Fire Technology

, Volume 53, Issue 1, pp 353–373 | Cite as

Engineering Variables to Replace the Concept of ‘Noncombustibility’

  • Vytenis Babrauskas


The concept of noncombustibility evolved in the early days of building codes, before quantitative methods of measuring and assessing components of fire hazard were available. ‘Noncombustible’ lacks a technical definition of general scope, but in the US codes, which are the primary focus of this study, it is defined as a material which meets the criteria of the ASTM E136 test. The hazard variables underlying the noncombustibility concept are examined in this study. In view of today’s state of the art, it is shown that noncombustibility requirements, in most cases, constitute a misapplication of fire safety principles and that their use should be discontinued, in preference of using variables that express quantitative fire safety principles. Heat release rate (HRR) is the primary variable which correctly establishes the relevant hazard. In recent years, some regulations have been promulgated which use bench-scale HRR test results directly for this purpose. The ultimate hazard to be addressed, however, is the full-scale HRR behavior. When the hazard involves fires which may spread over surface linings, however, the full-scale HRR is not simply directly scaled to the bench-scale HRR. To quantify this hazard properly, additional properties of the material which govern the flame spread behavior need to be considered. A simple, easy-to-use method for this purpose are described, which is based solely on data obtainable from the Cone Calorimeter (ASTM E1354; ISO 5660) test. Validation of the concept against room-scale data is provided and is shown to be successful.


ASTM E136 test Building codes Cone calorimeter Flame spread Heat release rate ISO 9705 test Noncombustibility Surface linings 

List of Symbols


ceiling area (m2)


nondimensional flame spread parameter (–)

\( \dot{q}^{\prime\prime}_{avg} \)

Cone Calorimeter average HRR, test-average (kW m−2)

\( \dot{q}^{\prime\prime}_{bs} \)

Cone Calorimeter average HRR, 60-s (kW m−2)


flashover time (s)


burnout time (s)


ignition time (s)


flashover speed (m s−1)

1 Introduction: The Historical Evolution

‘Noncombustibility’ is the oldest fire safety concept that exists for regulating the reaction-to-fire behavior of materials. Within codes and regulation, it is normally applied to components of buildings or structures, but not to occupant furnishings or movable goods, even though the heat content of the latter may dwarf the former. Surprisingly very few studies have sought to establish this partition, although a few examples exist [1, 2]. The original concept of noncombustibility was never clearly defined, but the general notion was that fires are likely to more quickly become severe if building components themselves are able to spread fire. Hundreds of years ago, some European cities laid down laws concerning building materials, when disastrous conflagrations destroyed blocks of wooden structures [3]. Non-combustibility, of course, was then not quantified as a technical concept, but merely as a categorical regulation.

To understand how the concept of non-combustibility is used today, it is important to examine some of the history, with the focus being on the US regulation history. Legislation for fire safety purposes goes back to the early days of American colonies, where examples of laws barring thatched roofs and similar prescriptions can be found. New York City already had building laws in the nineteenth century that, to some extent, resembled a building code [4]. During the twentieth century, however, while jurisdictions had (and still have) the prerogative to set up their own, unique code, the concept of ‘model building codes’ became established. A model building code does not have a force of law in itself. Instead, it is developed and published with the idea that it will be adopted by various jurisdictions as a regulation; thereby, it acquires the force of law in these jurisdictions. The first model building code in the US was published in 1905 [5] by the National Board of Fire Underwriters (NBFU), and, in its later editions became named the National Building Code. This code saw significant use through the 1960s, but was no longer issued1 after 1976 [6]. The 1905 National Building Code provided for four types of building constructions:
  • Fireproof buildings

  • Frame buildings

  • Mill construction

  • Ordinary construction.

The requirements in the 1905 code for fireproof buildings were wholly prescriptive. Certain materials (iron, brick, tile, etc.) were required to be used and certain minimum thicknesses were required for various members. Even though fire resistance testing was well established at that date, there were no performance requirements in the 1905 code for fire resistance, per se. Combustibility of the three other types of buildings was not controlled. The National Building Code was revised to a minor degree in 1907 and 1909. The 1915 code, however, constituted an entirely re-written document [7]. In this edition fire resistance tests were specified for the first time. Floors had to withstand 4 h, partition walls 2 h, etc.; however, the minimum thickness requirements were also maintained, although this would seem redundant.

In case the prescriptive requirements laid upon the various building members were not explicit enough, the 1905 code provisions for fireproof buildings contains a section entitled ‘Timber in Walls Prohibited,’ which stated: “No timber shall be used in any wall of any building where stone, brick, cement concrete or iron are commonly used, except lintels, as herein provided, and brace blocks not more than eight inches in length.”

Why was it necessary to regulate the materials to be used and to exclude combustible ones, if fire testing was mandated in order to determine fire resistance? The rationale was not explained in the code itself, but it was probably a well-meant, but not well-reasoned response to some severe urban conflagrations that occurred around the turn of the twentieth century [8]. In hindsight, the following technical reasons can be seen:
  1. (1)

    Combustibles add fuel to the fire, making it longer or more intense.

  2. (2)

    Burning surfaces are especially hazardous along an escape route.

  3. (3)

    The possibility exists for flame travel within void spaces of walls or floors that are comprised of combustible materials. This might lead to fire spreading story-to-story.

  4. (4)

    Combustible materials are ‘less good’ than non-combustible ones, and must, therefore, be excluded from the ‘best’ category of buildings. (This may date to the earlier part of the nineteenth century, before it was learned that unprotected steel or iron construction generally has very little fire resistance [4]. However, by the 1890s, numerous studies had already demonstrated this fact).


Reason #4 is clearly illogical, yet it appears to have been at least partly on the minds of regulators of that era. We will examine #1 later, since data were not readily available for it earlier. Reason #3 can be dealt with rather simply. If vertical void spaces exist within walls that span story-to-story, fire can travel through that channel, even if the walls have completely adequate fire resistance (when exposed from the face). The solution is, of course, trivial: Mandate fire stops. The value of fire stops was recognized even in the 1905 code (Sec. 51). This problem is by no means limited to building systems using combustible materials. One of the most significant fires demonstrating the consequence of lacking effective fire-stopping was the One New York Plaza fire [9]. In that high rise building, the entire construction was of a non-combustible type, but thin sheets of aluminum were used as fire stops. During the fire, major story-to-story fire spread occurred, whereby it was vividly demonstrated that thin sheets of aluminum do not form an effective barrier to fire.

At the turn of the twentieth century, a distinction generally was not being made between reason #1 and reason #2. Nowadays, the issues associated with reason #1 would be called ‘fuel load,’ while reason #2 comprises ‘surface spread of flame.’ The earliest scientific investigation on spread of flame likely dates to the Paris Charity Bazaar fire of 1897 [10], but for many decades this problem could not be addressed in a quantitative engineering way. The first-ever tool to become available in the US for measuring spread of flame was the Steiner tunnel, ASTM E84 [11], developed primarily by Al Steiner of Underwriters Laboratories (UL) and first described in 1943 [12, 13]. It has been suggested that this test was developed in response to the tragic 1942 fire at the Cocoanut Grove nightclub in Boston [14]. However, while the Cocoanut Grove fire did indeed precipitate much discussion on the topic, UL had been working on the development of tunnel testing from the late 1920s [15] and the work was already completed when the Cocoanut Grove fire occurred.

2 Tests for Noncombustibility

Until 1932, ‘non-combustibility’ was viewed as a simple categorical (‘yes/no’) question, with no need of testing. This was reasonable, in view of construction materials then in use. It required no testing to know that iron, steel, stonework, bricks, concrete, and plaster were non-combustible, and that wood was combustible. Few materials for construction existed apart from the above, especially ones whose status would be unclear. Fire-retardant-treated wood did exist, and was typically viewed as a special case, a treatment which continues in the North American building codes to this day.

Scientifically, neither ‘combustibility’ nor ‘noncombustibility’ has any specific meaning. Thus, recourse must be made to common understanding of the English language, which indicates that combustible materials have the ability to burn. This concept is innately categorical, yet it makes no sense when examined closely. Materials such as glass, marble, brick, and concrete indeed cannot burn under any conditions. But iron and steel can burn very well under certain circumstances, and laboratory demonstrations of burning steel wool are often done. In fact, combustibility has been so thoroughly established that the ignition temperatures of iron and steel have been compiled [16]. Thus, it was obvious even quite early that a categorical approach is misleading. A test method cannot establish whether a material is combustible in general; at best it could establish that it is or is not under certain circumstances.

In 1932, the first test method for non-combustibility (then called ‘incombustibility’) was published in the United Kingdom as part of BS 476 [17]. The test used a vertical tube furnace wherein a specimen was exposed to a temperature that rose from ambient to 750°C at a rate of 500°C per hour. A specimen passed if it did not flame or exhibited glowing combustion. In North America, the use of the British method was initially explored [18], but eventually both a different furnace and a different test procedure were evolved [19, 20, 21]. The resulting method was first described in 1957 [22], published as a tentative standard in 1958 and finally issued in 1965 as ASTM E136 [23]. The ASTM test uses a fixed 750°C exposure temperature and originally had two criteria: (a) neither the temperature on the surface nor in the middle of the sample may rise by more than 30°C above the furnace temperature, and (b) there be no flaming from the specimen after the first 30 s of exposure.

Internationally, ISO 1182 [24] was published in 1979; this is also based on a 750°C exposure but with, again, different details and criteria; its development was described by Herpol [25]. A comparison of some materials’ performance using the ASTM and the ISO tests was provided by Gross et al. [26]. All three of these methods, while different in their details, are functionally similar: materials which can withstand exposure to 750°C (for a certain length of time, under certain conditions) are deemed to be non-combustible. This concept was a crude expediency prior to the availability of fire safety science which could characterize some more quantitative concepts.

Because the concept is linguistically categorical, yet scientifically it is not, even the criteria for such tests have been confused and non-obvious. Signs of flaming obviously constitute ‘combustion.’ But, viewed more broadly, combustion is exothermicity which may not necessarily manifest as flaming. Thus, the primary test criterion in these tests has been a temperature rise. However, temperature rise is a quantitative variable and not a categorical determination. The only way to make a yes/no answer out of it is adopt an arbitrary temperature rise as comprising a failure in the test. After some experience at testing, it was then noticed that certain materials, e.g., polystyrene foam, may not show any flaming and show only such a small temperature rise as to pass the temperature criterion [27]. Obviously polystyrene foam is the epitome of a combustible material. What was happening physically is that the specimen would rapidly and completely vaporize, without registering significant exothermicity. Thus, the ASTM criteria were expanded to add a 3rd requirement that no more than 50% of the specimen’s mass could be lost during testing. This solved the polystyrene problem, but in a way which lacks a rational relation to fire hazard assessment.

3 Noncombustibility Requirements in Current US Building Codes

Noncombustibility concepts come into US building codes in several ways. At the top level, noncombustibility is one of the criteria delineating Types of Construction. This has been treated slightly differently in each code, but, in view of its widespread usage, it is instructive to consider the 2015 International Building Code (IBC) [28]. Five Types, Type I through V, are established. With certain (numerous!) exceptions, combustible materials are prohibited from Types I and II. Highest-level fire-resistance requirements are mandated for Type I, with lower ones for Type II. Type III is the traditional ‘ordinary’ construction, where the external shell is noncombustible (typically, masonry)2 and the internal components are unregulated (typically, wood framing). Type IV is heavy timber, while Type V is wood-frame construction. The latter may have no fire-resistance requirements (Type V-B; Type V-N under some other codes) or a 1-h fire resistance requirement (Type V-A; Type V-1-h in other codes). Apart from these global requirements, there are a large number of situations where combustible materials are prohibited specifically, and the term ‘noncombustible’ is mentioned 182 times in the Code. The actual requirements are extensive and lacking in any systematic organization, an overview summary is provided in the "Appendix".

The majority of these provisions are evidently to guard against the spread of fire over surfaces where this might be hazardous to the occupants. There are a few provisions in the above list that relate to fire resistance, but most of these are contained in the Code requirements pertinent to Types of Construction (not analyzed in detail here). These are all technically unsound, since fire resistance tests are true performance-based tests, and overlaying them with ad hoc provisions against combustibility dilutes the performance basis and achieves no safety gain. A few provisions are intended to prevent the ignition, rather than the spread of fire, and these are discussed below. Some other provisions are unique, but lacking in technical merit. Fire escapes are required to be non-combustible. But this should be a moot point, since they are prohibited in new construction. No known cases exist where exterior gutters were combustible and created or exacerbated a fire hazard. Plastic light diffusers and most skylights are made of plastics, and these are materials which inevitably can burn, and generally will burn rapidly. Thus, it makes little sense to require that they be hung on, or framed with, noncombustible materials—the hazard is in the bulk item and not in the trim.

4 Related Issues

4.1 Composite Materials

The concept of non-combustibility arose initially in connection with ‘elementary’ materials, i.e., materials which are homogeneous on a macroscopic scale. This was satisfactory until about 1950, when paper-faced gypsum wallboard started becoming a popular interior finish material. The gypsum core passes the ASTM E136 test, while paper does not. As a result, codes adopted definitions to add gypsum wallboard into the noncombustible category. A typical code provision [29] for this read “Materials having a structural base of noncombustible material as defined [above], with a surfacing material not over 1/8 inch [3.18 mm] thick which has [an ASTM E84] flame spread rating of 50 or less.” Another similar provision permits surface layers with a thickness no greater than 6.4 mm and a flame spread rating ≤25 (Sec. 803.13.4). This concept is appropriate, since it explicitly recognizes that the real purpose of the provisions is to control flame spread, however the criterion is not related to an engineering metric of actual flame spread, as further discussed below.

4.2 Role of Structural Fuel Load Component

Fuel load, and especially fuel load contained within the structural elements, does not have an effect on life loss in fires. This should not be surprising, since fire deaths, if they occur, are primarily due to toxic gas inhalation and secondarily due to thermal burns. Other causes, which would include trauma injuries due to collapse, among others, comprise only 2% of the total [30]. It is exceptionally rare to have occupants perish in a fire late, at a point where the structural components themselves are being threatened—or burning, if they are combustible. Christian [31] documented this in detail. Vogel [32, 33] examined for 84 fires in multi-family buildings, the factors that led to fires becoming large, instead of being confined to the area of origin. In 6% of the cases, this was due to deficiencies or improprieties in the fire propagation performance of interior finish materials. But there were zero cases attributed to deficiencies relevant to combustibility performance of structural members. In addition, it was demonstrated through measurements in standard fire resistance tests on gypsum-protected wood-frame wall assemblies that the heat release rate from the framing members was identical to that for fire retardant treated wood, within the experimental error [34]. Other researchers showed experimentally that, for 1-h-rated untreated wood stud wall assemblies, the heat release rate was negligible during the first 30 min of the test [35].

5 Relation Between HRR and Noncombustibility Tests

As early as the 1950s it was already anticipated that heat release rate (HRR) would be the way to quantify how materials produced heat in a fire [36]. But tools for properly measuring HRR did not come into general laboratory use until the 1980s. Such tools are now widely available and are used in almost any fire test laboratory of significance [37]. Today, fire safety science recognizes that the contribution of construction materials to a fire is properly quantified as HRR. Materials which show a low HRR will not make a fire worse. It is also important to realize the context of this. There is an irreducible hazard associated with occupant fuel load. For people to carry on normal life and business, a great deal of combustible material is required, which ranges from items as diverse as beds, rugs, books, and computers. Various fuel load surveys have established typical values for such occupant goods. There is also now a reasonable collection of data on room fires run in test cells involving non-combustible construction materials. The proper question which can now be posed quantitatively—which was impossible prior to the 1980s—is whether the HRR of construction materials potentially makes the situation worse, compared to the HRR of occupant goods alone.

But analyzing the role of HRR in building fires is complicated. In the vast majority of cases, the fire will go into ventilation-limited burning shortly after flashover, and stay in that mode until fuel supply is substantially depleted. Introducing some additional fuel from the building construction will not change this HRR, but will prolong the fire, and cause more unburned pyrolysates to be ejected. This means that a simple additive summation of HRR from occupant goods and construction materials would not be correct. Instead, it becomes necessary to consider what the actual hazard is, and how it may potentially be made worse. This reveals that there are two fire hazard issues to assess: (1) whether heat release from the construction materials may make the fire growth faster (worse) in the early stages of the fire, prior to when ventilation-limited burning is attained; and (2) whether a potentially longer-burning fire will adversely affect safety. The question of a longer-burning fire is normally irrelevant. For most building occupancies, the code-required times for fire resistance are much greater than what is required to withstand a complete burnout of the contents, including any contribution of the structure itself. But there are some occupancies, however, such as libraries or archives, where the fire resistance would have to be days-long in order for the structure to withstand the imposed fire [38]. Building codes, sensibly, do not mandate such fire resistance periods. Thus, it becomes clear that the function of a specific fire-resistance time period is to allow orderly evacuation of occupants and an adequate time for the fire department to conduct their operations without the hazard of a building collapse while they are doing that. Despite urgings of Ingberg many years ago [39], required fire-resistance rating periods have never been tailored towards the expected intensity of fires. With this in mind, the focus then becomes solely the avoidance of conditions that would exacerbate the growth of fire prior to the time that it becomes ventilation-limited.

In the fully general case, assessing the role of materials in fire spread is difficult. However, it is possible to take a simple, yet conservative strategy. If the characteristics of the construction material in question are such that it will not support a propagating fire over itself during the early stages of the fire, then its contribution to hazard will be insignificant, in comparison to that of the occupant fuel load. Using modern-day fire safety engineering concepts, large-scale testing using the ISO 9705 [40] test is done. (Much of the research with the ISO 9705 test was done in Europe; but ASTM E2257 [41], NFPA 286 [42], and ICBO/UBC Std. 8-2 [43] involve nearly identical procedures.) Materials are mounted on the walls and ceiling of a fire test room and exposed to a sizable, but not overwhelming, burner fire. HRR is recorded in this test, but the primary—and simplest—criterion is room flashover. Materials which do not show flashover in this type of test will not provide a hazard to escaping occupants, if any occupant goods burning at the same time do not constitute a severe fire. Occupant goods may, in fact, burn vigorously and lead to flashover. However, post-flashover fires are not survivable by occupants and likewise it is not expected that heat contribution from interior finish materials would be zero if such conditions are reached. This is not a new idea: the same logic is expressed in ASTM E136 testing. Materials which do not ignite and burn under a 750°C exposure and are consequently deemed ‘noncombustible’ may very well burn and contribute heat to a post-flashover room fire that is burning at 1000°C to 1100°C.

5.1 HRR-Based Criteria for ‘Noncombustibility’ or ‘Degrees of Combustibility’

Since routine testing in a large-scale room fire test would be onerous, a number of researchers and regulators have considered that noncombustibility can be expressed by means of bench-scale HRR testing. The most extensive use of bench-scale HRR criteria for noncombustibility has been in the Japanese national building code, i.e., their Building Standards Law (BSL). The BSL traditionally incorporated four degrees of combustibility: noncombustible, quasi-noncombustible, fire-retardant, and combustible. For a material to be deemed anything other than combustible, testing was done using unique tests found only in Japan. In a major revision to the BSL effective 1 June 2000, evaluation is now done using the bench-scale Cone Calorimeter HRR test (ISO 5660 [44], ASTM E1354 [45]), with specimens tested at a heat flux of 50 kW m−2. The new BSL criteria [46] are shown in Table 1. The ISO TS 17431 [47] permitted as an alternative test is the ISO version of the Japanese ‘model box’ test which has not seen significant use outside of Japan.
Table 1

The New HRR-Based Japanese Regulations for Noncombustibility


Test duration (min)

Peak HRR (kW m−2)

Total heat release (MJ m−2)

Alternative permitted test





ISO 1182





ISO TS 17431

Fire retardant




ISO TS 17431

A similar regulation has been proposed in Taiwan, however, the Taiwan regulation does not contain provisions for alternative testing with ISO TS 17431. Hermesky and Murrell [48] studied this regulation and concluded that the 8 MJ m−2 total HR criterion is unworkable, since the uncertainty of the measurement is ±6.9 MJ m−2. They also examined a data set of 19 products which had been tested both in large scale, using the ISO 9705 test, and with the Cone Calorimeter. An appropriate level of performance in the ISO 9705 test is if the product never leads to flashover during the 20-min test. Two products showed flashover, despite being classified as Noncombustible under the Japanese/Taiwanese criteria. Conversely, six other products were identified which did not lead to room flashover, yet failed the Japanese/Taiwanese criteria. Obviously, if 8 out of 19 products as misclassified, a classification scheme cannot be considered adequate. Thus, despite the fact that HRR is the single most important variable for quantifying fire hazard [49], the Japanese/Taiwanese scheme of regulating by means of two separate bench-scale HRR criteria, one for peak HRR and the second for total HR, does not rank products satisfactorily according to their actual fire hazard, as found in large-scale testing.

Underwriters’ Laboratories of Canada developed a standard, CAN/ULC-S135 [50], which uses the Cone Calorimeter at a heat flux of 50 kW m−2, but with a modified sampleholder which can test some difficult sandwich-type materials. Richardson [51, 52] reported some test results using this test standard, while Carpenter and Janssens provided additional information [53]. According to the 2010 National Building Code of Canada [54], materials are accepted for use in noncombustible construction if, when tested according to CAN/ULC-S135, their total HR and total smoke production over a 15 min period do not exceed 3 MJ m−2 and 1 m2, respectively. In view of the findings of Hermesky and Murrell, it must be concluded that a 3 MJ m−2 total HR criterion is even more unrealistic than the 8 MJ m−2 value in the Japanese and Taiwanese regulations.

Researchers in Australia working on revising the Australian building code recommended [55] that materials be treated as noncombustible if they are tested in the ISO 9705 room/corner test and do not cause the room to flash over. However, the research proposal was not adopted by the code [56], which has provisions broadly similar to those of the IBC.

The New Zealand Building Code had been requiring that, in some cases, exterior claddings be non-combustible. This is an exceedingly onerous requirement and most other countries utilize some milder type of flame spread testing, instead. In response to this, Wade [57] suggested that three degrees of combustibility be established, using Cone Calorimeter tests at a heat flux of 50 kW m−2. This was further evolved by Cowles and Soja [58], who proposed a series of three classes (Table 2). Class A would replace traditional noncombustibility requirements, while the other two classes would represent materials with greater heat production. The A and B classes were subsequently adopted by the New Zealand Building Code [59]. According to the database of Hermesky and Murrell, these criteria are also not satisfactory. Out of 19 products, 2 products pass the New Zealand criteria for Class A rating, yet led to room flashover. Conversely 9 products fail the New Zealand criteria, yet do not lead to flashover. With 11 out of 19 products being misclassified, this set of criteria is clearly even less successful than the Japanese/Taiwanese one. However, at least it should be noted that a 25 MJ m−2 total HR criterion is not unworkable on grounds of measurement uncertainty, the way that the 8 MJ m−2 Japanese/Taiwanese or the 3 MJ m−2 Canadian criteria are.
Table 2

HRR-Based Degrees of Combustibility Proposed in New Zealand


Test duration (min)

Peak HRR (kW m−2)

Total heat release (MJ m−2)













The ‘Euroclass’ used in the European Union [46] effectively includes some HRR-based material classifications. It is a highly complex system developed by regulatory fiat with only limited post hoc technical consideration of actual fire hazard. Classes A2 through D involve some HRR parameters. Somewhat similar to the Japanese system, in addition to a noncombustible class (A1) there is a class (A2) which may be considered quasi-noncombustible. For the A1 class, ISO 1182 governs. Classes B through D contain HRR criteria, without reference to ISO 1182. But class A2 may be complied with either by use of ISO 1182 with relaxed criteria (temperature rise < 50°C, flaming up to 20 s permitted), or else by HRR-based criteria.

Apart from the regulatory schemes, there have been a few physics-based research studies. Alpert and Khan [60] proposed that materials having a test-average HRR of \( \dot{q^{\prime\prime}}_{avg} \le 75\,{\text{kW}}\,{\text{m}}^{ - 2} \), when tested at an irradiance of 50 kW m−2, will not show sustained combustion. Ahonen et al. [61] equipped the ISO 1182 apparatus with oxygen consumption calorimetry instrumentation in order to determine the actual heat release being exhibited by specimens in that test. They noted that the test had “poorly controlled parameters resulting in a modest reproducibility” and showed that heat release measurements were feasible and, in fact, would be preferred. However, they found poor correlation between heat release rate and furnace temperature rise. Such work was not continued by other investigators, since measurements using the Cone Calorimeter were found to be more reliable.

6 Surface Spread-of-Flame Based Criteria

For flame spread over combustible surfaces to result in a fire hazard, there has to be a relatively rapid fire propagation over a sizable surface area. If the flame spread rate is slow, or if the extent of flame spread is small, then flame spread hazard will be insignificant. Scientists categorize flame spread as wind-aided or opposed, depending on whether flame spreads in the same direction as the air is moving, or opposite thereto. Significant hazard may be expected only if the flame spread is wind-aided, since opposed flow flame spread is always much slower. But unless the architectural details are prescribed to a high degree (which will not be the case, as concerns general regulation by a building code), the same material may experience both wind-aided and opposed flow flame spread. Consequently, it is the larger hazard—wind-aided flame spread—which must be directly addressed. It is not actually necessary to run a flame spread test to characterize the flame spread behavior, since computation of the flame spread behavior can be obtained by getting data from the Cone Calorimeter test, or a similar small-scale test where the HRR is determined, along with ignition time for exposure to a radiant flux.

In 1984, Babrauskas [62] provided the first expression for doing this. The hazard of fire spread over surface linings of a room can be tested in a large-scale room/corner fire test (e.g., ISO 9705). The primary hazard characteristic that emerges from this is the time to flashover. The longer the time to flashover, the less hazardous is the lining material/product. If the material never leads to flashover, this is the optimal outcome. To quantify this using bench-scale data, it was found that the variable v fo was useful:
$$ v_{{fo}} \propto \frac{{\dot{q}^{{\prime \prime }} _{{bs}} }}{{t_{{ig}} }} $$
where v fo = flashover speed (m s−1); \( \dot{q}^{{\prime \prime }} _{{bs}} \) = bench-scale HRR (kW m−2); and t ig = ignition time (s). In this original proposal, the bench-scale HRR was a 60 s average measured in the Cone Calorimeter when using a specimen irradiance of 25 kW m−2. Figure 1 shows that a reasonable prediction was achieved and it was possible to distinguish between products which showed a flashover hazard, versus ones which did not. The latter were characterized by \( \frac{{\dot{q}^{{\prime \prime }} _{{bs}} }}{{t_{{ig}} }} < 0.3 \). The role of ignition time was essential in this formulation, since flame spread can be viewed as effectively being a series of ignitions. More elaborate flame spread theories show that the same thermophysical variables which govern ignitability behavior are also the ones which govern wind-aided flame spread [63]. The correlation, as originally proposed, had limitations, however, since the 25 kW m−2 irradiance used was too low and some products which will show flame spread either do not ignite at this low heat flux, or else give erratic results.Later, Baroudi and Kokkala [64] provided an elegant, mathematically sophisticated theory for evaluating the potential for sustained flame propagation, but their work was not easily adoptable for building code use. Instead, it becomes useful to consider the analysis provided by Cleary and Quintiere [65], which gives a relationship that has good predictivity, yet is simple to use. They defined a parameter b:
$$ b = 0.01{\kern 1pt} \,{{\dot{q}}^{\prime\prime}}_{avg} - 1 - \frac{{t_{ig} }}{{t_{b} }} $$
where \( \dot{q}^{{\prime \prime }} _{{avg}} \) = average HRR (kW m-2), t ig = ignition time (s), and t b = duration of flaming (s). According to the theory, if b ≤ 0, then there is not a significant hazard of sustained, wind-aided flame propagation. This concept was examined using large-scale tests under the ISO 9705 method. The large-scale results [66, 67, 68] and corresponding Cone Calorimeter data (at an irradiance of 50 kW m−2) [69, 70, 71, 72], showed the principle to be correct. The author’s re-analysis shows that if b < −0.4, then when such materials are used to line the walls and ceiling of the ISO 9705 room/corner test, the severe exposure of the 300 kW burner fire in that test does not result in flashover of the room (Fig. 2). For the conditions prevailing in that test, this means that no significant flame spread occurs along materials that have b < −0.4. In can be seen in Fig. 2 that the original b = 0 criterion represents the mean trend of the data. However, there is some scatter and by adopting the criterion as b < −0.4, a conservative bound to the experimental data is obtained, which would be more appropriate for regulatory purposes.
Figure. 1

Validation of the original Babrauskas expression [62] (1984) for flashover hazard of combustible wall/ceiling linings

Figure. 2

Prediction of non-propagating fires using the Cleary/Quintiere b parameter

With the above results in mind, it is useful to reexamine regulatory concepts for using HRR criteria to see how consistent they are with the flame propagation/flashover hazard as discussed in this section. Gypsum wallboard is a product which will not propagate fire under reasonable fire scenarios (Table 3). Thus, it should be accepted as a low-hazard benchmark product. Yet, Table 3 shows that a criterion of 200 kW m−2 for peak HRR value would eliminate some acceptable grades of this product. Consequently, if peak HRR is used as a criterion, a higher value would have to be selected. The Japanese scheme of establishing three different levels of hazard, based solely on the time of test evaluation is also problematic. FR particleboard, FR chipboard, and FR PVC are three materials which would not meet the criterion of 8 MJ m−2 total heat release during a 20-min evaluation period. In the US building codes, FR wood products are permitted for use in noncombustible construction provided they obtain Class A (Flame Spread Index ≤ 25) when tested according to a 30 min extended version of the ASTM E 84 Steiner Tunnel test. In panel form these products can also be used as interior finish without much restriction. Thus, the b-parameter criterion provides a reasonable treatment of practical materials, whereas regulations based solely on controlling the peak HRR and total HR do not.
Table 3

HRR Parameters of Tested Materials Which Did not Lead to Flame Propagation/Flashover in Large-Scale ISO 9705 Room/Corner Test [66, 67, 68]



Peak HRR (kW m−2)

Total HR (MJ m−2)

Approx. duration (min)


Gypsum board





Paper covering on GWB





Gypsum board





Melamine-faced non-comb. Board





Plastic-faced steel sheet on mineral wool





FR particleboard





FR chipboard





Paper-faced GWB










3-Layered FR polycarbonate panelc




aNegligible HR until 8 min; roughly 8 MJ m−2 at 10 min time

bRoughly 5 MJ m−2 at 5 min time

cThis material is likely to have gone to flashover, had a slightly different burner or mounting method been used

The building codes already recognize the role of flame spread, but only in a partial and obsolete way. As discussed above, ASTM E84 testing is provided for composite products needing to be qualified as noncombustible. The E84 test is also used in numerous sections of the Code to regulate the flame spread of exposed materials directly. However, the E84 test is not an engineering test, and its results do not reflect flame spread hazards in actual buildings [73, 74]. The IBC also contains provisions (Sec. 806.1) that draperies and decorative trim in certain occupancies must pass NFPA 701 [75] or be noncombustible. Similar requirements pertain to membranes of air-supported structures (Sec. 3102.3.1). NFPA 701 is an ancient test which is used to examine the flame spread characteristics of textile type materials.

7 Necessary exclusions: High-temperature environments and hazardous chemicals

As can be seen in the list above, a few sections of the Code deal with two specific situations where prohibition of flammable materials is not due to flame spread issues or excessive conservativeness.
  1. (1)

    High-temperature environments. Areas in the vicinity of flues, chimneys, fireplaces, and similar high-temperature environments pose a danger of ignition if combustible materials, specifically, wood products, are used [76]. Thus, the Code takes the simplest and most reasonable approach here by prohibiting such uses, and engineering analyses are not needed.

  2. (2)

    Hazardous chemicals environments. The specific hazards of various hazardous chemicals are widely varied, so generalized Code prescriptions can be simplistic, but a conservative stance is appropriate. Solid oxidizer chemicals (e.g., calcium hypochlorite, ammonium nitrate) are a special situation where presence of fuels, such as wood or paper, can either create a hazard, or exacerbate an incident, for well-known chemical reasons [16, 77]. For example, it has been documented that ammonium nitrate dust impregnated into wood materials can increase the burning hazard [78].


Consequently, Code regulations in these two areas do not need to be replaced by engineering methods. However, the prescriptive Code provisions are not necessarily adequate. Sec. 415.8.4 specifies that “Floors in storage areas for organic peroxides, oxidizers, pyrophoric materials, unstable (reactive) materials and water-reactive solids and liquids shall be of liquid-tight, noncombustible construction.” This is necessary, but not sufficient. Under the Code, paper-faced gypsum board is accepted as a noncombustible material. While the paper presents no significant risk in normal usage, such as walls along an escape route, it would present an unnecessary hazard with regards to interaction with oxidizer chemicals, since paper reacts exothermically with these chemicals [79]. Thus, for materials directly proximate to oxidizer chemicals, it would be more prudent if the Code were to require that materials qualify only under ASTM E136 and not under the ASTM E84 alternative.

8 Conclusions and Recommendations

Noncombustibility was one of the earliest fire safety concepts and goes back well over a hundred years, to the era before any fire safety engineering (FSE) tests were developed. By the mid-twentieth century, several versions of noncombustibility tests were developed, but these are a categorical (pass/fail) tests, and the results from them are not usable for any FSE design or evaluation purposes. However, a plethora of fire safety engineering tests exist at this time, and it is urged that ancient provisions be supplanted by appropriate tests. To do this, noncombustibility objectives have to be reformulated in terms of modern FSE concepts. In terms of the modern engineering understanding, four purposes for noncombustibility requirements were evidently envisioned historically, albeit not ever explicitly stated: (1) prevention of fire spread within void spaces; (2) avoidance of excessive fuel load; (3) avoidance of flame spread hazard on the surfaces within the occupied space; and (4) prevention of ignition of materials adjacent to heat sources. The first purpose is simply and prescriptively achieved by requiring adequate firestopping provisions, a requirement already existing in any current building code. Avoidance of excessive fuel load has been shown to involve a misunderstanding of fuel loads in buildings and the role of occupant fuel load within the occupied space, versus potential fuel load not released during the time period when occupants are still able to escape the fire.

Flame spread limitation is the FSE objective which should become the replacement for traditional prescription of noncombustibility. Bench-scale HRR requirements, instead, has been the direction in which new code initiatives have tended to move. Such HRR requirements capture certain features of the hazard, but not fully enough to be an adequate solution. Flame spread and HRR are very closely related in fires involving burning wall/ceiling surfaces, but it is important to emphasize that the HRR which must be considered here is the large-scale HRR. Bench-scale HRR and large-scale HRR in such situations are not simply and linearly related. Instead, both the bench-scale HRR and a variable describing the flame spread propensity must be obtained in order to characterize either the flame spread or the HRR of the large-scale situation.

Existing engineering methods for relating the full-scale flame spread and HRR hazard to bench-scale test results were reviewed. It was found that the Cleary/Quintiere method is suitable due to relatively simple data analysis requirements and its good predictivity for actual large-scale room fire test results. It is recommended that this method be adopted in building codes and that it be used to replace the current provisions for noncombustibility. This recommendation is applicable to all building codes, not solely the IBC which has been the detailed focus of this paper.

The only current provisions which should remain are the ones which deal with use of materials in the vicinity of heat sources, e.g., fireplaces and chimneys. In these situations, it is not flame spread but ignitability that is the valid concern. Similarly, requirements for noncombustible materials in propinquity to certain stored chemicals are valid and should be continued. General prohibitions against use of combustible materials as a part of fire-rated assemblies have no legitimate engineering basis and should be abolished. A limited trend is emerging in this direction due to currently increasing use of midrise wood construction [80].


  1. 1.

    NBFU evolved into the American Insurance Association (AIA). After issuing the final edition in 1976, the latter sold the name—but not the text—to Building Officials and Code Administrators International, Inc., an entity which dissolved by merging into the International Code Council in 2003. Prior to the merger, several editions of the BOCA National Building Code were issued, but these had no relation to the NBFU code apart from the name.

  2. 2.

    Fire-retardant-treated (FRT) wood is also permitted as an exception.

  3. 3.

    The Code also uses noncombustibility concepts for items other than construction materials, for example, warehouses storing noncombustible liquids are treated differently from those storing combustible liquids.

  4. 4.

    Except that FRT wood, or ordinary wood no less than 25 mm thickness are permitted.

  5. 5.

    Except that asphalt paving is permitted at ground level.

  6. 6.

    Except that ordinary wood is permitted, of unrestricted thickness.

  7. 7.

    Except that ordinary wood is permitted, if it is of at least 45 mm thickness.

  8. 8.

    Fire-retardant-treated (FRT) wood and plastics meeting certain HRR requirements are also permitted as an exception.

  9. 9.

    Except that plastic pipe at least as heavy as Schedule 40 is permitted.

  10. 10.

    Fire-retardant materials is permitted as an alternative, but the term “fire-retardant materials” is not defined.

  11. 11.

    Fire escape provisions in the IBC existed only through the 2012 edition.



The contributions of Marc L. Janssens to this study are gratefully acknowledged.


  1. 1.
    Claret AM, Andrade AT (2007) Fire load survey of historic buildings: a case study. J Fire Prot Eng 17:103–112CrossRefGoogle Scholar
  2. 2.
    Narayanan P (1995) Fire severities for structural fire engineering design (Study Report No. 67), BRANZ, PoriruaGoogle Scholar
  3. 3.
    Bird EL, Docking SJ (1949) Fire in buildings. Adam & Charles Black, LondonGoogle Scholar
  4. 4.
    Babrauskas V (1976) Fire endurance in buildings (Ph.D. dissertation). University of California, BerkeleyGoogle Scholar
  5. 5.
    Building Code Recommended by the National Board of Fire Underwriters. James Kempster Printing Co., New York (1905)Google Scholar
  6. 6.
    The National Building Code (1976) Recommended by the American Insurance Association. American Insurance Association, Engineering and Safety Service, New YorkGoogle Scholar
  7. 7.
    Building Code Recommended by the National Board of Fire Underwriters, New York (1915) An ordinance providing for fire limits, and regulations governing the construction, alteration, equipment, repair or removal of buildings or structures. National Board of Fire Underwriters, New YorkGoogle Scholar
  8. 8.
    Lyons PR (1976) Fire in America, NFPA, BostonGoogle Scholar
  9. 9.
    Powers WR (1970) One New York Plaza Fire, New York, August 5, 1970, The New York Board of Fire Underwriters, New YorkGoogle Scholar
  10. 10.
    Sachs EO (1898) The Paris Charity Bazaar Fire (Publications of The British Fire Prevention Committee, No. 3), The British Fire Prevention Committee, LondonGoogle Scholar
  11. 11.
    Standard test method for surface burning characteristics of building materials (ASTM E84). ASTM, West ConshohockenGoogle Scholar
  12. 12.
    Steiner AJ (1943) Method of fire–hazard classification of building materials. ASTM Bull 19–22Google Scholar
  13. 13.
    Steiner AJ (1944) Fire hazard classification of building materials (Bull. of Research No. 32), Underwriters’ Laboratories, Inc., ChicagoGoogle Scholar
  14. 14.
    Cocoanut Grove Night Club Fire, Boston, November 28, 1942, National Fire Protection Association, Quincy (1943)Google Scholar
  15. 15.
    Steiner AJ (1964) Historical background of the Fire Protection Department, Underwriters’ Laboratories, Inc., pp 15–40. In: Brandon M, eds., The Departments of Underwriters’ Laboratories, Inc., Underwriters’ Laboratories, ChicagoGoogle Scholar
  16. 16.
    Babrauskas V (2003) Ignition handbook. Fire Science Publishers/Society of Fire Protection Engineers, IssaquahGoogle Scholar
  17. 17.
    Fire resistance, incombustibility, and noninflammability of building materials and structures (BS 476). British Standards Institution, London (1932)Google Scholar
  18. 18.
    Setchin NP, Ingberg SH (1945) Test criterion for an incombustible material, Proc ASTM 45:866–877Google Scholar
  19. 19.
    Shorter GW, Sumi K (1952) ASTM combustibility tests (Research Report 6). Division of Building, Research, National Research Council Canada, OttawaGoogle Scholar
  20. 20.
    Sumi K (1955) Comparison of combustibility test procedures (DBR Report No. 43), Division of Building, Research, National Research Council Canada, OttawaGoogle Scholar
  21. 21.
    Setchkin NP (1952) Combustibility tests of 47 ASTM material samples (NBS 1454), U. S. National Bureau Standards, WashingtonGoogle Scholar
  22. 22.
    ASTM (1957) Proposed method of test for defining noncombustibility of building materials. ASTM Bull 33–34Google Scholar
  23. 23.
    Noncombustibility of elementary materials (ASTM E136-65), ASTM, Philadelphia (1965). In the 1979 edition, this standard was renamed as: behavior of materials in a vertical tube furnace at 750°C.Google Scholar
  24. 24.
    Fire tests—building materials—non-combustibility test (ISO 1182), ISO, GenevaGoogle Scholar
  25. 25.
    Herpol GA (1972) Noncombustibility—its definition, measurement, and applications. In: Ignition, heat release, and noncombustibility of materials (ASTM STP 502), pp 99–111. ASTM, PhiladelphiaGoogle Scholar
  26. 26.
    Gross D, Lindauer RA, Willard R (1973) Evaluation of two test methods for noncombustibility: ASTM E136 and ISO R1182 (Report of Test No. FR 3846), U. S. National Bureau Standards, WashingtonGoogle Scholar
  27. 27.
    Standard test method for behavior of materials in a vertical tube furnace at 750°C (ASTM E136-2012). ASTM Intarnational, West Conshohocken (2012)Google Scholar
  28. 28.
    International Building Code. International Code Council, Inc., Country Club Hills (2006)Google Scholar
  29. 29.
    Uniform Building Code, 1970 edition, International Conference of Building Officials, Whittier CA (1970).Google Scholar
  30. 30.
    Hall JR, Jr. (2011) Fatal effects of fire, NFPA, QuincyGoogle Scholar
  31. 31.
    Christian WJ (1974) The effect of structural characteristics on dwelling fire fatalities. Fire J 68(1)22Google Scholar
  32. 32.
    Vogel BM (1977) A study of fire spread in multi-family residences: the causes, the remedies. Build Stand 46(2)56–59, 80Google Scholar
  33. 33.
    Vogel BM (1977) A study of fire spread in multi-family residences: the causes—the remedies (NBSIR 76–1194), U.S. National Bureau Standards, GaithersburgCrossRefGoogle Scholar
  34. 34.
    Chamberlain D, King E (1987) Heat release rates of construction assemblies by the substitution method (Technical Report No. 9). American Forest & Paper Associationn, WashingtonGoogle Scholar
  35. 35.
    Tran H, White R (1990) Heat release rate from wood wall assemblies using oxygen consumption method. In: Fire and polymers: hazards identification and prevention (ACS symposium series 425). American Chemical Society, Washington, pp 411–428Google Scholar
  36. 36.
    Thompson NJ, Cousins EW (1959) The FM construction materials calorimeter, NFPA Q 52:186–192Google Scholar
  37. 37.
    Babrauskas V, Grayson SJ (1992) Heat release in fires. E&FN Spon, LondonGoogle Scholar
  38. 38.
    Sharry JL, Walker E (1974) Military personnel records center fire, Overland, Missouri. Fire J 68:3:5–9, 65–70Google Scholar
  39. 39.
    Ingberg SH (1928) Tests of the severity of building fires. Q NFPA 22(3)43–61Google Scholar
  40. 40.
    International Standard—fire tests—full scale room test for surface products. ISO 970. ISO, Geneva (1993)Google Scholar
  41. 41.
    Standard test method for room fire test of wall and ceiling materials and assemblies (ASTM E2257). ASTM Interantional, West ConshohockenGoogle Scholar
  42. 42.
    Standard methods of fire tests for evaluating contribution of wall and ceiling interior finish to room fire growth (NFPA 286), National Fire Protection Assocaiation, QuincyGoogle Scholar
  43. 43.
    Standard test method for evaluating room fire growth contribution of textile wall covering (UBC Standard 8-2), Uniform Building Code, Vol. 3, International conference of Building Officials, Whittier (1997)Google Scholar
  44. 44.
    International Standard—fire tests—reaction to fire—part 1: rate of heat release from building products (Cone Calorimeter method). ISO 5660-1:2015(E). ISO, Geneva (2015)Google Scholar
  45. 45.
    Standard test method for heat and visible smoke release rates for materials and products using an oxygen consumption calorimeter (E1354). ASTM International, West ConshohockenGoogle Scholar
  46. 46.
    Hakkarainen T, Hayashi Y (2001) Comparison of Japanese and European Fire Classification Systems for Surface Linings. Fire Sci Technol (Jpn) 21(1)19–42CrossRefGoogle Scholar
  47. 47.
    Fire Tests—Reduced-scale model box test (ISO TS 17431), ISO, Geneva (2006)Google Scholar
  48. 48.
    Hermesky M, Murrell J (2009) Cone calorimeter—a cautionary tale. J ASTM Intl. 6:8CrossRefGoogle Scholar
  49. 49.
    Babrauskas V, Peacock RD (1992) Heat release rate: the single most important variable in fire hazard. Fire Saf J 18:255–272CrossRefGoogle Scholar
  50. 50.
    Standard test method for the determination of combustibility parameters of building materials using an oxygen consumption calorimeter (cone calorimeter), CAN/ULC-S135-04, Underwriters’ Laboratories of Canada, Toronto (2004)Google Scholar
  51. 51.
    Richardson L. R, Brooks ME (1991) Combustibility of building materials. Fire Saf J 15:131–136Google Scholar
  52. 52.
    Richardson LR (1994) Determining degrees of combustibility of building materials: National Building Code of Canada. Fire Mater 18:99–106CrossRefGoogle Scholar
  53. 53.
    Carpenter K, Janssens M (2005) Using heat release rate to assess combustibility of building products in the cone calorimeter. Fire Technol 41:79–92CrossRefGoogle Scholar
  54. 54.
    National Building Code of Canada (2010) Canadian Commission on Building and Fire Codes. National Research Council Canada, OttawaGoogle Scholar
  55. 55.
    Dowling VP, Blackmore JM (1998) Fire performance of wall and ceiling linings (Final Report Project 2, Stage A). Fire Code Reform Centre, SydneyGoogle Scholar
  56. 56.
    NCC 2015—National Building Code of Australia (2015) Australian Building Codes Board, Canberra, ACTGoogle Scholar
  57. 57.
    Wade C (1995) Fire performance of external wall claddings under a performance-based building code. Fire Mater 19:127–132CrossRefGoogle Scholar
  58. 58.
    Cowles GS, Soja E (1999) Flame spread classification method for exterior wall claddings. In: Interflam’99. Interscience Communications Ltd., London, pp 1021–1031Google Scholar
  59. 59.
    Approved Document for New Zealand Building Code (2006) Fire safety. Clauses C1, C2, C3, C4. Building Industry Authority/Standards New Zealand, Wellington (2001). Reissued as: Compliance Document for New Zealand Building Code. Clauses C1, C2, C3, C4. Fire Safety. Dept. of Building and Housing, WellingtonGoogle Scholar
  60. 60.
    Alpert RL, Khan MM (2003) A new test method for rating materials as noncombustible. In: Fire safety science—proceedings of 7th international symposium, International Association for Fire Safety Science, pp 791–802Google Scholar
  61. 61.
    Ahonen A, Weckman H, Yli-Penttilä M (1985) Application of oxygen-consumption calorimetry to non-combustibility testing. Fire Mater 9:135–144CrossRefGoogle Scholar
  62. 62.
    Babrauskas V (1984) Bench-scale methods for prediction of full-scale fire behavior of furnishings and wall linings (SFPE Technical Report 84-10). Society of Fire Protection Engineers, BostonGoogle Scholar
  63. 63.
    Drysdale DD (2011) An introduction to fire dynamics, 3rd edn. Wiley, ChichesterCrossRefGoogle Scholar
  64. 64.
    Baroudi D, Kokkala M (1992) Analysis of upward flame spread, Project 5 of the EUREFIC Fire Research Programme (VTT Publications 89). VTT, Technical Research Centre of Finland, EspooGoogle Scholar
  65. 65.
    Cleary TG, Quintiere JG (1991) Framework for utilizing fire property tests. In: Proceedings of 3rd international symposiums on fire safety science, International Association for Fire Safety Science, pp 647–656. Elsevier Applied Science, New YorkGoogle Scholar
  66. 66.
    Sundström B (1986) Full scale fire testing of surface materials (SP-RAPP 1986:45), BoråsGoogle Scholar
  67. 67.
    Söderbom J (1991) EUREFIC—large scale tests according to ISO DIS 9705. Project 4 of the EUREFIC Fire Research Programme (SP Report 1991:27), BoråsGoogle Scholar
  68. 68.
    Thureson P (1996) fire tests of linings according to room/corner test, ISO 9705 (Client report 95R22049), BoråsGoogle Scholar
  69. 69.
    Tsantaridis L, Östman B (1989) Smoke, gas and heat release data for building products in the cone calorimeter (Rapport I 8903013), Trätek, StockholmGoogle Scholar
  70. 70.
    Thureson P (1991) EUREFIC—cone calorimeter test results. Project 4 of the EUREFIC Fire Research Programme (SP Report 1991:24), BoråsGoogle Scholar
  71. 71.
    Dillon SE, Kim WH, Quintiere JG (1998) Determination of properties and the prediction of the energy release rate of materials in the ISO 9705 room-corner test. Appendices. (NIST-GCR-98-754), Nat. Inst. Stand. & Technol., GaithersburgGoogle Scholar
  72. 72.
    Tsantaridis L, Östman B (1999) Cone calorimeter data and comparisons for the SBI RR products (Report I 9812090), TrätekGoogle Scholar
  73. 73.
    Babrauskas V, White JA. Jr., Urbas J (1997) Testing for surface spread of flame: new tests to come into use. Build Stand 66(2):13–18Google Scholar
  74. 74.
    Babrauskas V, Lucas D, Eisenberg D, Singla V, Dedeo M, Blum A (2012) Flame retardants in building insulation: a case for re-evaluating building codes. Build Res Inf 40:738–755CrossRefGoogle Scholar
  75. 75.
    Standard Methods of Fire Tests for Flame Propagation of Textiles and Films (NFPA 701), National Fire Protection Association, QuincyGoogle Scholar
  76. 76.
    Babrauskas V, Gray BF, Janssens ML (2007) Prudent practices for the design and installation of heat-producing devices near wood materials. Fire Mater 31:125–135CrossRefGoogle Scholar
  77. 77.
    Babrauskas V (2016) Explosions of ammonium nitrate fertilizer in storage or transportation are preventable accidents. J Hazard Mater 304:134–149CrossRefGoogle Scholar
  78. 78.
    Torsuev NS (1936) Explosive fertilizer containing ammonium nitrate. J Chem Indus [Russian] 13:102–104Google Scholar
  79. 79.
    Babrauskas V (2016) UN Test O.1 errors in quantifying the behavior of solid oxidizers. J Loss Prevent Process Indus 39:1–6CrossRefGoogle Scholar
  80. 80.
    Su JZ, Lougheed GD (2014) Report to research consortium for wood and wood-hybrid mid-rise buildings (Report A1-004377.1), National Research Council of CanadaGoogle Scholar

Copyright information

© Springer Science+Business Media New York 2016

Authors and Affiliations

  1. 1.Fire Science and Technology Inc.San DiegoUSA

Personalised recommendations