Ceiling Jet Flows: David D. Evans
Ceiling Jet Flows: David D. Evans
Ceiling Jet Flows: David D. Evans
IN: DiNenno, P.J., et al., Editors, SFPE Handbook of Fire Protection Engineering,
2nd Edition, Chapter 4, Section 2, 2/32-2/39 pp
2-32
CEILING JET FLOWS 2-33
+————— R—————4 lations apply outside of this turning region as the flow moves
away from the impingement area. Certain constraints should
be understood when applying these correlations in the anal-
ysis of fire flows. The correlations apply only during times
after fire ignition when the ceiling flow may be considered
unconfined; i.e., no accumulated warm upper layer is
H present, Walls close to the fire affect the temperatures and
velocity in the ceiling jet. The correlations were developed
from test data to appIy in cases where the fire source is at
least a distance 1.8 times the ceiling height from the enclo-
sure walls. Ideally, for the special cases where burning fuel
is located against wall surfaces or two wall surfaces forming
a 90-degree corner, the correlations may be adjusted based
on method of reflection making use of symmetry to account
Fig. 2.4.I. Ceiling jet flow beneath m anconjined ceiling. for the effects of the walls blocking entrainment of air into
the fire plume. For the case of a fire adjacent to a flat wall, 2Q
is substituted for Q in the correlations. For a fire in a 90-
degree corner, 4Q is substituted for Q in the correlations.3
Alpert3 has also developed easy-to-use correlations to Experiments have shown that, unless great care is taken
quanti& the maximum gas temperature and velocity at a to ensure that the fuel perimeter is in contact with the wall
given position ~n a ceiling jet flow produced by a steady fire. surfaces, the method of reflection used to estimate the effects
These correlations are widely used in hazard analysis cal- of the walls on ceiling jet temperature will be inaccurate. For
culations. They have been employed by Evans and Stroup13 example, Zukoski et a114 found that a circular burner placed
in the development of a generalized program for prediction against a wall so that only one point on the perimeter con-
of heat detector response for the case of the detector totally tacted the wall, behaved almost identically to a fire far from
submerged in the ceiling jet flow. The correlations are based the wall with plume entrainment only decreasing by 3 per-
on measurements collected during test burns of fuel arrays cent. When using Equations 1 through 4, this fire would be
of wood and plastic pallets, cardboard boxes, plastic mate- represented by replacing Q with 1.05Q and not 2Q as would
rials in cardboard boxes, and liquid fuels with energy release be predicted by the method of reflections. The value of 2Q
rates ranging from 668 kW to 98 MW under ceiling hei hts would be appropriate for a semicircular burner with the
‘?
from 4.6 to 15.5 m. The correlations developed by Alpert for entire flat side pushed against the wall surface.
determining maximum ceiling jet temperatures and veloci- Consider the following calculations, which demon-
ties in S.1. units are strate typical uses of the correlations, using Equations 1
16.9Q213 through 4.
T- T.= Hsj3
for r/H s 0.18 (1)
(a) The maximum temperature rise under a ceiling 10 m
directly above a 1.0 MW energy release rate fire is calcu-
5.38(@r)2/3 lated using Equation 1 as
T–Tea= for r/H >0.18 (2)
H
16.9(1000)2/3
T– T.=
105/3
~=096Q
d)H
113
for rlH s 0.1.5 [3] = 16.9(100)
46.42
o.195Q~13w12 AT= 36.4”C
u= # for rlH >0.15 (4)
(b) The minimum energy release rate of a fire against non-
combustible walls in the corner of a building 12 m below
where temperature, T, is in “C: velocity, U, is in mfs; and total
the ceiling needed to raise the temperature of the gas
energy release rate, Q, is in kW; and ceiling height and radial
below the ceiling 50”C at a distance 5 m from the corner
position (r and H) are in m.
is calculated u.+ng Equation 2 and the symmetry substi-
Data from these tests were correlated using the total
energy release rate of the fire. Even though it is the convec- tution of LIQfor Q to account for the effects of the corner as
tive fraction of the total energy release rate that is directly 5.38( 4@r)2/3
related to the buoyancy of the fire, most available data is T- T.=
H
correlated using the total energy release rate. For common
materials, such as those used by Alpert, the convective en- so _ 5.38( 4Cj/5)2/3
ergy release rate, (&, is considered to be proportional to the 12
total energy release rate, Q.
The correlations for both temperatures and velocities
(Equations I through 4) are broken into two parts. One part
[1
Q=:qy3’2 -
applies for the ceiling jet in the area of the impingement Q = 1472 kw
point where the upward flow of gas in the plume turns to Q = 1.472 MkV
flow out beneath the ceiling horizontally. These correlations
(Equations 1 and 3) are independent of radius and are actu- (c) The maximum velocity at this position is calculated
.-
ally axial plume floW temperatures and velocities calculated from Equation 4, modified to account for the etlects of
at the ceiling height above the fire source. The other corre- the corner as
- 2-34 FIRE DYNAMICS
— HESKESTADAND
Y DELICHATSIOS(1978) , For time dependent fires, all estimates from the previous
\, --- ALPERT(1972) sectiop may still be used, but with the constant energy ~elease
\\ rate, Q, replaced by an appropriate time dependent Q(t). In
\\ making this replacement, a “quasisteady” flow has been as-
‘\ \ sumed. This assumption implies that when a change in energy
‘\ release rate occurs at the fire source, its full effects are felt
Q
“- everywhere in the flow field immediately. In a relatively small
L“ AT* \
\ \ room-size enclosure, under conditions where the fire is grow-
~ -----
~%, \ ing slowly, this assumption is reasonable. In large industrial
>0 1.0
3 -\\ facilities, where travel times of the fire gases from the burning
‘% \ fuel to a detector or sprinkler submerged in the ceiling jet flow
●&
k -x.
a -. may be 10s or longer, this may not be an appropriate assump-
‘\, tion, depending largely on the rate of fire growth end desired
s accuracy of the gas temperature and velocity predictions.
.~’ -
U*/(AT*)12 “.% Testing has shown that the energy release rate during
\
the growth phase of many fires can often be characterized by
simple time dependent polynomial or exponential func-
Q= K(t–ti)z (8)
o.197(4Q)~/31w2
u= @/6 Figure 2-4.3 shows one case where the energy release rate for
0.197(5888)1/3(12)1/2 a burning foam sofa during the growth phase of the fire more
. than 80 seconds (tf) after ignition17 can be represented by
55/6
the equation
U = 3.2 III/S
of the wood, which was found to be 20.9 MJ/kg. The result- A foam sofa, of the type analyzed in Figure 2-4,3, is burning
ing dimensionless correlations for maximum ceiling jet tem- in a showroom 5 m below a suspended ceiling. The show-
peratures and velocities are room temperature remote from the fire remains at 20”c at
floor level as the fire begins to grow. Determine the gas
temperature and velocity at the position of a ceiling-
o, t* s t; (lOa) mounted fire detector submerged in the ceiling jet flow 4 m
AT; = t;– t; away from the fire axis when the fire energy release rate first
4/3
t* > t; (lc)b) reaches 2.5 MW.
( 0.188+0.313 rlH ) Figure 2-4.3 shows that the energy release rate from the
[
sofa first reaches 2.5 MW (2500 kW) at about 200 s after
11~1~ = 0.59(r/HJ ‘0’63 (11) ignition. Using the analytic formula for the time dependent
energy release rate, Equation 9, the time from the virtual
where ignition of the sofa at 80 s to reach 2500 kW is
from the steady fire analysis, A7’0,(Equation 6) by the simple AT; = 61.9
relationship
AT2 t-tj-tf4/3 Equation 11 is used to calculate the dimensionless ceil-
() t–ti (18) ing jet velocity
%=
the correlations apply when the corridor half-width, tb, is Cooper23 has formulated an alternative calculation of
greater than 0.2 times the distance from the fire source to the substitute source energy release rate and position below the
ceiling, H, or (e@ > 0.2). In the case of beams, the flows ceiling that provides for generalization to situations in
must also be contained fully, so that a channel flow results which portions of the time averaged plume flow in the lower
without “spillage” over the beams. In order for this condition layer are at temperatures below the upper layer temperature.
to be satisfied, the beam depth, hb, must be greater than the In these cases, only part of the plume flow may penetrate the
quantity (H/10 )(/b/H_)-113 or upper layer sufficiently to impact on the ceiling. The remain-
ing portion at low temperature may not penetrate into the
hb/H > O.l(tb/~ -113 hotter upper layer. In the extreme, when the maximum tem-
perature in the lower layer plume flow is less than the upper
then layer temperature, none of the plume flow will penetrate
significantly into the upper layer. This could be the case
=029p3
A~mP “ ()
(b
exp[ – 0.20(Y/H)((b/Hj 1/3] during the decay phases of an enclosure fire, when the en-
ergy release rate is small compared to earlier times in the fire
for Y > tb growth and spread. In this calculation of substitute fire
source quantities, the first step is to calculate the fraction of
where ATimP is the temperature in the gas near the ceiling the plume mass flow penetrating the upper layer, m ~, from
directly over the fire, and Yis the distance along the channel Equations 27 and 28.
measured from the plume impingement point. * 1.04599u+0.36039@2
Generally, for large industrial or commercial storage (27)
‘2 = 1 +1.37748 cr+0.360391cr2
facilities, the analysis for unconfined ceiling jet flows will be
sut%cient for most purposes. In smaller rooms, or for very long where
times after &e ignition in larger industrial facilities, a quies-
cent warm layer of gas will accumulate in the upper portion of u = [&(&- 1)][(1 + @(&)2’3)/~) – 11 (28)
the enclosure. This warm layer can be deep enough to totally
submerge the ceiling jet flow, In that case, temperatures in the Then, analogous to Equations 24,25, and 26 of the previous
ceiling jet can be expected to be greater than if the ceiling jet method
was exdraining gas from the cooler room ambient temperature
layer. Quantitative methods for the prediction of temperature Z~,2 = Z~,1&3/5(mJ)2/5[(1 + u)/u]1\3 [29)
and velocity in a two-layer room environment in which the
ceiling jet is contained totally in a warm upper layer and the 0.,2 = Qc,l[Umj/(l + d] (30)
fire is burning totally in the lower cool layer have been formu-
lated. Contributions to this area have been made by Evans,21,22 H2 = HI – Z1,l + ZI,2 (31)
Cooper, 23 and Zukoski and Kubota. 11
In these methods, the flow of the ceiling jet within the After the substitute values of energy release rate and dis-
warm upper layer of the epclosure is imagined to result from tance to the ceiling are calculated, the warm upper layer gas
a fire totally contained in a uniform ambient environment temperature and density are used in the previous correla-
with temperature equal to that of the warm upper layer. This tions developed for ceiling jet flows in uniform ambient
substitute fire has an energy release rate, Q2, and location environments to predict ceiling jet temperature and velocity
below the ceiling, H2, differing from the original fire. Calcu- values.
lation of the substitute quantities Q2 and H2 depends on the Using a substitute fire source technique and a series of
energy release rate and location of the original fire as well as steady fires to represent growing fires in an enclosure, Evans
the depths and temperatures of the upper and lower layers has calculated the effects of warm upper layer depth on
within the enclosure. temperatures in the ceilin jet in an analysis of detector
Following the development by Evans, 22 the substitute response in compartments!2
source energy release rate and distance below the ceiling are TO demonstrate the use of the techniques, the previous
calculated from Equations 23 through 26. Originally devel- example in which a sofa was imagined to be burning in a
oped for the purpose of sprinkler and heat detector response showroom may be expanded. Let all the parameters of the
time calculations, these equations are applicable during the problem remain the same except that at 200 s after ignition
growth phase of enclosure fires. (t – ti = 120 s), when the fire energy release rate has reached
2.5 MW, a quiescent warm layer of gas at a temperature of
50”C is assumed to have accumulated under the ceiling to a
Q;2 = [(1 + G@ 2’3WT – lKT13’2 (23)
depth of 2 m. For this case, the two-layer analysis is needed
to determine the ceiling jet maximum temperature at the
same position as calculated previously (4 m radially distant
from the plume impingement point on the ceiling).
AU of the two-layer calculations presented assume qua-
sisteady conditions. Using Equation 18 and the values of
(25) parameters in the single-layer calculation, it can be shown
that AT2 = 0.85 ATo. So in the uniform ambient case, *e
quasisteady analysis should be adequate. It will be assumed
H2 = HI – Z1,l + ZI,2 (26) that this finding will carry over to the two-layer case.
Using Equations 23 through 26 from the work of
Further explanation of variables is contained in the nomen- Evans,22 values of the energy release rate and position of the
clature section. substitute fire source which compensates for the two-layer
CEILING JET FLOWS 2-37
-
~ec~ on the plume flow can be calculated, The dimension- AT= 106 K
less energy release rate of the fire source evaluated at the
~sition of the upper and lower layer interface is T=106K+ 323 K=429K=156”C
Q;I = O(P . Cp coTcog112@;) This is 52°C above the temperature calculated previ-
ously using the quasi-steady analysis and a uniform 200C
For an actual energy release rate of 2500 KW, ambient ambient.
temperat~e of 293 K, and distance between the fire source
ad ~ interface between the lower and upper layers of 3 m TRANSIENT CEILING JETS
this becomes
At the beginning of a fire, the initial buoyant flow from
& = 2500/(1.204 * 1 * 293 * 9.8112 * 351z) the fire must spread across the ceiling, driven by buoyancy,
to penetrate the cooler ambient air ahead of the flow. Re-
= 0.1452
search studies designed to quantify the temperatures and
velocities of this initial spreading flow have only recently
Using the ratio of upper-layer temperature to lower-
been started.24 At a minimum, it is useful to become aware
layer temperature ~ = 323/293 = 1.1024 and the constant of the many fluid mechanical phenomena embodied in a
CT = g.115, the dimensionless energy release for the sub-
description of the ceiling jet flow in a corridor up to the time
stitute fire source is
when the ceiling jet is totally submerged in a quiescent
warm upper layer. Borrowing heavily from a description of
(& = 0.1179.
this flow provided by Zukoski et al, 24 the process is as
follows.
Using the value for the constant (32 = 0.913, the posi-
Afire starts in a small room with an open door to a long
tion of the substitute fire source relative to the interface is corridor having a small vent near the floor at the opposite
end. As the fire starts, smoke and hot gases rise to form a
21,2 = 3.161
layer near the fire room ceiling. The layer is contained in the
small room by the door so ffit. [See Figure 2-4.4, part (a).] As
Using Equations 25 and 26, the dimensional energy
the fire continues, hot gas from the room begins to sp~ll out
release rate and position relative to the ceiling are found to under the soffit into the hallway. The fire grows to a rela-
be tively constant energy release rate.
The outflowing gas forms a short, buoyant plume [see
Q2 = 2313 kW
Figure 2-4.4, part (b)] that impinges on the hallway ceiling,
producing a thin jet that flows away from the fire room in the
172 = 5.161 m
same manner as the plume within the room flows over the
interior ceiling. The gas flow in this jet is supercritical, anal-
The analogous calculations for substitute fire-source
ogous to the shooting flow of liquids over a weir. The veloc-
energy release rate and position following the analysis of
ity of the gas in this flow is greater than the speed of gravity
Cooper23 Equations 27 through 31 are waves on the interface between the hot gas and the cooler
ambient air. The interaction of the leading edge of this flow
T = 23.60 m; = 0.962 ZI,2 = 3.176
with the ambient air ahead of it produces a hydraulic jump-
like condition, as shown in Figure 2-4.4, part (c). A substan-
so
tial amount of ambient air is entrained at this jump. Down-
stream of the jump, the velocity of the gas flow is reduced
~z = 2308 kW
and mass flow is increased due to the entrainment at the
jump. A head is formed at the leading edge of the flow.
Hz = 5.176 m
Mixing between this ceiling-layer flow and the ambient
cooler air occurs behind this head.
These two results are essentially identical for the purpose of
The flow that is formed travels along the hallway ceiling
ceiling jet flow analysis.
[see Figures 2-4.4, part (c) and 2-4.4, part (d)] with constant
The dimensionless maximum temperature in the ceil-
velocity and depth until it impinges on the end wall [see
ing jet flow, 4 m from the impingement point, is calculated
Figure 2-4.4, part (e)]. A group of waves are reflected back
from Equation 6, using the ceiling height above the substi-
toward the jump near the fire room, traveling on the inter-
tute source as
face. Mixing occurs during the wall impingement process
[see Figure 2-4.4, part (f)], but no significant entrainment
AT: = [0.188 + 0.313 (4/5.161 )]-4/3
occurs during the travel of the nonbreaking reflected wave.
When these waves reach the jump near the fire room door,
AT: = 3.076
the jump is submerged in the warm gas layer, eliminating the
entrainment of ambient lower layer air at this position. [See
Using the corresponding energy release rate for the sub-
Fi8ure 2-4.4,part (~).]After several wave reflections up and
stitute source and the upper layer ambient temperature, the
dimensional temperature elevation at the position in the down the corridor along the interface, the wave motion dies
ceiling jet is out, and a ceiling layer more uniform in depth is produced.
This layer slowly grows deeper as the hot gas continues to
ATs AT:TW (i&)z/3
flow into the hallway from the fire room.
It is clear from this description that quantification of
= 3.076 *323* [2313/( 1.092 *1*323 *9.81i2* 5.1615/2 )]2/3 effects in the transient ceiling jet flow is quite complex.
‘2-38 FIRE DYNAMICS
Subscripts
O based on steady fire
[d) 1 lower layer
2 upper layer
w ambient, outside ceiling jet or plume flows
convective fraction
~ associate with gas travel delay
I value at the interface position between the warm upper
layer and cool lower layer
i reference value at ignition
Superscripts
* dimensionless quantity
REFERENCES CITED
(f)
1. R.W. Pickard, D. Hird, and P. Nash, F.R. Note 247, Building
Research Establishment, Borehamwood (1957).
2. P.H. Thomas, Hf. Note 141, Building Research Establishment,
Borehamwood [1955).
3, R.L. Alpert, Fire Tech., 8, 181 (1972).
4. R.L. Alpert, Comb. Sci. and Tech., 11, 197 (1975).
(9)
5, H.Z. You, Fim and Matk., 9,46 (1985).
6. G. Heskestad and T. Hamada, FMRC J.I. OKOEI.RU 070(A),
Fig. 2-4.4. Transient ceilirrgjetflow inaroom rmdcorridor. z4 Factory Mutual Research Corp., Norwood (1984).
L.Y. Cooper, NBSIR 87-3535, J. ofHeat Trans., 104, 446 (1982).
7.
Analysis and experiments are under way to understand bet- L.Y. Cooper and A. Woodhouse, J. of Heat Trans., 108, 822
8.
ter the major features of a ceiling jet flow in a corridor. 25,26 (1986].
9. H.Z. You and G.M. Faeth, Fire and A4atk., 3, 140 (1979).
10. C.C. Veldman, T. Kubota, and E.E. Zukoski, NBS-GCR-77-97,
SUMMARY U.S. National Bureau of Standards, Gaithersburg [1977).
11. E.E. Zukoski and T. Kubota, NBS-GCR-77-98,National Bureau
Reliable means are available to predict the temperatures of Standards, Gaithersburg (1977).
and velocities of gases in fire-driven ceiling jet flows beneath 12, V. Motevalli and C.H. Marks, “Characterizing the Unconfined
unobstructed ceilings for both steady and t2fire growth. Ceiling Jet under Steady-State Conditions: A Reassessment,”
These predictive methods apply to quantifying the maxi- Fire Safety Science Proceedings of the 3rd International Sym-
mum temperature and maximum velocity at a given position posium, G. Cox and B. Langford, eds., Elsevier Applied Science,
in the ceiling jet flow and apply to situations where the flow New York, 301 (1991].
can be considered unconfined. These methods are the basis 13. D.D. Evans and D.W. Stroup, Fire Tech., 22,54 (1986).
14. E.E. Zukoski, T. Kubota, and B. Cetegen, F. Safety j., it, 107
for acceptable design methods exemplified by Appendix B of (1981).
NFPA 72, National Fire Alarm Code. 19 15. G. Heskestad and M.A. Delichatsios, The Initial Convective Flow
in Fire, 17th International Symposium on Combustion, Com-
NOMENCLATURE bustion Institute, Pittsburgh (1978).
16, C.L. Beyler, F. Safefy]., 11, 53 [1986).
17. R.P. Schifilliti, Use of Fire Pkme i’heory in the Design and
A g/(cpT~ p w ) Analysis of Fire Detector and Sprinkler Response, Thesis,
heat capacity at constant pressure Worcester Polytechnic Institute (1986).
?T constant related to plume- flow value 9.115’4 18, G. Heskestad and M.A. Delichatsios, NBS-GCR-77-86 and NBS-
g gravitational acceleration GCR-77-95, National Bureau of Standards, Gaithersb~g
H ceiling height above fire source (1977].
CEILING JET FLOWS 2-39
19. NFPA. 7?3 ~a~o:al Fire tiarm Code, Nationaf Fire protection 24. E.E. Zukoski, T. Kubota, and C..% Lim, NBS-GCR-85-493, Na-
Assoclab:n, Qumcy, MA (1993). tionaf Bureau of Standards, Gaithersburg [1985).
Zo. N4,A.Dehchatsios, Comb. and FJame, 43, I (1981). 25. H.W. Emmons, “The Ceiling Jet in Fires,” Fire Safe& Science,
~l. D.D. Evans, Comb. Sci. and Tech, 40, 79(1984). Proceedings of the 3rd International Symposium, G. Cox and B.
22. D.D. Evans, F. Safe& J., 9, 147 (1985). Langford, eds., ElsevierApplied Science, New York, 249 (1991].
23. L.Y. Cooper, A Buoyant Source in the Lower of 7w0, Ho- 26. W.R. Chan, E.E. Zukowski, and T. Kubota, “Experimental and
~ogeneous, Stably Stratified Layers, .20tlr International Symp- Numerical Studies on Two-Dimensional Gravity Currents in a
osium on Combustion, Combustion Institute, Pittsburgh Horizontal Channel,” MST-GCR-93-630, National Institute of
(1984). Standards and Technology, Gaithersburg (1993].