The Simulation of the Evacuation of the People in Tunnel Fires by alicejenny

VIEWS: 20 PAGES: 5

									                     Computer Model for People’s Safe Escape From Tunnel Fire
        YANG Gaoshang1, PENG Limin1, PENG Jianguo2, ZHANG Jinghua2, ZHAO Mingqiao1 & AN Yonglin1
               (1 School of Civil Engineering and Architecture, Central South University, Changsha 410075, Hu’nan, China;
                2 Hunan Province Communication Planning Survey and Design Institute, Changsha 410008, Hu’nan, China)

Abstract: The fundamental principle for tunnel fire fighting design is to ensure the safety of people’s escape from tunnel fire. Therefore the computer model
Tunev (tunnel evacuation) for the people's safe escape was developed. The model can predict the total escape time of various cases in tunnel, combines fire
simulation soft FDS which predict the fire dangerous approach time, the people's escape safety is estimated by comparing with the 2 kinds time, the escape
process also have a dynamic demo. The model contains simple database of the people’s behavior and structure parameters of tunnel fireproofing. The modeling
results can find out the deficiency of fire design and provide references for people’s safe escape scheme. Moreover, the model can also work as an adjutant tool
of fireproofing exercises. The article described some relevant concepts of the model in detail and an escape modeling of an object example—Xuefeng tunnel
was given out by using this computer model. Finally the model’s validation and actual application was described briefly.
Keywords: tunnel fire; people's safe escape; computer model


1      Introduction
     With the development of the express road construction, there are more and more tunnels. Many people and vehicles will pass or stay
in the tunnels. Once fire occurred, how to evacuate the people safely and quickly is not only a problem which should be considered
carefully and researched in tunnel fire design but also an important research project of the emergency scheme for tunnel operation.
Computer modeling is the most economical and feasible method to research and check the design scheme or emergency scheme[1].
     To avoid or reduce personnel injuries and deaths in tunnel fires, more and more attention is paid to how to evacuate safely when
fire occurred, except for adopting fire protection measures to prevent fire disaster. Research of people’s escape can quantify the
behavior of escape and make the design of the escape more scientific and rational [2,3]. An important criterion of escape design is that
whether people can escape to safe areas before danger comes [4]. So modeling computation of people’s escape is required, therefore
the computer model Tunev (Tunnel Evacuation) for people’s escape in tunnel fires is developed. Based on behavioral characteristic of
people in tunnel fire and structure parameters of tunnel fireproofing, the database of people’s behavior is established in this model.
By inputting each parameter needed and adopting analysis computing of qualitative and quantitative, the escape time is computed. By
comparing the escape time and the fire dangerous time, the safety of the people is estimated. The modeling results have important
value for tunnel fireproofing design, the escape direction system and the establishment of the emergency scheme. The concepts and
principles related to this model are illustrated, and the escape modeling of an object example is computed in this paper.
2      Critical Conditions of Fire Danger and Safe Criterion of the People’s Escape in Tunnel
2.1    Critical Conditions of Fire Danger in Tunnel
      In tunnel fires, smoke is the main factor that threatens people’s safety. It is dangerous to people when some parameters of smoke
increase to certain values. So we can determine the dangerous conditions based on some parameters related to the people’s escape
from tunnel fire, such as the CO concentration, the temperature, the visibility, etc. By integrating former research datum [5,6], critical
conditions of fire danger are put forward as follows.
      1) People are forced to prolong escape time in high temperature and irritant gas environment due to lower visibility in smoke.
On account of narrow and length of tunnel, 10 m is adopted as a critical visibility at characteristic height of human eyes where smoke
layer have fall from ceiling of tunnel.
      2) When smoke layer have fallen from ceiling of tunnel to characteristic height of human eyes, the harm to human would be
direct burn or inhaling hot gas. When temperature of smoke is lower than 80 ℃, humans would have possibility of survive.
Otherwise, humans would face potential hazard or death. Thus, 80 ℃ is adopted as a critical temperature at characteristic height of
human eyes.
      3) When smoke layer is lower than characteristic height of human eyes, whether hazardous conditions have been reached would
be also judged by concentration of toxic combustion products. For example, CO concentration would be adopted as a critical
concentration at characteristic height of human eyes.
      Take 1.5 m as the average height of human eyes characteristic height. Among three conditions mentioned above, which one is
reached firstly would be adopted as a critical condition of danger.
      In this paper, fire simulation software FDS4.0[7] was employed to predict the dangerous approach time of critical conditions in
tunnel fires by dynamic simulate.
2.2 Safe Criterion of the People’s Escape
      Whether people can escape safely in a tunnel fire depends on 2 kinds of time [8]. One is the Tfire that the danger approach time of
critical conditions; the other is Tevac that the escape time people needed. If people can escape to the safe passage prior to danger
comes, they can be considered as safe. Tfire can be obtained by the fire simulation software FDS and Tevac can be calculated by the
people’s escape model Tunev.
      People’s escape action can’t go with occurrence of the fire. There are three stages from beginning to end of people’s escape:
detection time Tdet, response time Tresp and moving time Ttrav. So the total time from beginning to end of people’s escape is showed as


    Project (20033179802) supported by the Science and Technology Program of China Western Transportation Development

                                                                             355
follows.
                                                      Tevac= Tdet + Tresp + Ttrav                                                     (1)
     The basic safe criterion of the people’s escape is showed as
                                                           Tfire > Tevac                                                             (2)
     For the people’s escape safety, Formula (2) should be satisfied in each location with any fire scenarios.
3   Establishment of People’s Escape Model in the Tunnel
      On the basis of the work [9-13] of predecessors, the people’s escape model Tunev applied specially to tunnel is put forward in this
article.
3.1 Flow of the Modeling
      Tunev model is a modeling computer program by inputting each parameter needed and combining with dynamic simulate of
tunnel fire through fire simulation software FDS4.0. It adopts proper qualitative and quantitative analysis methods based on database
of people’s behavior estimate the safety of people’s escape. With FLASH 7.0 which is powerful in processing the dynamic figures
and VB 6.0 as platform, the model can produce a plane animation of fire scene and display dynamical escape process. Tunev model
can simulate escape process of different fire scenarios in tunnel. The flow chart shown in Fig.1 intended to illustrate the program
process of Tunev model.
3.2 Input Parameters of Tunev Model
                                                                                                Begin
      1) Parameters of tunnel fireproof
                                                            Input parameters of                                 Input parameters of
design: geometry dimensions of tunnel,                       fireproofing design                               fire state at beginning
                                                                                               Input
number and width of escape passages, etc.                    Input parameters of
                                                                                             module
                                                                                                                Input parameters of
      2) Parameter of fire types and locations               people and vehicles                               fire location and type
                                                                                            Function
of fire source: location of fire source, fire                 Call FDS software
                                                               Calculate Tfire
                                                                                             module               Call escape design
                                                                                                                   Calculate Tevac
scale, fire power, etc.                                                 No               Whether escape is
                                                                                        satisfied to criterion?
      3) Parameters of people and vehicles in                                                     Yes

danger: types of vehicles and number of                                                 Display module
                                                                             display dynamical escape process
people in danger, width of the door of
                                                                                                End
vehicles.
      4) Parameters of fire state at beginning:                  Fig. 1 Block diagram of program process of model Tunev
detection and alarm time, response time, etc.
3.3 Problems Considered in the Model
      1) Characteristics of people’s behavior and their escape time: people’s behaviors in tunnel fire are mostly affected by some
factors, such as, physics environment of fire scene; smoke flow state; structure characteristics of tunnel; people’s distribution;
characteristics of psychology and physiology, etc. The database of people’s behavior with search and reasoning mechanism were
established based on the factors[14,15], which achieved to simulate the people’s escape. For example, escape velocity is one of
characteristics of people’s behavior. So the escape velocity is ascertained by the factors, such as power of fire source; people’s
density; smoke temperature and visibility. The simulation ratiocination is as follows.
      If fire power ≤ 20 MW and people’s density ≤ 1.0 person/m2 and smoke temperature ≤80 ℃ and smoke visibility  10 m then
people’s Evacuate Velocity = 1.3-1.5 m/s.
      People’s behaviors are complex in the fires, which mainly incarnate the escape velocities. To simplify the computing, Tunev
model take the crowd in the escape passage as people flow with certain density, velocity and flux instead of considering behavior of
each one. So only the average velocity of the flow is considered in the model. According to the research [16, 17] of predecessors, the
average escape velocity of the people flow can be set to 1.0-1.5 m/s, then the escape time of the people flow is only related to the
escape distances and velocity of people flow.
      2) Effective width of the escape passage: maybe some there are some barriers in both sides of the tunnel, so the people always
instinctively keep some spaces away the walls and between themselves when they walk along the tunnel. Additionally, their swung
arms are also required keep space away walls. So there is a boundary layer in the escape passages, which is not used by the people
flow. The width of the boundary layer should not be calculated in the escape computation. According to the research, the effective
width of the escape passage is measured from the place which is away from the side walls about 150-200 mm. It is assumed that the
width of the people flow is equal to the effective width of the escape passage in this model.
      3) Crowd congestion before the entry portal of escape passage: the effective width of escape passage is definitely narrower than
the width of the tunnel which could causes congestion at the entry portal of the escape passage. At the same time, the width of the
escape passage cannot be full used by the people. The effective width is narrower about 300-400 mm than that of the tunnel.
Supposing that number of people in the dangerous zone is p , people passing the unit width of the escape passage per unit time is r,
and the width of the passage is w , the number of the exits is n, then the time [18,19] all the people passing the escape passages is
described as follows.
                                                                        p
                                                           t1 
                                                                  nr ( w  0.4)                                                      (3)
Where r is 1.2-1.5 persons/m·s, here 1.2 persons/m·s is adopted as the most disadvantageous situation. Taking a passenger car as an
example, if there are 45 persons needing be evacuated, the width of escape passage is 2.2 m and the number of exits is 1, we can
obtain the time that all the people passed the entry portal is 21 s using the Formula (3). Obviously, compared with the one that the

                                                                  356
people flow escaped in the tunnel, this time is so small to be neglected, because there is no obvious congestion happening at the entry
portal of the passage.
     4) The escape time of the people who are furthest away from the doors of vehicles: considering the normal people and the
vehicle with many seats, it is difficult for people to move. The escape speed [19] of people in vehicle is about 0.2-0.4 m/s, here 0.2 m/s
is adopted as the most disadvantageous situation. Because people can directly find the fire in vehicle, the average time of detecting
and response is 60 s or so based on the statistics of behavior database. The actual escape time should include the detection and
response time. Taking a passenger car as an example, assuming the length of the car is 10 m, the escape time of the people who
moves most slowly from the back to the front door of the car is as follows: t2 =10/0.2+60=110 s.
     As mentioned above, assuming that the width of the door is 0.8 m, as the most disadvantageous situation is concerned, we can
compute the time all the people pass the front door is calculated to be 94 s with Formula (1). As to these two characteristic times,
people will not be retarded at the door of the vehicle for t1 is less than t2. So the longer time t2 is taken as the time all the people
escape from the passenger car in the case of fire.
4     Project Example
4.1    Outlines of the Project
       The Xuefeng Mountain Tunnel is about 7 km long, with unidirectional traffic, longitudinal ventilated, cross-section is about 62
m2, which is the domination project of Shaohuai section of the Shangrui freeway. The initial design of the average spacing between
cross-passages is 250 m. The space of every 3 passages is taken as a fireproof zone. Take the left line as an example, fire happens at
the first fireproof zone near to the entrance of tunnel, the escape behavioral characteristics such as habitualness, returning,
photokinesis are dominant which are of advantage to the escape than other zones. Adopt this zone as a modeling zone to illustrate the
computation steps of the people escape modeling
program. The sketch map of the fire position in                                                         fire source              Entrance
                                                                  Left line
tunnel is showed as Fig.2.                                                     Downstream                                 Upstream


      According to the statistics of the tunnel of                             Vehicle
                                                                               passage
                                                                                                 Personnel
                                                                                                  passage
                                                                                                                   Personnel
                                                                                                                    passage

Xuefeng Mountain Tunnel, there are many passenger
                                                                 Right line
cars and few tank trucks, nature wind speed is 1.5
m/s. Considering there are many people needed to be                          Fig. 2 Sketch map of fire position in tunnel
evacuated in case of fire, the fires of passenger car
and tank truck are adopted, which fire power is 20 MW and 50 MW separately. According to ‘Design Specification of Ventilation and
Lighting of Highway Tunnel (JTJ026-1-1999), the ventilation speed of is 2-3 m/s, which is taken to avoid the adverse current of the
smoke. It is safe for people to escape upstream the fire under 3 m/s ventilation speed which is calculated by software FDS. Because
the natural air speed in the tunnel is 1.5 m/s, so we only discuss the case that people escape downstream the fire with 4 scenes, which
are 2 fire powers go with 2 ventilation speeds separately.
4.2 Escape Time Tevac Computing
      2 stages are simulated the escape process in the Tunev Model: Stage 1: the people’s escape in the passenger car which is on fire.
Firstly people need to be evacuated from the burning car in the tunnel. This escape time has been obtained in section 2.3 of this paper,
which is 110 s; Stage 2: the people’s escape in the tunnel. All the people entering the tunnel from the passenger car and other vehicle
near the fire form a people flow which escape along tunnel.
      After inputting all the parameters, Tunev Model compute the escape time of different fire scenes based on behavior
characteristic of people in tunnel fire. Simple analysis is as follows.
      Considering the most disadvantageous situation, it is the longest escape distance when the fire occurs at the entry portal of the
cross-passage where the distance between the front or the back cross-passage and the fire source is 250 m. The entry portal position
is also the dividing line to estimate the safety of escape. If this position where the fire occurs can satisfy the conditions of safe escape,
then any position before or behind the dividing line can satisfy the conditions.
      This model determines the direction based on different fire scenes and people’s behavior characteristics for there are only 2
directions of escape. When a fire occurs in a passenger car at the entry portal of the cross-passage, the great majority people affected
by the most possible behaviors of habitualness and returning will safely escape upstream the fire; few people affected by the
behaviors of pyrophobia and babelism, who are too afraid of the high temperature flash and smoke to get across the fire will escape
downstream the fire, moreover the escape speeds (1.3-1.4 m/s) under natural air speed and critical ventilation speed have little
difference. When a fire occurs in a tank truck at the entry portal of the cross-passage, all the people in the passenger car which
upstream the fire affected by behaviors mentioned above will safely escape upstream the fire; but the people in the tank truck also
possibly escape downstream the fire affected by the most possible behaviors mentioned above, moreover the escape speeds (1.5-1.4
m/s) under natural air speed and critical ventilation speed have many difference. 4 fire scenarios were designed to study escape times
      Integrate the analysis mentioned above, the calculation results are showed as Table 1.
                                       Table 1     Average escape time under different fire scenes
                                     Fire scenes                                           Escape time Tevac /s
               1       20 MW, 1.5 m/s (ventilation invalidation)                                    302
               2       20 MW, 3 m/s (critical ventilation speed)                                    288
               3       50 MW, 1.5 m/s (ventilation invalidation)                                    268
               4       50 MW, 3 m/s (critical ventilation speed)                                    252

                                                                    357
4.3   Danger Time Tfire Calculating
     After inputting all the parameters, Tunev Model call the software FDS to calculate the danger time while computing the escape
time.
     Because the average space between adjacent cross-passages is 250 m which is the maximum escape distance. So adopt 300 m
length tunnel as the computing model length in order to reduce the calculation. The burning car was assumed to be located in the
middle lane of the tunnel, 20 m from the entry portal of the computing tunnel model. The dynamic simulation of 4 fire scenes in the
Xuefeng Mountain Tunnel is performed by using FDS software. Adopt the entry portal A of the cross-passages, where is nearest from
the entry portal of the computing tunnel model, to illustrate the variety law of the CO concentration, the temperature, the visibility
changing with time. The coordinate of entry portal A is: x=270 m (250 m away from the fire source), y=2.6 (the focus area of the
escape flow in the cross section), z=1.5 (average body height is 1.5 m). The simulation model is shown in Fig.3-a .b and the curve of
simulation results for position A is shown in Fig.4-6.

                                                                              z                              z



                                          fire source                                                                     fire source
                                                                                                                                                                                  A


                                                                          A
                           y                                                                                                                                                                 x

          Fig.3-a                  Cross section of simulation model /m                                      Fig.3-b        Longitudinal section of simulation model /m


                      60                                                                                     700
                                        20MW-1.5m/s                                                                             20MW-1.5m/s
                      50                20MW-3.0m/s                                                          600
                                                                                                                                20MW-3.0m/s
      Temperature/℃




                                        50MW-1.5m/s                                                          500
                      40                                                                                                        50MW-1.5m/s
                                        50MW-3.0m/s
                                                                                                    CO/ppm




                                                                                                             400
                      30                                                                                                        50MW-3.0m/s
                                                                                                             300
                      20
                                                                                                             200

                      10                                                                                     100

                       0                                                                                         0
                               0    60 120 180 240 300 360 420 480 540 600                                           0    60      120   180     240    300     360   420   480   540   600
                                                                                                                                                      Time/s
                                                  Time/s

              Fig. 4               Curve of temperature-time of position A                               Fig. 5          Curve of CO concentration-time of position A


                                                                       35.0
                                                                                                                            20MW-1.5m/s
                                                                       30.0                                                 20MW-3.0m/s

                                                                       25.0                                                 50MW-1.5m/s
                                                        Visibility/m




                                                                                                                            50MW-3.0m/s
                                                                       20.0

                                                                       15.0

                                                                       10.0

                                                                        5.0

                                                                        0.0
                                                                              0   60   120 180   240 300 360             420 480        540 600
                                                                                                   Time/s

                                                                         Fig. 6    Curve of visibility-time of position A

     From the Fig.4-6, combining the critical conditions of fire danger, the danger times Tfire under 4 fire scenes are derived. The
results are shown in the Table 2.

                                                                       Table 2     Danger time under different fire scenes
                                                 Fire scenes                                                                    Danger time Tfire /s
                      1             20 MW, 1.5 m/s (ventilation invalidation)                                                                 325
                      2             20 MW, 3 m/s (critical ventilation speed)                                                                 290
                      3             50 MW, 1.5 m/s (ventilation invalidation)                                                                 267
                      4             50 MW, 3 m/s (critical ventilation speed)                                                                 217


                                                                                                 358
4.4    Safety Estimation
      From the simulation results above, when a fire of 20MW occurs in the tunnel, it is little safe even unsafe to escape downstream
the fire under critical ventilation speed and ventilation invalidation situation, even if the escape speed is higher than 1.3-1.4 m/s.
      Likewise,when a fire of 50 MW occurs in the tunnel, it is unsafe to escape downstream the fire under critical ventilation speed
and ventilation invalidation situation, even if the escape speed is higher than 1.5-1.4 m/s.
4.5 Escape Animation
      After modeling of people’s escape, a plane animation is formed to illustrate the safety escape process by means of AutoCAD
and FLASH which has a powerful function on graphics processing.
5     Conclusions
     Based on the analysis of the results, when a fire of 20-50 MW occurs in the tunnel, it is unsafe to escape downstream the fire
under critical ventilation speed and ventilation invalidation situation, so some effective measures must be adopted to direct people’s
escape upstream the fire source; when a fire of 20-50 MW occurs in the tunnel, the ventilation velocity is at least 3 m/s to guarantee
the safe escape, which are in good agreement with specification and similar test [20-22], so Tunev model is proved to be validity.
                    the
     In the future, design of escape must be checked before construction and operation of the tunnel. Through modeling by Tunev
model, the defaults of the design is easily to be found out so that some measures can be taken,such as to shorten the distance
between every 2 cross-passages to reduce escape time; Increase the fire equipments to prolong the coming of danger time.
Furthermore, Tunev model can work as an aid to fire fighting exercises. However, the selection of the model parameters and the
completion of the behavior database should be researched further more and adjusted through continuous practical application, so as
to make the process of the modeling more concise and visual.

Acknowledgements
     The presented work was partially supported by Hunan province communication planning survey and design institute, which is
gratefully acknowledged.

References:
[1]   Xu Gao. Simulation model for crowd evacuation based on agent technology. Journal of Southwest Jiaotong University, 2003(6): 301-303(in Chinese)
[2]   Hu Zhongri. The current situation and development tendency of the fire evacuation. Fire Science and Technology, 2001(6): 6-7(in Chinese)
[3]   Gwynne S, Galea E R, Owen M, Lawrence P J, Filippidis L. A review of the methodologies used in evacuation modelling. Fire and Materials, 1999,
      23(6) :383-388
[4]   Bendelius A G. Tunnel fire and life safety within the world road association (PIARC). Tunnelling and Underground Space Technology, 2002(17):
      159-161
[5]   Hartell G E. Engineering analysis of hazards to life safety in fires: the fire effluent toxicity component. Safety Science, 2001, 38(2): 147-155
[6]   Dr Terje H Toften, Dr Bård Venås. Phoenics in safety analysis of offshore and underground constructions. Phoenics User Conference. Melbourne, 2004
[7]   McGrattan K B, Baum H R, Re h m R G, et al. Fire Dynamics Simulator (Version 3). NIST, 2002
[8]   Japan Construction Department. The Building Synthetic Fire Prevention Design. Sun Jinxiang, Gao Wei, tran. Tianjin: Tianjin Science and Technology
      Translation Press, 1994(in Chinese)
[9]   Michael M, Kostreva, Laura C. A comparison of two methodologies in HAZARD fire egress analysis. Fire and Materials, 1999, 23
[10] Rita F Fahy. Fire Analysis and Research Division. National Fire Protection Association
[11] Lo S M, Fang Z. A spatial- grid evacuation model for buildings. Journal of Fire Science, 2000, 18(5): 376-394
[12] Wang Zhigang, Zhang Yinghua, Bi Shaoying. Using computer model to evaluate the fire safety design of actual building. Fire Science and Technology,
      2001(2): 6-10(in Chinese)
[13] Hadjisophocleous G V, Proulx G, liu Q. Occupant Evacuation Model for Apartment and Office Buildings. Canada: Institute for Research in Construction,
      1997
[14] John L Bryan. Human behavior in fire: the development and maturity of a scholarly study area. Human Behavior in Fire Proceedings of the First
      International Symposium. University of Ulster, Northern Ireland, 1998
[15] Fraser- Mitchell J N. Modelling human behavior within the fire risk assessment tool CRISP. Fire and Materials, 1999, 23(6): 349-355
[16] PIARC, Fire and Smoke Control in Road Tunnels, PIARC Committee on Road Tunnels. Paris, France, 2002. 31-32
[17] Zhang Z D. Fire disaster ventilation of highway tunnel. Moden Tunnel Technology. 2003, 40(1): 3-637(in Chinese)
[18] Huo Ran, Hu Yuan, Li Yuanzhou. Safety Engineering Introduction. Beijing: China Science Technique University Press, 1999. 131-135(in Chinese)
[19] Huo Ran, Jin Xuhui, Liang Wen. Simulation analysis for people’s escape in large public building fire. Fire Safety Science, 1999, 8(2): 8-14(in Chinese)
[20] YAN Z G, YANG Q X. Study on temperature field distribution in qinling road tunnel by fire experiments. Undergroun Space, 2003, 23(2): 191-195(in
      Chinese)
[21] ZENG Y H, HE C, GUAN B SH. Study on simulation of temperature field in tunnel during fire. Underground Space, 2004, 24(1): 69-71(in Chinese)
[22] Jojo S M Li, W K Chow. Numerical studies on performance evaluation of tunnel ventilation safety systems. Tunnelling and Underground Space
      Technology, 2003, 18: 435-452




                                                                               359

								
To top