j ourna l ho me page: www.elsev ier .com/ locate /enbui ld
umerical investigation of leaking and dispersion of carbon dioxidendoor under ventilation condition
hirong Wang ∗, Yuanyuan Hu, Juncheng Jiangiangsu Key Laboratory of Urban and Industrial Safety, Institute of Safety Engineering, Nanjing University of Technology, Nanjing 210009, China
r t i c l e i n f o
rticle history:eceived 16 November 2012eceived in revised form 21 March 2013ccepted 28 June 2013
eywords:oom ventilationarbon dioxideelease and dispersion
a b s t r a c t
Numerical simulation of continuous release and dispersion process of carbon dioxide in a ventilated roomwas carried out by FLUENT with Computational Fluid Dynamic (CFD) method. The validation of the simu-lation method was made by comparison of simulation results with experimental data. The concentrationdistribution of carbon dioxide indoor was investigated. The effect of the release rate, wind speed andobstacle on gas dispersion was analyzed. Results showed that at the beginning carbon dioxide moved tothe top of room by the action of jet flow. Under the condition of wind, the concentration of carbon dioxideincreased along the leeward. Over a period of time, the concentration of carbon dioxide indoor kept con-stant. Gas-detecting alarm device should be installed at the downwind position above the leaking hole.
indbstacleeleasing rate
People should get away from the room before high concentration area forms. The concentration of carbondioxide downwind decreased with increment of wind speed, while increased with increment of releaserate. Vents or forced draft should be set to decrease the concentration of carbon dioxide. When therewas obstacle behind the leaking source, a region of high concentration appeared in windward side of theobstacle. When designing the layout of industrial workshop, it is necessary to minimize the quantity ofequipment indoor or put them near the wall, away from leakage sources.
With the development of modern city, chemical and petro-hemical industry, confined spaces have been developed rapidly,ncluding civil architectures, plants, underground tunnels andubways, etc. Because these buildings were relatively closed, acci-ents such as fire, explosion and poisoning may happen once theammable, explosive, virulent and harmful substances release. Itill cause not only wealth loss, but also casualties [1–3]. Carbonioxide is colorless, odorless gas which is widely used in many fieldsf industry such as industry, food, medicine and agriculture, etc.lthough carbon dioxide is non-inflammable and non-explosive,igh levels of carbon dioxide in air will lead to death of people dueo hypoxia and asphyxia. Currently, more research work is focusedn release and diffusion of carbon dioxide outdoor [4–6], but thetudy on release and diffusion of carbon dioxide indoor is less [7].nstalling toxic gas alarm system is an effective safety measure
or preventing the recurrence of similar accidents. But where thelarm system should be installed, and how to evacuate people inhe building if such accident occurs still need further investigation.
To solve these problems, the diffusion regularity of carbon dioxideindoor needs to be studied.
In this paper, the continuous release and diffusion process of car-bon dioxide indoor in a ventilated room was numerically simulatedwith FLUENT. The distribution of carbon dioxide in the room wasinvestigated. The effects of release rate, wind speed and obstacle onthe concentration distribution of carbon dioxide were also studied.The research results will provide theoretical basis and importantreference for accident prevention and emergency rescue.
2. Experiment
Carbon dioxide cylinder produced by Nanjing Special Gases Fac-tory was used in the experiment with the volume of 0.04 dm3 andthe pressure of 10 MPa, which was 1.32 m high, 0.219 m in diameter.The diameter of a circular leakage hole on the top of the steel cylin-der was 0.004 m. Pure carbon dioxide leaked into the room alongthe vertical direction at a continuous flow rate of 0.05 kg/s. Thesize of air inlet was 0.4 m in the longitudinal direction and 0.5 m inthe vertical direction. Concentration sensors were arranged in the
room (see Table 1). Firstly, the outlet pressure will be reduced toa steady value by adjusting pressure-reducing valve. The gas flowrate was obtained by adjusting the flowmeter. Then, the concen-tration of carbon dioxide detected by sensors will be transferred
o computer by data acquisition system. Finally, the experimentalesults were obtained by data acquisition system.
The sketch of the experiment was shown in Fig. 1. The volumef the room is 16 m3, which is 4 m in the lateral direction, 1.6 m inhe longitudinal direction and 2.5 m in the vertical direction. Theize of air outlet was 0.5 m in the longitudinal direction and 0.4 mn the vertical direction. The velocity of air inlet was zero in thexperiment.
. Modeling strategy
.1. Basic conservation equation
The leakage and diffusion of CO2 were described by usingasic equations of mass, momentum, energy and transportationquations. The solution of mathematical model was based on thequations of mass, momentum, energy and species transport con-ervation. Thus these differential equations used in the simulationould be expressed in the general form as follows [8,9]:
∂
∂�(��) + ∂
∂xj(�uj�) = ∂
∂xj
(�
∂�
∂xj
)+ S (1)
The four terms in the general equation are called time term, con-ection term, diffusion term and source term respectively. Where
is the time, s; � is the density of gas, kg/m3; � is the general vari-ble; uj is the velocity vector component (m/s); � is the diffusionoefficient; S is source term.
.2. Physical model
The physical model was shown as in Fig. 1. The three-imensional model was 4 m × 1.6 m × 2.5 m cuboids. Wind speedas 1.5 m/s.
.3. Calculation conditions
The following assumptions were made:
1) Carbon dioxide released at sonic constant rate. The change ofthe pressure of CO2 in cylinder was ignored.
2) There was no chemical reaction, phase change reaction anddroplet deposition.
3) Wind direction was along the x-axis horizontal. Wind speed andwind direction did not change with time, location and height.
4) There were no apparent difference between the temperature ofleaking carbon dioxide and air. So there was no heat exchange.
In this paper, the calculation domain was meshed into non-omogeneous hexahedrons. For the wall, the standard wall
unction method was adopted [10].Initial conditions: at the initial time, pure carbon dioxide was
pecified at the inlet. And 100% air was specified at the air inlet.
00 K was all specified at walls, fluid field, carbon dioxide and air.o choose separated solver and the implicit scheme model, thetandard � − ε turbulent model and component transport modelere used.
dings 66 (2013) 461–466
3.4. Validation of the calculation models
Fig. 2 shows the comparison of experimental data and simulatedvalues of four specified points. It was found that the simulatingresults were in good agreement with the experimental data. Itsuggested that the simulation method is feasible. However somedifference existed because of the error of instruments, selection ofparameters and assumptions of simulation.
4. Results and discussion
4.1. Concentration distribution
Fig. 3 shows the concentration distribution of carbon dioxideon the plane of y = 0.8 m at 10 s, 50 s, 100 s, 150 s, 200 s and 300 s,respectively. At the beginning of the leakage, carbon dioxide movedto the top of room with a high initial momentum. Thereafter, car-bon dioxide diffused along the downwind by the effect of wind,and the concentration downwind continued to increase. Whencarbon dioxide moved around the walls, some of them flowedfrom the outlet, and recirculation region formed near the wall. Theconcentration of carbon dioxide windward increased because ofaccumulation. After about 150 s, the flow field was stable and thedistribution of carbon dioxide in the room was unchanged.
Fig. 4 shows the relationship between concentration of car-bon dioxide and downwind distance on the plane of y = 1.0 m fordifferent heights. The concentration of carbon dioxide was lowat the bottom and high at the top of the room. The concentra-tion stratification below leaking inlet formed obviously, while theconcentration above the inlet between different heights was notstratified. After some distance from the leaking source, the con-centration increased obviously with the increasing of height, andstratification appeared. Subsequently, the concentration decreasedwith increasing distance, and stratification disappeared. After sometime, the concentration below the leaking point increased as shownin Fig. 5. However, the increment of the concentration above theleaking point was different as shown in Fig. 6. The concentrationincreased suddenly at the distance between about x = 1.0 m and1.5 m which is in corresponding to the area of gas exit. The jet fromthe leakage had an important effect on the dispersion. Carbon diox-ide diffused vertically upwards due to the high flow velocity fromthe leakage.
Because of the significant impacts of jet and wind, with noobvious influence by gravity, gas-detecting alarm device should beinstalled at the downwind position above the leaking hole. Peo-ple should get to the windward area away from leaking source toescape from the accident. And people must get away from the roombefore high concentration area forms.
4.2. Wind speed
Regarding different wind speed, the concentration of carbondioxide with the release time at one point is shown in Fig. 7. Fig. 8shows the change of the concentration along downwind distancewhen the wind speed varies. The result indicated that wind speedhad strongly effect on the diffusion of carbon dioxide. The concen-tration of carbon dioxide downwind decreased with increment ofwind speed. One reason for this was that the concentration wasdue to the enhancing of stratospheric transportation of gas cloudwith the increment of wind speed. Moreover, as the wind speedincreased, the velocity fluctuation and the turbulence intensity
strengthened, and carbon dioxide rapidly dispersed correspond-ingly. So in order to prevent suffocation accidents, vents should beset to decrease the concentration of carbon dioxide. When the windspeed is slow, forced draft should be used.
Z. Wang et al. / Energy and Buildings 66 (2013) 461–466 463
ch of t
4
bFi
Fig. 1. The sket
.3. Releasing rate
Regarding different releasing rates, the concentration of car-on dioxide with the release time at one point is shown in Fig. 9.ig. 10 is the concentration of carbon dioxide along the groundn the direction of wind in the room when the release rate was
0 50 0 100 0 150 0 200 0 250 0 300 00.000
0.002
0.004
0.006
0.008
0.010
0.012
co
ncen
trati
on
of
CO
2 /
vo
l
t/s
exper ime nt
simul atio n
sensor1
0 50 0 100 0 150 0 200 0 250 0 300 00.000
0.005
0.010
0.015
0.020
co
ncen
trati
on
of
CO
2 /
vo
l
t/s
exper ime nt
simul atio n
sensor3
Fig. 2. Comparison of experimental data a
he experiment.
0.01 kg/s, 0.02 kg/s, 0.05 kg/s and 0.1 kg/s respectively. The resultshows that release rate had also crucial effects on the concentra-
tion. With high release rate, more carbon dioxide has been diffusedindoor, and then the concentration of carbon dioxide was muchhigher. So when accidental release happens the valve of cylindermust be shut off as soon as possible.
0 50 0 100 0 150 0 200 0 250 0 300 00.000
0.005
0.010
0.015
0.020
OC
fo
noit
artn
ec
no
c2 /
vo
l
t/s
experiment
simul atio n
sensor2
0 50 0 100 0 150 0 200 0 250 0 300 00.000
0.002
0.004
0.006
0.008
0.010
0.012
0.014
co
ncen
trati
on
of
CO
2 /
vo
l
t/s
experiment
simul atio n
sensor4
nd simulated values of four points.
464 Z. Wang et al. / Energy and Buildings 66 (2013) 461–466
2 on t
4
t(Tt
Fp
into two flows by the obstacle. One flow accumulated largely in
Fig. 3. Contour for concentration of CO
.4. Obstacle
The release rate for the carbon dioxide was 0.02 kg/s, andhe wind speed was 1.5 m/s in x-direction, and an obstacle
0.5 m × 1 m × 1.5 m) was placed at x = 1 m from the leakage source.he design of physical model was shown in Fig. 1. Velocity vectors athe plane of z = 0.8 m are given in Fig. 11. Fig. 12 shows the change of
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.00.00
0.02
0.04
0.06
0.08
0.10
con
cen
trat
ion
of
CO
2
x/m
z=0.5m
z=0.8m
z=1.0m
z=1.5m
z=2.0m
ig. 4. Relationship between concentration of CO2 and downwind distance on thelane of y = 1.0 m plane for different heights.
he plane of y = 0.8 m at different time.
concentration of carbon dioxide with the release time for differentmonitoring points (2, 0.8, 0.5) and (3, 0.8, 0.5), whether the obsta-cle existed or not. The gas stream released into the room divided
windward side of the obstacle. Another reached in leeward side ofthe obstacle. A region of high concentration appeared between theobstacle and the steel cylinder. Gas concentration in leeward side
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.00.00
0.02
0.04
0.06
0.08
0.10
con
cen
trat
ion
of
CO
2
x/m
t=10s t= 30s t=50 s
t=100s t=150s
Fig. 5. Relationship between concentration of CO2 and downwind distance on theline of y =1.0 m, z = 0.5 m at different time.
Z. Wang et al. / Energy and Buildings 66 (2013) 461–466 465
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.00.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
OC
fo
noitar t
necn
o c2
x/m
t=10s
t=30s
t=50s
t=100 s
t=150 s
Fig. 6. Relationship between concentration of CO2 and downwind distance on theline of y = 1.0 m, z = 1.5 m at different time.
0 50 10 0 15 0 20 0 25 0 30 0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
con
cen
trat
ion
of
CO
2
t/s
1.5m/s
2.0m/s
3.0m/s
4.0m/s
Fig. 7. Relationship between concentration of CO2 and time at the point (2, 0.5, 0.5)for different wind speed.
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.00.00
0.02
0.04
0.06
0.08
0.10
conce
ntr
atio
n o
f C
O2
x/m
1.5m/s
2.0m/s
3.0m/s
4.0m/s
Fig. 8. Relationships between concentration of CO2 and downwind distance on theplane of y = 1.0 m for different wind speed.
0 50 100 150 200 250 3000.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
conce
ntr
atio
n o
f C
O2
t/s
0.01kg/s
0.02kg/s
0.05kg/s
0.1kg/s
Fig. 9. Relationship between concentration of CO2 and time at point (2, 0.5, 0.5) fordifferent release rates.
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.00.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
conce
ntr
atio
n o
f C
O2
x/m
0.01kg/s 0.02kg/s
0.05kg/s 0.1kg/s
Fig. 10. Relationship between concentration of CO2 and downwind distance on theplane of y = 1.0 m for different release rates.
Fig. 11. Velocity vector at the plane of z = 0.8 m.
Fig. 12. Relationship between concentration of CO2 and time for different monitor-ing points.
of the obstacle was less than that in the case of no obstacle. Basedon the above discussion, it is necessary to minimize the quantityof equipment indoor or put them near the wall, away from leakagesources.
5. Conclusions
By analyzing the concentration distribution and influence fac-tors of carbon dioxide indoor under ventilation condition, followingconclusions can be drawn:
(1) Gas-detecting alarm device should be installed at the down-wind position above the leaking hole. People should get away
from the room before high concentration area forms.
(2) The concentration of carbon dioxide downwind decreased withincrement of wind speed. Vents or forced draft should be set todecrease the concentration of carbon dioxide. When accidental
4 d Buil
(
A
y2B
R
66 Z. Wang et al. / Energy an
release happens the valve of cylinder must be shut off as soonas possible.
3) When designing the layout of industrial workshop, it is neces-sary to minimize the quantity of equipment indoor or put themnear the wall, away from leakage sources.
cknowledgements
The authors are grateful to the support given by “eleventh five-ear” National Science and Technology support plan of China (No.006BAK01B03) and Science Foundation of Jiangsu Province (No.K2007587).
eferences
[1] M.M. Hefny, R. Ooka, CFD analysis of pollutant dispersion around buildings:effect of cell geometry, Building and Environment 44 (2009) 1699–1706.
[2] J.M. Santos, E.V. Goulart, N.C. Reis Jr., et al., Numerical simulation offlow and dispersion around an isolated cubical building: the effect
[
dings 66 (2013) 461–466
of the atmospheric stratification, Atmospheric Environment 43 (2009)5484–5492.
[3] L. Ricciardi, C. Prévost, L. Bouilloux, et al., Experimental and numerical study ofheavy gas dispersion in a ventilated room, International Journal of Hydrogen152 (2008) 493–505.
[4] J.D. Posner, C.R. Buchanan, D. Dunn-Rankin, Measurement and prediction ofindoor air flow in a model room, Energy and Buildings 35 (2003) 515–526.
[5] S. Sklavounos, F. Rigas, Validation of turbulence models in heavy gas dispersionover obstacles, Journal of Hazardous Materials 108 (2004) 9–20.
[6] A. Robins, I. Castro, P. Hayden, et al., A wind tunnel study of dense gas dispersionin a neutral boundary layer over a rough surface, Atmospheric Environment 35(2001) 2243–2252.
[7] K.A. Papakonstantinou, C.T. Kiranoudis, N.C. Markatos, Numerical simulationof CO2 dispersion in an auditorium, Energy and Buildings 34 (2002) 245–250.
[8] D. Schmitidt, U. Krause, U. Schmidtchen, Numerical simulation of hydrogengas release between buildings, International Journal of Hydrogen 24 (1999)479–488.
[9] F. Gavelli, E. Bullister, H. Kytomaa, Application of CFD (Fluent) to LNG spills
into geometrically complex environments, Journal of Hazardous Materials 159(2008) 158–168.
10] K. Matsuura, H. Kanayama, H. Tsukikawa, et al., Numerical simulation of leakinghydrogen dispersion behavior in partially open space, International Journal ofHydrogen 33 (2008) 240–247.
