, Gaithersburg MD. 23. Evans, D.D., "Calculating Sprinkler Actuation Times in Compartments," Fire SafetyJournal, 9, pp 147-155, 1985. 3. [)avis., W .D .,and ..('looper: LY.7 "Estimating file Environmentand. the Response of Sprmkler Links tn Compartment Fires with Dr,fit 24. Evans, D.D., "Characterizing the Tbermal Response of Fusible Curtains and Fusible Link-Actuated Ceiling Vents ~ Part II: User Link Sprinklers," NBSIR 81-2329, National Bureau of Standm'ds, Cuide for the Colnputer Code LAVENT," NISTIR 89-4122, National Gaithersburg, MD, 1981. Institute of StandarcLs and Technology, Caithersburg MD, August 1989. 25. Cooper, L.Y. and Stroup, D.W., "Test Results and Predictions for the Response of Near-Ceiling Sprinkler Links in Full-Scale 4. Emmons, H.W., "The Flow of Gztses Through Vents," Harvard Compartment Fires," Fire Safa"3 Science -- Proceedings of th* Second University Home Fire Project Technical Report No. 75, March 16, International Symposium, Tokyo, June 13-17, 1988, pp 623-632, T. 1987. Wakumatsu et al, Eds., International Association of Fire Safety Science, Hemisphere Publishing Co., New York, 1989. 5. Thomas, P.H., eta/, "Investigations into the Flow of Hot Gases in Roof Venting," Fire Research Technic~'d Paper No. 7, HMSO, B-7 Nomenclature for Appendix B. London, 1963. A = plan area of single curtained space 6. Heskestad, G., "Smoke Movement andVentin. . g," Fire Safety Journal, 11, pp 77-83, 1986, and Appendix A: C,uzde for Smoke and Heat AE~ = effective area for heat transfer to the extended Venting. NFPA 204M, National Fire Protection Association, Quincy, lower ceiling surface, ffD~F /4 MA, 1982.

7. (looper, L.Y., "A Mathematical Model for Estimating Available A v = total area of open ceiling vents in curtained space Safe Egress Time in Fires, Fire and Materials," 6, 3?'4, pp. 135--144, 1982. AVOU7/. = total area of open vents to outside exclusive of A v 8. Heskestad, G., "Engineering Relations for Fire Plumes," Fire C = vent flow coeffident (= 0.68) Safe~.Joumal, 7, pp. 25-32, 1984. Cp = specific heat at constant pressure 9. Hinkley, P.L., "Rates of'Production' of Hot Gases in Roof Venting Experiments," Fire SafetyJournal, 10, pp. 57--64, 1986. CT = 9.115, dimensionless constant in plume model

10. Zukoski, E.E., Kuboha, T., and Cetegeu, B., Fire SafetyJournal, 3, Cg = specific heat at constant volume p 107, 1981. DEFF = effective diameter of AEFF I 1. (looper, L.Y., "A Buoyant Source in the Lower of Two, Homogeneous, Stably Stratified Layers," 20th International = effective diameter of fire source Symposium on Combustion. Comhustion Institute, pp. 1567-1573, DFIRE ( fiDe/RE /4 = area 0ffire source ) 1984.

12. Cooper, LY., "Convective Heat Transfer to Ceilings Above g = acceleration of gravity Enclosure Fires," 19lh Symposium (Interttational) on Combustion, Combustion Institute, pp. 933-939 (1982). H = distance below ceiling of equivalent source

13. Cooper, L.Y., "Heat Transfer fi'om a Buoyant Plume to ,an /~ = characteristic heat transfer coefficient iJnconfined Ceiling," Journal ¢fHazt Transfer, Vol. 104, pp. 446--451, Aug. 1982. h L, hU = lower, upper ceiling surface heat transfer coeffi- cient 14. Cooper, LY. and Woodhouse, A., "The Buoyant Plume-Drlven Adiatxttic, Ceiling Tern p(erarure Revlsited,"Journal of Heat Transfer, L = dlaracterisdc length of the plan area of curtained Vol. 108, pp. 822---826, Nov., 1986. space

15. Cooper, LY., and Stroup, D.W., "Thennal Response of LGURT = lengfll of the perimeter of area A connected to Unconfined Ceilings Above (;rowing Fires and the Importance of other curtained areas of the building Convective Heat Transfer,"Joumal of Heat Transfer. Vol. 109, pp. 172-178, Feb. 1987. LFLAME = flame length 16. Cooper, LY., "Ceiling Jet-Driven Wall Flows in Compartment 7hGURT = mass flow rate from below curtain to upper layer Fires," Condmstion .Scienceand Technolol~, Vol. 62, pp. 285--296, 1988. mENT = rate of plume mass entrainment between the fire 17. Hilsenratil,J., "Tables of Thermal Properties of Gases," and the layer interface Circular 564, National Bureau of Standards, Galthersburg, MD, Nov. 1955. 7nPLUME = mass fl0w rate of plume at interface 18. Yousef, W.W., Tarasuk,J.D., and McKeen, wJ., "Free Convec- tion Heat Transfer from IJpward-Facing, Isothermal, Horizontal

614 NFPA 204M ~ A97 ROP

m U = total m~ss of the upper layer TMA x (t) - ~,, (,: 0,,)-- r(z-- 0,,,,-- 0) ~hU = net mass flow rate to upper layer TS,I., TS,U = absolute temperature of lower, upper ceiling surfac e ~hVENT = mass flow rate through ceiling vents to upper layer Ts,L,n(t) rs,, T (z = N = number of equal-spaced nodes through the ceiling Tu, TA~ absolute temperature of upper layer, outside ambient NRAD = number of values of r n 7",, T(Z,t;r= %) P = length of perimeter of single curtained area t time

P,. Prandtl number, taken to be 0.7 v average flow velocity through all open Vents P "~ PAMB at floor elevation Vcj = velocity distribution of ceiling jet gas Pv , PA~ pressure in upper layer, outside ambient Vq,L Vcj at link

= energy release rate of fire VMAX = maximum value of VCj at a given r

Q, stren~h of continuation source in W = characteristic width of plan area of extended upper layer curtained space

Qn = dimensionless strengtla of plume at Wv = width of a single ceiling vent (or vent ceiling cluster) .,It QEQ = dimensionless strengtll of plume at Y, YC,EIL , YCURT , Y FIRE = elevation of: smoke layer interface, in te trace ceiling, bottom of curtain, fire above floor qCONV,L, qCONV,U = convective heat transfer fltLX tO lower, y" upper ceiling surface SOURCE = elevation of plume continuation point source in extended upper layer above q CONV.L,n = qCONV,L(r=rn,t) floor qCURT = endmlpy flow rate from below curtain Z = distance into the ceiling, measured to upper layer from bottom surface

= beat transfer rate to upper layer •Z, ='L = distance below lower ceiling surface, z at link q PLUME •, enthalpy flow rate of plume at interface O~ ru / T a2v~ q RAD- FIRE = radiation flux incident on lower = ratio of specific heat, Cp / C V surface of ceiling APc~L = cross-vent pressure difference • H qtt q RERAD ,L ' RERAI),U = re-radiation flux to lower, upper surface of ceiling APCURT = cross-curtain pressure difference

qv = net endmlpy flow rate plus heat = valueof z where vcj =VMA X/2 u,'ansfer rate to upper layer ~Z = distance between nodes through the • ss • ~u, ceiling thickness qL, qv •, net beat transfer fluxes to upper, lower ceiling surface E = constant, equation (8-18)

= enlbalpy flow rate through ceiling ~L' ~U' ~FLOOR' = emittance/absorptance of: lower, vent to upper layer e FAR upper, floor, and far field grey surfaces, all taken to be 1 R gas cooshant,

(~- qcp / ~ = cp - c.v 6) = normalized, dimensionless ceiling jet t(eTmperatur $ distribution, Re. H re)a~olds number of plume at ceiling ele~ttion q - ru) //T~x - ru ) Os = O at lower ceilinu surface, RT1 Response Time Index fTs.L-Tu)/I%x-Tu) I" .radial distance from plume-ceiling = fraction of Q radiated from combus- impmgement tion zone r L r at link 2coNy = fraction of Q transferred by convection from upper layer ~= discrete values of r • 'p 2coNy = fracuon of ~ transferred to the T absolute temperature of ceiling ceiling in a arcle of radius r, and material centered at r = 0, equation (B-56).

TAD adiabatic lower ceiling surface vv = kinematic viscosity of upper layer gas temperature Pv ,PAMB = density of upper layer, outside TcJ temperature distribution of ceiling jet ambient gas IT dimensionless variable, equation To] .L Tcj at llnk (B-28)

615 NFPA 204M -- A97 ROP

Appendix C User Guide for the LAVENT Computer Code (b) Ceiling/link separation distance; This Appendix is not a part of tlw r~omnwndations of this NFPA document (c) Link fuse temperature; and but is incbMed fi~r informational pu~pose.~ only. (d) The response time index (RTI) of the link. C-1 Overview. This appendix is a user gatide for tile LAVENT computer code (Link-Actuated VENTs), Version 1.1, and an For any particular run of LAVENT, the code outputs a summary of ,associated graphics code c=dled GRAPH. As discussed in Section 6-2 the input information and simulation results of die calculation, in and Appendix B, LAVENT has been developed to simulate tile tabular form, at uniform simulation time intervals requested by the environment and the response of sprinkler linLs in compartment user. The output results include: fires ,fith curtain boards and fusible-link-actuated ceiling vents. (a) Temperature of the upper smoke layer; A fire scenario simulated by LAVENT is defined by the following inl)ut parameters: (b) Height of the smoke layer interface;

(a) Area and height of tile curtained space; (c) Total mass in the layer;

(b) Floor-to-bottom-of-curtain separation distance; (d) Fire energy release rate;

(c) Length of the curtain (a portion of the perimeter of the (e) Radial distributions of the lower ceiling surface temperature; curctined space can include floor-to-ceiling w:,dls); (f) Radial distribution of heat transfer rates to the lower and upper (d) Thickness arid prol)erties of the ceiling material (density, ceiling surfaces; and thermal conductivity, and heat capacity); (g) For each link, the temperature, and the local velocity and (e) Constants that define a specified time-dependent energy temperature of the ceiling jet. release rate of tile fire; This appendix explains LAVENT using a series of exercises in (f) Fire elewation; which the reader reviews and modifies a default input data file that describes vent and sprinkler actuation during fire growth in an array (g) Area or characteristic energy rele:L~e rate per unit area of the of wood pallets located in a warehouse-type occupancy. Results of fire; tile default simulation are discnssed.

(h) Tohal area of ceiling vents whose openings are actuated by a LAVENT is written in FORTRAN 77. The executable code operates single filsible link (multiple vent area/link system combinations may on IBM PC-compatible computers and needs a minimum of 300 be permitted in any particular simulation); and kilobytes of memory.

(i) Identifying numbers of fusible links used to actuate single C-2 Introduction -- The Phenomena Simulated by LAVENT. Figure sprinkler heads or groups of sprinkler heads (multiple sprinkler C-2 depicts the generic fire scenario simulated by LAVENT. This links are permitted in any particular simulation). involves a fire in a building space with ceiling-mounted curtain boards and near-ceiling fusible-link-actuated ceiling vents and Tile characteristics of tile simulated fusible links are defined by the sprinklers. The curtained area can be considered as one of several following input parameters: such spaces in a single large building compartment. By specifying that file curtains be deep enough, they can be thought of as (a) Radial distance of tile link flom tile fire/ceiling impingement simulating file walls of a single uncurtalned compartment that is point; well-ventilated near the floor.

"•- Layerinterface Yj YFIreyI-,re---,~]-- i ~,~e,i ,~irCurt

Vent or sprinkler link

./

'1 - / \veloci~ Distancebelow ceiling

Figure C-2 Fire in a building space with curtain boards, ceiling vents, and fusible links. 616 NFPA 204M ~ A97 ROP

Tile fire generates a mixture of gaseous and solid-soot combustion D3 products. Because of high temperature, buoyancy forces drive the products upward towarcl tile ceiling, forming a plume of upward Draft curtain D2 moving hot gases and particulates. Cool gases are laterally entrained and mixed wid~ tile phmte flow, reducing its temperature as it continues its ,ascent to the ceiling. When file hot phnne flow impinges on the ceiling, it spreads under • • • • • • • it, forming a relatively thin, high-temperature ceiling,jet. Near- ceiling-deployed fusible lines engulfed by die ceilingjet are depicted L1 in Figure G-2. There is reciprocal convective cooling ,and heating of the ceiling jet and the cooler lower ceiling surface, respectively. The lower ceiling surface is also heated due to radiative transfer from tile combustion zone and cooled due to reradiadon to tile floor of the • • o.o compartment. Tile comparm)ent floor is ;tssnmed to be at ambient / temperature. The upper ceiling surface is cooled as a result of / Fire convection and radiation to a far-field, ambient temperature / environment. • • 0/3 • • • • V~qaen tile ceiling jet reaches a bounding vertical curtain board or / Vent wall surface, its flow is redistribttted across the entire curtained area • • /O • • \ • • and begins to form a relatively quiescent smoke layer (now some- (~,rin kl;r what reduced in temperatm'e) that submerges the continuing ceiling-jet flow activity. The vpper smoke layer grows in thickness. Away from bounding surf:tces, the time-dependent layer tempera- lure is assumed to be relatively uniform tlaroughout its thickness. It should be noted that the thickness and temperature of tile smoke layer ,affects the npper-phnne characteristics, the ceilingTjet characteristics, and the heat-transfer exchanges to die ceiling. D 1 = 12 ft L 1 = 6 R: 2 sprinklers If tile height of tile bottom of the smoke layer drops to the bottom ~ D 2 = 21 ft L 2 = 21.2 It: 2 vents of the curtain board and continues downward, rile smoke begins to D 3 = 42 ft L 3 = 44.3 ft: 2 vents flow below the curtain into tile adjacent curtained spaces. Tile L 4 = 13.4 ft: 4 sprinklers growth of the upper layer is retarded.

Fusible links that are designed to actuate tile opening of ceiling vents and tile onset of materflnw through sprinklers are deployed at specified distances below fl~e ceiling and at specified radial dist,-mces from the phnne/ceiling impingement point• These links are Figure C-3(a) Vent and sprinkler spacing and fire location for the submerged within the relatively high-temperature, high-velocity default simulation. ceiling-jet flow. Since the velocity and temperature of the ceiling jet varies widl location anti time, tile heat transfer to and time-of-fusing of any particular link design also varies. Fusible-link-actuated sprinklers are deployed on a square grid with | ~-~t spacing between sprinklers. The links have RTIs of 400 fit-s) Tile fusing of a ceiling-vent link leads to the opening of all vents ~/~ and fuse temperatures of 165°F. The sprinklers and links are "ganged" to that link. Once a ceiling vent is open, smoke flows out mounted 1 ft below the ceiling surface. of the curtained space. Again, :Ls in a case where smoke flows below tile curtains, growth of the upper layer tllickness is retarded. The simulation fire involves four abutting 5-ft higll stacks of 5-ft x 5- ft wood pallets. Tile combined grouj~ng of pallets makes up a The fusing of a sprinkler link initiates the flow of water tilrough the combustible array l0 ft x 10 ft (100 ft ~ in area) on the floor and 5 ft sprinkler. in height• It is assumed that other combustibles in the curtained compartment are far enough away from this array that they cannot All of these above phenomena, up to the time that waterflow be ignited in the time interval to be simulated. through a sprinkler is initiated, are simulated by LAVENT. Results cannot be used after water begins to flow through a sprinkler. The total energy release rate of the simulation fire, ~, is assumed to grow from ignition, at time t = 0, in proportion to t ~. According C-3 The Defauh Simulation. Tile use of LAVENT is discussed and is to the guidance in Table 4.2 of [ 1 ], in the growth phase of the fire, illustrated in tile following paragraphs where exercises in reviewing is taken specifically as follows: and modifying the LAVENT default-simulation input file are provided. To appreciate tile process more fully, a brief description Q = 1000 it/(130 s)l 2 Btu/s of tile default sitmdation is presented at the outset. The fire grows according to the above estimate until the combus- NOTE: As explained in G-4 Getting Sorted, the user can choose tibles are fully involved. It is then assumed that Q levels offto a to run LAVENT using either English or metric units. The relatively constant value. Following the guidance of Table 4.1 of [1] default simulation uses English units. Tile example in Appendix and Table 5-5.2(b), it is estimated that, at the fully developed stage D uses metric units. of the fire, the total energy release gate for the 5-ft high stack of wood,pallets will be 330 (Btu/s)/fi~, or 35,000 Btu/s for the entire Tile de,~ault scenario involves a 84 ft x 84 ft curtained compartment 100-ft~ array. The above equation leads to the result that the fully (7056 ft~in area) with the ceiling located 50 ft above tile floor. A developed stage of the fire will be initiated at tfd = 747 s. curtain board 15 ft in depth completely surrounds and defines the compartment, which is one of several such compartments in a larger A plot of the fire growth according to the above description is building space. The ceiling is constructed of a relatively thin sheet- shown in Figure C-3(b). In the actual calculation, the fire's steel lower surface that is well-insulated from above. [SeeFigure C- instantaneous energy release rate is estimated by interpolating 3(,,).1 linearly between a series of N input data points at times t n, n = 1 to N, on the fire-growth curve. These points are defined by-user- The curtained compartment has fou~ uniformly spaced, 48 ft2 specified values of [tn, 0 (tn) ]" For times larger than tN, the fire's ceding vents w~th a total area of 192 ft "~, or 2.7 percent of tile energy release rate is 5.ssfim~d to stay constant at Q (tN)~ The compartment ;trea. Opening of the ceiling vents~is.actuated by calculation fire-growth curve involves six input dat~ ptints (i.e., N = quick-response fusible links witll RTIs of 50 fit.s) J/~ and fuse 6). These points are plotted in Figure C-3(b). temperatures of 165°F. The lin~ are located at tile centers of the vents and 0.3 ft below the ceiling surface.

617 NFPA 204M -- A97 ROP

ff Option 1 or 2 is chosen, the program will ask for the name of the 40(103) I I I I I I I data file that will be used. If the chosen file resides on the hard disk, dais question should be answered by typing the path of the file name; for example, C~ksubdirectory~lename. If the file is on a floppy disk, type A:filename or B:filename, depending on whether the A or B 30(103 ) drive is being used. It is recommended that all data files use a common extender such as,in order to facilitate identification of these files.

A first-time user should select Option 4, RUN THE DEFAULT .~" 2o0o 3) CASE, by entering 4 [ret]. This will ensure that the code has been transferred intact. The defanh-case output is provided in Table C-4. This is discussed in Section C-8. As a point of information, the times needed to carry out the default simulation on IBM PC-compatible 486/33 MHZ and Pentium/90 MHZ computers were 40 s and 8 s, lO(lO 3) respectively. Now restart the code and, at dais point, choose Option 3, MODIFY THE DEFAULT CASE, to review and modify the default input dam. Enter 3 [ret]. OC 0 ! 200 400 600 t800 C-5 The Base Menu. Time (s) G5.1 Modifying the Default Case -- General. When Option 3, MODIFY THE DEFAULT CASE, is chosen, the following menu is displayed: Figure C-3(b) Energy release rate versus time for the fire of the default simulation. 1 ROOM PROPERTIES 2 PHYSICAL PROPERTIES 3 OUTPUT PARAMETERS The position of the center of the fire is identified in Figure G-S(a). 4 FUSIBLE LINK PROPERTIES In terms of thisplan view, the fire is assumed to be located at the 5 FIRE PROPERTIES midpoint ofa 12-ft line between two sprinkler links, at a distance of 6 SOLVER PARAMETERS 21,2 ~t from each of tile two closest equidistant vents (a total area of 0 NO CHANGES 96 fig), and at a distance of 44,3 ft from the remaining two ecluidis- t,ant vents (a total area of 96 ftg). Of the sprinklers and assocmted This will be referred to as the "base menu." links, two are closest and equidistant to the fire-plume ,axis at radial distances of 6 ft. Figure C-3(a) shows that the second ,and third Entering the appropriate option number of the base menu and closest groups of sprinklers and links are at radial distances of 13.4 ft then [ret] will always transfer the user to the indicated item on the (four sprinklers and links) and 18 ft (two sprinklers and links). In menu. Entering a zero will transfer the user to the file status portion the defauh calculation, the opening of each of the four vents occurs, of the input section discussed in Section C-6. and the flow out of the vents is initiated at the simulated time of fusing of their associated links, Also simulated in the default The next five sections discuss data entry under Options I through 6 calculation is the thermal res )unse, including time-of-fusing, of the of the base menu. pair of sprink er Inks c osest to the fire. Now choose Option 1, ROOM PROPERTIES, of the base menu to As a final specification of the fire, it is assumed that the characteris- review and modify the default room-property input dam. Enter 1 tic elevation of the fire remains at a fixed value, 2.5 ft above the [ret]. floor, at the initial mid-elevatlon of the array of combustibles. C-5.2 Room Properties. When Option 1, ROOM PROPERTIES, of For tile purpose of the default calculation, the simulation is carried the base menu is chosen, the following room properties menu is out to t = 400 s, with data output every 30 s. displayed:

Having described the defanlt simulation, the procedure for getting 1 S0.00000 CEILING HEIGHT (FT) started and using LAVENT follows. 2 84.00000 ROOM LENGTH (Fr) 3 84.00000 ROOM WIDTH (Fr) C-4 Getting Started. The executable code, LAVENT.EXE, is found 4 2 NUMBER OF VENTS, ETC. on the floppydisk. Before using it, backup copies should be made. 5 336.00000 CURTAIN LENGTH (FT) If the user has a hard drive, a separate directory should be created 6 15.00000 HEIGHT TO BOTTOM OF and the executable code should be copied into that directory. The CURTAIN (FF) code operates on an IBM PC or compatible computer containing a 0 TO CHANGE NOTHING mafia coprocessor. It is written in Fortran 77 andneeds a minimum of 300 kilobytes of memory. All input values are expressed in either Scientific Intemationale or English units, and the units are prompted on the input menus. To execute LAVENT, change to the proper directory or insert a floppy disk containing a copy of the executable code and enter Note that the default number of vents is 2 and not 4, since the LAVENT [retl. In this case [ret] refers to the ENTER or RETURN symmetry of tile default scenario, as indicated in Figure C-$(a), leads key. Tile first prompt is: to =ganged" operation of each of two pairs of the four vents involved.

ENTER 1 FOR ENGLISH UNITS, 2 FOR To change an input value in the above room properties menu (e.g., METRIC UNITS to change the ceiling height from 30 ft to 20 ft) the user would enter 1 [ret] and 20. [ret]. The screen would show revisions using the new The program has a urfit conversion function ~md transforms files value of 20 ft for the ceiling height. This or other values on this that are in one set of units to another set. The code executes in SI screen can be changed by repeating the process. units and so conversion is only done on input and output in order to avoid rounding errors. WARNING: THE USER IS WARNED THAT IT IS CRITICAL TO END EACH ENTRY NUMBER WITH A DECIMAL POINT WHEN A For the pttrposes of getting started, choose Option 1, ENGLISH NONINTEGER NUMBER IS INDICATED (I.E., WHEN THE UNITS. Enter 1 [ret]. The following menu will be displayed on the SCREEN DISPLAY SHOWS A DECIMAL POINT FOR THAT screen: ENTRY). THE USER IS WARNED FURTHERTHAT THE CODE WILL ATTEMPT TO RUN WITH ANY SPECIFIED INPUT FILE I READ AND RUN A DATA FILE AND THAT IT WILL NOT DISTINGUISH BETWEEN REALISTIC 2 READ AND MODIFY A DATA FILE AND UNREALISTIC INPUT VALUES. 3 MODIFY THE DEFAULT CASE TO CREATE A NEW FILE 4 RUN THE DEFAULT CASE

618 NFPA 204M -- A97 ROP

Table C,-4 (page 1 of 4 pages) The Default-Case Output

CEILING HEIGHT = 30.0 FT ROOM LENGTH = 84.0 FT ROOM WI DTH = 84.0 Fr CURTAIN LENGTH = 336.0 FT CIIRTAIN HEIGHT = 15.0 Fr MATERIAL = INSULATED DECK (SOLID POLYSTYRENE) CEILING CONDUCTIVrI'Y = .240EA)4 BTU/FF F S CEILING DENSITY = .655E+02 LB/FT3 CEILING HEAT CAPACITY = .277E+00 BTU/LB F CEILING THICKNESS = .500E+00 FT FIRE HEIGHT = 2.5 FT FIRE POWER/AREA = 0.3300E+03 BTU/S FT2

LINKNO= 1 RADIUS= 6.0FT DISTCEILING= 1.00FT RTI= 400.00 SQRT FUSION TEMPERATURE FOR LINK = 165.00 K LINKNO= 2RADIUS= 21.2FT DISTCEILING= 0.30FT RTI= 50.00 SQRT FUSION TEMPERATURE FOR LINK = 165.00 K LINKNO= 3 RADIUS= 44.3 FT DISTCEILING= 0.30 FT RTI= 50.00 SQRT FUSION TEMPERATURE FORLINK= 165.00 K VENT = 1 VENT AREA = 96.0 FT2 LINK CONTROLLING VENT = 2 VENT = 2 VENT AREA = 96.0 FT2 LINK CONTROLLING VENT = 3

TIME (S)= 0.000 LYR TEMP (F)= 80.0 LYR HT (FT)= 30.00 LYR MASS (LB)= 0.000E+00 FIRE OUTPUT (BTU/S)= 0.0000E+00 VENTAREA (FT2)= 0.00 LINK= 1 LINKTEMP (F)= 80.00 JET VELOCITY (FT/S)= 0.000JETTEMP (F) = 80.0 LINK = 2 LINK TEMP (F)= 80.00JET VELOCITY (FI'/S)= 0.000JET TEMP (F) = 80.0 LINK = 3 LINK TEMP (F)= 80.00JET VELOCXI'Y (FF/S)= 0.000JET TEMP (F) = 80.0 R (FT)= 0.00 TSL (F)= 80.0 QB (BTU/Fr2 S)= 0.000E+00 QT (BTU/FT2 S)= 0.000E+00 R (FT)= 12.41 TSL (F)= 80.0 QB (BTU/FT2 S)= 0.000E+00 QT (BTU/FT2 S)= 0.000E+00 R (FT)= 24.82 TSL (F)= 80.0 QB (BTU/FI'2 S)= 0.000E+00 QT (BTU/FT2 S)= 0.000E+00 R (FF)= 37.23 TSL (F)= 80.0 QB (BTU/FT2 S)= 0.000E+00 QT (BTU/FT2 S)= 0.000E+00 R (FT)= 49.64 TSL (F)= 80.0 QB (BTU/FT2 S)= 0.000E+00 QT (BTU/FT2 S)= 0.000E+00 R (FT)= 62.05 TSL (F)= 80,0 QB (BTU/FT2 S)= 0.000E+00 QT (BTU/FT2 S)= 0.000E+00

TIME (S)= 30.000 LYR TEMP (F)= 89.6 LYR HT (FT)= 28.90 LYR MASS (LB)= 0.562E+03 FIRE OUTPIJT (BTU/S)= 0.1776E+03 VENT AREA (FT2)= 0.00 LINK = 1 LINK TEMP (F)= 80.78JET VELOCrFY (Fr/s)= 1.866JET TEMP (F) = 94.9 LINK= 2 LINKTEMP (F)= 85.37JETVELOCI'IY (FI'/S)= 2.077JETTEMP (F) = 95.3 LINK= 3 LINKTEMP (F)= 81.83 JET VELOCKIY (FT/S)= 0.873 JET TEMP (F) = 87.4 R (FT)= 0.00 TSL (F)= 84.5 QB (BTU/FF2 S)= 0.312E-01 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 12.41 TSL (F)= 81.7 QB (BTU/FT2 S)= 0.122E-01 QT (BTU/FF2 S)= 0.847E-18 R (FT)= 24.82 TSL (F)= 80.8 QB (BTU/FT2 S)= 0.570E-02 QT (BTU/Fr2 S)= 0.847E-18 R (FT)= :¢7.23 TSL (F)= 80.4 QB (BTU/FT2 S)= 0.325E-02 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 49.64TSL (F)= 80.3 QB (BTU/FT2 S)= 0.212E-02 QT (BTU/Fr2 S)= 0.847E-18 R (FT)= 62.05 TSL (F)= 80.2 QB (BTU/FT2 S)= 0.152E-02 QT (BTU/FT2 S)= 0.847E-18

TIME (S)= 60.000 LYRTEMP (F)= 96.5 LYRHT (FT)= 27.34 LYR MASS (LB)- 0.134E+04 FIRE OUTtqJT (BTU/S)= 0.3552E+03 VENT AREA (FT2)= 0.00 LINK= 1 LINKTEMP (F)= 82.80 JET VELOCITY (FT/S)= 2.395JETTEMP (F) = 105.0 LINK= 2 LINK TEMP (F)= 95.13JETVELOCITY (FT/S)= 2.657JET TEMP (F) = 105.8 LINK= 3 LINKTEMP (F)= 85.76JETVELOCITY (FT/S)= 1.117JETTEMP (F) = 92.9 R (FT)= 0.00 TSL (F)= 92.7 QB (BTU/FT2 S)= 0.517E-01 QT (BTU/Fr2 S)= 0.847E-18 R (FF)= 12.41 TSL (F)= 85.2 QB (BTU/FT2 S)= 0.223E-01 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 24.82 TSL (F)= 82.5 QB (BTU/FT2 S)= 0.107E-01 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 37.23 TSL (F)= 81.4 QB (BTU/FF2 S)= 0.619E-02 QT (BTU/FT2 S)= 0.847E-18 R (FT)-- 49.64 TSL (F)= 80.9 QB (BTU/FT2 S)= 0.405E--02 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 62.05 TSL (F)= 80.6 QB (BTU/FT2 S)= 0.292E-02 QT (BTU/FT2 S)= 0.847E-18

619 NFPA 204M ~ A97 ROP

Table C~4 (cont'd, page 2 of 4 pages) The Default-Case Output

TIME (S)= 90.000 LYR TEMP (F)= 103.2 LYR HT (FT)= 25.65 LYR MASS (LB)= 0.216E+04 FIRE OI JTI'UT (BTU/S)= 0.5328E+03 VENT AREA (FT2)= 0.00 LINK = 1 IJNK TEMP (F)= 85.90JET VELOCITY (FT/S)= 2.809JET TEMP (F) = 114.5 LINK = 2 LINK TEMP (F)= 105.74JET VELOCITY (FT/S)= 3.104JET TEMP (F) = 115.8 LINK = .'4 LINK TEMP (F)= 90.66JET VELOCITY (FT/S)= 1.305JET TEMP (F) = 98.2 R (FT)= 0.00 TSL (F)= 102.4 QB (BTU/FT2 S)= 0.687E-01 QT (BTU/FT2 S)- 0.847E-18 R (FT)= 12.41 TSL (F)= 89.7 QB (BTU/FT2 S)= 0.317E-01 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 24.82 TSL (F)= 84.7 QB (BTU/FT2 S)= 0.156E-01 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 37.23 TSL (F)= 82.7 QB (BTU/FT2 S)= 0.908E-02 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 49.64 TSL (F)= 81.8 QB (BTU/FT2 S)= 0.598E-02 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 62.05 TSL (F)= 81.1 QB (BTU/FT2 S)= 0.987E-03 QT (BTU/FT2 S)= 0.847E-18

TIME (S)= 120.000 LYR TEMP (F)= 111.5 LYR HT (FT)= 23.85 LYR MASS (LB)= 0.301E+04 FIRE OUTt'IIT (BTU/S)= 0.9470E+03 VENT AREA (FT2)= 0.00 LINK = 1 LINK TEMP (F)= 90.30JET VELOCITY (FT/S)= 3.614JET TEMP (F) = 129.3 LINK = 2 LINK TEMP (F)= 118.43JET VELOCrIY (FT/S)= 3.966JET TEMP (F) = 132.1 LINK = 3 LINK TEMP (F)= 96.66JET VELOCITY (FT/S)= 1.667JET TEMP (F) = 106.2 R (FT)= 0.00 TSL (F)= 115.6 QB (BTU/FT2 S)= 0A13E+00 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 12.41 TSL (F)= 96.2 QB (BTU/FT2 S)= 0.543E-01 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 24.82 TSL (F)= 87.9 QB (BTU/FT2 S)= 0.266E-01 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 37.23 TSL (F)= 84,6 QB (BTU/FT2 S)-- 0.154E-01 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 49.~4 TSL (F)= 83.0 QB (BTU/FT2 S)= 0.101E-01 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 62.05 TSL (F)= 82.0 QB (BTU/FT2 S)= 0.728E-02 QT (BTU/FT2 S)= 0.847E-18

TIME (S)= 150.000 LYR TEMP (F)= 124.4 LYR HT (FT)= 21.85 LYR MASS (LB)= 0.390E+04 FIRE OUTIq.JT (BTU/S)= 0.1479E+04 VENT AREA (FT2)= 0.00 LINK = 1 LINK TEMI' (F)= 97.16JET VELOCITY (FT/S)= 4.364JET TEMP (F) = 149.2 LINK = 2 LINK TEMP (F)= 137.37JET VELOCITY (FT/S)= 4.754JET TEMP (F) = 153.4 LINK= 3 LINKTEMP (F)= 105.49JETVELOCITY (FT/S)= 1.998JETTEMP (F) = 117.4 R (FT)= 0.00 TSL (F)= 136.5 QB (BTU/FF2 S)= 0.158E+00 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 12.41 TSL (F)= 107.0 QB (BTU/FT2 S)= 0.810E-01 QT (BTU/FT2 S)= 0.847E-18 R (FT)--- 24.82 TSL (F)= 93.3 QB (BTU/FT2 S)= 0.405E-01 QT (BTU/FT2 S)= 0.847E-18 R (FT)--- 37.23 TSL (F)= 87.7 QB (BTU/FT2 S)= 0.236E-01 QT (BTU/FT2 S)-- 0.847E-18 R (FT)= 49.64 TSL (F)= 85.1 QB (BTU/FT2 S)= 0.155E-01 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 62.05 TSL (F)= 83.5 QB (BTU/FT2 S)= 0.112E-01 QT (BTU/FT2 S)= 0.847E-18

TIME (S)= 180.000 LYRTEMP (F)= 140.2 LYR HT (FT)= 19.77 LYR MASS (LB)= 0.477E+04 FIRE OI.~TPUT (BTU/S)= 0.2012E+04 VENT AREA (FT2)= 0.00 LINK = 1 L1NK TEMP (F)= 106.66JET VELOCXI3( (FT/S)= 5.008JET TEMP (F) = 171.4 LINK = 2 LINKTEMP (F)= 159.68JET VELOCITY (FT/S)= 5.414JET TEMP (F) = 176.5 LINK = 3 LINK TEMP (F)= 116.69JET VELOCITY (FT/S)= 2.275JET TEMP (F) = 130.2 R (FT)= 0.00 TSL (F)= 160.3 QB (BTU/FT2 S)= 0.195E+00 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 12.41 TSL (F)= 120.4 QB (BTU/FT2 S)= 0.106E+00 QT (BTU/FT2 S)= 0.84TE-18 R (FT)= 24.82 TSL (F)= 100.2 QB (BTU/FT2 S)= 0.545E-01 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 37.23 TSL (F)= 91.8 QB (BTU/FT2 S)= 0.322E-01 QT (BTU/FT2 S)= 9.847E-18 R (FT)= 49.64 TSL (F)= 87.8 QB (BTIJ/FT2 S)= 0.213E-01 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 62.05 TSL (F)= 85.3 QB (BTU/FT2 S)= 0.332E-02 QT (BTU/FT2 S)= 0.847E-18

TIME (S)= 210.000 LYR TEMP (F)= 158.7 LYR HT (FT)= 19.59 LYR MASS (LB)= 0.471E+04 FIRE OI 7TPUT (BTU/S)= 0.2722E+04 VENT AREA (FT2)= 96.00 LINK= 1 LINKTEMI' (F)= l18.85JETVELOCrrY(FT/S)= 5.605JETTEMP (F) = 196.8 LINK= 2 LINKTEMP (F)= 184.03JETVELOC1TY (FT/S)= 6.021JETTEMP (F) = 202.7 LINK= 3 LINKTEMP (F)= 129.71JETVELOCITY (FT/S)= 2.530JETTEMP (F) = 144.9 TIME LINK 2 OPENS EQUALq 186.7478 (S) R (FT)= 0.00 TSL (F)= 185.7 QB (BTU/FT2 S)= 0.239E+00 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 12.41 TSL (F)= 135.8 QB (BTU/FT2 S)= 0.137E+00 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 24.82 TSL (F)= 108.5 QB (BTU/FT2 S)-- 0.718E-01 QT (BTU/FT2 S)= 0.847E-18

620 t NgPA 204M a.- A97 ROP

Table C~ (cont'd, page 3 of 4 pages) The Default-Ca~ Output.

R (FI")= 37~23TSL(F)ffi 96.8 ~ (BTU/TT2 S)= 0.4ffTE.01 QT (BTU/FI~ S)= 0.847E-18 R (FT)ffi 49.64 TSL (F)ffi 91.1 QB (BTU/FT2 S)ffi 0.285E-01 QT (BTU/TT2 S)= 0.847E-18 R (FT)= 62.05 TSL (F)ffi 87.2 QB (BTU/FI~ S)= 0.210E-01 ~ (BTU/TT2 S)= 0.847E-18

TIME (S)ffi 240.000 LYR TEMP (F)ffi 184.9 LYgHT (FT)ffi 19.77 L-YR MASS (LB)= 0.444E+04 FIRE OUTPUT (BTU/S)ffi 0.3787E+04 VENT AREA (FT2)ffi 96.00 LINK= I LINKTEMP (F)= 134.89JETVELQCITY (FT/S)ffi 6.327.]ETTEMP OF) = 231.8 LINKffi 2 LINK TEMP (F)ffi 215.00JET VELOCITY (FT/S)= 6.741JET TEMP (F) •. 238.2 LINKffi 3 LINKTEMP (F)= 146.44JETVELOC/TY (FT/S)= 2.8~.JETTEMP (F) ffi 165.1 TIME ]:.INK 2 OPENS EQUALS 186.7478 (S) R (FT)= 0.00 TSL (F)= 218.6 QB (BTU/YT2 S)ffi 0.299E+00 QT (BTU/FT2 S)ffi0.847E-18 R (FT)= 12.41 TSL (F)ffi 156.6 QB (BTU/TT2 S)ffi0.180E+00 QT (BTU/TT2 S)ffi0.847E-18 R (FT)= 24.82 TSL (F)= 119.9 QB (BTU/FT2 S)= 0.971E~1 QT (BTU/Yr2 S)ffi 0.847E-18 R (FT)= 87.23 TSL (F)ffi !03,7 ~B (BTU/Fr2 S)= 0.582E-01 ~ (BTU/TT2 S)z 0.847E-18 R (FT)ffi 49.64 TSL (F)= 95:7 QB (BTU/FT2 S)= 0.389E-0i QT (BTU/FT2 S} 0.847E-18 R. (FT)= 62.05 TSL (F)= 90.3 QB (BTU/Yr2S)= 0.288E-01 QT (nTU/yr2 s)= 0.847~18

TIME (S)= 270.000 LYRTEMP (F)ffi 217.5 LYR HT (FT)= 20.17 LYR MASS (LB)= 0.407E+04 Free OUTPUT (BTU/S)ffi- 0.4852E+04 v~rr AREA (Fr2)ffi 192.00 LINK--. I LINKTEMP (F)= 155.49JETVEL~ (FT/S)= 6.854JKr TEMP (r) ffi 271.3 LINK = 2 UNKTEMP (F)= ~3.19JET V~LOCrrY (Fr/s)= 7.244JET TEMP (F) ffi 277.0 LINK ffi ~ LINKTEMP (F)= lfi7.24JET~qELoc~Y (FT/S)ffi s.o4sjKr ~ (F) ffi 188.5 TIME LINK 2 OPENS E~JALS .186.7478(S) TIME t2NK s OPENS EQUALS ~e.9820 (s). ¢ R (FT)ffi 0.00 TSL (F)w 254.40~B (BTU/YT2 S)= 0.339E+00.QT (BTU/TT2 S)= 0.847E-18 R (F'F)= !2.41 TSL (F)ffi 181.1 QB (BTU/FT2 S)ffi0.217E+00 QT (KgU/FT2 S)ffi0.847EA8 R (FT)= 24.82TSL (F)= 133.9 QB (BTU/TT2 S)ffi0.121E+00 QT (BTU/TT2 S)= 0.847E-18 R (In')ffi 87.28 TSL (F)ffi 112.2 QB (BTU/Fr2 S)= 0.7~E-01 ~ (BTU/Fr2 S)= 0.847E-18 R (FT) ffi 49,64 TSL (F)ffi 101.5 QB (BTu/Fr2 S)- 0.494F_~1 ~r~(BTU/F~_ S)ffi0.847E.18 R (FT) ffi 62.05 TSL (F)ffi 93.7 QB (BTU/FI~ S)ffi0,371E-01 ~YI' (BTU/TT2 S)= 0.847E-18

TIME (S)ffi 30O.000 LYRTEMP (F)ffi 253.4LYRHT (Fr)ffi 2284 LYR MASS (LB)= 0.281E+04 FIRE OUTPUT (BTU/S)= 0~9|8E+04 VENT .AREA (FT2)ffi t92:00 LINK= I LINKTEMP (F)ffi 179.59JETVELOGITY (lrf/S)ffi 6.901JET TE'MP (F) = 308.7 LINK = 2 LINKTEMP (F)ffi 289.67JET VELOCrIY (FT/S)ffi 7.195JETTEMP (F) = 311.3 LINK= 3 LINKTEMP (F)= 189.77JETVELOCrIY (FT/S)ffi S~O2SJET TEMP fF) = 211.4 TIME LINK 1 OPENS EQUALS 282.8710 (S) TIME LINK 2 OPENS EQUAI.~ 186.7478 (~. TIME I~INK 3 OPENSEQUALS 266.9820 (S) R (FT)ffi 0.00 TSL (F)= 287.1 QB (BTU/YT2 S)= 0.352E+00 ~ (BTU/TT2 S)ffi0.847E-18 R (FT)= 12.41 TSL (F)= 205.50.B (BTU/FT2 S)= 0.2.~8E+00 ~ (BTU/TT2 S)ffi0.847D18 R (FT)= 24.82 TSL (F)= 148.70~ (BTU/FT2 s)ffi O.138F.~0 Qrr (BTU/FT2 S)ffi0.847E-18 R (FT~= 37.23 TSL (F)ffi 121.5 QB (BTU/I~T2 S)ffi0.851E-01 QT (BTU/TT2 S)= 0.847E-18 R (FT)= 49.64 TSL (F)= 107.8 QB (BTU/TT2 S)ffi0.574F~1 QT (BTU/FF2 S)ffi0.847E-18 R (FT)ffi 62.05 TSL (F)= 98.8 QB (BTU/FF2 S)ffi 0.428E-01 QT (BTU/TT2 S)ffi0.847E-18

TIME (~= 330.000 LYR TEMP (F)ffi "284.4 LYR lit ~T)ffi 24.25 LYR MASS (LB)ffi0.216E+04 FIRE OUTPUT (nTU/S)ffi 0.00SSE+04VENTAREA (Fr2)ffi 192.oo LINK-= 1LINKTEMP(F)ffi 206.05JETVELOCnY(FT/S)= 7A09JETTEMP(F) = 842~ LINK = 2 LINKTEMP (F)= 322.58JET VEIX)CnY (FT/S)ffi 7.227JET TEMP (F)ffi 341.6 LINK= 3LINKTEMP(F)ffi 211.77JETVELOCrIY(FT/S)= 3.0.~JETTEMP(F) ffi 231.8 TIME LINK 1 OPENS EQUALS 282.8710 (S) TIME LINK 2 OPENS EQUALS 186.7478 (S) TIME LINK 3 OPENS EQUALS 266.9820 (S) R (FT)ffi 0.00 TSL (F)ffi 316.3 QB (BTU/FT2 S)ffi 0.366E+00 QT (BTUfFT2 S)ffi0.847E-18 R (FT)= 12:41 TSL (F)= 229.1 QB (BTU/Fr2 S)ffi0.25TE+00 QT (BTU/FT2 S)ffi 0.847E-18 R (FT)= 2~i.82 TSL (F)= 163.7~QB (BTU/FT2 S)= 0.153E+00 QT (BTU/FT2 S)ffi0.847E-18 821 NFPA 204M -- A97 ROP

Table C,-4 (cont'd, page 4 of 4 pages) The Default-Case Output

R (FF)= 37.23 TSL (F)= 130.9 QB (BTU/Fr2 S)= 0.952E-01 QT (BTU/FT2 S)= 0.847E-18 R (FF)= 49.64 TSL (F)= 114.2 QB (BTU/FT2 S)= 0.644E-01 QT (BTU/FT2 S)= 0.847E-18 R (FF)= 62.05 TSL (F)= 103.0 QB (BTU/FT2 S)= 0.481E-01 QT (BTU/Fr2 S)= 0.847E-18

TIME (S)= 360.000 LYR TEMP (F)= 307.3 LYR HT (Fr)= 24.77 LYR MASS (LB)= 0.191E+04 FIRE OUTPUT (BTU/S)= 0.8048E+04 VENT AREA (FT2)= 192.00 LINK = 1 LINK TEMP (F)= 233.80JET VELOCITY (FT/S)= 7.559JETTEMP (F) = 370.4 LINK = 2 LINKTEMP (F)= 351.11JETVELOC1TY (FT/S)= 7.461JETTEMP (F) = 367.4 LINK = 3 LINKTEMP (F)= 231.51JET VELOCITY (FT/S)= 3.134JETTEMP (F) = 248.9 TIME LINK 1 OPENS EQUALS 282.8710 (S) TIME LINK 2 OPENS EQUALS 186.7478 (S) TIME LINK 3 ()PENS EQUALS 266.9820 (S) R (VF)= 0.00 TSL (F)= 344.3 QB (BTU/FT2 S)= 0.380E+00 QT (BTU/Fr2 S)= 0.847Eo18 R (FT)= 12.41 TSL (F)= 252.3 QB (BTU/FI~2 S)= 0.275E+00 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 24.82 TSL (F)= 178.8 QB (BTU/FT2 S)= 0.167E+00 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 37.23 TSL (F)= 140.5 QB (BTU/FT2 S)= 0.105E+00 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 49.64 TSL (F)= 120.8 QB (BTU/FF2 S)= 0.709E-01 QT (BTU/Fr2 S)= 0.847E-18 R (FT)= 62.05 TSL (F)= 107.5 QB (BTU/FF2 S)= 0.530~01 QT (BTU/FT2 S)= 0.847E-18

TIME (S) = 390.000 LYR TEMP (F) = 327.0 LYR HT (FT) = 24.81 LYR MASS (LB)= 0.185E+04 FIRE OUTr'I_~T (BTU/S)= 0.9113E+04 VENT AREA (FT2)= 192.00 LINK = 1 LINKTEMP (F)= 262.32JET VELOCITY (FT/S)= 8.168JET TEMP (F) = 397.0 LINK = 2 LINKTEMP (F)= 376.92JETVELOCITY (FT/S)= 7.811 JET TEMP (F) = 392.0 LINK = 3 LINKTEMP (F)= 249.19JETVELOC1TY (FT/S)= 3.281JETTEMP (F) = 264.9 TIME LINK 1 OPENS EQUALS 282.8710 (S) TIME LINK 2 OPENS EQUALS 186.7478 (S) TIME LINK 3 ()PENS EQUALS 266.9820 (S) R (FT)= 0.00 TSL (F)= 372.0 QB (BTU/FT2 S)= 0.398E+00 QT (BTU/FT2 S)= 0.847G18 R (FT)= 12.41 TSL (F)= 275.6 QB (BTU/FT2 S)= 0.294E+00 QT (BTU/FF2 S)= 0.847E-18 R (FT)= 24.82 TSL (F)= 194.1 QB (BTU/FT2 S)= 0.181E+00 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 37.23 TSL (F)= 150.3 QB (BTU/Fr2 S)= 0.114E+00 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 49.64 TSL (F)= 127.5 QB (BTU/FT2 S)= 0.773E-01 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 62.05 TSL (F)= 113.2 QB (BTU/VI'2 S)= 0.574E-01 QT (BTU/FT2 S)= 0.847E-18

TIME (S)= 400.000 LYR TEMP (F)= 333.5 LYR HT (FT)= 24.77 LYR MASS (LB)= 0.185E+04 FIRE OUTPUT (BTU/S)= 0.9468E+04 VENT AREA (FF2)= 192.00 LINK = 1 LINK TEMP (F)= 271.98JET VELOCITY (FT/S)= 8.387JETTEMP (F) = 406.0 LINK = 2 LINK TEMP (F)= 385.32JET VELOCITY (FT/S)= 7.936JETTEMP (F) = 400.2 LINK = 3 LINK TEMP (F)= 254.85JET VELOCITY (Fr/s)= 3.333JET TEMP (F) = 270.2 TIME LINK 1 OPENS EQUALS 282.8710 (S) TIME LINK 2 ()PENS EQUALq 186.7478 (S) TIME LINK 3 OPENS EQUALS 266.9820 (S) R (FT)= 0.00 TSL (F)= 381.3 QB (BTU/FT2 S)= 0.403E+00 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 12.41 TSL (F)= 283.5 QB (BTU/FT2 S)= 0.300E÷00 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 24.82 TSL (F)= 199.2 QB (BTU/FT2 S)= 0.186E+00 QT (BTU/FT2 S)= 0.847E-18 R (FT)= 37.23 TSL (F)= 153.6 QB (BTU/FF2 S)= 0.117E+00 QT (BTU/FT2 S)= 0.847F~18 R (FF)= 4(L64 TSL (F)= 129.7 QB (BTU/FT2 S)= 0.794E-01 QT (BTU/FT2 S)= 0.847E-18 R (Fr)= 62.05 TSL (F)= 115.0 QB (BTU/Fr2 S)= 0.589E-01 QT (BTU/Fr2 S)= 0.847E-18

622 NFPA 204M ~ A97 ROP

Option 6, HEIGHT TO BOTTOM OF CURTAIN, of the room The user may now continue to modify or add additional ceiling roperties menu is used to define the height above file floor of the vents or return to the room-properties menu by entering 0 [ret]. If ~ ottom of tile curt;tin. /~.s can be seen, in the default data, this is 15 the user tries to associate a vent with a link not yet entered in the ft. Where this height is chosen to be identical to the ceiling heigilt, rogram, the code will warn the user, give the maximum number of the user should always define the very special idealized simulation ~inks available in file present data set, and request a new link value. ,associated with an extensive, unconfined ceiling fire scenario (i.e., If the user deletes a link that is assigned to a vent, the code will by whatever means, it is assumed that file flow of the ceiling jet is assign the link with the next smallest number to that veal The best extracted from tile compartment at tile extremities of the ceiling). method for assigning vents to links is to first use Option 4 FUSIBLE Under such a simulatiou, an upper layer never develops in file LINK PROPERTIES of the base menu (to be discussed in C-5.5) to compartment. The lower ceiling surface and fllsible links are assign the link parameters and then to use Option 1 ROOM submerged in and respond to an unconfined ceilingjet environ- PROPERTIES followed by the NUMBER OF VENTS, ETC. option to ment, which is unaffected by layer growth. This idealized fire assign vent properties. scenario, involving the unconfined ceiling, is used, for example, in [2] to simulate ceiling response and in [3] and [4] to simulate Now return to tile room-properties menu by entering 0 [ret], and sprinkler response. then to the base menu by entering O [ret] again. The choice of some oF, tions on a menu, such as Option 4, With the base menu back on the screen, choose Option 2 PHYSI- NUMBER OF VENTS, ETC. of the room properties menu, will lead CAL PROPERTIES to review and/or modify the detauh room- to a subsequent display/requirement of additional associated input property input data. Enter 2 [ret]. data. Menu options that necessitate multiple entries are indicated by the use of "ETC." In the c~xse of Option 4, NUMBER OF' VENTS, C-5.3 Physical Properties. When Option 2 PHYSICAL PROPER- ETC., three values are illvolved for each vent or group of vents TIES of the base menu is chosen, the following physical properties actuated by a fi~sible link. As indicated under Option 4, NUMBER menu is displayed: OF VENTS, ETC., the default data describe a scenario with two vents or groups of vents. MATERIAL = INSULATED DECK (SOLID POLYSTYRENE) Now choose Option 4, NUMBER OF VENTS, ETC., to review and HEAT CONDUCTIVITY = 2.400E-05 (BTU/S LB F) modify the defauh input data ,xssociated with these two vents or HEAT CAPACITY = 2.770E-01 (BTU/LB F) groups of vents. Enter ,t [ret]. The following is displayed on the DENSITY = 6.550E+01 (LB/Fr3) screen: 1 80.00000 AMBIENT TEMPERATURE (F) VENT NO. = 1 FUSIBLE LINK = 2 VENT AREA = 96.00000 FF2 2 0.50000 MATERIAL THICKNESS (t'T) VENT NO. = 2 FUSIBLE LINK = 3 VENT AREA = 96.00000 FF 3 MATERIAL = INSULATED DECK (SOLID POLYSTYRENE) 0 CHANGE NOTHING ENTER 6 TO REMOVE A VENT ENTER VENT NO., LINK NO., AND VENT AREA (FT2) TO The values in Options I and 2 are modified by entering the option ADD OR MODIFY A VENT number and then the new value. MAXIMUM NO. OF VENTS IS 5 ENTER 0 TO RETUPd~I TO THE MENU Now choose Option 3. Enter 3 [ret]. The following menu is displayed: This dispkty indicates that the two simulated vents or group-of-vents are numbered 1 (VENrl ' NO. = 1 ) and 2 (VENT NO. ffi 2), that these 1 CONCRETE are actuated by filsible links numbered 2 (FUSIBLE LINK = 2) and 3 2 BARE METAL DECK (FUSIBLE LINK = 3), respectively, and fllat each of the two vents or 3 INSULATED DECK (SOLID POLYSTYRENE) ~_TOt;.ps-of-vents have a total area of 96 ft" (VENT AREA = 96.00000 4 WOOD 5 OTHER

In tile default fire scenario it would be of interest to study the e~ect By choosing one of Options 1 through 4 of this menu, the user of g~mging the operation of all of tile four vents (total area 192 ft~) specifies the material properties of the ceiling according to the table to fusing of the closest vent link. To do so it would be necessary to of standard material properties in [5]. When the option number of first remove velar nmnber 2, as identified in the above menu, and one of these materials is chosen, file material name, thermal then to modify the area of vent number I; conductivity, heat capacity, and density are displayed on the screen as part of an updatedphysical properties menu. To remove vent rmmber 2 enter 6 [ret]. Tile following is now displayed on the screen: Now choose Option 5 OTHER. Enter 5 [ret]. The following screen is displayed: ENTER NUMBER OF VENT TO BE ELIMINATED ENTER 0 TO RETURN TO MENU ENTER MATERIAL NAME THERMAL CONDUCTIVITY (BTU/S FT F) Now enter 2 [ret]. This completes removal of vent 2, with the HEAT CAPACITY (BTU/LB F) following revised display on the screen: DENSITY (LB/FT3)

VENT NO. = ! FUSt[BLE LINK = 2 VENT AREA = 96.00000 FI'2 The four indicated inputs are required. After these are entered, the screen returns to an updated physical properties menu. ENTER 6 TO REMOVE A VENT ENTER VENT NO., LINK NO., AND VENT AREA (FT2) TO Now return to the default material, INSULATED DECK (SOLID ADD OR MODIFY A VENT POLYSTYRENE). To do so enter any arbitrary material name with MAXIMUM NO. OF VENTS IS 5 any threepropertyvalues (enter MATERIAL [ret], 1. [ret], 1., [ret], ENTER 0 TO RETURN TO THE MENU 1. [ret]); then choose Option $ MATERIAL from the menu displayed (enter 3 [ret]); and, from the final menu displayed, Now modify tile characteristic~ of vent number 1. To do this enter choose Option $ INSULATED DECK (SOLID POI.,YSTYRENE) 1 [ret], 2 [ret], 192. [ret]. The screen will now display: (enter 3 [ret]).

VENT NO. = 1 FUSIBLE LINK = 2 VENT AREA = 192.00000 FI'2 Nowreturn to the base menu. Enter 0 [ret]. Choose Option 3 OUTPUT PARAMETERS of the base menu to review and/or modify ENTER 6 TO REMOVE A VENT the default output-parameter data. Eater 3 [ret]. ENTER VENT NO., LINK NO., AND VENT AREA (FT2) TO ADD OR MODIFY A VENT G5.4 Output Parameters. When Option 3 OUTPUT PARAM- MAXIMUM NO. OF VENTS IS 5 ETER.S of the base menu is chosen, the following output-parameters ENTER 0 TO RETURN TO THE MENU menu is displayed:

To add or reim~lemecjt vent nnmlxr 2, actuated by link number 3, I 400.000000 FINAL TIME (S) and of area 96 ft-. enter 2 [ret], 3[retl, 96. [ret]. Now return to the 2 30.000000 OUTPUT INTERVAL (S) original defauh scedlario by bringing tile area of vent number 1 back 0 CHANGE NOTH][NG to its original 96 f~; valtle: enter I {ret], 2 [ret], and 96. [ret].

623 NFPA 204M ~ A97 ROP

The FINAL TIME represents tile ending time of tile calculation. Now remove link number 2 to return to the original default array The OI_ITPUT INTERVAL controls tile time interval between of links. To do so enter 11 [ret]. The following screen is displayed: successive outputs of tile calculation results. All times are in seconds. For example, assume that it is desired to run a fire scenario ENTER THE NUMBER OF THE LINK TO BE REMOVED for 500 s with an output of results each l0 s. Then first choose Option 1 wifll a ~tlue of 500 (enter l [ret], 500. [ret]), and then Enter 2 [ret] to remove link 2. Option 2 widl a value of 10. (enter 2 [ret], 10 [ret]). The following revised output-parameters meru! is displayed: Now return to the base menu from the fusible-link-properties menu by entering 0 [ret]. 1 500.000000 FINALTIME (S) 2 10.000000 OUTPUT INTERVAL (S) With the base menu back on die screen, choose Option 5 FIRE 0 CHANGE NOTHING PROPERTIES to review and/or modify the default fire-properties data. Enter 5 [ret]. Return to tile original default output-par,'mleters menu by entering 1 [ret], 400. [ret], followed by 2 [ret], 30. [ret]. C-5.6 Fire Properties. When Option 5 FIRE PROPERTIES from the base menu is chosen, the following fire-properties menu is displayed: Now return to the hase illenu from die output-parameters menu by entering 0 [ret]. 1 2.5 FIRE HEIGHT (FT) 2 330.0 FIRE POWER/AREA (BTU/S ET2), ETC. Widl the base menu back ou the screen, choose Option 4 FUSIBLE 3 FIRE OUTPUT AS A FUNCTION OF TIME LINK PROPERTIES to review and/o l" modify the default fusible-link- 0 CHANGE NOTHING properties data. Enter 4 [ret]. Tile value associated with Option 1 is the height of the base of the C-5.5 Fusible Link Properti~. Wheu Option 4 FUSIBLE LINK fire above the floor. Change this to 3 ft, for example, by entering l PROPERTIES of tile base inenu is chosen, the following fusible-link- [ret] and 3. [ret]. Then return to the default data by entering 1 properties menu is displayed: [ret] and 2.5 [ret].

The value associated with Option 2 is the fire-energy-release rate-per- TO ADD OR CHANGE A LINK, fire area. It is also possible to consider simulations where die fire ENTER LINK NO., RADIUS (FF), area is fixed by specifying a fixed fire diameter. The fire-energy- DISTANCE BELOW CEILING (Fr), release rate-per-fire area can be changed, or the fixed fire area-type RTI (SQRTIVI" SI), AND FUSE TEMPERATURE (F). of specification can he made by choosing Option 2. To do this enter MAXIMUM NUMBER OF LINKS EQUAL 10. 2 [ret]. This leads to a display of the following menu: ENTER 1 ! TO REMOVE A LINK. ENTER 0 TO RETURN TO THE MENU. 1 WOOD PALLETS,STACK, 5 F r HIGH 350 (BTU/S ET2) LINK# RADIUS D1STANCE (FI') RTI SQRT FUSE 2 CARTONS, COMPARTMENTED, STACKED 15 1¢1"HIGH (Fr) BELOW fiTS) TEMP 200 (BTU/S FT2) CEILING (F) 3 PE BO'Iq'LF~ IN COMPAR'IMENTED CARTONS 15 FF HIGH I 6.000 1.000 400.000 165.000 540 (BTU/S FT2) 2 21.200 0.300 50.000 165.000 4 PSJARS IN COMPARTMENTED CARTONS 15 I;T HIGH 3 44.300 0.300 50.000 165.000 1300 (BTU/S FT2) 5 GASOLINE 200 (BTU/S Fr2)

Each filsible link must be ,'t~sigrjed a link number (e.g., LINK # = 6 INPUT YOUR OWN VALUE IN (BTU/S ET2) 1), radial position from the phune-ceiling impingement point (e.g., 7 SPECIFY A CONSTANT DIAMETER FIRE IN FT RADIUS = 6.00 FF), ceiling-to-link separation distance (e.g., 0 CHANGE NOTHING DISTANCE BELOW CEILING = 1.00 vr), response-time-index (e.g., RTI = 400.00 SQRT[FT S]), and fixse temperature (e.g., FUSE TEMPERATURE = 165.00 F). Options 1 through 5 of the above menu are for variable-area fires. The Option 1-to-5 constants displayed above on the right are the Suppose that in tile default fire scenario it was desired to simulate fire-energy-release rate-per-unit fire area- Theyare taken from Table the thermal response of the group of (fou r) sprinkler links second 4.1 of [1 ]. ff one of these options is chosen, an appropriately- closest to tile fire. According to tile description of G-3 and Figure G- updated fire-properties menu is then displayed on the screen. 3(a), this would be done hy adding a fonrdl link, link number 4, at a Option 0 would lead to file return of the original fire-properties radi~ ¢J~tance of ! 3.4 It, I ftbel ow the ceiling, with an RTI of 400 menu. (ft-s) ~/~and a fnsion temperatore of 165°F. To do this enter 4 [ret], 13.4 [retl, 1. [ret], 400. [ret], 165. [ret]. Then the following Option 6 allows any other fire-energy-release rate-per-unit fire area screen is displayed: of the user's choice.

Option 7 allows the user to specify the diameter of a constant-area TO ADD OR CHANGE A LINK, fire instead ofa energy-release-rate-per-unit-area fire. ENTER LINK NO., RADIUS (FT), DISTANCE BELOW CEILING (Fr), Choice of Option 6 or 7 must be followed by entry of the appropri- RTI (SQRT[ Fr s]), AND FUSE TEMPERATURE (F). ate value. Then an appropriately updated fire-properties menu MAXIMUM NUMBER OF LINKS EQUAL I 0. appears on the screen. ENTER 11 TO REMOVE A LINK. ENTER 0 TO RETURN TO THE MENU. To try Option 7 SPECIFY A CONSTANT DIAMETER FIRE IN FEET, enter 7 [ret]. The following screen is displayed: LINK# RADIUS DISTANCE (FT) RTI SQRT FUSE (FF) BELOW (FI~S) TEMP ENTER YOUR VALUE FOR FIRE DIAMETER IN FT CEILING (F) 1 6.000 1.000 400.000 165.000 Assume the fire diameter is fixed at 5 ft. Enter 5. [ret]. Then the 2 13.400 1.000 400.000 165.000 following screen is displayed: 3 21.200 0.300 50.000 165.000 4 44.300 0.300 50.000 165.000 1 2.50000 FIRE HEIGHT (ET) 2 5.00000 FIRE DIAMETER (ET), ETC. Note that tile new link, which was entered as link number 4, was 3 FIRE OUTPUT AS A FUNCTION OF TIME sorted automatic~ally into tile list of tile original three links and that 0 CHANGE NOTHING all four links were renumbered according to radial distance from the fire. The original link-vent ,x~slgnments are preserved in dais Now return to the original default fire-properties menu. Enter 2 operation. Hence, the user need not return to Option 4 NUMBER [ret]. The previous menu will be displayed. In this, choose Option OF VENTS, ETC., unle~ it is desired to reassign link-vent combina- 1 WOOD PALLETS, etc. by entering 1 [ret]. tions. Option ~, FIRE OUTPUT AS A FUNCTION OF TIME of the fire- A MAXIMUM OF 10 LINK RESPONSF~S (2a2q BE SIMULATED IN ANY ONE SIMI.ILATI()N. 624 NFPA 204M ~ A97 ROP properties menu allows tile user to prescribe the fire as a function of Start the input part of the program get to the base menu. Then time. Tile prescription involves 1) linear interpolation between choose Option 6 SOLVER PARAMETERS. Enter 6 [ret]. The adjacent pairs of user-specified points with coordinates (time in s, following input options menu will be displayed: fire-energy-release rate in BTU/s), and 2) continuation of the fire to arlfitrarily large time at the fire-energy-release rate of the last clata I 0.6500E+00 GAUSS-SEIDEL RELAXATION point. 2 0.1000E-04 DIFF EQ SOLVER TOLERANCE 3 0.1000E-04 GAUSS-SEIDEL TOLERANCE Now choose Option 3 by entering 3 [ret]. Tile following screen 4 2.000000 FLUX UPDATE INTERVAL (S) associated with tile default fire-output data is displayed: 5 6 NUMBER OF CEILING GRID POINTS, MIN=2, MAX--50 6 0.1000E-07 SMALLEST MEANINGFUL VALUE I TIME(s) = 0.0O00 POWER(BTU/S) = 0.00O00E+00 7 CHANGE NOTHING 2 TIME(s) = I00.0000 POWER(BTU/S) = 0.59200E+03 3 TIME(s) = 200.0000 POWER(BTU/S) = 0.23670E+04 Tile solvers used in this code consist of a differential equation solver 4 TIME(s) = 400.0000 POWER(BTU/S) = 0.94680E+04 DDRIVE2, used to solve the set of differential equations associated 5 TIME(s) = 600.0000 POWER(BTU/S) = 0.21302E+05 with tile layer and the fusible links, and a Gauss-Seidel/Tridiagonal 6 TIME(s) = 747.0000 POWER(BTU/S) = 0.33000E+05 solver using the Crank-Nicolson formalism to solve tile set of partial differential equations associated with tile heat conduction calcula- ENTER DATA PO][NT NO., TIME (S), AND POWER (BTU/S) tion for the ceiling. Since two different solvers are being used in file ENTER 11 TO REMOVEA POINT code, there is potential for the solvers to become incompatible with ENTER 0 TO RETURN TO MENU each other, particularly if the upper layer has nearly reached a steady-state temperature but the ceiling is still increasing it's As discussed in G-3, with use of tile six above data points, the temperature. When this occurs, the differential equation solver will default simulation will egtimate the fire's energy-release-rate try to take time steps that are too large for the Gauss-Seidel solver to according to the plot of Figure C-3(b). handle and a growing oscillation in the ceiling temperature variable may occur. By reducing tile FLUX UPDATE INTERVAL, the Additional data points can be added to the fire-growth simulation OWing oscillation may be suppressed. The smaller the FLUX by entering the new data-point number, [ret], the time in seconds, DATE INTERVAL, the slower tile code will run. [ret], the energy-relea.se-rate in BTU/s, and [ret]. The GAUSS-SEIDEL RELAXATION coefficient may be changed to The maximum number of data points perufitted is 10. The points produce a faster running code or to handle a case that will not run may be entered in any order. A sorting routine will order the points with a different coefficient. Typical values of this coefficient should by time. One point must correspond to zero time. range between 0.2 and 1.0.

As an example of adding an additional clara point to the above six, The DIFF EQSOLVER TOLERANCE and the GAUSS-SEIDEL assume that a closer match to the "t-squared" default fire-growth TOLERANCE may also be changed. Decreasing or increasing these curve was d~sired between 200 s and 400 s. From Section 2 it ean be values may provide a faster running code for a given case and by verified that the fire energy-relea.se rate will be 5325 BTU/s at t = decreasing the value of the tolerances, the accuracy of the calcula- 300. To add this point to the data, thereby forcing the fire-growth tions may be increased. If the tolerance values are made too small, curve to pa.ss exactly through the "t-squared" curve at 300 s, enter 7 the code will either run very slowly or not run at all. Suggested [ret], 300. [ret], and 5325. [ret]. Tile following revised screen will tolerances would be in the range of 0.00001 to 0.0(~)001. be displayed: Consistent with file model assumptions, accuracy in the radial 1 TIME(s) = 0.000O POWER(BTU/S) = 0.00000E+00 ceiling temperature distribution around file plume/ceiling 2 TIME(s) = 100.0000 POWER(BTU/S) = 0.59200E+03 impingement point is dependent on tile NUMBER OF CEILING 3 TIME(s) = 200.0000 POWER(BTU/S) = 0.23670E+04 GRIDPOINTS. Relatively greater/lesser accuracy is achieved by 4 TIME(s) = 300.0000 POWER(BTU/S) = 0.53250E+04 using relatively more/fewer grid points. This leads, in turn, to a 5 TIME(s) = 400.0000 POWER(BTU/S) = 0.94680E+04 relatively slower/faster computer run. 6 TIME(s) = 600.~0N0 POWER(BTU/S) = 0.21302E+05 7 TIME(s) = 747.1)000 POWER(BTU/S) = 0.33000E+05 06 File Status - Running the Code. When Option 0 NO CHANGES of the base menu is chosen, the following file-status menu is ENTER DATA FT. NO., TIME (S), AND POWER (BTU/S) displayed: ENTER I I TO REMOVE A POINT 1 SAVE THE FILE AND RUN THE CODE ENTER fl TO RETURN TO MENU 2 SAVE THE FILE BUT DON'T RUN THE CODE 3 DON'T SAVE THE FILE BUT RUN THE CODE Note that the revised point, which was entered as point number 7, 4 ABORT THE CALCUlaTION has been resorted into the original array of data points and that all points have been renumbered appropriately. If one of the save options is selected, tile user will be asked to supply a file name to designate the file where the newly generated Now remove thepointjust added (wltich is now point number 4). input data is to be saved. Tile program will automatically create the First enter 11 [ret]. Then the following screen is displayed: new file. File names may be as long as 8 characters and should have a common extender such as .DAT, example MYFILE.DAT. The ENTER THE NUMBER OF THE DATA POINT TO BE RE- maximum len. ~ll that may be used for the total length of input or MOVED output files ~s 25 characters. For example, G:KSUBDIREGTWILENAME.DAT would allow a file named Now enter 4 [ret]. This brings the fire-growth-simulation cla.ta back FILENAME.DAT to be read from the subdirectory SUBDIRECT on to the origin,'d default set of values. file C drive. To read a file from a floppy disk in the A drive, use A:FILENAME.DAT. If Option 4 is chosen the program will end Now return to the fire.properties menu. Enter 0 [ret]. Then without anyfile being saved. return to the b;Lse menu by entering again 0 [ret]. A request for an output file name will appear on the screen. F'de With the base menu b:tck on the screen, it is assumed that imputing names may be as long as 8 characters and should have an extender of all data required to define the desired fire simulation is complete. such as .OUT such that tile output files can easily be recognized. To Now choose Option 0 NO CHANCES to proceed to the file-status output a file to a floppy disk in the A drive, name the file menu. Enter 0 [retl. AcFILENAMF_.OUT. To output a file to a subdirectory other than tile one which is resident to the program, use 05.7 Solver Paramete1,~. Users of the code will generally have no C:\$UBDIRECT~FILENAME.OUT for the subdirectory SUBDIRECT. need to refer to this section (i.e., especially when learning to use tile LAVENT code a user should now skip to G-6) since they are rarely, if Once tile output file has been designated, the program will begin ever, expected to run into a situation wlaere the code is not able to to execute. Tile statement PROGRAM RUNNING will appear on obutin a solution for a particular application or is taking an the screen. Each time the program writes to tile output file, a inordinate amount of time to produce tile solution. However, if this statement such as T = 3.0000E01 S will appear on the screen to does happen, there are a number of variations of the default solver provide the user with tile present output tame. parameter inputs which may resolve the problem. 07 The Output Varlables and the Output Options. The program

625 NFPA 204M -- A97 ROP

enerates two separate output files. An exarnple of die first output ~ le is appended :it tile end of tids document. This file is named by die user and consists of a listing of die input data plus all die 30 relevant output ~triables in a format where die output units are specified and the meaning of all but d/ree of tile output variables are clearly specified. These latter variables are TSL, QB, and QT which 29 . are the temperature of tile ceiling inside the enclosure, die- net heat transfer flux to die bottom surface of tile ceiling, and tile net heat transfer flux to tile top snrface of the ceiling. The variables are 28 output as a function of radius with R = 0 being tile center of tile fire plume projected on die ceiling. ()tiler abbreviations include LYR TEMP, LYR HT, LYR MASS,JET VELOCITY, and JET TEMP which 27 are die upper layer (layer adjacent to the ceiling) temperature, height of die upper layer interface above tile floor, mass of gas in die layer, ceiling.jet velocity and temperature at the position of each 26 fusible link. The VENT AREA is the total area of roof vents open at tile time of output. v ~ 25 The second olltpnt file, GRAPH.OUT, is used by die graphics program, GP, APH. GRAPH is a Fortran program which makes use of a graphics software package to produce graphical output of selected output variables.[6, 7] To use the graphics program, the file GRAPH.OUT must be in tile same directory as the program, GRAPH. GRAPH is a menu driven program which provides the user 23 widl tile ability to plot two sere of variables on the PC screen. An option exists which permits the user to print die plots from tile screen to a printer. If the user has all attached EPSON-compatible 22 printer, enter 'e' to produce a plot using tile printer. If the user wishes to generate a PostScript file for use on a laser printer, enter 'p' and provide a file name when the file name prompt appem's in 21 the upper left hand comer of the graph. To exit to screen mode from the graphics mode, enter 'c'. The file GRAPH.OUT will be destroyed each time the code LAVENT is run. If the user wishes to 2O save die graphics file, it must be copied using the DOS copy command into another file with a different file name. I I I I l i To demonstrate the use of GRAPH, start dae program by entering 0 1 O0 200 300 400 'graph' [ret]. GRAPH will read in dae graphics output file "rime Is) GRAPH.OUT and tile following screen will be displayed:

ENTER 0 TO PLOT POINTS, ENTER 1 TO PLOT AND Figure C~7(a) Plot of the height of the smoke layer ~aterface vs dme CONNECT POINTS for the default simulation.

The graphics presented in Figures C-7(a) to C-7(e) were done wida GRAPH using option 0. Enter 0 [ret] and die following graphics 340 I I I I I I I menu is displayed:

ENTER THE X AND Y VARIABLES FOR THE 320 DESIRED TWO GRAPHS 1 TIME 300 2 LAYER TEMPERATURE 3 LAYER HEIGHT 4 LAYER MASS 280 5 FIRE OUTPUT 6 CEILING VENT AREA 7 PLUME FLOW 260 8 LINK TEMPERATURE 9 JET VELOCITY AT LINK I 0 JET TEMPERATURE AT LINK 240 o v Two plots can be studied on a single screen. For example, from tile 220 default simulation ,x~sume that displays of tile plots of Figure C-7(a) ~- and G-7(b), LAYER HEIGHT vs TIME and LAYER TEMPERATURE vs TIME, respectively, are desired. Then enter 1 [ret], 3 [ret], 1 [retl, and 2 [ret]. The program will respond widl tile prompt:

ENTER THE TITLES FOR THE TWO GRAPHS, 16 CHARACTERS MAX. ,.J

The user might choose titles which would identify particular cases 160 such ,as LY HT RUN 100 [ret] and LY TEMP RUN 100 [ret]. If die fide is chosen to be longer dlan 16 characters, it will be truncated to 140 16 characters. After the titles have been entered the program will respond widl: 120 ENTER 1 FOR DEFAULT SCALING, 2 FOR USER SCALING.

100

80 I i i , , I 0 100 200 300 400 Time (s)

Figure C-7(b) Plot of the temperature of the smoke layer vs time for the default simulation. 626 NFPA 204M w A97 ROP

400 , , , , w , , 260

240 350

220 300

200 o v E sso E"160 E 2oo 160

.I¢ .G ..J 150 140

120 100

100

50 | Sl i t | | t 0 1 O0 200 300 400 Time (s) 80 • I I I I I i i 0 100 200 300 400 F'~;ure C-7(c) Plot of the closest (R = 21.2 ft) vent-link temperatures Time (s) vs time for the default simulation. Figure C-7(e) Plot of the closest (R = 6 ft) sprinkler-link tempera- 260 tures vs time for the default simulation. If the user chooses option 1, tile desired plots will appear on the screen with an internal scaling for the X and Y axis of each graph. If 260 the user chooses option 2, the program will respond with the following prompt:

ENTER THE MINIMUM AND MAXIMUM VALUES FOR THE X 240 AND Y AXIS OF EACH GRAPH. ENTER 0 FOR THE MINIMUM AND MAXIMUM VALUES OF EACH AXIS WHERE DEFAULT SCALING IS DESIRED. FOR EXAMPLE, VALUES SHOULD BE ENTERED AS 220 0.,100.,0.,200.,10.,50.,20.,100. [RET] FOR XI (0-100), Yl(0-200), X2(10-50), Y2(20-100).

E" 200 Use of this option allows a number of different cases to be compared o v usingsimilar values for the X and Y axis of each graph. All eight numbers must be entered and separated with commas before -I entering [ret]. Once the entry is made, the plots will appear on the 160 screen. Note that this option permits a mixture of default scaling and user specified scaling.

~ lOO Once a pair of plots are displayed on the screen, the user would have the choice of entering "p' or 'e', to obtain a hard-copy plot of -J the graphs, or of entering 'c' to exit the graphics mode.

140 To plot a second pair of graphs, the user would exit the graphics mode by entering 'c' and then repeat the above process by entering 'graph' [ret], etc.

120 D If the user selects plots which involve variables defined by Options 8, 9, or 10, then, following the entry 8 [ret], 9 [ret] or 10 [ret], the following prompt for identifying the desired link number (in the 100 default simulation with 3 simulated links) will be displayed immedi- ately: ENTER LINK NUMBER, MAXIMUM NUMBER = 3 80 " le i i i i i i 0 1 00 200 300 400 The user then enters the desired link number followed by [ret], and Time (s) continues entering the remaining input data which define the desired plots. Figure C-7(d) Plot of the far (R = 44.3 It) pair of vent-link tempera- tures vs time for the default simulation.

627 NFPA 204M i A97 ROP

As an example of generativg link-related plots, consider displaying Appendix D Sample Problem Using Engineering Equations (Hand tile pair of plots LINK TEMI'EP~TURE vs TIME and JET VELOC- Calculations) and LAVENT ITYAT LINK~ TIME for lirtk number 3 in the default simulationo First enter 1 [ret] (for TIME on file X axis) and 8 [ret](for LINK Abstract TEMPERATURE on the Y axis). At tiffs point, "ENTER LINK NI.IMBER ..." would be displayed on the screen. Continue by The following example problem illustrates the use of the informa- entering 3 [ret] (fl)r link number 3). This would complete the data tion, engineering equations, hand calculations and computer model entry for file first of the two plots. For the second plot enter 1 [ret] described in this document. The impact of a fire on a non- (for TIME on the X axis) and 9 [ret] (for LINK TEMPERATURE on sprinklered retail storage building and its occupants is assessed. The the Y axis). At this point, "ENTER LINK NUMBER ..." would be effects of an anticipated fire on the subject building are predicted, displayed a second time. Then conch*de data input for tile pair of and the impact of smoke and beat vents are illustrated. plots by entering 3 [ret] (for link number 3). At this point the desired pair of plots would be displayed on the screen. Design goals and objectives were developed and a high challenge fire, likely to occur in the subject building, was identified. C-8 An Example Simulation - The Defauh Case. This section presents and reviews briefly the simulation of tim default case. The fire impact was assessed using d~ree different methods: • Hand calculations assuming a quasi-steady fire The tabular output of the default simulation is presented in Table • Hand calculations assuming a continuous growth (t-squared) G-4. Plots of tile layer-interface height and of the layer temperature fire as fimctions of time are plotted in Figures C-7(a) and C-7(b), • Tile computer model LAVENT respectively. Plots of the thermal resl)onse of the two pairs of vent litlKS and tile pair nf sprinkler links closest to die fire are presented Hand calculations are useful for quick estimates of the impact of in Figures C-7(c) to C7(e), respeclively. vents on fire effects. However, hand calculations are not able to assess time-varying events. A number of simplifying assumptions From Table (i;-4 and Figures C-7(c) to C7(e) it is seen that the have been used to facilitate problem-solving via algebraic equations. sequence of link fi~sing (at 165 F) is predicted to be the near pair of Hand-calculated results are considered valid, but produce somewhat vents at 187 s, tile l.u" pair of vents at 267 s, ztnd tile pair of closest conservative estimates of fire effects such as upper layer tempera- sprinklers at 283 s. Although tile sprinkler links are closer to the ture. A computer model, like LAVENT, will generally provide a fire than any of the watt links, and although all links bave the same more complete analysis of the fire-produced effects and, in some flise temperatures, the sinullatiou predicts that the sprinkler links instances, is preferable over hand calculations. fi~se after all of the veto links. There are two reasons for this. First, the RTI of tile sprinkler links are larger than those of the vent links Introduction and, therefore, slower to respond thermally. Second, tile two sprinkler links simulated :ire far etmugh from tile ceiling ,as to be The following example problem illustrates the use of engineering below tile peak temperature of the ceiling jet which is relatively thin equations and a computer model to assess the impact of a fire in a at the 6 ft radial position (see tile lower sketch of Figure C~2). nonsprinklered retail-storage building. The problem illustrates the impact of vents and predicts the effect of the anticipated fire on the The effect on layer growth of fusing of die two pair of vent links building. and opening of their correspondittg vents at 187 s and 267 s can be noted in Figure C7(a). Note that tile opening of the first pair of Goal vents effectively stops tile rate-nf-increase of layer thickness and opening of the second pair of vertts leads to a relatively rapid rate-of- Develop a vent design for the subject building which will maintain a decrease in tile layer thickness. All of this is of course occurring at tenable environment for a period of time at least equal to the time times when tile energy-release-rate of the fire is growing rapidly. required to evacuate the building, and to maintain the hot upper layer a minimum of 3 meters above floor level until the local .&s can be seen it* Figure G-7(a), lip to tile 400 s of simulation time Fire Department enters the building. tile smnke is still contained in the original curtained compartment and h:ts not "spilled over" to ac!jaceut spaces. From this figure it Objective appears that with no venting, the layer would have dropped below tile bottom of the curtain boards prior to fitsing of the first sprinkler Determine the vent area required to maintain the smoke layer at lin "1"1~. This could be confirmed with a second simulation run of least $ meters above floor level for 300 seconds following detection LAVENT, where :ill refit action w:ts removed from the default data. of the fire by an automatic detection syste~2n. Also, limit the heat flux at floor level to a maximum of 2.5 kW/m , the threshold irradiance C-9 References for Appendix C. causing severe pain to exposed skin [ 1 ], during the time required for evacuation of the building occupant,s. 1. GuideJbrSmoke and t-k~,~t Venting, NFPA 204M, National Fire Protection ~ssocialion, Qoirtcy MA, 1982. Building Details

2. Cooper LX. and Stroup, D.W., "Thermal Response of The building is 73 m wide, 73 m long, and is 9.1 m high. The l.]ncnnfined Ceili~,gs Above (;rowing Fires and the building is not subdivided nor is itprovided with a sprinkler system. Importance of (;onvective Heat Transfer,"Journal of Heat The roof is an insulated deck (solidpolystyrene). A complete fire Tran.~m 109, pp. 172-178, 1987. alarm system is to be installed using heat detectors spaced 15.2 m on center and 6.1 m from walls. ])t~ectors have an activation tempera- 3. Evans. D.D., "'Cah:ulating Sprinkler Activation Time in ture of 74 o C, RTI of 55 (m • s) */ , and are located 0.3 m below the Compartments," Fire St~e~..[ournal, 9, pp. 147-155, 1985. roof. Sixteen vents ,are proposed, with vents spaced 18.3 m on center. Vents are located 9.05 m from walls. The vents are activated 4. Stroup, D.W. and Evaus, D.D., "Ilse of Computer Fire by fusibleI IlL)ks having an activation temperature of 74°C, an RTI of Models for Anal~ing Thermal Detector Spacing," Fire 28 (m s) "-, and located 0.3 m below the roof. Inlet air openings SafetyJourruzl. 14, pp. 33-45, 1988. are equal to 1.5 the total vent area. See Figure D-1. 5. (?;ross, D., "Data Sources for Parameters I lsed in Predictive Occupancy Details Modeling of Fire Growth and Smoke Spread," NBSIR 85- 3223, Natio,'tal Bureau of Standards (presently National The building is to be occupied for retail storage. This analysis deals Institute of Standards and Teclmology), Gaithersbnrg with a fire in rack storage of sofas in the center of the building. The MI), September 1985. sofas are stored in two racks. The racks are each 9.75 m long, 1.2 m wide, and are separated from each other by 2.4 m. Distance to 6. Kaharter, D., Moiler, C., and N:tsh, S., NumaicalMethods combustibles surrounding the racks is sufficient to prevent fire a,ut &~war~;.Prentice Hall, 1989. spread to those combustibles during the time period covered by this analysis. The sofas are identified as specimen F32 contained within 7. K~tbaner, D., NIST, p,-ivate communication. Table 5-5.3(d). Data for the same sofasrare contained within a data base of Hazard I [2], where the sofas are identified as specimen UPS001. Each sofa contains 51.5 kg of combustible mass. The sofas are wrapped in polyethylene. Each rack has four tiers of storage, four sofas per tier, and a total storage height of 7.6 m.

628 NFPA 204M ~ A97 ROP

Accordingly, fire growth in the first sofa ignited may be approxi- i mated by a fast ( 0~ = 0.044 kW/s 2) "t-squared" tire. Fu~her, according to 6-1.4.~3, 0tg is directly proportional to the storage height. Therefore, the fire grox~h~ constant ((X_) for sofas stacked four high is 4 times 0.044 kW/s or O~g equals (~.18 kW/s 2 and / m m m initial tire growth is approximated as Q = O~gt 2, where: Of,g = 0.18kW/s 2

Fire growth in the first rack of sofas results in radiant heat transfer to a second rack of sofas separated from the first rack hy 2.4 m. It must be determined when the second rack of sofas ignite. The fire size, m i / m when ignition of the second rack of sofas occurs, is determined using 73 m equation 6-10 with its terms rearranged:

Q = (W/0.042) 2 where: m m m m W = aisle width (meters) Q = tire output (kW)

~ Smokeand heat ven~ Q= (2.4/0.042) 2 = 3265 kW ----- 3250 kW / Im II Next, the time of ignition of the second rack is computed using Q = O~g t 2 t = (Q//(gg )1/2 = (3250/0.18)1/2 = 134 seconds 73 m > When the second rack of sofas is ignited at 134 seconds, the tire growth coefficient, (~g, for the two racks burning together is assumed to double the value for the first rack burning alone ( O~g = 0.36 kW/s2). At that time, the fire appears to have originated at Figure D-I Vent plan view effective ignition time, tOg. For t > 134 seconds:

Q = 0.36 (t- tog) 2 kW Ignition Determine tOg as follows: An ignition is :~sumed to occur in a sofa on the first tier of one of the racks. Ignition of a sofa o11 tile first tier is a probable worst-case 3250 = 0.36 (134- tOg) 2 scenario and, ms a practical matter, is a location where ignidon may be expected. Also, placing the fire near floor level results in near- tog = 39seconds m,xximum smoke production (entrainment). Then, for t > 134 seconds: Fire Growth Q = 0.36 (t-39) 2 First, an estimate of file anticipated fire growth must be developed. A "t-squared" fire will be assumed -- see 6-1.4.6.1 and A-6-1.4.6.1. In The maximum fire size is now estimated. Sofa UPS001 from the a "t-squared" fire Hazard I database [2] (Specimen F-32 in Table 5-5.Bd) has a peak Q = (Zgt 2, burning rate of 3120 kW. Maximum fire size, Qmax, is based on the where assumption that all 32 sofas are burning at their individual peak rates, 3190 kW. Q = total heat rele,'x~e rate (kW) Qmax = 32 (3120) ~ 100 MW 0~g = tire growth coefficient Now, the time, tmax, to reach 100 MW must be determined using: t = time (seconds) Qmax = 0.36 (t- 39) 2 when Q = 100,000 kW The dam. base within Hazard I [2] contains data from furniture calorimeter tests of sofas. A sofa (UPS001) was tested and demon- 100,000 = 0.36 (tmax-B9) 2 strated a growth time (tg) to 1 MW of approxirmately 200 seconds. The fire in tile sofa in tiIfis example is assumed to have a growth time tm,-Lx = (100,000/.3611/2 + 39 = 566 seconds of 150 seconds to 1 M~Ar as a reasonable, conservative, approximation of the anticipated fire in tile sof,xs stored in the example building. If An estimate of fire duration, tend, is now made using data from the a more precise estimate of the burning characteristics of an Hazard I [2] database for sofa UPS001: individual sofa is necessary, the exact sofa to be stored in the building could be tested in a calorimeter. A fire growd/dmeLpf 150 Individual sofa combustible mass = 51.5 kg seconds results in an 0~, for tile individual sofa of 0.044 kW/s (see Sofa effective heat of combustion = 18,900 kJ/kg equation 6-I 1b). That%: Maximum tire size = 100,000 kW

(Zg = 1,000/tg 2 = I,~)00/1502 = 0.044kW/s 2 The mass consumed from t = 0 to t = 134 seconds is determined from the total heat release as follows: 134

. Total hea~ release ~34~0Qdt134 = -g--0.18 t 3 = (0.18/3) (134/3 = 144,366kJ trom t to t = 0

620 NFPA 204M -- A97 ROP

be calculated using equation 6-14. DETACT-QS (see 6-1.4.7.3.2) is a Since Q = fiah -- see Equation 5-1 -- mass loss, A m, for t = readily available computational tool that performs this calculation. 134 seconds, .ts determJc . ned ,xs follows: A complete fire alarm system is to be installed using heat detectors m m = 144,366 kJ/18,O00 kJ/kg = 7.6 kg or ----- 8 Kg spaced 15.2 m on center (6.1 m from walls), having an activation temperature of 74°C and an RTI of 55 (m*s)l/2. Assuming the The mass consumed from t = 134 seconds to tmax, the time when anticipated fire is as described above, the maximum distance from a tile maximum fire size is reached, is similarly determined from the detector to the fire axis is the diagonal [2 (15.2/2)211/2 = 10.7 m, total heat release rate ,alter 134 seconds, as follows: ambient temperature is 21°C, andthe fire is 0.5 m above floor level, 527 DETACT-QS predicts the activation of a heat detector at 230 t max 566 IQdt = I 0.36 (t _ 39)2dt = 566-39i seconds. In the event quicker detection is judged necessary, smoke 0.36132d1~ = 0-36/3 (t) 31 detector activation can be predicted by DETACT-QS using the 134 134 134-39 guidance provided in 6-1.4.7.2.2. Detection time for smoke 95 detectors ~s based on the gas temperature rise at the detector site. Smoke detector activatio n can be approximated using DETACT-QS, assuming die smoke detector will respond like a heat detector which Total heat release from t = 134 to t = 566 has a small RTI [e.g., 1 (m.s) 1/2] and a certain activation tempera- = 0.12 [(527) 3- (9~31 = 17.460,697 kJ, ture above ambient (see 6-1.4.7.2.2). Tests, involving burning of the a~ad tbe m~s lost, /..x m, is sofa upholstery with the actual detector to be installed, have /_Am = 17,460,697 kJ/18,.CKKl kJ/kg = 923.8 _~ 924 kg determined that 10°C above ambient is a representative activation condition. Assuming smoke detectors are spaced 9.1 m on center Approximately (924 + 8) kg = 9:?;2 kg is consumed during the 566- (located a maximum of 6.5 m from the axis of tiae fire), smoke second time h iterval required to reach Qmax- The total combustible detector activation is predicted by DETACT-QS at 48 seconds. m;Lss is 51.5 kg x 32 = 1648 kg. Therefore, around (1648 -932) kg = 716 kg is available to I)m'n at Q = ")L~lax : 100 MW, ,after t = 566 Using DETACT-QS, vent operation is predicted using fusible links seconds, from wbich the fire duration can be calculated as follows: having an activation temperature of 74°C and an RTI of 28 (re°s) 1/2. Assuming the anticipated fire is located in the center of the Qmax (tend- 566) = 100,000 (ten d -566) = 716 (18,900) building, the ambient temperature is 21°C, and assuming the fire is 0.5 m above floor level, activation of the first vents (equidistant from tend = 566+716 (18,9001 / 100,000 = 701.3seconds _= 700 the fire) separated [2(18.3/21211/2 = 12.9 m from the fire is seconds predicted by DETACT-QS at 228 seconds. The next set of vents (equidistant from the fire at 28.9 m) are predicted to open at 317 The coml)ustible m,xss of the sofas alone is able to support the seconds. Similarly, the third set of four vents, 38.8 m from the fire anticipated fire for approximately 700 seconds. In reality, the fire in axis, open at 356 seconds. All 16 vents are open at 356 seconds. tile sofas would reach a m,'Lxhnum of 100 MW at 550-600 seconds Alternatively, if fusible links having the same RTI as the heat and burn briefly at the 100-MW peak until the combustible mass detectors [55 (re°s)1/2] are used, all vents are predicted to be open available began to be consumed, at wlfich time the fire's rate of beat at 384 seconds. release would begin to decline. Using a ten d of 700 seconds is conservative. Vent Design

In summary, the analysis to this point leads to dae following Of main concern in this example is the temperature of the smoke estimate for the anticipated fire: layer, which governs the heat flux radiated to the floor. Assuming 9 an emissivity of 1 and a configuration factor of 1, the radiant heat Q = o.18t" for O 566 seconds T = temperature of the layer (°K)

See FigaJre D-2. k = Stefan Boltzmann constant = 5.67 E-11 kW/m 2 K4 Fire Detection E = emissivity = 1 The time of fire detection is now calculated given tile fire and O = configuration factor = i building a.s described. Tile time of detection will be estimated Ixlsed upon tile actual composite fire described above. Detection time can Fluxfl = (5.67 E-I1) T 4 kW/m 2

120

A 100

X 80 q = 0.36 (' 60

e-. 40 "6 q = 0.36 (t) 2 for t< 134 and 20 n- q =0.36(t - 39) 2 for t > 134

O~ 0 100 200 300 400 500 600 700 800 Time (s)

Figure D-2 Fire output.

630 NFPA 204M i A97 ROP

9 For a flux limit of 2.5 kW/m-, ;is stated in the objective, rile 2.5 kW/m 2. Using tile equation for radiar~heat flux to the floor temperature of the smoke layer is calculated as 458°K, or 164°K presented previously, the value 29.7 kW/m is calculated for a above the ambient teml~eratnre of 294°K. smoke layer temperature of 557 + 294 = 851°K. Steady Fire Not only is rile smoke layer temperature, 557 + 21 = 578°C, so high that it produces unacceptable levels of radiant flux at the floor, but it Smoke Layer Temperature is also close to tile level, 600°C, where fire can flash over all the combustibles under rile smoke layer (see 6-1.5.1.1). Furthermore, it First, conditions foil owing attainment of the maximum beat release exceeds the value, 540°C, where unprotected steel begins loosing rate of 100 MW can be exanfined, i.e., at times greater rimn 566 strength (see 6-1.5.1.2). Directly over the fire (see 6-1.5.1.2) the seconds, assumir~g a smoke layer at tile lowest acceptable height, 3 m temperatures may locally reach 1155°C (from equalion 6-9 with r = above the floor. (The heat detector installation contemplated was 1 ), far in excess of the threshold for steel damage. calculated to provide alarm at 230 seconds; 300 seconds following detection places the time of interest ,~ 550 seconds, close to the Sizing of Vents attainmentof the maxirmnn beat release rate.) This building arrangement will not meet design objectives. The effective diameter of the fire is required for the calculations. However, it may be instructive to investigate the venting require- This diameter can be determined with the aid of equation 6-13, ments in order to illustrate general procedures which might be used setting Q = 100,000 kW and selecting an appropriate value for the to develop alternative designs. heat release rate per unit floor area Q". The two racks facing each other across the 2.4 m wide aisle are ~75 m long and 1.2 m wide - All 16 vents are predicted tO be open prior to 566 seconds -- the see Figure [)-3. The he'.tt release rate per unit area is taken as the time of interesL fully-involved heat release rate, 100,000 kW, divided by the com- bined area of the two racks pins the aisle, or 9.75 • 1.2 • 2.2 + 9.75 * Tile aerodynamic vent area, Ava, is determined with the aid of 2.4 = 46.8 m ~. Accordingly, the heat release rate per unit area is: equation 6-8: (,)~2_xl/2rT0 Arp] l/2A .11/2 Q" = 100,000/46.8 = 2136 kW/m z fiav =,,.vor,) [ a.j ,-,vu At equilibrium, the mass 0ow through the ve~ts is equal to the sqloke production rate, In,. Substituting m° = a2.6 kg/s for 9 m v in equation 6-8, togetfier with r o = 1.2 k~/m", g = 9.81 m/s-, Effective fire diameter T O = 294°K, DT = 559°K, T = 294+559 = 85~°K, andd = 9.1 -3 = 6.1 m, the equation can be solved for the aerodynamic vent area. Fire area Tim result is:

Ava = 10.04 m 2 1.2 m Rack 1 ] T The vents are assumed to have a discharge coefficient of 0.61, and therefore, tile corresponding actual vent area is (see 6-1.4.2):" 2 Effective Av = 10.04/0.61 = 16.46 m (geometric vent area) ---- 16.5 m 2 fire diar~ 2.4 m Tile building design contemplates that inlet air openings will be 1.5 times the vent area. Equation 6-6 is used to calculate a correction, ! ,l M, for file limited inlet air openings: Rack 2 [ 1.2 m M = [1 + (Av/Ai) 2 (To/T)] 1/2 ! M = [1+(1/1.5) 2(294/853)] 1/2 = 1.07 -I The corrected actual vent area is: 1.07• 16.5 = 17.66m 2

Figure D-3 Effective fire diameter. Distributed among tile 16 vent locations, tile actual area per vent is: 2 Tiffs value can be assumed to be representative of most of the fire 17.66/16 = 1.10 m history, except for the initial st-age. The effective diameter of the fire at 100,000 kW is then, nsing equation 6-13: The nearest commercial vent size equal to or larger than this unit vent area would be selected. !/2 D = [(4 • 100,0O0])/ (n•2136)1 = 7.72m Equation 6-17 is used to check for Qfeasiblo where H = 9.1 - 0.5 = Equation 6-9 is nsed to estimate the smoke layer temperature rise. 8.6 m and d = 6.1 m Tlie mass flow rate in the plume ~¢ it enters the smol~e layer, mp, is calculated fiom equations 6-2 or 6-3, depending on whether rile Qfeasible = 229,265 kW -~ 230 MW flame height is smaller or larger than the height of the smoke layer above tile base of the fire, 3- 0.5 = 2.5 m. The flame height is Tiais value is higher than the projected heat release rate, 100 MW, calculated from equation 6-1 : and by itself is not of direct concern (see 6-1.5.2). L = (-1.02 • 7.72) + (0.235 * 100,0002/5) = 15.6 m Increased Height of Smoke Interface which is greater than the height of the smoke layer. (It is even Inspection of equation 6-3 indicates that the larger the height of the smoke interface above rile base .of the fire, the larger the vaqu e of greater than the ce ng be gbt so that the flames will impinge on the mass entrained in rile plume, m_, and eouation~5-9 indicates that ceding and flow radiall/outward.) Therefore, the mass flow rate m the temperature rise in the smok~' layer will be reduced. The the plume as it enters the smoke layer is calculated from equation 6- calculat*onsjust completed for a smoke layer height of 3 m above 3,~s follows (.'L~smning ~).2c = 0.7 Q): the floor may be repeated for oriler smoke layer fieighl~ in search of acceptable alternative designs. The two additional smoke layer l:flp = [o.0056 (0.7-100.000) 1 [2.5/15.61 = 62.8 kg/s heights of 6 and 7.3 m have been investigated, the latter near the maximum associated with tile minimum recommended curtain Now the tempe,ztture rise in the smoke layer can be estimated using depth for the 9.l-m-high building (see 4-3). The final results of equation 6-9, with Cp = 1.00 kJ/kg•K andthe value ofr = 0.5 rilese additional calculations indicate values of temperature rise in recommended in 6-'1,3.4. the smoke layer of 253°K for the 6 m high level and205°K for the 7.3 m high level. Although these values of smoke layer temperature AT = 0.5 * 70,000/(I.00- 62.8) = 557°K rise are still a little high compared to tile target of 164°K, riley represent a major improvement. Furthermore, rile temperatures are This value is considerably above 164~K, and therefore the floor low enongh so as not to represent a flashover hazard or endanger radiant beat flux cart be expected to be much higher than the limit structural steel. 631 NFPA 204M I A97 ROP

Table D-I. Results of calculations for vent design.

Case T i m e Q D L (I~m_566 100 7.7 15.6 6.0 5.5 3.1 253 5.1 137.8 1.11 53.8 3 >_566 1O0 7.7 15.6 7.3 6.8 1.8 205 3.5 170.4 1.12 89.2

4 530 86.8 7.2 14.9 3.0 2.5 6.1 531 26.4 57.2 1.08 16.1 5 530 86.8 7.2 14.9 6.0 5.5 3.1 241 4.7 125.9 1.12 49.7 6 530 86.8 7.2 14.9 7.3 6.8 1.8 195 3.3 155.7 1.13 82.6

7 348 34.4 4.5 10.7 3.0 2.5 6.1 383 11.8 31.4 1.09 8.6 8 348 34.4 4.5 10.7 6.0 5.5 3.1 174 2.7 69.0 1.13 28.3 9 348 34.4 4.5 10.7 7.3 6.8 1.8 141 2.0 85.3 1.14 47.8

The calculations for the three smoke layer heights at die maximum A complete smoke detection system is to be installed with detectors heat release rate are summarized in Table D-l, entered ,as Cases 1-3. spaced 9.1 m on center. Detectors are located a maximum of 6.5 m In the ~able, Hf represents the height of die ceiling above the floor; from the fire axis, i.e., one-half the diagonal distance between Hf-d is the height of the smoke interface above d~e floor; H - d is die "detectors. As noted in 6-1.4.7.2.2, detectors have an activation height of the smoke interface above file base of the fire. In cases 1- temperature of 31°C (10°C above ambient), and are located 0.1m 3, the radiant heat flux at floor level, Fluxfl, is seen to decrease to 5.1 below the ceiling. mad 3.5 kW/m 2 ,xs the smoke interface is raised, but still remains above 2.5 kW/m2. The total required vent area (Corrected Av) The vent design will use sixteen 1.76-m2 vents located 18.3 meters incre;Lses sharply ,xs the smoke layer interface is raised. For file on center. All vents automatically open upon activation of the first largest interface heigltt, the total vent area of 89.2 m2 corresponds smoke detector. to an area per vent of 89.2/16 = 5.57 m2, which is ~till small~ than fl~e maximum vent area discussed in 3-4(a), i.e., 2d- = 2 • 1.8- = LAVENT predicts the upper layer temperature will be 377°C and the 6.48 m2. upper "hot" layer will be 4.6 m above floor level at 600 seconds. A 3-m clear layer is maintained throughout the 600-second time Growing Fire interval. However, heat flux at floor level is projected to be approximately 10 kW/m2 at 600 seconds, and file desitzn objective of Cases 4-6 in Table D-I correspond to the growing fire with detection limiting heat flux to 2.5 kW/m2 at floor level is exceeded. At 342 at 230 seconds using heat detectors. The state of the fire is seconds, the time of detection plus 300 seconds, however, the design represented at a time 300 seconds following detection with heat objectives are met. At 360 seconds LAVENT predicts the upper layer detectors, i.e., at 930 + 300 = 530 seconds. It is ,assumed that die 16 temperature as 444°K (171°C), with the layer being 7.3 m above the vents are all operated togedler at the alarm of the first heat detector, floor. The predicted 1500K temperature rise is limited to less than alternatively, the vents are actuated individually with fusible links of the target value of ] 64°K, and heat flux at floor level is predicted to the same RTI and activation temperature ,as the heat detectors, for be 2.2 ~W/m2. Therefore, the design objectives are satisfied for a which it may be confirmed with DETAG'I'-QS that all vents open time interval greater than the time of detection plus 300 seconds. prior to 530 seconds. The calculations are parallel to Cases 1-3, except that the fire is slightly smaller, ms determined from: Inlet air is 1.5 times the vent area. To maintain the vent flow predicted by LAVENT, inlet air net free area should be maintained at a minimum of twice the open vent area. Although the net free Q = 0.36 (t- 39) 2 = 0.36 (530- 39) 2 = 86,800 kW inlet air area is less than required, the inlet area is sufficiendy large dlat LAVENT predictions may be assumed to be reasonably valid. In (21ses 476, the smoke layer temperatures (AT) arid radiant fltlxes However, consideration should be ~iven to increasing the vent area to the floor are only slighdy reduced front the corresponding steady to account for die restrictions in inlet air. fire situations, Cases I-3. Also, there is little change in die required ve n t areas. See Figures D-4 through D-11 and Table D-2. (2Lses 7-9 in Table 1-)-1 correspond to the growing fire, with detection at 48 seconds using smoke detectors. Again, the state of 650 the fire is represented at a time 300 seconds from detection, Le., at 348 seconds. It is :Lssmned that die 16 vents are operated together at the alarm of the first smoke detector. The calcukttions are executed at a state of fire development from: 600 9 = Q = 0.36 (t- 39)- = 0.36 (348- 39) 2 34,400 kW 550 It is seen that C:~e 9 meets the design objective of heat fluxes to the floor that are ~=dculmed ;Ls being lower than 2.5 kW/m2, :rod Case 8 nearly does so. The required vent are,x~ are 28.3 m2and 47.8 m2 for C;Lses 8 and 9, resl)ectively, corresponding to unit vent areas (16 50O vents) of 1.8 and 2~,.0 In2, hod/of which are well below their respective maxima, 2d 2, I)n.sed on 3-4(a). ~. 450 It will be noted the (~k~se 8 solutiorJ using "hand calculations" E provides a close, somewhat couserwative approximation of die LAVENT predictions, which are summarized below. ,,J LAVENT Analysls

The Table 1-)-1, (2ase 8 vent design will now be analyzed using the 350 computer program LAVENT [3]. LAVENT is able to assess the time- ~,,u'ying events associated with the predicted fire. The fire has been previously determined as follows: 300 Q = o.18t 2 for o 566 seconds 2% 150 300 450 600 The values for the above-described fire will be used as input for Time (s) LAVENT. The fire is ,x~sumed to start in the center of the building. Figure D-4 Temperature.

632 NFPA 204M -- A97 ROP

9.5 30 28 9.0 26 8.5 24 22 8.0 20 7.5 ~ 18 2 Jr- ~ 16

.J 6.5 12 10 6.0 8 5.5 614 f 5.0 2 4.5 150 300 450 600 00 150 300 450 600 Time (s) Time (s)

Figure D-5 Layer height. F'tgure D-7 Vent area.

0.10E+09

0.90E+08 13000 12000 0.80E+08 1100o

0.70E+08 10000 A 90oo ~- 0.60E+~ 8000

0 .~ 0.50E+OO 7o00 LI,. -J 0.40E+08 / 6ooo 50oo 0.30E+08 40oo 3oo0

0.10E+08

o.oo o 150 300 450 600 0 o 150 300 450 600 Time Is) -nine (s)

Plgure D-6 F}re output. Figure D-.8 Layer mass.

633 NFPA 204M m A97 ROP

120- 1050 - 1000 110 • 95O 100 900

90 85O 8OO 750 7O =E 60 ,,¢ 61111 a. ,.J 55O 4O

3O

2O 350 10

25O o 0 o 150 300 450 150 300 450 6OO Tm~e (s) Time (s)

Figure D-9 Plume flow. F'qgure D-I 1 Jet temperature.

1050- 1000 950 9OO 850 8O0 750 700 650 600 v, .c .,J 550 500 450 400 350 300

2500 150 300 450 600 Time (s)

Figure D-10 Detector temperature.

6~4 NFPA 204M -- A97 ROP

Table D-2

CEILING HEIGHT = 9.1M ROOM LENGTH = 73.0 M ROOM WIDTH = 73.0 M CURTAIN I_~NGTH = 292.0 M CU RTAIN HEIGHT = 0.0 M MATERIAL = INSULATED DECK (SOLID POLYSTYRENE) CEILING CONDUCTIVrI'Y= .149E+00 W/M K CEILIING DENSITY = .116E+04 KG/M3 CEILING HEAT CdkPACITY = .105E+04J/M K CEIL)[NG THICKNES, S = .152E+00 M FIRE HEIGHT = 0.5-M FIRE POWER/AREA = 0.2136E+07 W/M2

LINK NO = 1 RADIUS = 6.5 M DIST CEILING = 0.1 M RTI= 1.00SQRT(MS) FUSION TEMPERATURE FOR LINK = 304.00 VENT= 1 VENTAREA= 28.2 M2 LINKCONTROIJJNGVENT= 1 TIME; (S)--- 0.0000LYRTEMP (K)= 294.0LYRHT (M) = 9.10 LYR MASS (KG)=0. 000E+00 FIRE OUTPUT (W) = 0.0000E+00 VENT AREA (M2) = 0.00 LINK = 1 LINK TEMP (K) = 294.00JET VELOCrI'Y (M/S) = O.O00JET TEMP (K) = 294.0 R (M) = 0.00 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 1.74 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 3.48 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 5.22 TSL (K) = 294.0 QB 0N/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (m) = 6.95 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 8.69 TSL (g) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 10.43 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 12.17 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 13.91 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 15.65 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 17.39 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 19.12 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 20.86 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 22.60 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 24.34TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 26.08 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 27.82 TSL (K) = 294,0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 29.56 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 31.29 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 33.03 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 34.77 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 36.51 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 38.25 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 39.99 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 41.73 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 43.46 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 45.20 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 46.94 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2)-- 0.000E+00 R (M) = 48.68 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 R (M) = 50.42 TSL (K) = 294.0 QB (W/M2) = 0.000E+00 QT (W/M2) = 0.000E+00 TIME (S)= 60.0000 LYRTEMP (K)= 301.4LYRHT (M) = 8.99 LYR MAKS (KG)=0.657E+03 FIRE OUTPUT (W) = 0.6480E+06 VENT AREA (M2) = 28.20 LINK= 1 LINKTEMP (K) = 309.95JETVELOCITY (M/S) = 1.104

6S5 NFPA 204M -- A97 ROP

JETTEMP (K) = 310.2 TIME LINK 1 OPENS EQUALS 41.7098 (S) R (M) = 0.00 TSL (K) = 302.1 QB (W/M2) = 0.834E+03 QT (W/M2) = 0.000E+00 R (M) = 1.74 TSL (K) = 299.5 QB (W/M2) = 0.587E+03 QT (W/M2) = 0.000E+00 R (M) = 3.48 YSL (K) = 297.8 QB (W/M2) = 0.417E+03 QT (W/M2) = 0.000E+00 R (M) = 5.22 TSL (K) = 296.6 QB (W/M2) = 0.287E+03 QT (W/M2) = 0.000E+00 R (M) = 6.95 TSL (K) = 295.8 QB (W/M2) = 0.205E+03 QT (W/M2) = 0.000E+00 R (M) = 8.69 TSL (K) = 295.4 QB (W/M2) = 0.153E+03 QT (W/M2) = 0.000E+00 R (M) = 10.43 TSL (K) = 295.0 QB (W/M2) = 0.117E+03 QT (W/M2) = 0.000E+00 R (M) = 12.17 TSL (K) = 294.8 QB (W/M2) = 0.925E+02 QT (W/M2) = 0.000E+00 R (M) = 13.91 TSL (K) = 294.7 QB (W/M2) = 0.748E+02 QT (W/M2) = 0.000E+00 R (M) = 15.65 TSL (K) = 294.6 QB (W/M2) = 0.619E+02 QT (W/M2) = 0.000E+00 R (M) = 17.39 TSL (K) = 294.5 QB (W/M2) = 0.522E+02 QT (W/M2) = 0.000E+00 R (M) = 19.12 TSL (K) = 294.4 QB (W/M2) = 0.448E+02 QT (W/M2) = 0.000E+00 R (M) = 20.86 TSL (K) = 294.3 QB (W/M2) = 0.389E+02 QT (W/M2) = 0.000E+00 R (M) = 22.60 TSL (K) = 294.3 QB (W/M2) = 0.343E+02 QT (W/M2) = 0.000E+00 R (M) = 24.34 TSL (K) = 294.3 QB (W/M2) = 0.305E+02 QT (W/M2) = 0.000E+00 R (M) = 26.08 TSL (K) = 294.2 QB (W/M2) = 0.274E+02 QT (W/M2) = 0.000E+00 R (M) = 27.82 TSL (K) = 294.2 QB (W/M2) = 0.248E+02 QT (W/M2) = 0.000E+00 R (M) = 29.56 TSL (K) = 294.2 QB (W/M2) = 0.226E+02 QT (W/M2) = 0.000E+00 R (M) = :~1.29 TSL (K) = 294.2 QB (W/M2) = 0.207E+02 QT (W/M2) = 0.000E+00 R (M) = 33.03TSL (K) = 294.2 QB (W/M2) = 0.191E+02 QT (W/M2) = 0.000E+00 R (M) = 34.77 TSL (K) = 294.2 QB (W/M2) = 0.177E+02 QT (W/M2) = 0.000E+00 R (M) = 36.51 TSL (K) = 294.1 QB (W/M2) = 0.165E+02 QT (W/M2) = 0.000E+00 R (M) = 38.25 TSL (K) = 294.1 QB (W/M2) = 0.154E+02 QT (W/M2) = 0.000E+00 R (M) = 39.q9 TSL (K) = 294.1 QB (W/M2) = 0.144E+02 QT (W/M2) = 0.000E+00 R (M) = 41.73 TSL (K) = 294.1 QB (W/M2) = 0.136E+02 QT (W/M2) = 0.000E+00 R (M) = 43.46 TSL (K) = 294.1 QB (W/M2) = 0.128E+02 QT (W/M2) = 0.000E+00 R (M) = 45.20TSL (K) = 294.1 QB (W/M2) = 0.121E+02 QT (W/M2) = 0.000E+00 R (M) = 46.94 TSL (K) = 294.1 QB (W/M2) = 0.115E+02 QT (W/M2) = 0.000E+00 R (M) = 48.68 TSL (K) = 294.0 QB (W/M2) = 0.122E+01 QT (W/M2) = 0.000E+00 R (M) = 50.42 TSL (K) = 294.0 QB (W/M2) = 0.110E+01 QT (W/M2) = 0.000E+00 TIME (S)= 120.0000 LYRTEMP (K)= 317.2 LYRHT (M) = 8.83 LYR MASS (KG)=0.162E+04 FIRE OUTPUT (W) = 0.2743E+07 VENT AREA (M2) = 28.20 LINK = 1 LINK TEMP (K) = 339.83JET VELOCrIY (M/S) = 1.761 JET TEMP (K) = 340.2 TIME LINK 1 OPENS EQUALS 41.7098 (S) R (M) = 0.00 TSL (K) = 332.0 QB (W/M2) = 0.242E+04 QT (W/M2) = 0.000E+00 R (M) = 1.74 TSL (K) = 322.4 QB (W/M2) = 0.188E+04 QT (W/M2) = 0.000E+00 R(M) = 3.48TSL (K) = 314.9 QB (W/M2) = 0.142E+04 QT (W/M2) = 0.000E+00 R (M) = 5.22 TSL (K) = 308.8 QB (W/M2) = 0.102E+04 QT (W/M2) = 0.000E+00 R (M) = 6.95 TSL (K) = 304.7 QB (W/M2) = 0.753E+03 QT (W/M2) = 0.000E+00 R(M)= 8.69TSL(K)= 302.1QB (W/M2) = 0.569E+03 QT (W/M2) = 0.000E+00 R (M) = 10.43TSL (K) = 300.2 QB (W/M2) = 0.441E+03 QT (W/M2) = 0.000E+00 R (M) = 12.17TSL (K) = 298.9 QB (W/M2) = 0.350E+03 QT (W/M2) = 0.000E+00 R(M)= 13.91TSL(K)= 298.0QB (W/M2) = 0.285E+03 QT (W/M2) = 0.000E+00 R(M)= 15.65TSL(K)= 297.3QB (W/M2) = 0.236E+03 QT (W/M2) = 0.000E+00 R(M)= 17.39TSL(K)= 296.8QB (W/M2) = 0.199E+03 QT (W/M2) = 0.000E+00 R(M)= 19.12 TSL (K) = 296.4QB (W/M2) = 0.171E+03 QT (W/M2) = 0.000E+00 R(M)= 20.86TSL(K)= 296.1QB (W/M2) = 0.149E+03 QT (W/M2) = 0.000E+00 R(M)= 22.60TSL(K)= 295.8QB (W/M2) = 0.131E+03 QT (W/M2) = 0.000E+00 R (M) = 24..'+4 TSL (K) = 295.6 QB (W/M2) = 0.117E+03 QT (W/M2) = 0.000E+00 R(M)= 26.08TSL(K)= 295.5QB (W/M2) = 0.105E+03 QT (W/M2) = 0.000E+00 R(M)= 27.82TSL(K)= 295.3QB (W/M2) = 0.951E+02 QT (W/M2) = 0.000E+00 R(M)= 29.56 TSL (K) = 295.2QB (W/M2) = 0.867E+02 QT (W/M2) = 0.000E+00 R(M)= 31.29TSL(K)= 295.1QB (W/M2) = 0.795E+02 QT (W/M2) = 0.000E+00

636 NFPA 204M ~ A97 ROP

R (ld) = "53.03 TSL (K) = 295.0 QB (W/M2) = 0.734E+02 QT (W/M2) = 0.000E+00 R (M) = 34.77 TSL (K) = 294.9 QB (W/M2) = 0.680E+02 QT (W/M2) = 0.000E+00 R (M) = 36.51 TSL (K) = 294.9 QB (W/M2) = 0.633E+02 QT (W/M2) = 0.000E+00 R (M) = 38.25 TSL (K) = 294.8 QB (W/M2) = 0.592E+02 QT (W/M2) = 0.000E+00 R (iYl) = 39.99 TSL (K) = 294.8 QB (W/M2) = 0.555E+02 QT (W/M2) = 0.000E+00 R (M) = 41.73 TSL (K) = 294.7 QB (W/M2) = 0.522E+02 QT (W/M2) = 0.000E+00 R (M) = 43.46TSL (K) = 294.7 QB (W/M2) = 0.492E+02 QT (W/M2) = 0.000E+00 R (M) = 45.20 TSL (K) = 294,6 QB (W/M2) = 0.466E+02 QT (W/M2) = 0.000E+00 R (M) = 46.94 TSL (K) = 294.6 QB (W/M2) = 0.442E+02 QT (W/M2) = 0.000E+00 R (M) = 48.68 TSL (K) = 294.1 QB (W/M2) = 0.504E+01 QT (W/M2) = 0.000E+00 R (M) = 50.42 TSL (K) = 294.1 QB (W/M2) = 0.455E+01 QT (W/M2) = 0.000E+00 TIME (S)= 180.0000 LYR TEMP (K)= 339,8 LYRHT (M) = 8.60 L'~q~ MASS (KG)=0.276E+04 FIRE OUTPUT (W) = 0.7483E+07 VENT AREA (M2) = 28.20 LINK = 1 LINK TEMP (K) = 385.73JET VELOCrIY (M/S) = 2.493 JETTEMP (K) = 386.3 TIME LINK 1 OPENS EQUALS 41.7098 (S) R (M) = 0.00 TSL (K) = 386.4 QB (W/M2) = 0.514E+04 QT (W/M2) = 0.000E+00 R (M) = 1.74 TSL (K) = 367.0 QB (W/M2) = 0.421E+04 QT (W/M2) = 0.000E+00 R (M) = 3.48 TSL (K) = 349.7 QB (W/M2) = 0.329E+04 QT (W/M2) = 0.000E+00 R (M) = 5.22 TSL (K) = 334.5 QB (W/M2) = 0.244E+04 QT (W/M2) = 0.000E+00 R (M) = 6.95 TSL (K) = 324.0 QB (W/M2) = 0.183E+04 QT (W/M2) = 0.000E+00 R (M) = 8.69 TSL (K) = 316.7 QB (W/M2) = 0.140E+04 QT (W/M2) = 0.000E+00 R(M)= 10.43TSL(K)= 311.6QB(W/M2)=O.IO9E+O4QT(W/M2)= 0.000E+00 RIM) = 12.17 TSL (K) = 308.0 QB (W/M2) = 0.864E+03 QT (W/M2) = 0.000E+00 RIM) = 13.91 TSL (K) = 305.3 QB (W/M2) = 0.702E+03 QT (W/M2) = 0.000E+00 R qM) = 15.65 TSL (K) -- 303.4 QB (W/M2) = 0.582E+03 QT (W/M2) = 0.000E+00 R (M) = 17.39 TSL (K) = 301.9 QB (W/M2) = 0.491E+03 QT (W/M2) = 0.000E+00 R (M) = 19.12 TSL (K) = 300.8 QB (W/M2) = 0.420E+03 QT (W/M2) = 0.000E+00 R (M) = 20.86 TSL (K) = 299.9 QB (W/M2) = 0.365E+03 QT (W/M2) = 0.000E+00 R ~IM) = 22.60 TSL (K) = 299.2 QB (W/M2) = 0.321E+03 QT (W/M2) = ff000E+00 R (M) = 24.34 TSL (K) = 298.6 QB (W/M2) = 0.286E+03 QT (W/M2) = 0.000E+00 R (M) = 26.08 TSL (K) = 298.1 QB (W/M2) = 0.256E+03 QT (W/M2) = 0.000E+00 R (M) = 27.82 TSL (K) = 297.7 QB (W/M2) = 0.232E+03 QT (W/M2) = 0.000E+00 R (M) = 29.56 TSL (K) = 297.4 QB (W/M2) = 0.211E+03 QT (W/M2) = 0.000E+00 R (m) = 31.29 TSL (K) = 297.1 QB (W/M2) = 0.193E+03 QT (W/M2) = 0.000E+00 R (M) = 33.03 TSL (K) = 296.9 QB (W/M2) = 0.178E+03 QT (W/M2) = 0.000E+00 R (M) = 34.77 TSL (K) = 296.7 QB (W/M2) = 0.165E+03 QT (W/M2) = 0.000E+00 R (M) = 36.51 TSL (K) = 296.5 QB (W/M2) = 0.154E+03 QT (W/M2) = 0.000E+00 R (M) = 38,25 TSL (K) = 296.3 QB (W/M2) = 0.143E+03 QT (W/M2) = 0.000E+00 R (M) = 39.99 TSL (K) = 296.2 QB (W/M2) = 0.134E+03 QT (W/M2) = 0.000E+00 R (M) = 41.73 TSL (K) = 296.0 QB (W/M2) = 0.126E+03 QT (W/M2) = 0.000E+00 R (M) = 43.46 TSL (K) = 295.9 QB (W/M2) = 0.I19E+03 QT (W/M2) = 0.000E+00 R (M) = 45.20 TSL (K) = 295.8 QB (W/M2) = 0.113E+03 QT (W/M2) = 0.000E+00 R (M) = 46.94 TSL (K) = 295.7 QB (W/M2) = 0.107E+03 QT (W/M2) = 0.000E+00 R (M) = 48.68 TSL (K) = 294.2 QB (W/M2) = 0.136E+02 QT (W/M2) = 0.000E+00 R (M) = 50.42 TSL (K) = 294.2 QB (W/M2) = 0.123E+02 QT (W/M2) = 0.000E+00 TIME (S)= 240.0000 LYR TEMP (K)= 371.5 LYR HT (M) = 8.28 LYR MASS (KG)=0.414E+04 FIRE OUTPUT (W) = 0.1541E+08 VENT AREA (M2) = 28.20 LINK= 1 LINKTEMP (K) = 447.57JETVELOCrIY (M/S) = 3.186 JET TEMP (K) = 448.2 TIME LINK 1 OPENS EQUALS 41.7098 (S) R (M) = 0.00 TSL (K) = 469.7 QB (W/M2) = 0.816E+04 QT (W/M2) = 0.000E+00 R (M) = 1.74 TSL (K) = 439.3 QB (W/M2) = 0.700E+04 QT (W/M2) = 0.000E+00 R (M) = 3.48 TSL (K) = 408.8 QB (W/M2) = 0.570E+04 QT (W/M2) = 0.000E+00 R (M) = 5.22 TSL (K) = 380.2 QB (W/M2) = 0.439E+04 QT (W/M2) = 0.000E+00 R (M) = 6,95 TSL (K) = 359.0 QB (W/M2) = 0.335E+04 QT (W/M2) = 0.000E+00

637 NFPA 204M -- A97 ROP

R (M) = 8.69 TSL (K) = 343,8 QB (W/M2) = 0.259E+04 QT (W/M2) = 0.000E+00 R(M)= 10.43TSL(K)= 332.8 QB (W/M2) = 0.203E+04 QT (W/M2) = 0.000E+00 R(M)= 12.17TSL(K)= 324.9 QB (W/M2) = 0.162E+04 QT (W/M2) = 0.000E+00 R(M)= 13.91TSL(K)= 319.1 QB (W/M2) = 0.132E+04 QT (W/M2) = 0.000E+00 R(M)= 15.65TSL(K)= 314.8 QB (W/M2) = 0.109E+04 QT (W/M2) = 0.000E+00 R(M)= 17.39TSL(K)= 311.6 QB (W/M2) = 0.922E+03 QT (W/M2) = 0.000E+00 R(M)= 19.12TSL(K)= 309.1 QB (W/M2) = 0.790E+03 QT (W/M2)'= 0.000E+00 R(M)= 20.86TSL(K)= 307.1 QB (W/M2) = 0.687E+03 QT (W/M2) = 0.000E+00 R(M)= 22.60TSL(K)= 305.5 QB (W/M2) = 0.604E+03 QT (W/M2) = 0.000E+00 R (M) = 24.34TSL (K) = 304.2 QB (W/M2) = 0.536E+03 QT (W/M2) = 0.000E+00 R(M)= 26.08TSL(K)= 303.2 QB (W/M2) = 0.481E+03 QT (W/M2) = 0.000E+00 R(M)= 27.82TSL(K)= 302.3 QB (W/M2) = 0.435E+03 QT (W/M2) = 0.000E+00 R(M)= 29.56TSL(K)= 301.6 QB (W/M2) = 0.396E+03 QT (W/M2) = 0.000E+00 R(M)= 31.29TSL(K)= 300.9 QB (W/M2) -- 0.363E+03 QT (W/M2) = 0.000E+00 R(M)= 33.03TSL(K)= 300.4 QB (W/M2) = 0.334E+03 QT (W/M2) = 0.000E+00 R(M)= 34.77TSL(K)= 299.9 QB (W/M2) = 0.309E+03 QT (W/M2) = 0.000E+00 R(M)= 36.51TSL(K)= 299.5 QB (W/M2) = 0,288E+03 QT (W/M2) = 0.000E+00 R(M)= 38.25TSL(K)= 299.1 QB (W/M2) = 0.269E+03 QT (W/M2) = 0.000E+00 R(M)= 3iL99TSL(K)= 298.8 QB (W/M2) = 0.252E+03 QT (W/M2) = 0.000E+00 R(M)= 41.73TSL(K)= 298.5 QB (W/M2) = 0.237E+03 QT (W/M2) = 0.000E+00 R(M)= 43.46TSL(K)= 298.3 QB (W/M2) = 0.223E+03 QT (W/M2) = 0.000E+00 R(M)= 45.20TSL(K)= 298.0 QB (W/M2) = 0,211E+03 QT (W/M2) = 0.000E+00 R(M)= 46.94TSL(K)= 297.8 QB (W/M2) = 0,200E+03 QT (W/M2) = 0.000E+00 R(M)= 48.68TSL(K)= 296.6 QB (W/M2) = 0.198E+03 QT (W/M2) = 0.000E+00 R(M)= 50.42TSL(K)= 294.5 QB (W/M2) = 0.250E+02 QT (W/M2) = 0.000E+00 TIME (S)= 300.0000 LYR TEMP (K)= 406.7 LYR HT (M) = 7.86 LYR MASS (KG)=0.575E+04 FIRE OUTPUT (W) = 0.2452E+08 VENT AREA (M2) = 28.20 LINK = 1 LINK TEMP (K) = 511.85JET VELOCITY (M/S) = 3.699 JETTEMP (K) = 512.4 TIME LINK 1 OPENS EQUALS 41.7098 (S) R (M) = 0.00 TSL (K) = 561.4 QB (W/M2) = 0.962E+04 QT (W/M2) =-0.297E-11 R (M) = 1.74 TSL (K) = 523.2 QB (W/M2) = 0.859E+04 QT (W/M2) =-0.297E-11 R (M) = 3.48 TSL (K) = 481.7 QB (W/M2) = 0.731E+04 QT (W/M2) = -0.297E-11 R (M) = 5.22 TSL (K) = 439.7 QB (W/M2) = 0.588E+04 QT (W/M2) =-0.297E-11 R (M) = 6.95 TSL (K) = 406.1 QB (W/M2) = 0.464E+04 QT (W/M2) =-0.297E-11 R (M) = 8.69 TSL (K) = ,'481.0 QB (W/M2) = 0.365E+04 QT (W/M2) = -0.297E,-11 R (M) = 10.43 TSL (K) = 362.4 QB (W/M2) = 0.289E+04 QT (W/M2) = -0.297E-11 R (M) = 12.17 TSL (K) = 348.8 QB (W/M2) = 0.234E+04 QT (W/M2) =-0.297E-11 R (M) = 13,91 TSL (K) = 338.7 QB (W/M2) = 0.191E+04 QT (W/M2) =-0.297E-11 R (M) = 15.65 TSL (K) = 331.1 QB (W/M2) = 0.159E+04 QT (W/M2) = -0.297E-11 R (M) = 17.39 TSL (K) = 325.4 QB (W/M2) = 0.135E+04 QT (W/M2) = -0.297E-11 R (M) = 19.12 TSL (K) = 320.9 QB (W/M2) = 0.116E+04 QT (W/M2) = -0.297E-11 R (M) = 20.86 TSL (K) = 317.4 QB (W/M2) = 0.101E+04 QT (W/M2) =-0.297E-11 R (M) = 22.60 TSL (K) = 314.6 QB (W/M2) = 0.887E+03 QT (W/M2) = -0.297E-11 R (M) = 24.34 TSL (K) = 312.3 QB (W/M2) = 0.789E+03 QT (W/M2) = -0.297E-11 R (M) = 26.08 TSL (K) = 310.4 QB (W/M2) = 0.708E+03 QT (W/M2) = -0.297E-11 R (M) = 27.82 TSL (K) = 308.8 QB (W/M2) = 0.640E+03 QT (W/M2) = -0.297E-11 R (M) = 29.56 TSL (K) = 307.5 QB (W/M2) = 0.583E+03 QT (W/M2) = -0.297E-11 R (M) = 31.29 TSL (K) = 306.4 QB (W/M2) = 0.535E+03 QT (W/M2) = -0.297E-11 R(M)= 33.03 TSL (K) = 305.4 QB (W/M2) = 0.493E+03 QT (W/M2) = -0.297E-11 R (M) = 34.77 TSL (K) = 304.6 QB (W/M2) = 0.456E+03 QT (W/M2) =-0.297E-11 R (M) = 36.51 TSL (K) = 303.8 QB (W/M2) = 0.425E+03 QT (W/M2) =-0.297E-11 R (M) = 38.25 TSL (K) = 303.2 QB (W/M2) = 0.397E+03 QT (W/M2) = -0.297E-11 R (M) = 39.99 TSL (K) = 302.6 QB (W/M2) = 0.372E+03 QT (W/M2) =-0.297E-11 R (M) = 41.73 TSL (K) = 302.1 QB (W/M2) = 0.350E+03 QT (W/M2) = -0.297E-11

638 NFPA 204M -- A97 ROP

R (M) =, 43.46 TSL (K) = 301.6 QB (W/M2) = 0.330E+03 QT (W/M2) =-0.297E-11 R (M) = 45.20 TSL (K) = 301.2 QB (W/M2) = 0.312E+03 QT (W/M2) = -0.297E-11 R (M) = 46.94 TSL (K) = 300.8 QB (W/M2) = 0.296E+03 QT (W/M2) =-0.297E-11 R (M) = 48.68 TSL (K) = 299.8 QB (W/M2) = 0.286E+03 QT (W/M2) = -0,297E-11 R (M) = 50.42 TSL (K) = 294.9 QB (W/M2) = 0.390E+02 QT (W/M2) =-0.297E-11 q]ME (S)= 360,0000 LYR TEMP (K)= 443.6 LYR HT (M) = 7,31 LYR MASS (KG)=0.760E+04 FIRE OUTPUT (W) = 0.3795E+08 VENT AREA (M2) = 28.20 [,INK= 1 LINKTEMP (K) --- 590.31JETVELOCITY (M/S) = 4.317 JET TEMP (K) = 590.9 TIME LINK 1 OPENS EQUALS 41.7098 (S) R (M) = 0.00 TSL (K) = 658.1 QB (W/M2) = 0.117E+05 QT (W/M2) = -0.297E-11 R (M) = 1.74 TSL (K) = 614.7 QB (W/M2) = 0,107E+05 QT (W/M2) =-0.297E-11 v, (M) = 3.48 TSL (K) = 564.3 QB (W/M2) = 0.939E+04 QT (W/M2) = -0.297E-11 tt (M) = 5.22 TSL (K) -- 510,0 QB (W/M2) = 0.780E+04 QT (W/M2) =-0.297E-11 R (M) = 6.95 TSL (K) = 463.8 QB (W/M2) = 0.631E+04 QT (W/M2) = -0.297E-11 It (M) = 8.69 TSL (K) = 427.5 QB (W/M2) = 0.505E+04 QT (W/M2) =-0.297E-11 R (M) = 10.43 TSL (K) = 399.9 QB (W/M2) = 0.405E+04 QT (W/M2) =-0.297E-11 R (M) = 12.17 TSL (K) = 379.3 QB (W/M2) = 0.329E+04 QT (W/M2) =-0.297E-11 R (M) = 13.91 TSL (K) = 363.9 QB (W/M2) = 0.271E+04 QT (W/M2) =-0.297E-11 R (M) = 15.65 TSL (K) = 352.2 QB (W/M2) = 0.226E+04 QT (W/M2) =-0,297E-11 R (M) = 17.39 TSL (K) = 343.2 QB (W/M2) = 0.192E+04 QT (W/M2) =-0.297E-11 R (M) = 19.12 TSL (K) = 336.2 QB (W/M2) = 0.165E+04 QT (W/M2) =-0.297E-11 R (M) = 20.86 TSL (K) = 330.7 QB (W/M2) = 0.143E+04 QT (W/M2) = -0.297E-11 R (M) = 22.60 TSL (K) = 326.3 QB (W/M2) = 0,126E+04 QT (W/M2) =-0.297E-11 R (M) = 24.34TSL (K) = 322.7 QB (W/M2) = 01112E+04 QT (W/M2) =-0.297E-11 R (M) = 26.08 TSL (K) = 319.8 QB (W/M2) = 0.101E+04 QT (W/M2) =-0,297E-11 R (M) = 27.82 TSL (K) = 317.3 QB (W/M2) = 0.910E+03 QT (W/M2) =-0.297E-11 R (M) = 29.56 TSL (K) = 315.2 QB (W/M2) = 0.828E+03 QT (W/M2) =-0.297E-11 R (M) = 31.29 TSL (K) = 313.4 QB (W/M2) = 0.759E+03 QT (W/M2) =-0.297E-11 R (M) -- 33.03 TSL (K) = 311.9 QB (W/M2) = 0.699E+03 QT (W/M2) =-0.297E-11 R (M) = 34.77 TSL (K) = 310.6 QB (W/M2) = 0.647E+03 QT (W/M2) =-0.297E-11 R (M) = 36.51 TSL (K) = 309.4 QB (W/M2) = 0.602E+03 QT (W/M2) =-0.297E-11 R (M) = 38.25 TSL (K) = 308.4 QB (W/M2) = 0.562E+03 QT (W/M2) =-0,297E-11 R (M) = 39.99 TSL (K) = 307.5 QB (W/M2) = 0.527E+03 QT (W/M2) =-0.297E-11 R (M) = 41.73 TSL (K) = 306.7 QB (W/M2) = 0.495E+03 QT (W/M2) =-0.297E-11 R (M) = 43.46 TSL (K) = 306.0 QB (W/M2) = 0.467E+03 QT (W/M2) =-0.297E-11 R (M) = 45.20 TSL (K) = 305.3 QB (W/M2) = 0,442E+03 QT (W/M2) =-0.297E-11 R (M) = 46.94 TSL (K) = 304.7 QB (W/M2) = 0.419E+03 QT (W/M2) =-0.297E-11 R (M) = 48.68 TSL (K) = 303.7 QB (W/M2) = 0.402E+03 QT (W/M2) =-0.297E-11 R (M) = 50.42 TSL (K) = 295.4 QB (W/M2) = 0,597E+02 QT (W/M2) =-0.297E-11 TIME (S)= 420.0000 LYRTEMP (K)= 483.7 LYRHT (M) = 6.66 LYR MASS (KG)=0.949E+04 FIRE OUTPUT (W) = 0.5283E+08 VENT AREA (M2) = 28.20 LINK = 1 LINK TEMP (K) = 677.18JET VELOCITY (M/S) = 4.879 ,JET TEMP (K) = 677.9 TIME LINK 1 OPENS EQUALS 41.7098 (S) R (M) = 0.00 TSL (K) = 747.8 QB (W/M2) = 0.129E+05 QT (W/M2) =-0.297E-11 R (M) = 1.74 TSL (K) = 701,8 QB (W/M2) = 0.120E+05 QT (W/M2) = -0,297E-11 R (M) = 3.48 TSL (K) = 646.0 QB (W/M2) = 0.108E+05 QT (W/M2) =-0.297E-11 R (M) = 5.22 TSL (K) = 583.0 QB (W/M2) = 0.920E+04 QT (W/M2) =-0.297E-11 R (M) = 6.95 TSL (K) = 526.3 QB (W/M2) = 0.764E+04 QT (W/M2) = -0.297E-11 R (M) = 8.69 TSL (K) = 479.6 QB (W/M2) = 0.625E+04 QT (W/M2) = -0.297E-11 R (M) = 10.43TSL (K) = 443.0 QB (W/M2) = 0.510E+04 QT (W/M2) =-0.297E-11 R (M) = 12.17TSL (K) = 414.9 QB (W/M2) = 0.419E+04 QT (W/M2) =-0.297E-11 R (M) = 13.91 TSL (K) = 393.6 QB (W/M2) -- 0.347E+04 QT (W/M2) ---0,297E-11 R (M) = 15.65 TSL (K) = 377.2 QB (W/M2) = 0,292E+04 QT (W/M2) =-0.297E-11 R (M) = 17.39 TSL (K) = 364.6 QB (W/M2) = 0.249E+04 QT (W/M2) =-0.297E-11

6.~9 NFPA 204M ~ A97 ROP

R (M) = 19.12 TSL (K) = 354.7 QB (W/M2) = 0.214E+04 QT (W/M2) = -0.297E-11 R (M) = 20.86TSL (K) = 346.8 QB (W/M2) = 0.187E+04 QT (W/M2) = -0.297E-11 R (M) = 22.60 TSL (K) = 340.5 QB (W/M2) = 0.165E+04 QT (W/M2) = -0.297E-11 R (M) = 24.34 TSL (K) = 335.4 QB (W/M2) = 0.147E+04 QT (W/M2) =-0.297E-11 R (M) = 26.08 TSL (K) = 331.1 QB (W/M2) = 0.132E+04 QT (W/M2) = -0.297E-11 R (M) = 27.82 TSL (K) = 327.6 QB (W/M2) = 0.119E+04 QT (W/M2) = -0.297E-I 1 R (M) = 29.56TSL (K) = 324.6 QB (W/M2) -- 0.108E+04 QT (W/M2) =-0.297E-11 R (M) = 31.29 TSL (K) = 322.0 QB (W/M2) = 0.994E+03 QT (W/M2) =-0.297E-11 R (M) = 33.03 TSL (K) = 319.8 QB (W/M2) = 0.916E+03 QT (W/M2) =-0.297E-11 R (M) = 34.77 TSL (K) = 317.9 QB (W/M2) = 0.849E+03 QT (W/M2) = -0.297E-11 R (M) = 36.51 TSL (K) = 316.2 QB (W/M2) = 0.790E+03 QT (W/M2) =-0.297E-11 R (M) = 38.25 TSL (K) = 314.8 QB (W/M2) = 0.737E+03 QT (W/M2) = -0.297E-11 R (M) = 39.99 TSL (K) = 313.5 QB (W/M2) = 0.691E+03 QT (W/M2) =-0.297E-11 R (M) = 41.73 TSL (K) = 312.3 QB (W/M2) = 0.650E+03 QT (W/M2) = -0.297E-11 R (M) = 43.46TSL (K) = 311.3 QB (W/M2) = 0.613E+03 QT (W/M2) =-0.297E-11 R (M) = 45,20 TSL (K) = 310.3 QB (W/M2) = 0.580E+03 QT (W/M2) = -0.297E-11 R (M) = 46.94 TSL (K) = 309.5 QB (W/M2) = 0.549E+03 QT (W/M2) = -0.297E-11 R (M) = 48,68 TSL (K) = 308.3 QB (W/M2) = 0.525E+03 QT (W/M2) = -0.297E-11 R (M) = 50.42 TSL (K) = 296.2 QB (W/M2) = 0.820E+02 QT (W/M2) =-0.297E-11 TIME (S)= 480.0000 LYR TEMP (K)= 530.8 LYR HT (M) = 5.94 LYR MA,qS (KG)=0.112E+05 FIRE OUTPUT (W) = 0.7059E+08 VENTAREA (M2) = 28.20 LINK = 1 LINKTEMP (K) = 784.41JETVELOCITY (M/S) = 5.462 JETTEMP (K) = 785.2 TIME LINK 1 OPENS EQUALS 41.7098 (S) R (M) = 0.00 TSL (K) = 837.6 QB (W/M2) = 0.137E+05 QT (W/M2) =-0.297E-11 R (M) = 1.74 TSL (K) = 789.0 QB (W/M2) = 0.128E+05 QT (W/M2) = -0.297E-11 R (M) = 3.48 TSL (K) = 729.0 QB (W/M2) = 0.117E+05 QT (W/M2) = -0.297E-11 R (M) = 5.22 TSL (K) = 659.2 QB (W/M2) = 0.103E+05 QT (W/M2) =-0.297E-11 R (M) = 6.95 TSL (K) = 593.8 QB (W/M2) = 0.876E+04 QT (W/M2) =-0.297E-11 R (M) = 8.69 TSL (K) = 537.8 QB (W/M2) = 0.736E+04 QT (W/M2) =-0.297E-11 R (M) = 10.43 TSL (K) = 492.4 QB (W/M2) = 0.613E+04 QT (W/M2) =-0.297E-11 R (m) = t2.17 TSL (K) = 456.6 QB (W/M2) = 0.511E+04 QT (W/M2) =-0.297E-11 R (M) = 13.91 TSL (K) = 428.8 QB (W/M2) = 0.429E+04 QT (W/M2) =-0.297E-11 R (M) = 15.65 TSL (K) = 407.2 QB (W/M2) = 0.363E+04 QT (W/M2) =-0.297E-11 R (M) = 17.39 TSL (K) = 390.4 QB (W/M2) -- 0.311E+04 QT (W/M2) =-0.297E-11 R (M) = 19.12 TSL (K) = 377.0 QB (W/M2) = 0.270E+04 QT (W/M2) = -0.297E-11 R (M) = 20.86 TSL (K) -- 366.4 QB (W/M2) = 0.236E+04 QT (W/M2) =-0.297E-11 R (M) = 22.60 TSL (K) = 357.9 QB (W/M2) = 0.209E+04 QT (W/M2) = -0.297E-11 R (M) = 24.34 TSL (K) = 350.9 QB (W/M2) = 0.186E+04 QT (W/M2) =-0.297E-11 R (M) = 26.08 TSL (K) = 345.1 QB (W/M2) = 0.167E+04 QT (W/M2) =-0.297E-11 R (M) = 27.82 TSL (K) = 340.2 QB (W/M2) = 0.152E+04 QT (W/M2) =-0.297E-11 R (M) = 29.56 TSL (K) = 336.1 QB (W/M2) = 0.138E+04 QT (W/M2) =-0.297E-11 R (M) = 31.29 TSL (K) = 332.6 QB (W/M2) = 0.127E+04 QT (W/M2) ---0.297E-11 R (M) = 33.03TSL (K) = 329.6 QB (W/M2) = 0.117E+04 QT (W/M2) =-0.297E-11 R (M) = 34.77 TSL (K) = 327.0 QB (W/M2) = 0.109E+04 QT (W/M2) =-0.297E-11 R (M) = 36.51 TSL (K) = 324.7 QB (W/M2) = 0.101E+04 QT (W/M2) =-0.297E-11 R (M) = 38.25 TSL (K) = 322.7 QB (W/M2) = 0.944E+03 QT (W/M2) =-0.297E-11 R (M) = 39.99 TSL (K) = 320.9 QB (W/M2) = 0.886E+03 QT (W/M2) =-0.297E-11 R (M) = 41.73 TSL (K) = 319.3 QB (W/M2) = 0.833E+03 QT (W/M2) =-0.297E-11 R (M) = 43.46 TSL (K) = 317.8 QB (W/M2) = 0.786E+03 QT (W/M2) =-0.297E-11 R (M) = 45,20 TSL (K) = 316.5 QB (W/M2) = 0,743E+03 QT (W/M2) =-0.297E-11 R (M) = 46.94 TSL (K) = 315.4 QB (W/M2) = 0.705E+03 QT (W/M2) =-0.297E-11 R (M) = 48.68 TSL (K) = 313.9 QB (W/M2) = 0.673E+03 QT (W/M2) =-0.297E-11 R (M) = 50.42 TSL (K) = 297.1 QB (W/M2) = 0.108E+03 QT (W/M2) =-0.297E-11 TIME (S)= 540.0000 LYR TEMP (K)= 586.5 LYR HT (M) = 5.20

640 N't~A ~ -- A97 ItOP

LYR MASS (KG)f0.125E+05 FIRE OUTPUT ON) = 0.9073E+08 VENT AREA (M2) = 28.20 LrNK= 1 IJNKTEMt)(K) ffi 915,64JETVELOCrIY (M/S) = 6.041 JETTEMPCK) ffi 9.16.6 TIMELINK 10PENSEQUAI~ 41;7098(S) R (M) = 0,00 TSL (K)= 9`>1.9 Qs (W/~ = o.t46E+05 QT (W/M`>) ffi -0~q97E-11 R (M) = ].~74TSL (K)= 870.2 QB (W/M2) ~0A~QT (W/M2) --0.297E-11 , R (M) = 3:48TSL (K)= 806.7 QB ('W/M2) ~0.1~i~05 QT 0g/M2) =-0.297E.11 R(M) ffi 5~.`>2TSL (K);ffi 731.6"QB(W/M2) ffi 0.11~E~0~,QT (W/M~) ~-0.297E-11 R (M) = 0~5 TSL(K) -~ 66~0 Q}~ (w/try) =,0.9~E+O4QT Or/M2) =,0.~97E-11 R (M) = ~'9 TSL.(k3= ~97.00~(W/M2) f~.~Qr,~WlM2) ==O.~m_AI R 0a) =, I0.4~ TSL-tK)~= 5.44,`>QBOV/t~)= 0,70~)e4 Q,r.(w/tm) =,o.'ts~-,.H g (M) =. 12.1~ ~ (K),~'r,oLS:~ OVl)~)= o,eom~4 QT(WlM~) =~.29~11 R (M) = 13.9i TSL(10 ='467,5 QB (W/IVLg.)= 0,511F,~O4~rF (W/M2,) = =0,297F=,.11 R (M) = 15.f~-TSL (I0,= ,140.7 QB CW/M2)~= 0;43T£+04 QT (36r/M`>) =~.O;297E-II R (M) = 17.39 TSL (K) = 419.5 QB OV/M2) = 0.377E+04 QT (W/M2) = -0.`>97E-11 R (M) = 19.1`>TSL (K) = 402.6 QB (W./M2) = 0.3~gEqq~t QT (W/M2) •-0.`>97E-11 R (M) = 20~6 TSL (K) ffi ssg.00~ (W/M2) = 0.t~OE+04 QT (W/M~) =-0.t~E-11 g (M) = 2`>.60TSL 00: = 378.0 QB OV/M~) = 0.t~7~+o4 ~ (W/M~) =-0~97g-ll R (~) = `>4.~ TsL (10 = ~)S.9 QB 0V/M~) = 0.~S0~+04 QT 0V/M~) ==0.~7~H R (M) = 26.08~ (K} = 361.4 QB OV/M`>) = 0.~0WE+04 QT OV/M~) = -0.297E-I1 R (M) ffi 27~) TSL (K) = 355.0QB OV/M~) ffi 0.188E+04 QT OV/M~) ~-O.~/E-11 R (M) = ~9.56TSL (K) ffi 349.7 QB OV/M~) = 0,172E+O4 QT (W/M`>) =-0~9~TA1 R (M) ffi S~.~ TSL' (i~ ffi ~45.1 QS (W/M`>) = 0~I.'~+O4 QT 0V/M~) ffi>~l~ R(M) = ~03TSL ~)t) =. ~4LI Q~ 0V~M~) = 0.]4OE+O4 QT 0V/M~) =-0.~7E-11 R (M) = M.7"7TSL(K)'= 337,7QB (W/M~ = 0.13~4)4QT t3~/M2) =-0.29~g-II

R (M) = S.q.00'rSL (K) =- ~.6 ~ (W/M~ = 0AHE+04 QT (W/M`>) = =0.~7~4~

• " R{M)= 45,20 TSL (IO-,= .S`>S.9 Q~ (W/~M2) ffi 0.953E+03 QT (W/M2)' ~ =0.29,7E=II R (My= 46.94 TSL(K)= 3`>~.4QB (W~/M~) = 0.~85E+03 QT (W/M2) = .0.`>9~AI R (,M):= .48.68TSL {K) =~,~0~7,Q~(W/hi2) = 0.84~+0~ QT(,W,L/M2) =-0,`>97E-II R(M) •- 50.4`>TSL (I0 ffi ~98,`>QB (W/M2) = 0.138E-t03 QT OV/M2) =.0.297FA1 .TIME (S)ffi 600.0000LYRTEI~P (K)ffi f)49;9 LYR HT (M). ffi 4.57 LYR MASS (~KG)f0.i~)|g+0/~ HRE OUTPUT(W) ffi 0_q999E+08 VENT AREA (M`>) = 28.`>0 LINK= 1 LINKTEMP {K)= 10`>9.11JETVEIX)Crl~ (M/S)= 6.'>47 t JETTEMF ~K) ffi i:|,029:6" ~ ~ ! ~OPEI~ EQUAI~ 41.7098 (2" R (~) = 0.00TSL(~) = 976.80_~Og/m) =0,~`>S~O~ ~T(W/M~) -0,~T,-11 R ~M)= - 1.74 TSL (K) = 923.1 ~ 0At/M2) = 0.115~+05 QT (W~M2) ==0,297E=II- - • R 0~) = S.4S~L (K)-~a Oil'/m) =0.10~+0S ~T 0~/~). =.O.'~U

R{M) -~ 6.95 TSL'(K)' = 710~'i QB (~/M2) = 0.861E+04-QT (WIM~) ==0.`>97E-11 R (M) = 8.69 T~'L 0g0 ffi fi~4,7 Q~ (Wl,bL2) = 0.761E-t04 ~ (W/~I2) ==0.297E-1-I At 0a) -= -~i0.43 ~L (~ = .~.5 0,S .tW/m) = 0.~+O4 QT 0Vlm) =.0~n R (M) = ~9.~7TSL (K) = ~41.7 QB0V/m)= o.r)mm+o4 Q~r 0V/UL~) =.O~Y~.I~ ":R (M) = IS.9i TSL (K)~=-5o~,6~ (wIM~,) ffi o.~ QT (W/I~ ~.0,297E=i1 R (M) ffi " 15.6S TSL00 = 472.9 ~,~W,/M~) ffi 0.~t4~)E+04 QT (W/M~) = =0.`>97E-11 R (M) = 17.S9 TSL:(K) ffi 448A O ~q/m) ~O.387E+04~ (W/M~) •-0.297E-ll

R 0~)= • 19,1~ .TSL (K) =~4~8.~ QS;(w/m) = 0.S42E+O4 QT (W/m) =.0.~7~ ~ R (M) = 20:86 TSL (K) ffi 41t.9 QI] (W/M2) = 0.304E+0~ QT (W/M2) = ~0:~F~7E=II R (M) ffi 22.60 TSL (If.) = 3~8.6 QB OV/M2) = 0.27`>E+04 QT (W/,M2) = .0.297E-11 • R (M) = 24.$4 TSL (K) = 387,6 ~B (W/M`>) = 0.245E+O4 ~T (W/M`>) ffi -0.297E-11 R (M) = `>6.08TSL (I0= ~'78:4 Q B 0v/M~) ffi o.~+O4QT (w/M`>) =..0.297F_AI .R (M) ffi `>7.8`>TSL:(10 = 370.6 ~!~ (WtM~) = 0.`>03E+O4QT (W/M~[) = ;0.~7E..ll

t

641 NFPA 204M -- A97 ROP

g (M) = 29.56 TSL (K) = 364.0 0..8 (W/M2) = 0.187E+04 QT (W/M2) = -0.297E-11 R (M) = 31.29 TSL (K) = 358.4 QB (W/M2) = 0.172E+04 QT (W/M2) = -0.297E-11 R (M) = 33.03 TSL (K) = 353.5 QB (W/M2) = 0.160E+04 QT (W/M2) = -0.297E-11 g (M) = 34.77 TSL (K) = 349.2 QB (W/M2) = 0.149E+04 QT (W/M2) = -0.297E-11 R (M) = 36.51 TSL (K) = 345.4 QB (W/M2) = 0.139E+04 QT (W/M2) = -0.297E-11 R (M) = 38.25 TSL (K) = 342.1 QB (W/M2) = 0.130E+04 QT (W/M2) = -0.297E-11 R (M) = 39.99 TSL (K) = 339.2 QB (W/M2) = 0.123E+04 QT (W/M2) = -0.297E-11 R (M) = 41.73 TSL (K) = 336.5 QB (W/M2) = 0.116E+04 QT (W/M2) = -0.297E-11 R (M) = 43.46 TSL (K) = 334.1 QB (W/M2) = 0.1091/:+04 QT (W/M2) = -0.297E-11 R (M) = 45.20 TSL (K) = 332.0 QB (W/M2) = 0.104E+04 QT (W/M2) = -0.297E-11 R (M) = 46.94 TSL (K) = 330.1 QB (W/M2) = 0.986E+03 QT (W/M2) = -0.297E-11 R (M) = 48.68 TSL (K) = 328.0 QB (W/M2) = 0.9411/:+03 QT (W/M2) = -0.297E-11 R (M) = 50.42 TSL (K) = 299.4 QB (W/M2) = 0.147E+03 QT (W/M2) = -0.297E-11

References Borehamwood, UK, May 11, 1992. 1. Purser, David A., "ToxicityAssessment of Comhistion Products," Heskestad, G., "Model Studies of Automatic Smoke and Heat Vent Section 2/Chapter 8, The SFPE Handbook of Fire Protection Engineering, Performance in Sprinklered Fires," Technical Report FMRC Serial second edition, Society of Fire Protection Engineers and National No. 21935RC74-T-29, Factory Mutual Research Corp., Norwood, MA, Fire Protection Association, 1995. September 1974. 2. Peacock e t al., Software User's Guide for the Hazard I Fire Hazard Heskestad, G. and Bill, R.G., "Modeling of Thermal Responsiveness Assessment Method, Version 1.1. NIST Handbook 146, Volume I, of Automatic Sprinklers, ~ Fire Safe0 Science- Proceedings ofthe Second United States Department of Commerce, National Institute of International Symposium, Hemisphere Publishing corporation, New StancLards and Technology, 199 !. York, 1989(A), pp 603--612. 3. Cooper, Leonard Y. and [)avis, William D., Estimating the Heskestad, G. and Delidaatsios, M.A., "Update: The Initial Environment a~ut tlw Response of Sprinkler Links in Compartment Fires with Convective Flow in Fire," Fire Safety Journal, Vo115 [1989(B)], pp Draft. Curtains and Fusible Link-Actuated Ceiling Vents ~ Part 1I: User 471-475. Guide for the C~mpnter C~,de Lavent. NISTIR 89-4122, United States Department of Commerce, National Institute of Standards and Heskestad, G. "Venting Practices," in Fire Protection Handbook, Technology, July 1989. seventh edition, ed. by A.E. Cote, National Fire Protection Associa- tion, Quincy, Massachusetts, 1991, pp 6-104 to 6-116. Appendix E Referenced Publications Heskestad, G., "Fire Plumes," Section 2, Chapter 2 of SFPE E-I General. The following documents or portions thereof are Handbook of Fire Protection Engine~rin~ second edition 1995, pp. 2-9 to referenced within this guide for informational purposes only and 2-19. thus are not considered part of the recommendations of dais document. There are additional lists of references at the end of Hinldey, P.L., Hansell, G.O., Marshall, N.1L and Harrison, R., Appendices B, C, and D. "Experiments at the Multifunctioneel Trainingcentrum, Ghent, on dlelnteraction Between Sprinklers and Smoke Venting," Fire E-2 Bibliography. Research Station, Building Research Establishment, Borehamwood, Hefts, 1992. Alpert, R.L. and Ward, E.J., "Evaluation of Unsprinklered Fire Hazards," Fire Safi~. Journal. Vol 7 (1984), pp 127-143. Hinkley, P.L., "Smoke and Heat Venting," Section 2, Chapter B of SFPE Handbook of Fire Protection Engineering~ second edition, 1995, pp Babrausk~, V., "Burning Rates," Section 3, Chapter 1 of SFPE 3-160 to 3-173. Ha~uibook of Fire Protection Englneming. second edition 1995, pp. 3-2 to 5-4. Kanury, A.M., "Flaming Ignition of Solid Fuels," SFPEHandbook of Fire Protection Engineetng, DiNenno, P.J., ed., National Fire Protection Carslaw, H.S. and Jaeger, J.C., Go~utuction of Heat Solids, Oxford, Association, Boston, MA, 1988. 1959. Miller, E. E., A Position Paper to NFPA 204 Subcommittee, "Fwe DiNenno p.J., et at, e~., Table B-7 of SFPE Handbook of Fire Protection Venting of Sprinklered Properties," 1980. En~neming. second edition 1995, pp. A-35 to A-36. Nelson, H. E. and Forssell, E. W., "Use of Small-Scale Test Data in Drysdale, D., An Introduction to D3namics. Wiley, 1985. Hazard Analysis, Fire Safety Science ~ Proceedings of the Fourth International .Symposium, International Association for F'we Safety Evans, ILl). and Stroup, D.W., "Methods to Calculate the Response Science, pp 9']1-982. Time of Heat and Smoke Detectors Installed Below Large Unob- structed Ceilings," Fire Technolog3. Vol. 22, No. 1, February 1985, p. Notarianni, ILE., "Predicting the Response of Sprinlders and 54. Detectors in Large Spaces," extended abstracts from the SFPE Seminar "Large Fires: Causes and Consequences," November 16-18, Gust.-ffgson, N-E, "Smoke Ventilation and Sprinklers -- A Sprinkler 1992, Dallas, Society for Fire Protection Engineers, Boston. Specialist's View," Seminar at the Fire Research Station,

642 NFPA 204M -- A97 ROP

uintiere, J.G. and Harkleroad, M., "New Concepts for Measuring uter Models," NISTIR 4947, National Institute of Standards and ,'une Spread Properties," NBSIR 84-2945, Nation,'d Bureau of ethnology, Gaithersburg MD, 1995. Standards, Gaithersburg. MD, 1984. Waterman, T. E. et al., Fire Venting of Sptinklered Buildings, IITRI Tewarson, A, "Generation of Heat and Chemical Compounds in ProjectJ08585 for Vendng Research Committee, lIT Research Fires," Sectiort 3, Chapter 4 of A'FPEHandbook of Fire Protection Institute, Chicago, IL 60616, July 1982. Engine~Mng. second edition, 1995, pp 3-55 to 3-124. Yu, H-Z and Stavrianidis, P., "Tile Transient Ceiling Flows of Thomas, P.H. and Hinkley, P.L., "Design of Roof-Venting Systems Growing Rack Storage Fires," Fire Safety Science- Proceedings of the for Single-Story Buildings," Fire Research Technical Paper No. 10, Third International Symposium, Elsevier Applied Science, London, Department of Scientific m)d Industrial Research and Fire Offices' 1991, pp 281-290. Committee,Joint Fire Research Organization, London: H.M. Stationery Office, 1964. E-3 Computer Programs. Tien, C.L., Lee, K.Y. :rod Stretton, A.J., "Radiation Heat Transfer," Section I, Chapter 4 of SFPE Handbook of Fire Protection Engln~ring, DETACr-QS computer code... second edition 1995, pp 1-65 to 1-79. DETACr-T2 computer code..: Troup, J.M.A., Large Scale Fire Tests of Rack Stored Group A Plastiea in Retail Operation Scgnarios Protected ~. Extra Large Orific, (ELO) Sprinklers, LAVENT (Link-Activated VENTS) computer code... FMRC Serial No.J.l. 0XIR0.RR for Group A Plastics Committee, Factory Mutual Research Corp., Norwood, MA, November 1994. NOTE TO REVIEWER: Details on above three programs will be added during ROC preparation to permit document user to Walton, W.D. and Notarianni, K.E., "A Comparison of Ceiling Jet access these tools. Temperatures Me.x~ured in an Aircr,-fft H,'mger Tests Fire Widl Temperatures Predicted by the DETACT-QS and LAVENT Corn-

645