Skip to content
From Fire Safety Academy Wiki
Peter (talk | contribs)
FSE wiki: Fact-check: qualify spill-rate, duration, 95 L boundary, flame spread and historical foam rates (written by Claude Opus 5.5)
Lisa (talk | contribs)
Add 4 watermarked NFPA SVG illustrations
Line 7: Line 7:


Budnick, Hunt and Wright use the same distinction in the handbook. Confined fuel fires are fires in containers or curbed areas. Unconfined spills are those in which the liquid is free to spread and its depth depends only on the liquid and its surface tension.<ref name="B3149" /> On concrete, the unconfined depth is about 2.8 mm or less, and they consider it reasonable to treat pools deeper than 2.8 mm as confined.<ref name="B3149" />
Budnick, Hunt and Wright use the same distinction in the handbook. Confined fuel fires are fires in containers or curbed areas. Unconfined spills are those in which the liquid is free to spread and its depth depends only on the liquid and its surface tension.<ref name="B3149" /> On concrete, the unconfined depth is about 2.8 mm or less, and they consider it reasonable to treat pools deeper than 2.8 mm as confined.<ref name="B3149" />
[[File:fse-spill-fire-v1.svg|thumb|upright=1.5|alt=Cross-sections of a thin unconfined spill on concrete losing heat downward and of a deeper confined pool inside a curb, each with flames above.|Small unconfined spills on concrete lose heat to the floor and can burn at about one-fifth of a deep pool's rate per unit area. The depths shown are conservative estimates; large continuous spills may approach pool rates. Based on Gottuk and White (2002), pp. 2-300 and 2-311–2-312, and Budnick et al. (2008), p. 3-149.]]


Spills are further divided by the moment of ignition. In a static or instantaneous spill, the fuel spreads to a nearly maximum area and is then ignited, so that the flame spreads across the fuel surface.<ref name="GW298">Gottuk and White (2002), p. 2-298.</ref> In a continuously flowing spill, ignition occurs while fuel is still flowing from the source. The flame may first spread across the liquid, but the fire front is then controlled by the spread of fuel over the surface until steady burning is reached.<ref name="GW298" /> A burning continuous spill settles at a size where the volume of fuel supplied each second equals the volume burned.<ref name="GW298" />
Spills are further divided by the moment of ignition. In a static or instantaneous spill, the fuel spreads to a nearly maximum area and is then ignited, so that the flame spreads across the fuel surface.<ref name="GW298">Gottuk and White (2002), p. 2-298.</ref> In a continuously flowing spill, ignition occurs while fuel is still flowing from the source. The flame may first spread across the liquid, but the fire front is then controlled by the spread of fuel over the surface until steady burning is reached.<ref name="GW298" /> A burning continuous spill settles at a size where the volume of fuel supplied each second equals the volume burned.<ref name="GW298" />
Line 34: Line 35:
== Fire growth and flame spread ==
== Fire growth and flame spread ==
Once a spill is ignited, the fire grows as flame spreads over the liquid. Gottuk and White identify liquid temperature relative to its flash point as the main factor in spread rate.<ref name="GW300" /> Liquid-controlled spread is roughly 1 to 12 cm/s; gas-phase spread reaches 130 to 220 cm/s.<ref name="GW300" /> For many pure liquids the regimes lie below and above the flash point, respectively. In JP-5 and JP-8 tests, however, the transition was near 18 °C above the closed-cup flash point.<ref name="GW300" /><ref name="GW304">Gottuk and White (2002), p. 2-304.</ref> On porous fuel-soaked surfaces, spread is much slower, in centimetres per minute.<ref name="GW300" />
Once a spill is ignited, the fire grows as flame spreads over the liquid. Gottuk and White identify liquid temperature relative to its flash point as the main factor in spread rate.<ref name="GW300" /> Liquid-controlled spread is roughly 1 to 12 cm/s; gas-phase spread reaches 130 to 220 cm/s.<ref name="GW300" /> For many pure liquids the regimes lie below and above the flash point, respectively. In JP-5 and JP-8 tests, however, the transition was near 18 °C above the closed-cup flash point.<ref name="GW300" /><ref name="GW304">Gottuk and White (2002), p. 2-304.</ref> On porous fuel-soaked surfaces, spread is much slower, in centimetres per minute.<ref name="GW300" />
[[File:fse-spill-fire-v2.svg|thumb|upright=1.5|alt=Semi-log plot of flame spread rate against liquid temperature minus flash point, low band 1 to 12 cm/s, high band 130 to 220 cm/s, jump near plus 18 degrees C.|For the tested JP jet fuels, flame spread changed from liquid control at about 1–12 cm/s to gas-phase spread at 130–220 cm/s near 18 °C above the closed-cup flash point. Based on Gottuk and White (2002), pp. 2-300 and 2-304.]]


The depth of the layer affects only the slower, liquid-controlled regime. Flame spread rates over decane increased with depth up to about 3 cm, and flames did not spread away from the ignition source in layers less than about 1.5 mm deep.<ref name="GW302">Gottuk and White (2002), p. 2-302.</ref> Depth has no effect on spread through the vapour phase.<ref name="GW302" /> The flash point of a fuel mixture can also change after a release. In jet fuel tests, delaying ignition by 3 to 460 s gave results consistent with a rise in flash point of about 10 °C.<ref name="GW304" /> Gottuk and White therefore treat instantaneous ignition of a released fuel as the conservative assumption.<ref name="GW306">Gottuk and White (2002), p. 2-306.</ref>
The depth of the layer affects only the slower, liquid-controlled regime. Flame spread rates over decane increased with depth up to about 3 cm, and flames did not spread away from the ignition source in layers less than about 1.5 mm deep.<ref name="GW302">Gottuk and White (2002), p. 2-302.</ref> Depth has no effect on spread through the vapour phase.<ref name="GW302" /> The flash point of a fuel mixture can also change after a release. In jet fuel tests, delaying ignition by 3 to 460 s gave results consistent with a rise in flash point of about 10 °C.<ref name="GW304" /> Gottuk and White therefore treat instantaneous ignition of a released fuel as the conservative assumption.<ref name="GW306">Gottuk and White (2002), p. 2-306.</ref>
Line 64: Line 66:


A continuously flowing spill behaves differently. Its burning area settles where the volumetric flow of liquid equals the volumetric burning rate, so the steady diameter follows from a balance:<ref name="GW312" />
A continuously flowing spill behaves differently. Its burning area settles where the volumetric flow of liquid equals the volumetric burning rate, so the steady diameter follows from a balance:<ref name="GW312" />
[[File:fse-spill-fire-v3.svg|thumb|upright=1.5|alt=Log-log line of equilibrium burning area against leak rate, 0.74 square metres at 3.8 L/min rising to 74 square metres at 380 L/min.|Equilibrium burning area of a continuous spill at a burning rate of 0.3 m per hour: the area grows in proportion to the leak rate. Based on Slye (2008), p. 6-206.]]


<math display="block">\dot{V}_L = A_{ss}\,\dot{y} \quad\Rightarrow\quad D_{ss} = \left(\frac{4\,\dot{V}_L}{\pi\,\dot{y}}\right)^{1/2}</math>
<math display="block">\dot{V}_L = A_{ss}\,\dot{y} \quad\Rightarrow\quad D_{ss} = \left(\frac{4\,\dot{V}_L}{\pi\,\dot{y}}\right)^{1/2}</math>
Line 108: Line 111:


For comparison, the same 154 kg held in a bund of an assumed 4 m² would form a confined pool about 52 mm deep. At 2400 kW/m², it would burn at about 9.6 MW.<ref name="B3149" /> With the total energy of about 6.73 GJ, the constant-rate model gives 6.73 × 10<sup>6</sup> kJ / 9600 kW ≈ 700 s, or nearly 12 minutes. This illustrates the point made in NUREG-1805. For a given amount of fuel, a large spill burns with a high heat release rate for a short time, and a smaller pool burns with a lower rate for longer.<ref name="IS35">Iqbal and Salley (2004), p. 3-5.</ref>
For comparison, the same 154 kg held in a bund of an assumed 4 m² would form a confined pool about 52 mm deep. At 2400 kW/m², it would burn at about 9.6 MW.<ref name="B3149" /> With the total energy of about 6.73 GJ, the constant-rate model gives 6.73 × 10<sup>6</sup> kJ / 9600 kW ≈ 700 s, or nearly 12 minutes. This illustrates the point made in NUREG-1805. For a given amount of fuel, a large spill burns with a high heat release rate for a short time, and a smaller pool burns with a lower rate for longer.<ref name="IS35">Iqbal and Salley (2004), p. 3-5.</ref>
[[File:fse-spill-fire-v4.svg|thumb|upright=1.5|alt=Two idealised rectangles of heat release rate against time: spill at 56 MW for 121 s and bunded pool at 9.6 MW for about 700 s.|Constant-rate models for 208 L of gasoline: unconfined spill (about 56 MW for 121 s) and an assumed 4 m² bund (about 9.6 MW for about 700 s). Both release about 6.7 GJ. Actual fires grow and decay. Based on Gottuk and White (2002), pp. 2-313–2-314, and Budnick et al. (2008), p. 3-149.]]


A continuous leak of 38 L/min (10 gpm) can be treated with the handbook rule. At a regression rate of 0.3 m per hour, it reaches an equilibrium burning area of about 7.4 m².<ref name="S6206" /> Since this release rate exceeds about 10 L/min, Gottuk and White would apply pool burning rates.<ref name="GW312" /> At steady state, all fuel supplied is burned, so for gasoline the heat release rate is the mass flow, 0.000631 m³/s × 740 kg/m³ ≈ 0.467 kg/s, times 43.7 MJ/kg, which is about 20 MW. NUREG/CR-6850 uses the same approach of mass flow times heat of combustion for continuously flowing spills.<ref name="N05G10" />
A continuous leak of 38 L/min (10 gpm) can be treated with the handbook rule. At a regression rate of 0.3 m per hour, it reaches an equilibrium burning area of about 7.4 m².<ref name="S6206" /> Since this release rate exceeds about 10 L/min, Gottuk and White would apply pool burning rates.<ref name="GW312" /> At steady state, all fuel supplied is burned, so for gasoline the heat release rate is the mass flow, 0.000631 m³/s × 740 kg/m³ ≈ 0.467 kg/s, times 43.7 MJ/kg, which is about 20 MW. NUREG/CR-6850 uses the same approach of mass flow times heat of combustion for continuously flowing spills.<ref name="N05G10" />

Revision as of 07:41, 8 November 2026

A spill fire is a fire in a thin layer of flammable or combustible liquid that has spread freely over a floor, the ground or another surface after a release. The Fire Protection Handbook (2008) classes spill fires, together with running fires, as fires in liquid layers of 6.4 mm (¼ in.) depth or less.[1] In the SFPE Handbook, Gottuk and White contrast a spill, a thin layer from an unconfined release, with a pool fire, in which the fuel is confined and typically more than 1 cm deep.[2]

Liquid spill fires and trash fires are the transient exposure fires most often postulated in the fire hazard analysis of United States nuclear power plants.[3] In tests of unconfined spills on concrete, heat loss to the surface reduced the burning rate per unit area to about one-fifth of a deep pool's rate; large, continuously fed spills can approach pool rates.[4][5][6] A model of a 208 L gasoline drum on a warehouse floor gives about 56 MW for roughly 2 minutes if peak burning is assumed throughout.[7]

Definition and types

Gottuk and White describe the nature of a spill fire as highly variable. It depends on the source of the release, on the surface on which the fuel lands, such as concrete, ground or water, and on where and when ignition occurs.[2] A confined release, held by the walls of a room or by obstructions on a floor, forms a smaller and deeper pool than the same liquid would form if free to spread.[2]

Budnick, Hunt and Wright use the same distinction in the handbook. Confined fuel fires are fires in containers or curbed areas. Unconfined spills are those in which the liquid is free to spread and its depth depends only on the liquid and its surface tension.[6] On concrete, the unconfined depth is about 2.8 mm or less, and they consider it reasonable to treat pools deeper than 2.8 mm as confined.[6]

Cross-sections of a thin unconfined spill on concrete losing heat downward and of a deeper confined pool inside a curb, each with flames above.
Small unconfined spills on concrete lose heat to the floor and can burn at about one-fifth of a deep pool's rate per unit area. The depths shown are conservative estimates; large continuous spills may approach pool rates. Based on Gottuk and White (2002), pp. 2-300 and 2-311–2-312, and Budnick et al. (2008), p. 3-149.

Spills are further divided by the moment of ignition. In a static or instantaneous spill, the fuel spreads to a nearly maximum area and is then ignited, so that the flame spreads across the fuel surface.[8] In a continuously flowing spill, ignition occurs while fuel is still flowing from the source. The flame may first spread across the liquid, but the fire front is then controlled by the spread of fuel over the surface until steady burning is reached.[8] A burning continuous spill settles at a size where the volume of fuel supplied each second equals the volume burned.[8]

The handbook chapter on fire extinguisher use names three general types of flammable liquid fires. The first are fires in liquids deeper than 6.4 mm, such as dip tanks and quench tanks. The second are spill or running fires in layers of 6.4 mm or less, and the third are fires of liquids or gases escaping under pressure.[1] Scheffey distinguishes such horizontal surface fires from three-dimensional fires of falling fuel and from pressure fires.[9] On offshore platforms, a broken fuel line can produce a spray fire or a three-dimensional spill fire with a pool fire in the bilge below.[10]

Where spill fires occur

Spill fires are a recurring scenario wherever liquids are stored, transferred or used. Ordile notes that storage tank fires may start with an internal explosion or with a spill fire exposing the tanks.[11] The emergency venting requirements for aboveground tanks that Ordile describes are based on the tank area wetted by the liquid and exposed to an external ground fire, such as a spill.[12]

One transport example is a postcrash aircraft fuel spill. Kreckie cites an early fire test on a C-97 aircraft in which the cabin stayed at 21.1 °C while the surrounding spill fire burned for 2 minutes. Burnthrough occurred at 130 seconds, and conditions in the cabin became unsurvivable at 138 seconds.[13]

Spills also matter for life safety in industry. NFPA records show that most fires or explosions in manufacturing facilities that caused 10 or more deaths involved explosions or, to a lesser extent, flash fires or fuel spills.[14] In nuclear power plants, NUREG-1805 lists leaks from reactor coolant pump seals, oil spills from transformers and pumps, and liquids used for cleaning and painting as sources of liquid fires.[15]

Spill size

The area of a spill correlates with the size of the resulting fire, so the first step in an analysis is to estimate the area that the liquid covers.[2] For a confined pool, the area is set by the boundaries and the depth follows from the volume. For an unconfined spill, the area is usually found from the same relation with an estimated depth:[8]

δ=VAS

In the past, engineers estimated the depth conservatively as the minimum depth that supports flame spread. This gives the largest possible spill area and therefore the largest possible fire.[8] Measured spill depths range from 0.7 to 4 mm, depending mainly on the amount released.[16] For releases of a few litres the area per litre varies widely, from 0.3 to 1.6 m² per litre, because the release height, the surface and the fluid properties matter more. For releases of 95 L or more, the area per litre approaches about 0.3 m² per litre.[8]

Gottuk and White therefore give two conservative rules that yield the maximum spill area. A spill of less than 95 L is taken as 0.7 mm deep, which is 1.4 m² per litre. A spill of 95 L or more is taken as 2.8 mm deep, which is 0.36 m² per litre.[17] Experiments also showed that a burning spill grows after ignition. For releases of 2 to 190 L, the fire increased the initial spill area by 22 to 89 percent. For the maximum fire size, they suggest a fire area of about 155 percent of the initial spill area:[8]

A=1.55AS

The surface also matters. Porous materials such as sand or carpet change both the spill size and the burning rate.[2] The 2005 U.S. nuclear-plant fire risk method uses the same depth values, but puts exactly 95 L in the smaller-spill category.[18] Its printed area values of 0.8 and 0.05 m² per litre were corrected by an errata sheet to 1.4 and 0.36 m² per litre.[19]

Fire growth and flame spread

Once a spill is ignited, the fire grows as flame spreads over the liquid. Gottuk and White identify liquid temperature relative to its flash point as the main factor in spread rate.[17] Liquid-controlled spread is roughly 1 to 12 cm/s; gas-phase spread reaches 130 to 220 cm/s.[17] For many pure liquids the regimes lie below and above the flash point, respectively. In JP-5 and JP-8 tests, however, the transition was near 18 °C above the closed-cup flash point.[17][20] On porous fuel-soaked surfaces, spread is much slower, in centimetres per minute.[17]

Semi-log plot of flame spread rate against liquid temperature minus flash point, low band 1 to 12 cm/s, high band 130 to 220 cm/s, jump near plus 18 degrees C.
For the tested JP jet fuels, flame spread changed from liquid control at about 1–12 cm/s to gas-phase spread at 130–220 cm/s near 18 °C above the closed-cup flash point. Based on Gottuk and White (2002), pp. 2-300 and 2-304.

The depth of the layer affects only the slower, liquid-controlled regime. Flame spread rates over decane increased with depth up to about 3 cm, and flames did not spread away from the ignition source in layers less than about 1.5 mm deep.[21] Depth has no effect on spread through the vapour phase.[21] The flash point of a fuel mixture can also change after a release. In jet fuel tests, delaying ignition by 3 to 460 s gave results consistent with a rise in flash point of about 10 °C.[20] Gottuk and White therefore treat instantaneous ignition of a released fuel as the conservative assumption.[22]

In the largest spill fire tests reviewed by Gottuk and White, 114 L of JP-8 jet fuel spilled on concrete covered about 30 m² at ignition. The free-burning spill reached about 37 m², and the fuel burned out in roughly 2 minutes.[23] The growth of the burning area can be estimated from the spread rate. For a circular spill ignited at the centre, which gives the fastest involvement, the burning radius grows as r=vt and the area as A=πv2t2.[24] With a constant burning rate per unit area, the heat release rate then grows with the square of time. The whole spill is burning after a time equal to the spill radius divided by the spread rate.[24]

Burning rate and heat release rate

The heat release rate of a spill fire is the product of the burning rate per unit area, the burning area and the heat of combustion of the fuel.[5] The handbook writes the same relation with an energy release rate per unit area, which depends on the fuel type, the spill area and the spill depth.[6]

Q˙=m˙″AΔhc=q˙″A

Confined pools larger than about 1 m in diameter burn down at a nearly constant 4 mm per minute for the fuels tested by Blinov and Khudiakov.[25] Unconfined spills burn much more slowly. In tests with JP-8 and JP-5 jet fuel on smooth coated concrete, the burning rate of spills larger than about 1 m was roughly one-fifth of the maximum pool burning rate.[4] The averages were 0.007 and 0.010 kg/m²s for fixed-quantity and continuous JP-8 spills, 0.010 kg/m²s for JP-5 and 0.011 kg/m²s for gasoline.[4]

The lower rate is attributed to heat loss into the surface below. The spill fires had fuel depths of about 1 mm, against 1 cm or more in most pan tests. Concrete, with a thermal conductivity of about 1.4 W/m·K, conducts heat away from the thin layer much better than a thick layer of fuel at about 0.11 W/m·K.[4][5] On the basis of limited small-spill tests, Gottuk and White suggest estimating the burning rate of an unconfined spill as one-fifth of the maximum pool burning rate.[5] The handbook table of Budnick and co-authors applies the same reduction of about 80 percent.[6]

Energy release rate per unit area of optically thick liquid fires (Budnick et al., 2008, Table 3.9.2)
Fuel Confined liquid fire (kW/m²) Unconfined liquid fire (kW/m²)
Gasoline 2400 480
Kerosene 1680 340
Diesel 1950 390
Transformer oil 1790 360
Heavy fuel oil 1390 280

A continuously flowing spill behaves differently. Its burning area settles where the volumetric flow of liquid equals the volumetric burning rate, so the steady diameter follows from a balance:[5]

Log-log line of equilibrium burning area against leak rate, 0.74 square metres at 3.8 L/min rising to 74 square metres at 380 L/min.
Equilibrium burning area of a continuous spill at a burning rate of 0.3 m per hour: the area grows in proportion to the leak rate. Based on Slye (2008), p. 6-206.

V˙L=Assy˙⇒Dss=(4V˙Lπy˙)1/2

This assumes that all fuel burns and none is lost, for example into a porous surface.[5] Introduction to Fire Safety Engineering treats many fire calculations as such balances, in which the change in a system equals what flows in minus what flows out.[26] Slye gives the handbook's rule of thumb. At a burning rate of 0.3 m per hour, each 3.8 L/min of spillage reaches an equilibrium burning area of about 0.74 m², so 38 L/min burns over about 7.4 m².[27] Large continuous JP-5 spills of 568 to 2271 L/min on concrete burned at about 0.055 kg/m²s, close to pool fire rates. Gottuk and White therefore recommend pool burning rates for release rates above about 10 L/min and one-fifth of the pool rate for smaller releases.[5][28]

Symbols
Symbol Description Unit
δ depth of the spilled liquid m
V volume of liquid released m³ (or L)
AS initial spill area m²
A burning area of the fire m²
r, v, t burning radius, flame spread rate, time since ignition m, m/s, s
Q˙ heat release rate kW
m˙″ mass burning rate per unit area kg/m²s
Δhc heat of combustion kJ/kg
q˙″ energy release rate per unit area kW/m²
V˙L volumetric flow rate of the release m³/s
y˙ regression rate of the liquid surface m/s
Ass, Dss steady burning area and diameter of a continuous spill m², m

The rules apply to spills on level surfaces in the open, under well-ventilated conditions, and are based on limited data.[28] The median intermittent flame height can be estimated with the Heskestad correlation, Lf=0.23Q˙2/5−1.02D, with heat release rate in kW and fire diameter in m. For jet fuel spills, it predicted heights about 17 percent lower than those measured.[28][7]

Worked example

Gottuk and White consider a 208 L (55 gal) drum of gasoline that ruptures in a warehouse. The fuel spreads quickly over the floor and is ignited by faulty equipment.[7] The release is 95 L or more, so the spill area is 0.36 m²/L × 208 L ≈ 74.9 m². The burning area is then 1.55 × 74.9 ≈ 116 m², equal to a circle of 12.2 m diameter.[17][8][7]

For gasoline they use a heat of combustion of 43.7 MJ/kg, a density of 740 kg/m³ and a maximum pool burning rate of 0.055 kg/m²s. One-fifth of this rate, 0.011 kg/m²s, is used in this spill model.[7] The heat release rate is 0.011 × 116 × 43,700 ≈ 55,800 kW, or about 56 MW. The handbook value for unconfined gasoline, 480 kW/m², gives almost the same result, 480 × 116 ≈ 55.7 MW.[6] The Heskestad correlation gives a 50th-percentile intermittent flame height of 0.23 × 55,8000.4 − 1.02 × 12.2 ≈ 5.8 m.[28][7]

The 208 L of fuel has a mass of 0.208 × 740 ≈ 154 kg. Assuming the whole area burns at its peak rate throughout, the model gives 154 / (0.011 × 116) ≈ 121 s, or about 2 minutes. Gottuk and White warn that the actual fire would last longer because flames need time to spread across the spill.[28][7] At a warehouse temperature of 20 °C, gasoline is far above its flash point of about −43 °C. The flame therefore spreads through the vapour at up to about 2 m/s.[7][29] From a central ignition point, it covers the 6.1 m radius in 6.1 / 2.0 ≈ 3 s.[7]

For comparison, the same 154 kg held in a bund of an assumed 4 m² would form a confined pool about 52 mm deep. At 2400 kW/m², it would burn at about 9.6 MW.[6] With the total energy of about 6.73 GJ, the constant-rate model gives 6.73 × 106 kJ / 9600 kW ≈ 700 s, or nearly 12 minutes. This illustrates the point made in NUREG-1805. For a given amount of fuel, a large spill burns with a high heat release rate for a short time, and a smaller pool burns with a lower rate for longer.[30]

Two idealised rectangles of heat release rate against time: spill at 56 MW for 121 s and bunded pool at 9.6 MW for about 700 s.
Constant-rate models for 208 L of gasoline: unconfined spill (about 56 MW for 121 s) and an assumed 4 m² bund (about 9.6 MW for about 700 s). Both release about 6.7 GJ. Actual fires grow and decay. Based on Gottuk and White (2002), pp. 2-313–2-314, and Budnick et al. (2008), p. 3-149.

A continuous leak of 38 L/min (10 gpm) can be treated with the handbook rule. At a regression rate of 0.3 m per hour, it reaches an equilibrium burning area of about 7.4 m².[27] Since this release rate exceeds about 10 L/min, Gottuk and White would apply pool burning rates.[5] At steady state, all fuel supplied is burned, so for gasoline the heat release rate is the mass flow, 0.000631 m³/s × 740 kg/m³ ≈ 0.467 kg/s, times 43.7 MJ/kg, which is about 20 MW. NUREG/CR-6850 uses the same approach of mass flow times heat of combustion for continuously flowing spills.[18]

Control and suppression

Prevention begins with keeping spills small and away from exposures. Schwab recommends that large flammable liquid spills be confined where they occur. Diking may not stop the liquid near the source from burning, so a better approach is drainage to an impounding basin where it can burn out harmlessly.[31] Spills are kept small by limiting the amount of liquid in a process and by fitting fire-safe isolation valves on large containers.[31] For tank storage, Ordile describes sloping ground, directional dikes, intermediate dikes at least 457 mm high and dikes sized for the largest tank.[32] In Ordile's 2008 account of NFPA 30, aboveground tanks with compliant secondary containment need not meet these diking rules within specified capacities.[32]

Ordile notes that the ability to drain away spilled liquid in a fire is critical, especially where sprinkler protection uses water only. Fire tests have shown that sprinklers can fail to prevent the rupture of high-piled liquid containers, whose released contents add to the fire.[33] Scheffey adds that water discharge may have limited effectiveness for controlling or extinguishing large spill fires.[34]

Firefighting foam is the principal extinguishing agent for spill fires. In Scheffey's chapter, low-expansion foam is the principal agent for burning liquid spill and tank fires, and the only permanent extinguishing agent for these fires. Fuel spills are quickly made safe by a foam blanket.[35] The fire must be a horizontal surface fire: falling or pressurised fuel cannot be extinguished by foam unless the liquid has a relatively high flash point.[9] Aqueous film-forming foam (AFFF) also forms a film of solution on hydrocarbon fuel. Its fluidity and film strength on kerosene make it particularly suitable for jet fuel spill fires.[9]

The foam application rate is the volume of foam solution reaching each square metre of fuel surface per minute.[36] Rates above the recommended minimum shorten extinguishment, but little is gained above three times the minimum. If foam is lost to heat and fuel attack as fast as it is applied, the fire will not be controlled.[36] In Scheffey's 2008 account, AFFF from standard sprinklers must extinguish test fires at 4.1 L/min/m² (0.10 US gpm/ft²). The NFPA 16 edition he describes specifies a minimum design rate of 0.16 US gpm/ft² (about 6.5 L/min/m²) for all foam agents.[37] For aircraft fires, the International Civil Aviation Organization in 1972 found 0.13 US gpm/ft² (5.30 L/min/m²) suitable for AFFF and confirmed 0.15 gpm/ft² (6.11 L/min/m²) for protein foam.[38]

Other methods complement foam. Dry chemical gives very fast flame control but no cooling, so reflash can occur. Combined, or twinned, dry chemical and AFFF equipment therefore knocks down the flames and then secures the surface.[39] With portable extinguishers, a side-to-side sweep from the near edge may work better on spill fires than an overhead application.[40] Foam from a portable extinguisher may be bounced off the floor just in front of a burning spill, and the operator should never turn away because unburned fuel may reignite.[41]

Limits and uncertainty

The data on unconfined spill fires are limited, and Gottuk and White stress the need for more tests, especially at large diameters.[5] Their rules come from fires on level surfaces. On inclined or cluttered surfaces, fuel can pool, form channels or wet larger areas, and no published studies had evaluated the effect of three-dimensional fuel flow on spill burning rates.[28] Fires in enclosures can burn faster once a hot upper layer forms, or slower when the air supply is restricted, and wind can raise or lower burning rates or push a spill fire downwind.[28]

The usual assumption that the peak burning rate lasts for the whole fire is conservative for fire size. In reality the peak in an unconfined spill lasts a short time, and the fire burns longer at a lower average rate.[28] The worked example above shows this. The predicted 2 minute burn is probably too short, because the flame takes time to spread across the spill.[7] Budnick and co-authors caution that maximising the heat release rate is not always conservative, for example when estimating when detectors or sprinklers will operate.[6]

A spill fire is also linked to other hazards. Heavier-than-air vapours from a spill can flow along the floor, be ignited at a distance and flash back to the liquid.[42] Liquids burn only once they have evaporated, and the fire hazard of a spill comes from its vapour rather than from the liquid itself.[43][44]

See also

Notes

  1. ↑ 1.0 1.1 Conroy (2008), p. 17-72.
  2. ↑ 2.0 2.1 2.2 2.3 2.4 Gottuk and White (2002), p. 2-297.
  3. ↑ Iqbal and Salley (2004), p. 3-3.
  4. ↑ 4.0 4.1 4.2 4.3 Gottuk and White (2002), p. 2-311.
  5. ↑ 5.0 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 Gottuk and White (2002), p. 2-312.
  6. ↑ 6.0 6.1 6.2 6.3 6.4 6.5 6.6 6.7 Budnick et al. (2008), p. 3-149.
  7. ↑ 7.00 7.01 7.02 7.03 7.04 7.05 7.06 7.07 7.08 7.09 Gottuk and White (2002), p. 2-314.
  8. ↑ 8.0 8.1 8.2 8.3 8.4 8.5 8.6 8.7 Gottuk and White (2002), p. 2-298.
  9. ↑ 9.0 9.1 9.2 Scheffey (2008), p. 17-46.
  10. ↑ Mawhinney (2008), p. 16-137.
  11. ↑ Ordile (2008), p. 7-23.
  12. ↑ Ordile (2008), p. 7-20.
  13. ↑ Kreckie (2008), p. 13-101.
  14. ↑ Hall and Cote (2008), p. 3-26.
  15. ↑ Iqbal and Salley (2004), p. 3-6.
  16. ↑ Gottuk and White (2002), p. 2-299.
  17. ↑ 17.0 17.1 17.2 17.3 17.4 17.5 Gottuk and White (2002), p. 2-300.
  18. ↑ 18.0 18.1 Nowlen et al. (2005b), p. G-10.
  19. ↑ NUREG/CR-6850, EPRI TR 1011989 Errata Sheet (2006), p. 2.
  20. ↑ 20.0 20.1 Gottuk and White (2002), p. 2-304.
  21. ↑ 21.0 21.1 Gottuk and White (2002), p. 2-302.
  22. ↑ Gottuk and White (2002), p. 2-306.
  23. ↑ Gottuk and White (2002), p. 2-305.
  24. ↑ 24.0 24.1 Gottuk and White (2002), p. 2-308.
  25. ↑ Blinov and Khudiakov, as cited in Gottuk and White (2002), p. 2-309.
  26. ↑ Van Rossum (2027), §3.7.
  27. ↑ 27.0 27.1 Slye (2008), p. 6-206.
  28. ↑ 28.0 28.1 28.2 28.3 28.4 28.5 28.6 28.7 Gottuk and White (2002), p. 2-313.
  29. ↑ Slye (2008), p. 6-203.
  30. ↑ Iqbal and Salley (2004), p. 3-5.
  31. ↑ 31.0 31.1 Schwab (2008a), p. 8-22.
  32. ↑ 32.0 32.1 Ordile (2008), p. 7-22.
  33. ↑ Ordile (2008), p. 7-27.
  34. ↑ Scheffey (2008), p. 17-63.
  35. ↑ Scheffey (2008), p. 17-45.
  36. ↑ 36.0 36.1 Scheffey (2008), p. 17-49.
  37. ↑ Scheffey (2008), p. 17-64.
  38. ↑ Kreckie (2008), p. 13-103.
  39. ↑ Scheffey (2008), p. 17-65.
  40. ↑ Conroy (2008), p. 17-78.
  41. ↑ Conroy (2008), p. 17-82.
  42. ↑ Slye (2008), p. 6-207.
  43. ↑ Van Rossum (2027), §4.1.
  44. ↑ Slye (2008), p. 6-197.

References

  • Budnick, E. K., Hunt, S. P., & Wright, M. T. (2008). Closed form enclosure fire calculations. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 3-145–3-166). National Fire Protection Association.
  • Conroy, M. T. (2008). Fire extinguisher use and maintenance. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 17-71–17-92). National Fire Protection Association.
  • Gottuk, D. T., & White, D. A. (2002). Liquid fuel fires. In P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, R. L. P. Custer, J. R. Hall Jr., & J. M. Watts Jr. (Eds.), SFPE handbook of fire protection engineering (3rd ed., pp. 2-297–2-316). National Fire Protection Association; Society of Fire Protection Engineers.
  • Hall, J. R., Jr., & Cote, A. E. (2008). An overview of the fire problem and fire protection. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 3-3–3-30). National Fire Protection Association.
  • Iqbal, N., & Salley, M. H. (2004). Fire Dynamics Tools (FDTs): Quantitative fire hazard analysis methods for the U.S. Nuclear Regulatory Commission Fire Protection Inspection Program (NUREG-1805). U.S. Nuclear Regulatory Commission, Office of Nuclear Reactor Regulation.
  • Kreckie, J. (2008). Aircraft rescue and fire fighting (ARFF). In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 13-99–13-115). National Fire Protection Association.
  • Mawhinney, J. R. (2008). Water mist fire suppression systems. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 16-135–16-180). National Fire Protection Association.
  • Nowlen, S. P., Najafi, B., Wyant, F. J., Forester, J., Kazarians, M., Kolaczkowski, A., Joglar, F., Funk, D., Hannaman, G., & Anoba, R. (2005b). EPRI/NRC-RES fire PRA methodology for nuclear power facilities: Vol. 2. Detailed methodology (NUREG/CR-6850; EPRI 1011989). Electric Power Research Institute; U.S. Nuclear Regulatory Commission, Office of Nuclear Regulatory Research.
  • NUREG/CR-6850, EPRI TR 1011989 errata sheet. (2006).
  • Ordile, A. M. (2008). Storage of flammable and combustible liquids. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 7-15–7-33). National Fire Protection Association.
  • Scheffey, J. L. (2008). Foam extinguishing agents and systems. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 17-45–17-70). National Fire Protection Association.
  • Schwab, R. F. (2008a). Chemical processing equipment. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 8-17–8-29). National Fire Protection Association.
  • Slye, O. M., Jr. (2008). Flammable and combustible liquids. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 6-197–6-208). National Fire Protection Association.
  • Van Rossum, E. (2027). Introduction to Fire Safety Engineering. Fire Safety Academy.