INTRODUCTION
Combustible interior finishes, which include the exposed ceiling, wall, and floor linings in buildings, are large continuous surfaces over which fires can spread. 1 These finishes, along with combustible furnishings and contents, provide the fuels that can permit the development of enclosure fires, in many cases, to flashover conditions. Depending on the flammability characteristics of the interior finishes and the fire scenarios in which they are involved, interior finishes may serve as the primary fuel driving a fire to flashover or as a secondary fuel acting as a "fuse" to spread a fire between primary fuel packages. Once flashover occurs and all exposed combustible surfaces within an enclosure ignite, interior finishes may represent the most significant fuel package contributing to the post-flashover fire because of their large surface areas and total energy content.

Because of their potential to serve as the primary fuel driving an enclosure fire to flashover, the flammability characteristics of interior wall and ceiling finishes have been regulated for more 50 years. In this paper, the roles of interior finishes in fire development are addressed. An historical of the significant fires that shaped the regulation of interior in the United States is presented. scientific understanding of fire spread over interior finishes has developed over the past 25 years, with the quantitative methods needed evaluate the fundamental flammability properties of materials. Theoretical associated with flame spread are presented in the following These concepts demonstrate to a large extent, flame spread on finishes can be viewed as a race the ignition and burnout of surface elements. Finally, a new way evaluating, and perhaps eventually the flammability characteristics combustible interior finishes is presented. This methodology provides way to move away from the current basis for the regulation of finish flammability to a more quantitative scientific basis.

HISTORICAL PERSPECTIVE
As a result of a number of major widely publicized building fires in United States during the 1940s, including the Cocoanut Grove fire2 in Boston, the LaSalle Hotel fire 3 in Chicago, and the Hotel Winecoff fire 4 in Atlanta, the role of interior finish in fire development became more widely recognized in the fire protection engineering and building regulation communities than it had been previously. That these fires occurred in buildings of so-called fireproof construction highlighted the contribution of the interior finishes and decorations to these fires.

Following the Cocoanut Grove fire in 1942, but before the LaSalle Hotel and Hotel Winecoff fires in 1946, A. J. Steiner of Underwriters Laboratories published a test method to classify the hazards of building materials. 5 As noted by Steiner 6 apparently in reference to the Cocoanut Grove fire, Public concern is aroused periodically when a rapidly spreading fire kills a large number of people or produces an extraordinary property loss. This concern prompted the development of a test method whereby the fire hazards of materials could be measured and classified with reference to the rate of spread of fire, the amount of fuel contributed to the fire, and the production of objectionable smoke while burning. This test method is now widely known as the tunnel test because of the duct-like configuration of the fire test chamber or as the Steiner tunnel test in honor of its principal developer.

In 1950, ASTME84-50T, Tentative Method of Fire Hazard Classification of Building Materials, 7 was first approved by the American Society for Testing and Materials as a tentative standard. This test method was adopted by all the model building codes in the United States and by the NFPA Building Exits Code (now the Life Safety Code), resulting in wide-spread regulation of the flame spread and smoke development of interior wall and ceiling finishes based on tunnel test results. The tunnel test remains the primary fire test method used to regulate the flammability of interior wall and ceiling finishes in the United States more than 50 years later, despite recognition of its technical shortcomings and the development of more realistic fire test methods for interior wall and ceiling finishes.

In 1950, Factory Mutual Laboratories (FM) published a report 8 describing a room fire test method to evaluate the life hazard of interior finishes. In the after-math of the large life-loss fires of the 1940s identified above, this report noted that, There is considerable agitation at the present time to write regulations governing the use of interior wall and ceiling finish materials, in the interest of reducing the life hazard in public areas where these materials are used in quantity. Before adequate and equitable regulations can be established, fire conditions constituting a life hazard will, of necessity, need to be defined and materials tested under such conditions of exposure. FM developed a test room approximately 4.2 m (14-ft.) by 6.1 m (20-ft.) by 3.7 m (12-ft.) high. FM experimented with a number of ignition sources consisting of wood cribs weighing from 2.3 kg to 13.6 kg (5 lb to 30 lb) and/or ethyl alcohol weighing from 0.2 kg to 3.4 kg (1 lb to 7.5 lb) placed in a corner of the room.

The conclusion of the FM report was that the ignition source consisting of 7.5 lb of wood and 0.75 lb of alcohol was considered to be "the most suitable exposure in this enclosure for establishing the extent to which interior wall and ceiling finish materials contributed to produce a life hazard. Several factors influenced the selection of this exposure: 1) It was of sufficient intensity to ignite materials causing them to burn and contribute to the rise of temperature within the enclosure. 2) Its location in one corner of the room adjacent to two walls produced a maximum exposure condition to wall and ceiling material. 3) It was the largest test exposure that could be used without producing a life hazard in the test enclosure by the burning of the enclosure itself. 4) This exposure ... produced a temperature of 155F (68C) at the breathing level, which was sufficiently below the chosen life hazard temperature of 300F (150C) to determine to what extent the wall or ceiling material would contribute a life hazard."

This FM report is significant for a number of reasons. It represents one of the first systematic efforts to evaluate the flammability of interior wall and ceiling finishes in an end-use configuration. It recognizes that fires located in corners represent a realistic worst-case exposure geometry for wall and ceiling linings. It establishes a selection process for ignition sources that challenge the materials being evaluated but do not overwhelm their performance. Unfortunately, the room fire test method developed by FM to evaluate the life hazard of interior finishes never gained the widespread acceptance within the building regulatory community that the tunnel test did.

Through the 1950s, the tunnel test method became more firmly entrenched as the standard for regulating the flammability characteristics of interior finish materials despite the fact that it only had tentative status under ASTM. During this period, the use of plastics in building construction also started to grow tremendously. Both Steiner9 at UL and Wilson10 at FM voiced concern with the small-scale laboratory procedures, such as ASTM D635 and D1692, and the terminology, such as "self-extinguishing," "slowburning," and "nonburning," being used to evaluate and describe the flammability performance of plastic building products.

Wilson noted that these small-scale laboratory tests are "intended solely for comparing the relative flammability of various plastic materials," and that they "are neither designed nor appropriate for the rating of plastic products as building materials." Steiner noted that "the tests which classify plastics as self-extinguishing and slow-burning do not correlate with the Fire Hazard Classification. To illustrate, some time ago a plastic which had been classified as slow-burning was subjected to the tunnel test, and the results were disastrous. The material burned so fiercely and created so much smoke and molten residue that it took days to clean up and repair our furnace. Need for action by a fire protection group is essential to control the fire hazard being created." Steiner went on to say that "the value of results of a test are dependent on their significance as related to their use, based on actual field fire experience."

Steiner was a proponent of small-scale tests as "effective instruments for development and research, as well as tools for inspection," but he also recognized their limitations: "The small-scale tests can be used in the examination of products to determine whether they provide the same properties as other materials tested in the same manner ..., but they do not provide fire protection information on the behavior of the product, or of assemblies employing it, under actual use conditions in buildings." He goes on to say that "the same fire protection engineering considerations must be given to all tests, whether small or large. The results must be representative of actual conditions, the classifications must be realistic and the requirements consistent." It is interesting to note that Steiner11 viewed the tunnel test as a large-scale test, while others12 have viewed the tunnel test as a small-scale test.

In 1961, Wilson13 reviewed a number of test methods then being used to evaluate the surface flammability of materials. Wilson noted that "None of the agencies developing these test methods has reported any relation between their test results and actual fire conditions. ... There has been nothing reported to indicate that four of the test methods (including the tunnel test) have ever been directly compared with any form of actual fire condition." Both Steiner and Wilson seemed to agree that the results of fire tests should be representative of actual conditions to be valid.

Through the 1960s, some of the technical shortcomings associated with the tunnel test began to be recognized more widely when the tunnel test was used to evaluate the flammability characteristics of newly developed foam plastic insulation products that were starting to be used in buildings. Some of these products received low flame-spread ratings in the tunnel test, yet rapidly spread fires when installed in buildings. This anomalous propensity for rapid flame-spread and fire development on exposed foam plastics despite low flame-spread ratings was demonstrated by newly developed open-corner fire tests14 that more realistically simulated the dynamics of enclosure fires than the tunnel test did. An example of this anomalous behavior is illustrated in Figure 1, which shows an open-corner fire test of a polyurethane foam insulation product with a low reported flame-spread rating.

As a consequence of the little-known Childress residence fire15 in which two children died as a result of a fire involving exposed polyurethane foam insulation installed in their home, the Federal Trade Commission (FTC) filed a proposed complaint16 against 27 respondents, including 25 manufacturers of foam plastic products and 2 trade organizations, the Society of the Plastics Industry (SPI) and the American Society for Testing and Materials (ASTM), claiming that the respondents were knowingly marketing foam plastic insulation products with misleading representations that such products were " nonburning" and "self-extinguishing" on the basis of inadequate test methods, including the tunnel test.

There was a great deal of activity during the year after the FTC proposed complaint was issued, which culminated in the "Complaint and Decision" of November 4, 1974, that included a Consent Decree signed by 24 companies and the SPI17. As part of the Consent Decree, the respondents agreed to perform many activities, which ranged from notifying all prior purchasers of foam insulation products of the dangers of the products to sponsoring and conducting research into the proper ways to protect foam plastic insulation products. These activities are summarized in the 1980 Final Report of the Products Research Committee, 18 which was formed to administer a $5 million trust fund established as part of the Consent Decree.

Between the time when the FTC Consent Decree was signed in 1974 and the PRC Final Report was issued in 1980, the use of thermal barriers to separate foam plastic insulation products from occupied spaces in buildings became the standard practice. For example, the 1973 edition of the Uniform Building Code (UBC) did not make any reference to foam plastics while the 1976 edition of the UBC included a new section (Section 1717) devoted exclusively to foam plastics. This new section generally required foam plastics to be separated from the interior of a building by a thermal barrier, such as 1/2 in. (13 mm) thick gypsum wallboard, having a finish rating of not less than 15 minutes unless specifically approved on the basis of " approved diversified tests," including "fire tests related to actual end-use such as a corner test." The details of a diversified test to be used for evaluating foam plastics were not specified until 1982.

Room fire test methods were used increasingly during the mid-to late-1970s as an alternative to the open-corner fire tests that had been used during the 1960s and early 1970s. In 1975, Underwriters Laboratories reported19 on a series of flammability studies of interior finishes that included room fire tests. In 1977, ASTM first published ASTM E603, Standard Guide for Room Fire Experiments. This document noted that, "There is no standard room fire test at the present time, and this report does not define one. It does set down many of the considerations for such a test: for example, room size and shape, ventilation, specimen description, ignition source, instrumentation, and safety considerations which must be decided upon in the design of a room fire experiment."

In 1979, Williamson and Fisher20 described efforts then underway at the University of California, Berkeley, to develop a standard room fire test method. They subsequently reported21 on their efforts to evaluate this room fire test method. They used an enclosure with dimensions of 2.4 m (8-ft.) wide by 3.7 m (12-ft.) long by 2.4 m (8-ft.) high, which was becoming the most typical enclosure size for room fire tests. This work and related work at other fire research laboratories resulted in a proposed ASTM standard room fire test method for wall and ceiling materials and assemblies22 in 1982, but this proposed standard was never adopted by ASTM.

In 1982, Uniform Building Code Standard No. 17-5, Room Fire Test Standard for Interior of Foam Plastic Systems, was first published to "detail a test method to evaluate the burning characteristics of foam plastic assemblies in a standard room configuration" and thus to serve as an approved diversified test for foam plastics under the UBC. This standard specified a room 2.4 m (8-ft.) wide by 3.7 m (12-ft.) long by 2.4 m (8-ft.) high with a doorway 0.8 m (2-ft. 6-in.) wide by 2.1 m (7-ft.) high centered in one of the 2.4 m (8-ft.) long walls of the enclosure. The ignition source specified for this test method was a 13.6 kg (30 lb) wood crib located 25 mm (1 in.) from a corner opposite the doorway opening.

During the 1980s, another series of hotel fires occurred that was reminiscent of those in the 1940s, except that these hotel fires involved modern high-rise buildings with interior finish materials that should have met modern regulatory requirements. The first of these hotel fires was the November 1980 fire at the MGM Grand Hotel23 located along the Las Vegas Strip in Clark County, Nevada. The early development of the MGM Grand fire was on the interior wall and ceiling finishes of a service side station in the deli restaurant on the casino level. 24 Once the fire flashed over the side station, it quickly enveloped the deli restaurant, feeding on the combustible interior finishes and furnishings in the restaurant. The deli restaurant then flashed over, and the fire spread into and along the length of the casino, which was roughly the size of a football field. The fire was confined to the casino level, but 85 people died as a result of this fire, with approximately 68 of the victims located on the upper floors of the high-rise portion of the building above the casino.

Three months after the MGM Grand Hotel fire, the Las Vegas Hilton Hotel25 suffered a devastating fire that killed 8 people. This fire started in the 8th floor elevator lobby in the east wing of the 30-story building. The walls and ceiling of this elevator lobby, as well as all the other elevator lobbies on floors served by these elevators, were lined with a textile carpet material. The fire in the 8th floor elevator lobby developed to flashover, then spread from the 8th floor to the 28th floor of the building via the exterior windows located in each elevator lobby. The fire did not reach the 29th floor because of an architectural detail that deflected the flame out and away from the lobby windows.

The Las Vegas Hilton Hotel fire and other less-publicized fires involving textile materials motivated the textile industry to sponsor research at the University of California, Berkeley, to evaluate how well the tunnel test predicts the performance of textile wall coverings. 26 As a result of this research project, a room fire test method for textile wall coverings was developed. This room fire test method was adopted as UBC Standard 42-2 in 1988 and is also currently designated as NFPA 265, which is referenced by the Life Safety Code and the International Building Code.

The fire at the DuPont Plaza Hotel27 in San Juan, Puerto Rico, occurred on December 31, 1986. This fire, which claimed the lives of 99 people located in the hotel's casino, started in a ballroom located across a covered foyer from the casino. The fire in the ballroom developed to flashover conditions on the new furniture being stored in the ballroom as well as on the textile wall material and foam-insulated movable partitions lining the walls of the ballroom. The combustible ceiling in the foyer also contributed to the fire development.

With the exception of the Las Vegas Hilton Hotel fire leading to the development of the room fire test method for textile wall coverings, the hotel fires of the 1980s did not inspire significant changes to interior finish requirements in the building regulations. Instead, these fires motivated the widespread use of automatic sprinkler protection in high-rise hotels and other residential and commercial buildings where sprinkler protection had not traditionally been installed.

The fire at the Station nightclub in West Warwick, Rhode Island, in February 2003 provides the latest extreme example of the role of interior finish in fire development. This fire, which claimed the lives of 100 victims and injured hundreds more, spread very quickly, primarily on the exposed convoluted flexible polyurethane foam material that had been installed on the walls and ceiling of the bandstand in the nightclub. This foam plastic product reportedly was intended for use as a packing material and therefore did not incorporate even a nominal amount of fire retardants. In light of the widespread recognition of the fire hazards associated with exposed foam plastic interior finishes and the regulation of the application of these products since the 1970s, it is difficult to comprehend how this application could have existed in 2003. It should serve as a reminder to fire safety professionals everywhere of the need for continual diligence.

Much of the focus on the Station fire has been on the lack of automatic sprinkler protection in the nightclub rather than on the exposed foam plastic interior finish that ignited so easily and spread the fire so quickly. Recent largescale experiments conducted at the National Institute of Standards and Technology28 with a wet-pipe sprinkler system and quick-response sprinklers suggest that the presence of similar automatic sprinkler protection in the Station may have significantly improved the outcome of the fire there. While automatic sprinkler protection is widely recognized to be beneficial for both life safety and property protection, it should not be considered as an acceptable trade-off for unsafe and improper installations of foam plastic materials as interior finishes. Where such installations of exposed foam plastics exist, they should be removed, regardless of the presence of automatic sprinkler protection.

FLAME SPREAD THEORY AND MODELING
Concurrent with the development of room fire test methods that more accurately portray the performance of building materials under actual fire conditions, the scientific understanding of flame spread on solid surfaces has advanced, and models of the flame spread process have been developed. For example, Quintiere29 has developed a fairly comprehensive yet relatively simple simulation model for flame spread that has been incorporated in the BRANZFIRE zone fire model. 30

The ultimate objective of research on flame spread is to be able to predict the development of fire under a full range of scenarios based on fundamental material flammability properties obtained from quantitative small-scale tests, such as the Cone Calorimeter, 31 the LIFT apparatus, 32 and the FM Fire Propagation Apparatus. 33 While considerable progress has been made, there is still a need for large-scale testing, both to verify model predictions and to evaluate performance characteristics of some materials and assemblies that cannot yet be modeled accurately, such as melting, dripping, delamination, and warping.

Consider the scenario depicted in Figure 2, which is representative of the scenario used in most room fire tests. The walls and/or ceiling of an enclosure are lined with a combustible interior finish material. A section of the lining material is subjected to an imposed heat flux from an ignition source fire, which is normally selected to represent a typical incidental fire, such as a small trash receptacle fire. 34 Such ignition sources are normally selected to realistically challenge the lining materials but not overwhelm the performance of the lining materials. In room fire tests, such ignition sources are typically located near the corner of two walls because this represents a realistic " worstcase" ignition scenario, as noted in the 1950 FM room fire tests.

The section of lining material directly behind the ignition source will be the first to ignite. The flame on this section then may spread vertically and beneath the ceiling, as indicated by the orange arrows in Figure 2, as well as laterally and downward, as indicated by the black arrows in Figure 2. In general, the upward flame spread and spread beneath a ceiling are known as wind-aided spread because the flame is spreading in the same direction as the buoyant flow of gases. This windaided spread is generally much faster than the lateral and downward spread because of the larger sections of wall and ceiling being heated by the advancing flame front.

Flame spread on a fuel surface can be considered as a sequence of ignitions, as illustrated in Figure 3. An exposure fire or the flame from a segment of the material that is already burning imposes a heat flux on a fuel element that has not yet ignited. The temperature of this fuel surface element increases under the imposed heat flux. When a fuel element reaches its ignition temperature, the flame spreads to that fuel element, and it begins to burn. With this fuel element now burning, the flame grows longer and imposes a heat flux on the next fuel surface element. Some materials, such as thin combustible surface coatings or materials adhered to noncombustible substrates, burn out relatively quickly once ignited. Other materials, such as some wood products, char and consequently have a burning rate that decreases with time. Under some exposure conditions, such materials may not burn with sufficient intensity long enough to ignite subsequent fuel elements.

Upward flame spread on a fuel surface generally requires two conditions to occur:

  1. The flame from the currently burning area of the fuel surface must extend beyond the burning area to expose the adjacent area to a heat flux high enough to ignite the adjacent area; and
  2. The heat flux must be applied long enough to ignite the adjacent fuel surface.

To satisfy the first condition, the heat release rate per unit area of the burning fuel must be high enough to cause the flame to extend beyond the burning area. In general, the length of a flame along a vertical burning surface will be proportional to its heat release rate per unit width, 35 which in turn is proportional to the heat release rate per unit area. This can be expressed as:

(1)

where xf is the length of the flame (m), measured from the base of the pyrolysis zone, xp is the length of the pyrolysis zone (m), kf is an appropriate flame length coefficient ((m/(kW/m)n), is the heat release rate per unit width (kW/m), and is the heat release rate per unit area (kW/m2) of the burning area of the fuel surface. According to this simple theory, the flame length must be greater than the pyrolysis length in order for flame spread to occur. Mathematically, this means that for flame spread to occur the following relation must hold true:

(2)

Expressed differently, this also establishes the minimum heatrelease rate per unit area for upward or wind-aided flame spread to occur:

(3)

For example, Cleary and Quintiere36 have suggested that kf =0.01 m2/kW and n=1 can be used to represent the flame length relationship, with a linear relationship between the flame length and the pyrolysis length. Based on these values, a heat release rate per unit area of 100 kW/m2 would be needed for upward flame spread to occur. Tu and Quintiere37 have also suggested that kf =0.067 m5/3/kW2/ 3 and n=2/3 are appropriate values to represent this flame-length relationship. Based on these values, the minimum heat-release rate per unit area needed for upward flame spread would be kW/m2. Note that this value is a function of the pyrolysis zone length, with larger heat-release rates per unit area needed to sustain upward flame spread for longer pyrolysis zone lengths. This is one reason why some fires may burn out after spreading some distance up a wall. These relations are shown in Figure 4.

To satisfy the second condition, the burning duration, tb , of the burning region must be greater than the ignition time, ti g , of the exposed region. More specifically, the burning duration should be evaluated as the period of time that the burning region burns at a rate sufficient to achieve the first condition. In other words, the burning duration for the second condition would be the period of time during which the heat release-rate per unit area causes the flame length to exceed the pyrolysis zone length. In general, the burning duration can be evaluated as:

(4)

where is the energy content of the fuel surface per unit area (kJ/m2), is the average heat-release rate per unit area (kW/m2), is the combustible mass per unit area (kg/m2), L is the effective heat of gasification of the combustible mass (kJ/kg), and is the net heat flux to the fuel surface (kW/m2) in the pyrolysis zone. For thermally thick surfaces, the time to ignition is generally represented, for a constant net heat flux at the fuel surface, as:

(5)

where the product is the thermal inertia of the solid surface ((kW/m2K)2s), is the difference between the ignition temperature and the initial surface temperature (K), and is the net heat flux to the fuel surface in the flame region (kW/m2). In general, the net heat flux terms in Equations 4 and 5 will not be equal to each other, but for this discussion they are assumed to be proportional to each other, i.e., . In general, the net heat flux in the pyrolysis zone is expected to be greater than the net heat flux in the flame zone, in which case the proportionality factor, , will have a value of less than one.

The burning duration expressed by Equation 4 can be equated with the ignition time expressed by Equation 5 to determine the minimum flame heat flux needed to cause ignition before burnout occurs. After some manipulation, this can be expressed as:

(6)

Equation 6 would be difficult to evaluate quantitatively, particularly since the value of the proportionality factor is not known. Nonetheless, Equation 6 is useful for a number of reasons. First, it demonstrates that there is expected to be a minimum heat flux for flame spread for materials where fuel burnout is significant. Thus, it is important that such materials be tested under exposure conditions sufficient to exceed this minimum heat flux; otherwise, anomalous test results can occur when compared with actual field performance. This behavior has been observed for textile wall coverings, as noted above. Second, Equation 6 shows how different material properties are expected to influence the minimum heat flux for flame spread. Higher thermal inertias and larger ignition temperatures would be expected to increase the minimum heat flux for flame spread, while more fuel per unit area would be expected to lower it. Third, Equation 6 demonstrates the critical nature of flame spread, where a slight change in the heat flux or in the combustible mass per unit area (e.g., another coat of paint) can spell the difference between burnout and flame propagation. Finally, Equation 6 also shows that preheating of a fuel surface will tend to decrease the minimum heat flux for flame spread by decreasing the temperature rise needed to ignite the surface.

The relatively simple theoretical analysis presented here has identified a number of material properties and environmental conditions that are expected to influence flame spread on interior wall and ceiling finishes. The material properties include:

  • The thermal inertia of the material. As shown in Equation 5, the thermal inertia of a material is directly proportional to the ignition time. Low-density materials tend to also have low thermal conductivities and consequently have very low thermal inertias. This is the primary reason why flame spread can be very rapid on exposed foam plastic products.
  • The ignition temperature of the material. Although Equation 5 shows that the time to ignition varies with the square of the ignition temperature rise, ignition temperatures for most building materials fall within a relatively small range, so differences in ignition temperatures among materials do not affect flame spread nearly as much as the order of magnitude differences in thermal inertia do.
  • The combustible mass per unit area of the material. This parameter is most significant for relatively thin coatings and materials on noncombustible substrates, such as painted or unpainted paper facers on gypsum wallboard or textile wall coverings adhered to gypsum wallboard, but is also important for materials that tend to char. Such materials are more likely to burn out locally and not spread a fire than materials with more combustible mass per unit area.
  • The ratio between the heat of combustion and the heat of gasification of the material. As demonstrated in Equation 4, this " combustibility ratio" is directly proportional to the heat-release rate per unit area of a material and consequently has an influence on the flame length as well as on the total heat-release rate of the fire, which will have an influence on the preheating of fuel surfaces as well as the potential for flame extension beyond the room of origin.
  • The heat of gasification of a material. While this property individually is not as significant as the "combustibility ratio," it does have an influence on the burning duration and consequently on the minimum heat flux for flame spread, as demonstrated by Equation 6.

The environmental parameters that influence-flame spread on wall and ceiling finishes include:

  • The heat flux imposed on the fuel surface by an exposure fire. This will influence the burning rate and the size of the fuel area first ignited, and consequently the flame length extending from this area and exposing adjacent fuel elements. By influencing the burning rate of the fuel, this parameter also influences the burning duration of this area. Ironically, a higher imposed heat flux may cause earlier burnout that a lower heat flux and consequently not cause flame spread under some conditions that a lower heat does.
  • The heat flux imposed by burning surface flames on adjacent fuel elements. This will influence the time to ignition of these adjacent fuel surface elements and consequently the speed of flame spread.
  • The gas temperatures within the enclosure. The accumulation of hot gases beneath the ceiling as a result of a fire causes preheating of the fuel surfaces in contact with the hot gases. As these surfaces heat up, the temperature rise needed to cause ignition decreases, resulting in shorter ignition times and lower minimum heat fluxes for fire spread. This effect will be most pronounced for materials that are good insulators, such as foam plastics and other low-density materials, because their good insulating qualities will result in higher gas temperatures as well as higher surface temperatures than more conductive materials will.

Based on this analysis, it should be apparent that flame spread on interior wall and ceiling finishes involves a numberof complex interrelated processes, even for relatively simple geometries, homogeneous fuels, and well-characterized exposure conditions. It is for this reason that individual fire tests of interior finish materials may not be able to characterize their performance under a full range of field-use conditions.

PROPOSED EVALUATION METHODOLOGY
In 1978, Williamson and coworkers34 suggested that "a standard room fire test could be used both as a development tool and a performance evaluation method until such time as a series of smaller, less expensive tests has been proven. Even then, new materials and systems which are different in principle from those already validated for smallscale fire tests would still require the full-scale test to show the applicability of small-scale tests." This is similar in concept to the evaluation methodology proposed here.

The evaluation methodology proposed here includes a preliminary screening/qualification step, followed by a more detailed analysis step. In the screening/qualification step, the flammability characteristics of a material are evaluated using a quantitative smallscale fire test method, such as the Cone Calorimeter or the FM Fire Propagation Apparatus. One of three outcomes will occur, depending on the performance of the material in the bench-scale test. These outcomes include:

  • the material will be screened from any further consideration if it exhibits flammability characteristics recognized
  • the material will be qualified as acceptable if it exhibits flammability characteristics considered to be fully acceptable for the anticipated conditions of use;
  • the material will need to be subjected to additional large-scale testing, such as room fire testing, if its expected field performance cannot be adequately judged based on its bench-scale performance.

This concept is illustrated in Figure 4. The specific values of the different parameters to be used for screening or qualification will need to be evaluated along with the exposure conditions under which these parameters are evaluated. The values identified in Figure 4 are intended only as examples, although they are consistent with expected performance.

REFERENCES

  1. Belles, D.W., "Interior Finish," Section 12/Chapter 3, Fire Protection Handbook, 19th Edition, National Fire Protection Association, Inc., 2003.
  2. Moulton, R.S., "The Cocoanut Grove Nightclub Fire, Boston, Massachusetts, November 28, 1942," National Fire Protection Association, 1962.
  3. McElroy, J.K., "The LaSalle Hotel Fire," Quarterly of the National Fire Protection Association, Vol. 40, No. 1, July 1946, pp. 4-18.
  4. McElroy, J.K., "The Hotel Winecoff Disaster," Quarterly of the National Fire Protection Association, Vol. 40, No. 3, January 1947, pp. 140-159.
  5. Steiner, A.J., "Fire Hazard Tests of Building Materials," Quarterly of the National Fire Protection Association, Vol. 37, No. 1, July 1943, pp. 69-78.
  6. Steiner, A.J., "Fire Hazard Classification of Building Materials," Bulletin of Research No. 32, Underwriters' Laboratories, Inc., September 1944.
  7. ASTM E-84-50T, Tentative Method of Fire Hazard Classification of Building Materials, American Society for Testing and Materials, 1950.
  8. "Life Hazard of Interior Finishes (Development of Method)," Laboratory Report No. 11760, Factory Mutual Laboratories, June 1, 1950.
  9. Steiner, A.J., "Testing for Fire Safety," Quarterly of the National Fire Protection Association, Vol. 50, No. 3, January 1957, pp. 195-204.
  10. Wilson, J.A., "A Different View on Plastic Fire Hazard Classifications," Quarterly of the National Fire Protection Association, Vol. 56, No. 2, October 1962, pp. 162-164.
  11. Steiner, A.J., "Fire Tests and Fire Protection Engineering," Quarterly of the National Fire Protection Association, Vol. 52, No. 3, January 1959, pp. 209-220.
  12. "Fire Research on Cellular Plastics: The Final Report of the Products Research Committee," 1980.
  13. Wilson, J.A., "Surface Flammability of Materials: A Survey of Test Methods and to be unacceptable for the anticipated conditions of use; Robert Brady Williamson is with the University of California at Berkeley, and Frederick W. Mowrer is with the University of Maryland. Comparison of Results," ASTM Special Technical Publication No. 301, Symposium on Fire Test Methods, American Society for Testing and Materials, 1961, pp. 61-82.
  14. Williamson, R.B., and Baron, F.M., "A Corner Test to Simulate Residential Fires," Journal of Fire and Flammability, Vol. 4, April 1973, pp. 99-105.
  15. Childress vs. Cook Paint & Varnish Company, In the Circuit Court of Carlyle County, Missouri, Case No. 11077, October 6, 1971.
  16. Federal Trade Commission Complaint on the Flammability of Plastic Products, File No. 732-3040, May 1973.
  17. Docket C-2596, Complaint and Decision, Nov. 4, 1974, printed on pages 1253 to 1279 in the "Federal Trade Commission Decisions."
  18. Fire Research on Cellular Plastics: The Final Report of the Products Research Committee, 1980.
  19. Castino, G.T., Beyreis, J.R., and Metes, W.S., "Flammability Studies of Cellular Plastics and Other Building Materials used for Interior Finishes," Subject 723, Underwriters Laboratories, Inc., June 13, 1975.
  20. Williamson, R.B., and Fisher, F.L., "Fire Growth Experiments Toward a Standard Room Fire Test," Paper No. 79-48, 1979 Fall Meeting, Western State Section of the Combustion Institute.
  21. Fisher, F.L., and Williamson, R.B., "Intralaboratory Evaluation of a Room Fire Test Method," UCB FRG 82-1, Final Report on NBS Grant NB80NADA1072, University of California, Berkeley, August 1982.
  22. "Proposed Standard Method for Room Fire Test of Wall and Ceiling Materials and Assemblies," 1982 Annual Book of Standards, Part 18, American Society for Testing and Materials, 1982.
  23. Best, R., and Demers, D.P., "Investigation Report on the MGM Grand Hotel Fire," National Fire Protection Association, January 15, 1982.
  24. Mowrer, F.W., Williamson, R.B., and Fisher, F.L., "Analysis of the Early Fire Development at the MGM Grand Hotel," Proceedings of the Second International Conference on Fire Research and Engineering, Society of Fire Protection Engineers, 1997.
  25. "Investigation Report on the Las Vegas Hilton Hotel," Fire Journal, Vol. 76, No. 1, January 1982, p. 52.
  26. Belles, D.W., Fisher, F.L., and Williamson, R.B., "How Well Does ASTM E-84 Predict Fire Performance of Textile Wallcoverings?," Fire Journal, January/February 1988, p. 24.
  27. Klem, T.J., "Investigation Report on the Dupont Plaza Hotel Fire, December 31, 1986, San Juan, Puerto Rico," National Fire Protection Association, undated.
  28. Madrzykowski, D., Bryner, N., Grosshandler, W., and Stroup, D., "Fire Spread Through a Room with Polyurethane Foam-Covered Walls", NIST Special Publication 1000-101, National Institute of Standards and Technology, Gaithersburg, MD, June 2004.
  29. Quintiere, J.G., "A Simulation Model for Fire Growth on Materials Subject to a Room-Corner Test," Fire Safety Journal, Vol. 20, 1993, pp. 313-339.
  30. Wade, C.A., "BRANZFIRE Technical Reference Guide," BRANZ Study Report 92 (revised), Building Research Association of New Zealand, 2003.
  31. ASTM E1354, "Standard Test Method for Heat and Visible Smoke Release Rates for Materials and Products Using an Oxygen Consumption Calorimeter," ASTM International, West Conshohocken, PA, 2002.
  32. ASTM E1321, "Standard Test Method for Determining Material Ignition and Flame Spread Properties," ASTM International, West Conshohocken, PA, 1997.
  33. ASTM E2058, "Standard Test Methods for Measurement of Synthetic Polymer Material Flammability Using a Fire Propagation Apparatus," ASTM International, West Conshohocken, PA, 2002.
  34. Van Volkinburg, D.R., Williamson, R.B., Fisher, F.L., and Hasegawa, H., "Toward a Standard Ignition Source," Paper No. 78-64, 1978 Fall Meeting, Western States Section of the Combustion Institute, 1978.
  35. Tu, K.M., and Quintiere, J.G., "Wall Flame Heights with External Radiation," Fire Technology, Vol. 27, No. 3, 1991, pp. 195-203.
  36. Cleary, T.G., and Quintiere, J.G., "A Framework for Utilizing Fire Property Tests," Proceedings of the 3rd International Symposium on Fire Safety Science, pp. 647-656.
  37. Tu, K.M., and Quintiere, J.G., "Wall Flame Heights with External Radiation," Fire Technology, Vol. 27, No. 3, 1991, pp. 195-203.