当前位置:首页 >> 电力/水利 >>

Risk assessment for fire safety considering characteristic evacuees and


J Mar Sci Technol (2005) 10:147–157 DOI 10.1007/s00773-005-0193-2

Risk assessment for ?re safety considering characteristic evacuees and smoke movement in marine ?res
Nobuyoshi Fukuchi and Teruyuki Imamura
Department of Marine Systems Engineering, Faculty of Engineering, Kyushu University, 6-10-1 Hakozaki, Higashi-ku, Fukuoka 812-8581, Japan

Abstract In the safety design of marine systems, the matter of human life in the event of a marine ?re must be considered. It may be dif?cult to establish a safe evacuation system because a characteristic behavioral pattern based on human factors is elusive. This study consists of three analyses on (1) the state of smoke diffusion, calculated by the use of a two-layer zone model, (2) evacuation movements, simulated by a group behavioral model, and considering the occurrence of panic and the smoke spreading phase, and (3) the risk index for safety assessment using the results of smoke diffusion and escape movements. This risk index can be used for comparisons between various safety systems, and its validity is con?rmed by an evacuation model of common spaces on a typical cruise ship. Key words Risk assessment · Fire safety · Psychological model · Smoke diffusion · Escape simulation · Risk index

1 Introduction It may be dif?cult to establish a methodology for safe evacuation in the case of an outbreak of marine ?re owing to the intricacies of ?re-spreading phenomena and uncertain escape behaviors based on human factors which are not uniform. Suitable measures have to be taken to cope with the rapid spread of smoke that is apt to endanger even personnel away from the source of the ?re outbreak. In addition, it is necessary to establish a redundancy system for safe evacuation because of the existence of factors which cannot be measured, such as the ambiguity of human behavioral patterns. Accord-

Address correspondence to: N. Fukuchi (fukuchi@nams.kyushu-u.ac.jp) Received: December 19, 2003 / Accepted: January 28, 2005

ingly, the design criteria for the safety of an evacuation system, which consists of escape routes and human ?ow control, should preferably be instituted with an understanding of personnel behavior based on simulations of smoke diffusion and evacuation. There have been a number of studies attempting to simulate ?re phenomena, and a recent review has been provided by Tieszen.1 Zone models, in which roomaveraged quantities are predicted for multiroom problems, have been used extensively for engineering applications. Recently, ?eld models, which make a simulation possible in much ?ner spatial and temporal resolution, have been used in the study of ?res. In the case of a ?eld model approach, the Boussinesq approximation that is generally used is not suitable for the ?re problem.2,3 However, this dif?culty has been resolved by Erlebachar et al.4 The smoke ?ow model is required to account for the effect of unresolved smallscale movements. Here, the calculation of smoke movement was carried out using a two-layer zone model5,6 in which the temperature and yields of product species in each zone are microscopically assumed to be uniform, and the balance equations of all the state variables are deduced based on the relationship between neighboring zones. In addition, the evacuation movement7 could be predicted numerically by the proposed method using a group behavioral model which considers (1) action ability according to the ages of evacuees, (2) the decrease in walking speed with a lowering smoke layer, and (3) the occurrence of self-isolation in psychological action. The risk index for a safety assessment of the evacuation system was de?ned by using the results of a simulation of smoke diffusion and escape behavior simultaneously. This risk index may be useful for choosing a suitable escape route and for comparison with various supported systems for safe evacuation. Its validity can be con?rmed using numerical examples from the threedecked common spaces on a typical cruise ship.

148

N. Fukuchi and T. Imamura: Risk assessment for ?re safety

2 Safe evacuation system and human behavior An evacuation system involves the interdisciplinary consideration of smoke transport, heat transfer, toxicology, geometrically complex spaces, and human behavior with psychological intelligence during emergencies.7 Planning for safe evacuation consists mainly of designing appropriate escape routes with complete knowledge of the shipboard layout and the various support systems, such as those designed to prevent ?re and smoke. 2.1 Propensities of evacuation action Evacuees behave in various ways during a marine ?re, and tend to act by instinct.8 In particular, the following instinct and the homing instinct are apt to lead to mistakes in selecting an escape route. In addition, incorrect judgment of the escape procedure generally tends to occur when passengers are in a state of panic on a cruise ship. To avoid this, clear indications of escape routes and adequate guidance on evacuation are extremely important. A swift escape depends on walking speed, which varies with age, and therefore walking speed is divided into three groups: children (about 1.0 m/s), youths/ adults (about 1.3 m/s) and the elderly (about 1.0 m/s). Further, walking speed is generally altered by the walking posture adopted under ?re smoke. The walking speed on a stairway is taken to be 50% of the usual rate.

2.2 Mental state in an emergency and the occurrence of a state of panic 2.2.1 The psychological intelligence process in a state of emergency The psychological intelligence process in emergency conditions, such as marine ?res or a sinking ship, is shown in the ?ow chart9 in Fig. 1a, and is composed of recognition of the circumstances, judgment of the situation, action with comprehension, and external correspondence. During dangerous conditions, the comprehending script and the action script are activated to rede?ne the circumstances, and to produce the expectation of action. These scripts are the ?xed form of the individual’s knowledge about judging a speci?ed situation, and normative action depends on the quantity of education and training taken in the life of the person concerned. In a case of severe expectancy of the circumstances, mental self-isolation, like a panic state, occurs as a consequence of an emotional reaction beyond the threshold of fear. To prevent this self-isolation, one needs to increase the quality and quantity of both scripts by undergoing suitable training and obtaining adequate education on emergency states. 2.2.2 A mathematical model of the intelligence process A block diagram of the psychological intelligence processes in an emergency is shown in Fig. 1b as a

Fig. 1. Model of psychological intelligence processes in a state of emergency

N. Fukuchi and T. Imamura: Risk assessment for ?re safety

149

mathematical model10 of control engineering, in which u is the stimulus by an accident (input), y is the reaction of emotions and action (output), A is the coef?cient of the action script, B is the coef?cient of the monitoring ability, and 1/s means an integral operation. The transfer function of this arithmetic model and the state equation can be expressed as

capacity is shown in Table 1. Based on many case studies,9,11,12 it is assumed in this calculation that human abilities will decrease by as much as 15% during an outbreak of self-isolation. 3 Escape simulation considering smoke diffusion 3.1 Modeling evacuation

( )= U ( s) s
Y s

C , 2 + As + B

d2 y dy +A + By = Cu dt 2 dt

(1)

In order to fully assess the potential evacuation ability of a ship, it is useful for the evacuation model7,13,14 to be

in which t denotes time, and C is a constant governed by the comprehending script. The emotional reaction for input stimulus can be simulated as time proceeds in the condition of the given magnitudes of the coef?cients A, B and C. The control factors for a psychological model may be determined by the analytical results of psychological experiments11 and a case study9 on accident instances. The magnitude of the stimulus is classi?ed into grades I–IV, and the grades for each case of marine ?re are decided based on several case studies, as follows8,9: —Grade I, hearing the emergency ?re alarm; —Grade II, exposure to a smoke layer under 1.5 m high; —Grade III, exposure to a smoke layer under 0.9 m high; —Grade IV, (i) exposure to a smoke layer under 0.6 m high, or (ii) perception of a ?ame and exposure to a smoke layer under 0.9 m high.

2.2.3 Mental self-isolation When an accident stimulus goes beyond the terror threshold, an outbreak of self-isolation which is like a state of panic occurs, and various types of human ability are greatly reduced. Further, the probability of the occurrence of human error is known to be about 50% in a state of panic. Some examples from the evacuation simulation of the reduced capacity for various personal abilities as time proceeds are shown in Fig. 2. The relations between the occurrence of self-isolation corresponding to various stimulus grades and personal

Fig. 2. Variations in the diminishing capacities of various personal abilities

Table 1. Occurrences of mental self-isolation according to various stimulus grades Comprehending script Action script Monitoring ability Grade Grade Grade Grade I II III IV High High n n n n Low n n n n High Low High n n n o Low n n n o High High n n n o Low n n n o Middle Low High n n o o Low n n o o High High n n o o Low n n o o Low Low High n o o o Low n o o o

n, no occurrence of self-isolation; o, occurrence of self-isolation

150

N. Fukuchi and T. Imamura: Risk assessment for ?re safety

Fig. 4. Evacuation paths to an exit in various spaces. a Obstacle type. b Clear type

3.2 Escape simulation When evacuees are in an objective space with many furnishings, the escape path is apt to be of the L-type, as shown in Fig. 4a. The number of evacuees arriving at the exit until time t s can be expressed as

Fig. 3. Walking postures and speeds with various smoke layer heights. a Walking postures for various ceiling heights. b The relation between walking speed and ceiling height

N = n0
established by considering the geometric, environmental, behavioral, and procedural aspects of the evacuation process. Here, escape movements are simulated using the group behavioral model by referring to 2). Procedural and behavioral aspects8,10 are taken to be the actions of the crew, the passengers’ prior knowledge of the ship’s interior, emergency signing, etc., and the evacuees are grouped into three types corresponding to walking speed and decision-making ability. Also, the reduction in walking speed caused by a lowering smoke layer must be considered in the escape movement.10 Figure 3 shows the relation between posture and walking speed with various heights of smoke layer. These data were obtained by experiments of restricted walking using 20 adults. The geometric considerations include enclosure layout, the number and type of exits, corridor widths, travel distances, etc. Then the evacuation spaces are divided into four types: (1) objective spaces such as a cabin, theater, etc., with one or more exits, (2) passageways, (3) merging spaces with many entrances and exits, and (4) ?nal disembarkation spaces. In all cases, environmental aspects such as spreading smoke need to be considered. The stagnation of evacuees in narrow exits and stairways are expressed as con?ned ?ow coef?cients, e.g., 1.5 persons/m/s (horizontal way) and 1.3 persons/m/s (stairway).6

(ut )
2

2

- n0
2

(ut - a)
2

2

? a? Hv?t - ˙ ? u?
(2)

- n0

(ut - b)
2

? b? Hv?t - ˙ ? u?

where a and b are the lengths of the sides (m) of objective spaces, u is the walking speed (m/s), n0 is the initial evacuee density (person/m2), and Hv[x] means the Heaviside step function. In the case of a centripetal escape path in objective spaces without obstacles, as in Fig. 4b, the number of evacuees arriving at the exit can be obtained by the formula a? ˙ {(ut) q - aut sinq }Hv???t - u ? ? b? n - {(ut ) q - but sin q }Hv?t - ˙ 2 ? u? 4 2

N = n0

p (ut )

2

n0 2
2

2

1

1

0

2

(3)

where q1 = cos-1 (a/ut), and q2 = cos-1 (b/ut). If an evacuee chooses the crammed exit in an objective space with many exits, the shortest time needed to dispease the crowed state is

t 0,min = ? N exit,i
i

? mexit,i
i

(4)

in which mexit,i and Nexit,i denote the out?ow rate (p./s) and the initial number (p.) of evacuees con?ned at exit- i, respectively. Hence, the shared number N*exit,i (p.) for quickly dispeasing the crowd is expressed as

N. Fukuchi and T. Imamura: Risk assessment for ?re safety

151

N *exit,i = mexit,i ? t 0 ,min

(5)

Evacuees are assumed to move uD t per time-step D t toward exits in passageways. Moreover, it is presumed that they walk along the shortest route connected the entrance to the exit in merging spaces. The escape movement is ?nished when evacuees ?nally arrived at disembarkation spaces. 3.3 Smoke diffusion simulation It is necessary to simulate escape movements together with calculations of smoke diffusion to show the reduction in walking speed and the occurrence of selfisolation (panic) caused by a lowering smoke layer. The calculation of smoke diffusion is carried out using a twolayer zone model,5 in which the temperature and yields of product species in each zone are microscopically assumed to be uniform, and the balance equations for each state variable are deduced based on the relation between mutual neighboring zones. 3.3.1 The mass balance equation of composite gas Accompanying the combustion of ?ammable articles, oxygen is consumed and carbon dioxide, carbon monoxide, hydrooxide, etc., are created in proportion to the reaction rate Finally, the mass balance of composite gases in each ventilated zone has the equation

quantity (kW) of heat transferring to the objective zone, and D H denotes the difference in enthalpies between the original state and the created state in proportion to . the reaction rate mb of combustion. 3.3.3 The volume equation based on the thermal capacity of the smoke zone It can be assumed that the physical properties of combusted air-rich gases are almost the same as those of fresh air, and the pressure in the objective zone is considered to be approximately the same as the pressure of enclosed air. In the case of the objective zone having leaks of air/gases, the following volume equation in the smoke-?lled upper zone is found using the balance of thermal capacity.

c p r sTs

dVs ˙ + DHm ˙ b , s + c p ? Tj m ˙j =Q H, s dt j ?s

(8)

in which subscript s means the quantity related to the smoke (upper) zone. The volume in the lower zone can be obtained by subtracting Vs from V.

4 Risk index for safety assessment The safety of evacuees depends on the appearance of smoke diffusion and establishing escape routes with various support systems for safe evacuation. Consequently, in the case that G groups of evacuees escape from N divisions, the risk index Rev(t) for a safety assessment of the evacuation system is de?ned by the following formula, obtained from simulation results of escape behavior associated with a simultaneous calculated smoke movement.
N G ? mi , j t Rev t = ? ? 1 - ri , j t ? M ? i =1 j =1 ?

rV

dwk ˙ j +n ˙k = ? wk , j - wk m dt j

(

)

(6)

where wk is the mass concentration (kg/m3) of compos. ite gas (k = O2, N2, CO2, CO, or H2O), mj is the mass . transfer velocity (kg/s) from zone j, nk is the creation rate (positive) or the consumption rate (negative) by combustion, and r and V are the density and volume, respectively of a zone. Subscript ( j) means the neighboring zone, and

? denotes
j

()

[

( )]

?x ? ( )? ˙ ? ? L ˙ ?? ?
i

(9)

the summation of ?ow

quantities from/to neighboring zones. 3.3.2 The thermal state equation in an objective zone Because the difference in the speci?c heats of composite gases is small, and the variation in pressure in compartments with air leakage, such as cabins, is generally unimportant, the thermal state equation in objective zones can be expressed by the heat balance as
c p rV dT ˙ + DHm ˙ b + c p ? Tj - T m ˙j =Q H dt j

(

)

(7)

where T and Tj are the temperatures in the objective zone and neighboring zone, respectively, cp is the aver. age speci?c heat of air (kJ/kg/K), QH is the rate of the

where ri,j(t) is the reduction ratio of the walking speed of the j-group evacuees, and {1 - ri,j(t)} means the degree of apprehension concerning the smoke. Moreover, mi,j(t) is the number of j-group evacuees in the i division for M in all groups, and xi is the distance between the i division and the embarkation space for the total length L of the escape path. The risk index Rev denotes the degree of danger in the situation for the evacuation support systems, such as the escape routes and the measures for preventing smoke, and Rev = 1 means the worst condition, in which all evacuees are dead from exposure to smoke at their locations before escape.

152

N. Fukuchi and T. Imamura: Risk assessment for ?re safety

Fig. 6. Scenarios of the evacuation initiation process

Table 2. Composure ratio and ordinary walking speed of each group Evacuating persons Fig. 5. General arrangement of the common spaces of a passenger ship Item Composure ratio (%) Walking speed (m/s) Self-isolation Stimulus Grad III Stimulus Grad IV A group 30.8 1.3 None Break out B group 52.3 1.3 Break out Break out C group 16.9 1.0 None Break out

5 Application to evacuation on a passenger ship 5.1 An analytical model The smoke diffusion and the evacuee movements on three deck spaces of a passenger ship, as shown in Fig. 5, were simulated numerically, and our proposed method was applied. Here, a ?re accident was assumed to start in a cabin on C-deck during “day boat” in the condition that the passengers/crew (422 persons) were located as shown in Fig. 5. Everyone evacuated toward two ?nal exits, guided by supposed evacuation scenarios. The computational parameters for the simulation of smoke diffusion used were the usual physical properties of heat and mass transfer, and the ceiling height of all spaces except the atrium was presumed to be 2.2 m. The escape scenarios were set as shown in Fig. 6 by referring to time study data of actual ?re accidents on land. Scenario 1 is the ideal case, in which the passengers on C-deck are aware of the dangerous situation and escape immediately. Scenario 2 supposes that the passengers on C-deck notice the ?re 2 min after the outbreak. This long delay in sensing danger and starting the evacuation occurs on all decks in scenario 3.

From Table 1, it can be seen that panic commonly occurs with a Grade IV stimulus, but persons with a poor level of comprehension are apt to enter a selfisolation (panic) state for a Grade III stimulus. By investigating the actions and the psychological responses of evacuees in the records of large-scale building ?res15,16 and the statistical composition of passenger ages in Japan,17 the passengers/crew on a ship were divided into three groups according to their walking speed and decision-making ability. The youths/adults were in Group A (with self-possession) or Group B (without composure), and the elderly were in Group C, which tends to consist of slow walkers with self-possession depending on the richness of their life experiences. These groups have ordinary walking speeds of 1.3 m/s or 1.0 m/s, and the occurrence or nonoccurrence of selfisolation at a smoke height of 0.9 m (Grade III stimu-

N. Fukuchi and T. Imamura: Risk assessment for ?re safety

153

lus), as shown in Table 2. Consequently, the evacuating speeds of the three groups are presumed to be like those shown in Table 3. 5.2 Results of simulations Figure 7 shows the variation in the heights of the smoke layers on three decks as time proceeds when these areas are ventilated naturally at the tops of the atrium and the halls. From the calculated results, the ?ow of ?re smoke can be found to affect the evacuation in the hall/ corridor of A deck, the hall of B deck, and the corridor of C deck. In this case, the success or failure of the evacuation depends on the rapidity of the smoke movement toward A deck. The height of the smoke layer in the A-deck hall is 1.5 m (Grade II stimulus) at 150 s after the ?re starts, and the smoke layer descends to 0.9 m (Grade III stimulus) after 220 s. Also, after 330 s, the smoke layer in the A-deck hall descends below 0.6 m (Grade IV stimulus), and then evacuees cannot move without crawling. The escape simulations were carried out on the hypotheses of (1) a constant walking speed, (2) a decreasing walk speed, and (3) breaking out of the self-isolation (panic) caused by a lowering smoke layer. In these examples, the times at the outbreak of panic are t = 220 s for the B group only, and t = 330 s for all groups on the

A deck. Based on these simulations, Figs. 8, 9, and 10 show the variations in evacuee numbers in the halls of every deck and on the stairways as time proceeds. The layout of the evacuee locations at three different times in the case of Scenario 3 is shown in Fig. 11. As these ?gures show, the evacuation time is apt to get much longer with the outbreak of self-isolation.

5.3 Risk assessment for safe evacuation The risk indices Rev(t) obtained by Eq. 9 using the simulation results in every scenario are shown in Fig. 12, which is a comparison of indices for all scenarios in the case of a constant walking speed, and in Fig. 13, which is

a

Table 3. Ratio of evacuation speeds and ordinary walking speeds Evacuating persons (%) Height of smoke 1.5 m 0.9 m 0.6 m A group 88 72 10 B group 88 10 10 C group 88 72 10
b

c

Fig. 7. Variations in smoke layer heights as time proceeds

Fig. 8. Variations in evacuee numbers at speci?ed spots with time (Scenario 1). a At a constant walking speed. b Decreasing walking speed with a lowering smoke layer. c Decreasing walking speed and the outbreak of self-isolation (at 220 s) with a lowering smoke layer

154

N. Fukuchi and T. Imamura: Risk assessment for ?re safety

Number of Evacuees

a

Number of Evacuees

c

Fig. 9. Variations in evacuee numbers at speci?ed spots with time (Scenario 2). a At a constant walking speed. b Decreasing walking speed with a lowering smoke layer. c Decreasing walking speed and the outbreak of self-isolation (at 220 s) with

Number of Evacuees
d

Number of Evacuees
b

a lowering smoke layer. d Decreasing walking speed and the outbreak of self-isolation (at 330 s) with a lowering smoke layer

a

b

c

d

Fig. 10. Variations in evacuee numbers at speci?ed spots with time (Scenario 3). a At a constant walking speed. b Decreasing walking speed with a lowering smoke layer. c Decreasing

walking speed and the outbreak of self-isolation (at 220 s) with a lowering smoke layer. d Decreasing walking speed and the outbreak of self-isolation with a lowering smoke layer

N. Fukuchi and T. Imamura: Risk assessment for ?re safety

155

Fig. 11. Simulation results of evacuee behavior at 0 s, 420 s, and 450 s (Scenario 3). a 0 s (before the start of evacuation. b 420 s. c 450 s

Fig. 12. Risk indices of the escape route in the cases with a constant walking speed

the variation in the aggregated risk indices as time proceeds. From these ?gures, it is seen that the aggregated risk and the cumulative risk indices indicate the occurrence of unsafe states if there is any delay in beginning the escape action. Furthermore, a decrease in walking speed and the outbreak of self-isolation (panic) during escape have a considerable in?uence on the evacuation time taken and the critical conditions for human life. The severe outcomes in these simulations occurred in Scenarios 2 and 3 but not Scenario 1, in which many evacuees had to pass through the hall of A deck when it was full of smoke. In order to establish a safe evacuation system, one needs to take measures to (1) prevent smoke diffusion, and (2) institute a proper escape support system. Here, smoke screens and an additional emergency exit are given as examples.

Smoke screens at the top of corridors. When these are smoke screens (0.4 m high) hanging from the ceiling at the border of corridors in the living space on C deck, the risk index Rev(t) changes from the solid line to the dotted line in Fig. 14a. This ?gure shows that smoke screens are effective in the early stages of evacuation. An additional emergency exit on A deck. Quick smoke movement toward A deck is a distinguishing factor of the dangerous state occurring in this simulation. If an additional emergency exit with passage to the disembarkation deck is installed, the risk index Rev(t) decreases considerably, as shown in Fig. 14b. This shows that an additional exit is very effective for avoiding the critical phase. Accordingly, it can be useful to select the optimal escape route and evaluate the various support systems for a safe evacuation by using the risk index Rev(t).

156

N. Fukuchi and T. Imamura: Risk assessment for ?re safety

a

b Fig. 14. Examples of the effects of an escape support system (in the case of a constant walking speed). a Smoke screens. b Additional emergency exit

Fig. 13. Variations in the aggregated risk index as time proceeds

6 Conclusions Numerical escape simulations taking account of smoke diffusion and human factors were carried out for a marine ?re. Furthermore, the risky situation of evacuation was evaluated quantitatively by using the proposed risk index Rev(t) upon these simulation results. The results are itemized below. 1. Escape simulations with smoke diffusion are helpful for assessing safety matters in any attempted evacuation plan. 2. A delay at the beginning of escape action tends to bring about an unsafe state. 3. As the result of a lowering smoke layer, evacuees have a tendency to decrease their walking speed and experience mental self-isolation. These situa-

tions have a considerable in?uence on the escape time and the occurrence of critical conditions. 4. The risk index Rev(t) is useful for establishing a safe evacuation plan, in combination with an optimal escape route and the evaluation of the various support systems. References
1. Tieszen SR (2001) On the ?uid mechanics of ?res. Annu Rev Fluid Mech 33:67–92 2. Gray DD, Giorgini A (1976) The validity of the boussinesq approximation for liquids and gases. Int J Heat Mass Transfer 19:545–551 3. Fukuchi N, Hu C (2004) A pseudo?eld model approach to simulate compartment-?re phenomena for marine ?re safety design. J Mar Sci Technol 8:177–184 4. Erlebachar G, Hussaini MY, Speziale CG, Zang TA (1992) Toward the large-eddy simulation of compressible turbulent ?ows. J Fluid Mech 238:155–1851 5. Weaver S (2000) A comparison of data reduction techniques for zone model validation. Fire Eng Res Rep 2000/12, USA, pp 3– 18 6. Tanaka T (2000) Necessity of design methodology in the framework of a performance-based ?re safety design system. In: 15th

N. Fukuchi and T. Imamura: Risk assessment for ?re safety
Meeting of the UJNR Panel on Fire Research and Safety, NSTIR 6588, vol 1, pp 435–440 Takahashi K, Tanaka (1988) An evacuation model for use in ?re safety design of buildings. In: Proceeding of the 2nd International Symposium on Fire Safety Science Rasmussen J (1982) Human errors: a taxonomy for describing human malfunction in industrial installations. J Occup Accid 4:311–335 Ikeda K (1986) Intelligence process in an emergency (in Japanese). Publishing Association of Tokyo University, pp 115– 149 Fukuchi N, et al (1999) An analysis of evacuation behavior by using the walking model with psychological intelligence process in an emergency (in Japanese). J Soc Nav Archit Jpn 184:545–558 Kugihara N (1995) Panic experiment-social psychology (in Japanese). Nakanishiya Press

157
12. Nagane M (1987) The stress index based on an examination of psychological stress using the MFF test (in Japanese). Jpn J Psychol 157:383–386 13. Thompson PA (1995) A computer model for the evacuation of large building populations. Fire Safety J 24 14. Katsuhara M, et al (1996) Simulations and demonstrations of human escape on board. RINA Conference on Escape, Evacuation and Rescue 15. Jones WW, Quinteire JG (1984) Prediction of corridor smoke ?lling by zone models. Combus Sci Technol 835:239–243 16. Saito H (1977) The report on action and psychological response of building exodus (in Japanese). In: Proceeding of the Society of Architects of Japan (Chugoku District) 17. Ministry of Transport of Japan, Transport Policy Bureau (1997) A transport white paper, Part 2. The trend of public transportation (in Japanese). pp 712–714

7.

8.

9.

10.

11.


相关文章:
Risk assessment for fire safety considering characteristic ....pdf
Risk assessment for fire safety considering characteristic evacuees and - J Mar Sci Technol (2005...
...Avoidance Method for Mobile Robot Considering Motion and ....unkown
Taiwan Collision Avoidance Method for Mobile Robot Considering Motion and Personal Spaces of Evacuees Takeshi OHKI, Keiji NAGATANI and Kazuya YOSHIDA Abstract...
...Avoidance Method for Mobile Robot Considering Motion and ....pdf
Collision Avoidance Method for Mobile Robot Considering Motion and Personal Spaces of Evacuees_英语考试_外语学习_教育专区。The 2010 IEEE/RSJ International ...
更多相关标签: