This document is downloaded from CityU Institutional Repository,

Run Run Shaw Library, City University of Hong Kong.

Title Fire safety provisions and evacuation strategy for supertall

residential building in Hong Kong

Author(s) Ho, Ka Ming (何嘉銘)

Citation

Ho, K. M. (2015). Fire safety provisions and evacuation strategy for supertall residential building in Hong Kong (Outstanding Academic Papers by Students (OAPS)). Retrieved from City University of Hong Kong, CityU Institutional Repository.

Issue Date 2015

URL http://hdl.handle.net/2031/8311

Rights This work is protected by copyright. Reproduction or distribution of the work in any format is prohibited without written permission of the copyright owner. Access is unrestricted.

FIRE SAFETY PROVISIONS AND

EVACUATION STRATEGY FOR

SUPERTALL RESIDENTIAL BUILDING

IN HONG KONG

By

Ka Ming HO

Submitted in partial fulfilment of the requirements for

The degree of Bachelor of Engineering (Honours) in Building Engineering

(Building Services Engineering)

Department of Architecture and Civil Engineering

City University of Hong Kong

March 2015

i

ABSTRACT

Staircase is the most common and important means of escape in many buildings. A safe

and efficient evacuation is a key issue to the fire safety of occupants, especially in a

supertall building. Currently, empirical calculation and computational software are

widely adopted to examine the total required evacuation time of each building. The

maximum movement speed or evacuation flow depends much on the configuration of

staircase and the assumptions of human walking speed. As there are more and more

concern on the effect of evacuation time due to various human behaviors during the

crowd and emergency situation, the actual time required will longer than the value

obtained by the simulation. Furthermore, when considering the effect of ageing problem,

physical challenge of an elderly travelling long distance will further prolong the

evacuation time. As fire cases more commonly happened in residential building, hence,

this report reviews the problems of the current evacuation strategy for a supertall

residential building in Hong Kong by both empirical calculation and computer

simulation. Regarding to the problems found, feasible evacuation strategies – phased

evacuation and defense-in-place are proposed to minimize the number of occupants

involved in the evacuation, and thus the population density along the evacuation routes.

A noticeable reduction of evacuation time is achieved by adopting the new strategy

comparing with current strategy. Upon the new strategy, automatic sprinkler system is

proposed to be installed in the corridor and staircases of residential building in order to

minimize the fire spread and enhance the safety level of occupants which carry out

defense-in-place approach in the unaffected part of the building. Based on the findings

of this paper, it was highly recommended to develop a more reliable and appropriate

evacuation strategy for residential buildings to minimize effect of human behavior on

evacuation time and consider the needs of society in the future.

ii

ACKNOWLEDGEMENTS

To be able to complete this final year project, I would like to thank my supervisor, Dr.

Nadia Chow for her support, sharing of knowledge and guidance during our numerous

meetings. I would also like to thank Mr. John Chan from Financial Office of City

University of Hong Kong for his great support in purchasing the VE modules Annual

License which facilitate the progress of my project.

iii

Table of Contents

Abstract i

Acknowledgements ii

Table of Contents iii - v

List of Figures vi - vii

CHAPTER 1: INTRODUCTION 1

CHAPTER 2: RESEARCH OBJECTIVE 4

CHAPTER 3: LITERATURE REVIEW

3.1 Fire Risk in Super-tall Residential Building 4

3.1.1. Fire Load Density 4

3.1.2. Stack Effect 5

3.1.3. Large Window Pane in High-rise Building 5

3.2 Fire Safety Provision in High-rise Building 6

3.3 Evacuation Model 7

3.3.1. Total Evacuation 8

3.3.2. Phased Evacuation 8

3.3.3. Defense-in-Place 9

3.3.4. Delayed Evacuation 10

3.4 Means of Escape 10

3.4.1. Staircase 10

3.4.2. Refuge Floor 11

iv

3.5 Human Behavior in Evacuation 12

3.5.1. Effect of Fatigue and Physical Ability 12

3.5.2. Group Behavior 13

3.5.3. Non-adaptive Behavior 14

3.5.3.1. Panic 15

3.5.3.2. Decision-making 15

3.6 Pre-evacuation Time 16

CHAPTER 4: METHODOLOGY

4.1 Empirical Calculation 18

4.1.1. Effective Width 19

4.1.2. Movement Speed 20

4.1.3. Specific Flow 21

4.2 Computational Analysis 23

4.2.1. Assumption 24

4.2.2. Building Population 24

4.2.3. Body Shape 25

CHAPTER 5: DESCRIPTION OF CASE STUDIES 26

CHAPTER 6: RESULTS 29

6.1 Empirical Calculation 29

6.2 Computer Simulation 30

6.2.1. Case 1 30

6.2.2. Case 2 33

6.2.3. Case 3 37

v

6.2.4. Case 4 40

CHAPTER 7: RESULT ANALYSIS 44

7.1 Comparison between Empirical Calculation 44

and SIMULEX Result

7.2 Comparison between Total Evacuation and 46

Phased Evacuation Strategy

7.2.1. 5 Occupants inside each flat 46

7.2.2. 12 Occupants inside each flat 48

CHAPTER 8: DISCUSSION 49

CHAPTER 9: CONCLUSION 54

CHAPTER 10: REFERENCE 55

APPENDIX A1 – Empirical Calculation Result for Case 1 59

APPENDIX A2 – Empirical Calculation Result for Case 2 64

APPENDIX B1 – SIMULEX Result of Case 1 70

APPENDIX B2 – SIMULEX Result of Case 2 73

APPENDIX B3 – SIMULEX Result of Case 3 78

APPENDIX B4 – SIMULEX Result of Case 4 81

vi

List of Figures

Figure 1.1 Number of Fire Cases Recorded by HKFSD 2

Figure 4.1.1 Effective Width and Clear Width 19

Figure 4.1.2.1 Relationship between Constant and Evacuation Speed 21

Figure 4.1.2.2 Relationship between Density and Movement Speed 21

Figure 4.2.3.1 Maximum Specific Flow 22

Figure 4.1.3.2 Relationship between Density and Specific Flow 22

Figure 4.2.1.1 Variation of Movement Speed between Types 24

and Gender of Occupant in Simulex

Figure 5.1.1 Information of Selected Super-tall Residential Building 26

Figure 5.1.2 Floor Layout Plan of Selected Super-tall Residential 27

Building

Figure 6.2.1.1 Total Cumulative Persons Arriving Refuge Floor - Case 1 30

Figure 6.2.1.2 Cumulative Persons Arriving Refuge Floor via Staircase 1 31

Figure 6.2.1.3 Cumulative Persons Arriving Refuge Floor via Staircase 2 31

Figure 6.2.1.4 Population Density Recorded at Measuring Points 32

(Staircase 1) – Case 1

Figure 6.2.1.5 Population Density Recorded at Measuring Points 32

(Staircase 2) – Case 1

Figure 6.2.2.1 Total Cumulative Persons Arriving Refuge Floor - Case 2 33

Figure 6.2.2.2 Population Density Recorded at Measuring Points 34

(Staircase 1) – Case 2

Figure 6.2.2.3 Population Density Recorded at Measuring Points 35

(Staircase 2) – Case 2

Figure 6.2.2.4 Cumulative Persons Arriving Refuge Floor via. Staircase 1 36

vii

Figure 6.2.2.5 Cumulative Persons Arriving Refuge Floor via. Staircase 2 36

Figure 6.2.3.1 Total Cumulative Persons Arriving Refuge Floor – Case 3 38

Figure 6.2.3.2 Population Density Recorded at Measuring Points 39

(Staircase 1) – Case 3

Figure 6.2.3.3 Population Density Recorded at Measuring Points 40

(Staircase 2) – Case 3

Figure 6.2.4.1 Total Cumulative Persons Arriving Refuge Floor - Case 4 41

Figure 6.2.4.2 Population Density Recorded at Measuring Points 42

(Staircase 1) – Case 4

Figure 6.2.4.3 Population Density Recorded at Measuring Points 42

(Staircase 2) – Case 4

Figure 8.1 Summary of Results in Different Cases 51

1

1. INTRODUCTION

In recent decades, many buildings had been built around the world with its building

height around 250m, while some of them even higher than 500m. In the view of the

construction requirement or building height, these buildings are not simply classified

as “high-rise” building anymore. For the building with its height over 300m will be

classified as “Supertall Building” by the definition of Council on Tall Building and

Urban Habitat (CTBUH) (Jan Klerks, 2011).

The definition of high-rise building is varies in different countries or cities. In Dubai,

the building with structure having occupancies at more than 23 meters above the lowest

level of Fire Services Access is categorized as high-rise building. (UAE, 2011) In New

York, high-rise building means that the structure with more than 75 feet in height. (NYC,

2008) In Hong Kong, local authority classified the “High Rise Building” as the building

with the uppermost storey exceeds 30m above the point of discharge at ground floor

level. However, there are no official definition carried out by the authorities in different

countries including Hong Kong. There is no any definition for supertall building even

many skyscrapers already exist in Hong Kong. For the supertall buildings in Hong

Kong, most of them was designed for commercial or hotel usage which can be

commonly found in Central Business District. However, number of supertall buildings

for residential purpose have already been built in some Asian cities such as Dubai,

Busan and Shenzhen in recent years (Jan Klerks, 2011).

2

Types of Fire 2009 2010 2011 2012 2013

Public area 425 518 478 382 385

Squatter area 123 89 123 101 94

Vehicle 293 285 310 223 227

Housing estates 1341 1347 1285 1185 1258

Factory buildings 122 62 123 95 100

Institutional buildings 190 250 192 191 165

Non-domestic buildings 464 371 419 420 376

Domestic buildings 1588 1562 1460 1438 1531

Vegetation 1294 819 1335 818 810

Vessel 24 15 18 23 26

Others 1580 1468 1612 1306 1417

Total 7444 6786 7355 6182 6389

Figure 1.1: Number of Fire Cases Recorded by HKFSD [34]

Most of the residential buildings including public housing and private building still with

the building height around 150m, taking 3m as the floor to floor height of a 45 storey

building in Hong Kong nowadays. As residential building is the major type of building

in Hong Kong, fire safety concern is always the great concern. According to the Year

2014 Statistic Report of Fire Services Department (FSD) [34], the percentage of

residential building-related fire case is ranging from 37.3% to 43% between 2009 and

2013, while the fire case of non-domestic building is only ranging from 8.3% to 10%

in the same period. By comparing two sets of value, it implies that fire case has a higher

chance happening in residential building which might be due to overcooking or electric

fault. In the view of urban development in other Asian cities, there is more and more

opportunity to have a supertall residential building be built in Hong Kong in future,

hence, fire safety for the occupants inside the building and the fire safety management

should be highly considered.

3

Currently, the highest working height of the aerial ladder platform equipped by HKFSD

is 53m, while a super-tall building usually with the building height over 300m, which

already around six times more than the accessible height of the fire engine. Concerning

on the fire safety of a building, preventing fire occur is obviously an desirable goal to

attain the highest safety level to occupants, however, accidents usually happen

unexpectedly. Once the fire case happens in a building, staircase is the only available

mean of escape for the occupants since it is the only prescriptively accepted means of

escape by local authority. Although there are lot of requirements concerning the

reliability and configuration of staircases to ensure the safety of evacuees, the

effectiveness and efficiency of evacuation is not guaranteed since these are dependent

on various factors. Hence, the effectiveness of the evacuation will be a key factor in

determining the safety of evacuees or occupants, which people can escape from the

building to an ultimate place of safety.

Regarding to the current practice, all floors between refuge areas will be alerted for the

evacuation in fire case of a high-rise residential building. [9] For this evacuation

approach, it will involve number of occupants during the evacuation and thus result in

inefficient evacuation and potential risk in crowd movement. However, a zoned

evacuation is adopted for sprinkler-covered commercial buildings and hotel buildings,

this approach can significant reduce the number of evacuees being alerted. Some

argument had been voiced out after the tragedies about the insufficient fire safety

provisions inside residential buildings comparing with the safety standards of other

countries.

4

2. RESEARCH OBJECTIVE

This study is aimed to review the evacuation performance of total evacuation approach

in a selected super-tall residential buildings by case study basis. In the city, there were

voices arguing the adequacy of fire safety provisions in residential building after

tragedy happened. On the other hand, the long evacuation time in super-tall buildings

is also getting more awareness in the world. The evacuation time will get worse when

considering the human behavior and physical challenge of elderly due to the ageing

problem in the society. By using empirical calculation and computational tools, the

efficiency of current evacuation approach under different scenarios will be analyzed.

Focused on the shortcoming of current strategy, new evacuation strategy and fire safety

systems will be proposed as the solution which aims to enhance the efficiency of

evacuation procedure and safety level to all occupancy inside the premises.

3. LITERATURE REVIEW

3.1. Fire Risk in Super-tall Residential Building

3.1.1. Fire Load Density

Hong Kong is well-known in dense population density among the world. Due to the

storage of land, many residential buildings were built as more flats as possible. For

some residential flats with relatively small area, fire load density is quite high since

many combustible content was stored inside the flat such as book, magazine, clothes,

curtain or furniture. These combustible materials generally placed crowded together

due to the limitation of the area, which result in a high fire load density inside a

residential flat. Some studies revealed that the fire load density of a residential flat

might reach 1400MJ/m2, which already exceed the value of 1135MJ/m2 as the limitation

stated by Building Department and local fire code (Chow and Chow, 2010)

Although the development of fire is not solely depend on the fire load, other factors

such as ventilation rate should also be taken into consideration. When a compartment

5

only with a small opening get with fire, a large fire might not be resulted if the

ventilation rate is not sufficient to sustain a ventilation-controlled fire. However, the

majority of ventilation is coming from the window opening in a flat which is capable

to provide adequate oxygen for the combustion, a large fire will be resulted especially

for supertall building with the aid of stack effect.

3.1.2. Stack Effect

Apart from the fire load density inside the flat, stack effect and wind velocity also take

important roles in determining the fire growth rate. The ventilation rate depends on the

total volumetric flow rate flowing into the compartment or the flat in this study. There

are three major elements contributing into the total volumetric flow rate, volume flow

rate due to fire, volume flow rate through leakage opening area due to stack effect and

volume flow rate due to wind action (Chow and Chow, 2010). For stack effect, the

greater the difference of between compartment temperature and ambient temperature

and also the height difference, the greater volumetric flow rate would be resulted, hence,

more fresh air can be supplied to support the combustion. For wind action, it is also

proportional with the elevation height. Wind velocity will be gained from zero at the

ground floor to a higher value at upper level of a building. For a supertall building, the

building height is generally at least 300m. For that building height, the wind velocity

acting on the flat will be very high.

3.1.3. Large Window Pane in High-rise Building

In the view of the luxury high-rise residential buildings built along the seafront of

Victoria Harbour such as West Kowloon District, most of the buildings built with a

large window pane in the living room and the bedroom in order to let the residence

enjoy the harbour view at the home. Despite the large window pane give the benefit to

the occupants, it also conceal the risk during the fire case. Even a narrow opening is

6

kept for the general ventilation purpose in the flat, this opening will still act as a

sufficient source of fresh air to sustain the fire growth once fire happens in the living

room, thus, the compartment temperature might reach a high level. In general, he

average breakage temperature of a glass usually lower than 100oC, it implies that there

will be a high opportunity that the temperature resulted from the fire will cause cracking

or breakage of the window pane. A study found that cracking will be formed if the

temperature difference between heated glass and the edge reach 70 OC. Besides, when

the value of HRR reach 100kW and 500kW will already cause the cracking and fall out

of the single glazing window respectively (Lau and Chow, 2014). Once the window

pane cannot withstand the larger cracking, an additional opening will be created when

breakage of glass occur.

The large opening led by the breakage will foment the stack effect and wind action

acting the combustion since it provide sufficient oxygen content to sustain the burning,

thus, a larger fire size would be resulted. Moreover, the consequence of the window

pane breakage is accelerating the growth phase of the fire, while the chain effect would

bring the flashover happening forward to an earlier time, hence, the available

evacuation time would be shortened. During the flashover inside the flat, the

compartment temperature will significantly increase to around 600 oC, any occupants

remaining inside will be extremely danger.

3.2. Fire Safety Provision in High-rise Building

In Hong Kong, different classifications of premises and area of usage will be required

for different fire services systems and installations. A building will be classified as a

high-rise building as the floor of uppermost storey exceeds 30m above the point of

discharge in staircases relative to the ground floor level. For a high-rise commercial

7

building, there are sixteen types of systems or installations prescriptively required

including audio/visual advisory system, fire alarm system, fire detection system and

sprinkler system. These system should be well-prepared in both public, landlord and

tenant area within the commercial building. Similarly to a high-rise hotel building, the

same sixteen types of systems or installations are required in the premises while fire

detection system and sprinkler system should also be provided in all guest rooms to

ensure the fire safety level to the visitors. In the same classification of high-rise building,

there are only seven types of systems or equipment are required for a domestic premises.

Local authorities have no legislative requirement on fire services equipment within

private domestic flats since the fire load and fire hazard are generally low in a domestic

building. As a result, local fire code only requires the essential fire services systems

such as fire hydrant and hose reel system, but not any fire detection system or sprinkler

system. In foreign countries including United States, United Arab Emirates and

Australia, sprinkler system are already required in their local regulations. [2,30]

3.3. Evacuation Model

Despite of the level of fire safety and protection system provided in the high-rise

building, evacuation is always considered as best way to let occupants to the ultimate

place of safety. An evacuation strategy is not entirely suitable to all different types of

building due to the occupant characteristic and hazard level of the premises. That is the

reason why the preparation of an appropriate evacuation strategy is necessary to the

buildings, especially for super-tall buildings. The increase of occupant load of high-rise

buildings might cause the extra burden to the egress components. In general, occupant

will evacuate downwards in a building rather than evacuate towards the roof in fire case.

Evacuation to roof would only be considered in case of untenable fire happen in lowest

level of the building which cause the blockage of exit of staircase. This strategy might

8

involve number of problems such as the air turbulence

Currently, there are four high-rise building evacuation strategy widely adopted,

including total evacuation, phased evacuation, defend-in-place and delayed evacuation.

3.3.1. Total Evacuation

Total evacuation is the traditional single stage building evacuation method which

adopted for low-rise building in the past. All occupants inside the building are planned

to evacuate towards the ground level by using the staircase after the recognition of fire

alarm signal. In Hong Kong, the building code is regulating the configuration of means

of escape provided in the building including the minimum number and width of exit

door and staircases. [5] In the past experience, total evacuation performed well in low-

rise buildings, the capacity of evacuation route is sufficient for all occupants initiate

their evacuation simultaneously.

However, when the same strategy applied for a higher elevation building with larger

number of occupants, queuing might be experienced to the evacuee along the

evacuation route including corridor, exit door and staircase before reaching the ground

floor level. The congestion will lead to an extra time required for the evacuees to finish

the evacuation process, evacuees could be exposed to different kinds of danger situation

such as mental disorientation and physical tiredness. (Luo and Wong, 2006)

3.3.2. Phased Evacuation

Owing to the inefficient and non-practical for adopting single stage total evacuation in

high-rise building, phased evacuation is then developed which aims to evacuate the

occupant at most critical floors such as fire floor and floor nearby in prior. For the

9

commercial buildings and hotel buildings in Hong Kong, once the fire occur at a floor,

the fire alarm will only be actuated at fire floor, one floor below and two floors above

the floor. Those floors will initiate the evacuation first, while the remaining floors can

start the evacuation subsequently once it is necessary. Hence, the population density

and the induced queuing time can be reduced at the first stage of evacuation. (Ronchi

& Nilsson, 2013)

Since a majority of occupants will start the evacuation in the later stage once be alerted,

fire compartmentation and adequate fire services installations are essential to prevent

the fire spread to adjacent rooms to ensure the safety level of remaining occupants

before the arrival of fire brigades.

3.3.3. Defense-in-Place

Defense-in place can also be an appropriate evacuation strategy for high-rise building

especially in residential usage. Apartment buildings usually have the occupants with

elderly and disabilities who might not be able to evacuate without the assistance. Some

literature stated that this approach actually is the most suitable to a high-rise residential

building if the building provide enough fire compartmentation ability and fire services

system. In many serious tragic fire cases, number of casualties were found in staircases

or corridor rather than the flat (Ronchi and Nilsson, 2013). During the evacuation in the

staircase, some reviews found that apart from the human behavior such as panic,

evacuee could not take any effective measure to deal with the emergency situation, for

example, the infiltration of fire flame or dense smoke into the staircase, it might result

in the tragic ending. In contrast, Proulx proposed that defense-in-place can maintain the

survival rate inside the residential flat, since occupants can take active measure such as

tightly close the door and seal the cracks before having the rescue by fire fighters.

10

3.3.4. Delayed Evacuation

Delayed evacuation is meant that the evacuees will take a temporary assembly in the

designated area to wait the rescue by fire fighters. In Hong Kong regulation, refuge

floors are required for the building exceeding 25 storeys above the lowest ground storey.

For the residential building, refuge floor should be designed except those exceeding 25

storeys but not exceeding 40 storeys above the lowest ground level. [5] In the view of

supertall residential buildings, it is generally designed with more than 80 storeys, hence,

more refuge floor is then required to comply with local requirement. Refuge floor is a

protected area for the occupants to assemble in case of fire. This strategy might be

useful to the occupants with injured or disabilities. After those occupants travel a certain

number of floors, most of them actually cannot keep on the evacuation by using the

staircase. Refuge floor can provide a relatively safe area to the evacuee for having a

rest in a short period of time. Besides, evacuation elevators might also be provided to

serve between the refuge floor and ground floor level. It might significantly help the

occupants to reach the ultimate place of safety in a shorter period of time.

3.4. Means of Escape

3.4.1. Staircase

The use of staircase is the traditional method to evacuate from a building. Staircase

should be carefully designed to provide sufficient capacity for all occupants undertake

the evacuation. In view of the evacuation flow, the configuration of staircase contributes

the most influences. Thus, the local regulation listed out the requirements including the

minimum width of treads, minimum height of risers, minimum of staircase provided

and the minimum and maximum width of staircase.

11

Apart from the construction requirement, the efficiency of an evacuation also depends

on physical strength of occupants. In a high-rise building, occupants might need to

travel around 20 floors to reach the refuge floor or ground floor, fatigue might have

higher probability happen to a child or elderly. Each stop for rest will prolong the entire

process and even causing blockage to the evacuee in behind.

Merging stream is an important issue to a high-rise building. In the lower part of the

evacuation flow, the staircase is usually packed with number of evacuee, merging

stream at each floor will negatively affect the movement speed of evacuee. Higher the

population density in the staircase, the more difficulty to completely merge two streams

of evacuees together, which might the influence to the evacuation time.

3.4.2. Refuge Floor

The aims of refuge floor is not only provide a place for rest of evacuees, it also provide

a platform for fire fighters setting up the command point in order to assist the evacuation

and those in injuries. Some study proved that refuge floor is able to keep away from

smoke to keep evacuees safe even in the case of 25MW fire and strong wind condition.

(Chow and Chow, 2010) Although refuge floor is recognized as the place of safety,

some uncontrollable factors might still lead to the failure situation of refuge floor

concept such as overcrowding, under-utilization and variety of human behavior in a

threatened condition.

According to the local fire code, refuge floor should be provided with the net area larger

than 50% of the total gross floor area of refuge floor. [5] It is questionable about the

ability to accommodate all occupants from around twenty floors within a refuge floor.

The occupant density of an office building or residential building is 9 m2 /person, the

12

occupant density in refuge floor will be significantly reduced to around 0.4 m2 /person.

The over-crowd situation may result in people misusing the refuge floor finally.

3.5. Human Behavior in Evacuation

Effective and safe evacuation is always the major issue regarding the fire safety in a

building as fire case can never be avoided in the real situation. Only a safe evacuation

can bring the occupants to a place of safety. There are lots of methods including

empirical calculation and computational tools now available to analyze and calculate

the total evacuation time for all occupants inside the building. However, the effect of

large occupant loading in multi-storey high-rise buildings and psychological effects of

occupants have not been taken into consider, and some of the tools are highly rely on

the assumption of the individual human behavior. Human are mostly treated as

homogeneous, without any individual emotion and physical needs. Once considering

the human behavior in emergency evacuation situation, the longer evacuation and more

complicated conditions inside the staircase will be resulted.

3.5.1. Effects of fatigue and physical ability

It is easy to understand that the movement speed of an evacuee cannot keep as constant

during travelling a long distance pathway. The situation is also the same for evacuation

in staircase. Some studies found that the degree of physical fatigue is in proportional to

the travelled distance of evacuees. In contrast, the travelled distance will be in

proportional to the evacuation speed. An experiment found the deceleration rate of the

movement speed of exhausted evacuees by comparing with a normal occupants which

have not start the evacuation. (Koo and Kim, 2014) The result of experiment found that

the full speed can only be maintained in the first 100m travelling distance. After

travelling 100m, the movement speed of a person without disability started decreasing

13

to 89% of the initial speed. Moreover, the deterioration of movement speed of a person

with low stamina and mobility limitation will be more significant. For travelling the

same distance with a normal evacuee, the decrease of evacuation speed of person in

low stamina and disability is around twice faster than normal evacuees. After travelling

100m distance, the evacuation speed can only attain around 78% of the initial speed.

Upon the deceleration rate, the evacuation speed will further affected by the blockage

situation in narrow area such as the staircase. (Koo and Kim, 2014) It is because the

deceleration of exhausted evacuee in low stamina and disability will cause the

evacuation flow on behind. Besides, people who assist them will move relatively slowly

and then block the narrow pathway of other evacuees.

In real situation, it is noticeable that the variances of movement speed exist between

different agent types such as male, female, child, elderly. Helbing and Monlar (1995)

also found that the variances of evacuation speed in same agent type also cause a

significant blockage effect at narrow area. Jeongin Koo (2014) also observed that the

longer travelling distance will increase the variances of evacuation speed, the blockage

problem will get more significant which resulting in an additional time required for

completing the evacuation.

3.5.2. Group Behavior

In evacuation study, crowd movement with different groups are important especially in

residential building. It is important to distinguish the difference between crowd

movement and group behavior. Group behavior is related to the psychology of persons

within the groups. For a residential building, Hofinger et al. (2014) found that the

occupants will try to spend time informing the neighbors and help them to leave before

initiate the evacuation towards the staircase. Affiliation, leadership, trust and helping

14

behavior, all are the aspects related to group behavior. Hence, Pan et al. (2006) observed

that the crowd of evacuation flow is frequently formed with different small groups in

size of around 6 persons rather than taking the evacuation individually.

Although evacuees might show better controlled and rational behavior when evacuating

in small groups by having a closer proximity with group members, sometimes it might

cause other impact on the evacuation flow. During the crowd situation such as the

merging flow at floors, some group members might loss the proximity with each other,

some of them might stop at the evacuation route in order to wait for the reconstitution

with the separated group members in the narrow area such as staircases. This situation

is mostly occur in groups which in hierarchically structure such as a family with parents

and the children. Either the stop or the contrary movement will impede or block the

evacuation flow of other evacuees in the staircase.

3.5.3. Non-adaptive behavior

In real, every person is considered as individual, each of the person will have the own

decision-making, social interaction and behavior. It is the reason why human behavior

is difficult to be predicted, especially in emergency or crowd situation. For a premises

with large occupancy loading such as super-tall residential building, once the

evacuation is required under emergency cases, the non-adaptive crowd behavior may

cause serious consequences even in well-designed buildings. In the history, some

tragedies are caused by the human behavior in the crowd rather than the fire. From the

past accidents show that non-adaptive behavior such as flight or pushing might occur

in crowd during emergency situation. The non-adaptive behavior is related to panic and

decision-making.

15

3.5.3.1. Panic

Comparing with the normal situation, people might feel fear and stressed during

emergency situation. When occupants recognize fire case happen in the situated

building after hearing the alarm bell buzzing, they will start behave irrationally and

panic. (Pan et al., 2006)

3.5.3.2. Decision-making

In an evacuation, everyone would like to have a higher chance to exit the building faster.

Every decision made by an evacuee will be the course of action to defense their chance

of reaching a place of safe. For example, if another person pushing on an individual,

the individual might feel the chance of earlier evacuation might be threatened, hence,

the individual would decide to push each other to keep the chance of evacuation safely.

Pan el al. (2006) found that most of these quick decisions made during a high stress and

emergency situation based on the instinct and psychology of each individual, while

some of the decisions might lead to inefficient evacuation.

Competitive behavior is commonly found during an evacuation since all occupants

would like to compete for their chance to get evacuation first. Evacuee in the staircase

always compete with the evacuee from the floors during the merging flow since they

usually think that they have the priority to continue the evacuation. They would not

prefer to avoid any collision to fight for the evacuation chance. (Aghabayk et al., 2014)

Hence, only few persons from the floors can enter the staircase to travel downwards a

safe area. For a buildings with more storeys, more merging flow the evacuees might

experience during evacuation. The cumulative effect would cause an effect to the total

evacuation time for the building.

16

3.6. Pre-evacuation Time

In a successful evacuation, the required safety egress time (RSET) should be always

less than the available safe egress time (ASET), while the pre-movement time is an

important part, because the longer the pre-movement time will delay the evacuation

process, the higher risk the occupants might face the dangerous situation. However, the

length of pre-movement time is subjected to many factors such as the psychological

process of a person before initiating the evacuation.

A survey investigated that only 9.5% of occupants will start the evacuation at once after

receiving fire cue. The other occupants had other reactions to the fire cue rather than

evacuate immediately from the origin. These human behavior has been classifies into

five aspects including not taking any movement, trying to extinguish the fire, alerting

the neighbor and trying to confirm more information of fires. (Mu et al., 2013) It should

be concerned that a portion of occupants would like to ignore the alarm and maintain

the original status. The reasons of ignoring the alarm might due to the past experience

of false alarm and the alertness of occupants.

In the human psychological aspect, the behavior of an individual is based on the

decision-making process. Wong (2004) found that past experience will affect one’s

decision making in the future since human usually rely on the past experience to make

any decision. If false alarm happened lots of time in the building, occupants would

consider the current fire alarm as false alarm even a real fire was ignited in the building.

A survey about the fire alarm system of high-rise residential buildings in Hong Kong

reveals that a high number of false alarm were reported in the estate within five years.

The false alarm were due to vandalism, improper maintenance as well as environmental

impact. No matter the causes of false alarms, these unwanted fire alarm would affect

the credibility with the occupants in real fire situation. Literature shows that a certain

17

number of occupants in Hong Kong would like to ignore the fire alarm unless they can

recognize any abnormal condition on their own. The “cry wolf syndrome” might lead

to occupants prefer to ignore the alarm which resulting in life-threatening situation.

Apart from the false alarm, the alertness of occupants is also a major concern in the

view of pre-movement time, especially in residential buildings. In commercial

buildings such as office, employees are normally in conscious status during their work,

they can receive the fire alarm in a fast response time and then start the evacuation

process immediately. However, occupants in the residential buildings might fall asleep

or taking a rest when get sick. Occupants might fail to recognize the fire alarm in

unconscious state such as the midnight. Occupants keep on sleeping instead of

evacuating from the premises will lead to a dangerous situation. As the fire growth

continues in the building, the later the occupants realize the fire alarm, the higher

chance the occupants will face the risk, especially in residential buildings as no any

active fire safety installation is provided.

18

4. METHODOLOGY

4.1. Empirical Calculation

Empirical equations used for numerical calculation is aims to determine the time

required for total evacuation by staircase. There are number of empirical methods such

as Pauls, Togawa or Russian Method (Cai and Chow, 2010), while the method

established by Nelson and Mowrer was adopted for this study.

Calculation of Evacuation Time Required by Using Staircase

The approach in predicting the evacuation time for a group of evacuees is based on the

relationship of movement speed and population density of the evacuation flow.

Several assumption are made to facilitate the calculation process. [25,26]

a. All occupants stayed inside the building will start the evacuation at the same time

b. No any interruption will be involved in the evacuation stream of evacuees

c. No mobility impaired person who affecting the movement of group will be involved

in the evacuation

d. Occupants evacuate from the original location to the staircase via the nearest exit

door

e. The evacuation rate depends on the density and width of each egress components

Nelson and Mowrer Method

There are several approaches to calculate the evacuation duration derived from Nelson

and Mowrer method, while two common approaches are First-Order Hydraulic Model

and Second-Order Hydraulic Model. Both approaches are used to assess the time

between the initiation of the movement for evacuation and the arrival of place of safety.

For first-order hydraulic model, it is a relative simplified approach since it is not

consider the effect from all elements throughout the escape route such as population

19

density or calculated flow. It only considers the component which causing the most

significant effect to the evacuation flow, for example, the maximum limitation or

constraint of the specific flow. After identify the controlling factor, it will control the

calculation procedure of the latter part.

Second-order hydraulic model which adopted for this study is a relatively

comprehensive approach. It takes all structural characteristic that evacuees will face

during the evacuation into consideration. These characteristics are not only some

physical constraint related to the area, but also include the merging and narrowing

situation between corridor and exit door. Thus, it is a more realistic calculation with the

actual environment.

4.1.1. Effective Width

Effective width of the escape route is the clear width of the escape route minus the

boundary layer. Boundary layer usually exist when a person walk along the exit route

because people always keep a clearance distance away from wall or other obstacles.

Hence, the effective width accommodates the person passing is not equal to the actual

width of the escape route. (Nelson and Mowrer, 2004)

Figure 4.1.1. Effective width and Clear Width

20

4.1.2. Movement Speed

The movement speed of evacuees depends on the density and configuration in the

escape route, hence, the movement pace of an evacuee is categorized into three

conditions based on the population density in the escape route.

For the population density is lower than 0.54 persons/m2, evacuees travel with their

own pace without being affected by each other in order to attain a higher moving speed.

For the population density exceeds 3.8 persons/m2, it means that there is too crowd in

the escape route, each evacuee is no longer able to have any movement. If the

population density is between 0.54 and 3.8 persons/m2, the movement speed will

depend on the equation, (Nelson and Mowrer, 2004)

S = k – akD

where k is a constant depending on the type of evacuation route

a is a constant with a value of 0.266

From the above equation, the types of staircase will affect the efficiency of evacuation.

When the staircase has a lower height of riser and longer tread, it can reach a higher

movement speed comparing the staircase with higher riser and shorter tread, since the

staircase will be constructed in greater slope, evacuees might face more challenge to

move with a high speed.

21

Figure 4.1.2.1. Relationship between Constant and Evacuation Speed

Figure 4.1.2.2. Relationship between Density and Movement Speed

4.1.3. Specific Flow

Specific flow is defined as the number of person passing a point of an exit route per

unit time per unit of effective width.

Fs = SD

= (1 – aD)kD

Due to limitation of the exit route, there is the corresponding maximum specific flow

regarding different types of escape route including corridor and staircase (Nelson and

Mowrer, 2004).

22

Figure 4.1.3.1. Maximum Specific Flow

Figure 4.1.3.2. Relationship between Density and Specific Flow

23

4.2. Computational Analysis

Computer modelling or simulation method is required to analyze the evacuation

situation in certain buildings such as high-rise building, shopping malls due to its

geometrical complexities of the area and the large building population. Currently, many

computer programs are available to model the evacuation flow in crowd situation. Some

simulation packages are developed to analyze the movement of individual evacuee by

using a fine network of nodes in a defined area. (Ng, 2009)

Simulex is a computation tool which utilize the particle tracking technique to study the

evacuation movement of large population of individuals in a building. By inserting the

CAD floor plans into the program with all well-connected staircases and exits, the

‘distance mapping’ technique is used to determine all travel distances to the final exit

and routing in floors for each individual after the initial starting positions have been set

by the user. The ‘distance map’ consists with a fine mesh of 0.2m x 0.2m spatial block

which covering the entire area of each floor, while the numerical value assigned to each

block is the travel distance between the center of block and the nearest exit. (IES, 2014)

In this study, Version 2012.0.1.1 of the Simulex package is used to evaluate the

evacuation process of the selected super-tall residential building.

24

4.2.1. Assumption

In the user guide of the software stated that the simulation of evacuation process us

based on number of principles and assumptions related to both building geometry and

occupant characteristics. (Thompson & Marchant, 1995)

a. Each person is evacuating with a normal and unimpeded movement speed.

b. Walking speed will be reduced as shorter the inter-person distance

c. Each person will move towards the exit by using a right direction which shown on

the distance map of each case.

d. Overtaking, body rotation, side-stepping and even back-stepping would be

considered in the simulation process.

4.2.2. Building Population

There are different occupant types can be selected in order to suit with the occupant

characteristic of the buildings. Different occupant types such as office staff, commuters

or shoppers will have their own proportion of male, female, child and elderly, which

can directly affect the resultant total evacuation time. For example, the database of

Simulex has preset that male occupants can travel with the highest velocity while

children will travel with a lower speed. [7]

Figure 4.2.1.1. Variation of Movement Speed between Types

and Gender of Occupant in Simulex

25

In this study, the profile of shoppers will be adopted for the simulation in order to

reflect the evacuation time for families.

4.2.3. Body Shape

The human body shape is presented mathematically as three circles (one for main body

and two for the shoulders) which is widely accepted approach to show the shape of

body from the plan view.

Each person will have the same average body size, but the body dimensions will be

varied between different body types such as the gender and age. The average body size

used in the software is 0.5m x 0.5m, which is an average value based on the literature

and studies. (Thompson & Marchant, 1995)

26

5. DESCRIPTION OF CASE STUDIES

An all-residential building with 75-storey in West Kowloon of Hong Kong is selected.

There are several blocks within the residential development project, one block in the

project is chosen for the case study. Up to 2015, the selected building is the fourth

highest all residential building in Hong Kong. [27] In the whole building, 63 storeys in

total are designed with domestic flats, while the remaining floors are used for general

usage such as mechanical floor and also the refuge floors.

There are six domestic units on each floor (Flat A-F), while all of them are designed

with around 100m2 in average usable area. All the flat area and dimension used for

numerical calculation and computer simulation are based on the sales brochure

provided by the developer. [26] The floor layout plan, building information and the

geometry of major egress components are listed below.

Total Number of floor 75

Floor with residential flats 7-13/F, 15-23/F, 25/F, 27-33/F, 35-43/F, 45-52/F, 55-

63/F, 65-73/F, 75-76/F, 77&78/F, 79&80/F

Number of flats on each floor 6

Location of Refuge Floor 26/F , 53/F

Floor-to-floor Height 3m

Number of Staircase 2

Corridor Clear Width 1200 mm

Staircase Clear Width 1100 mm

Exit Door Clear Width 800 mm

Dimension of (Corridor) 14.2m (L) x 1.2m (W)

Figure 5.1.1. Information of Selected Super-tall Residential Building [26]

27

Figure 5.1.2. Floor Layout Plan of Selected Super-tall Residential Building [25]

Four cases with different number of occupants (5 and 12 occupants inside each flat) and

evacuation strategies (Total evacuation and phased evacuation) will be studied. 5

occupants inside each flat is based on a typical four-person families with an elderly or

maid, while12 occupants inside each flat is based on the assessment of occupant density

in local regulation. For a typical flat, the occupant factor is 9 m2 usable floor area per

person. With the domestic flat in 100 m2 in average, there will be 12 occupants in each

flat. Although 12 occupants inside a flat is seems a large value, actually, there is the

possibility to have large occupant density in a flat especially for Chinese people. During

the celebration of festivals such as Lunar Chinese New Year or Mid-Autumn Festival,

it is the custom for the family member and relatives gathering to have a dinner or

different activities. Evacuation of those visitors is also a concern in safe evacuation.

Staircase – 2

Staircase – 1

Corridor – 2

Corridor – 1

28

Two evacuation strategies are used to analysis the effectiveness of evacuation process,

total evacuation method and phased evacuation method. Total evacuation means that

all occupants live between refuge floors or roof are required to have the evacuation. In

case study, all residence between 55/F and 79/F will need to evacuate towards the refuge

floor at 53/F in total evacuation strategy. In other hand, only occupants live in four

storeys (fire floor, nearest two floors above and the nearest floor below) are notified to

take the evacuation towards the nearest refuge floor in phased evacuation strategy. In

the case studies, fire is assumed to happen on 77/F, occupants live in four topmost floors

(76/F to 79/F) will need to evacuate to refuge floor with the longest travelling distance.

Case 1: 5 occupants inside each flat; Total evacuation strategy adopted in building

Case 2: 12 occupants inside each flat; Total evacuation strategy adopted in building

Case 3: 5 occupants inside each flat; Phased evacuation strategy adopted in building

Case 4: 12 occupants inside each flat; Phased evacuation strategy adopted in building

For all cases, the total travelling time and population density along the staircases will

be analyzed by using the computational tool. For the result of population density, there

are six measuring points selected in the staircase with 30 seconds interval. The

measuring points are listed as follows:

1. Staircase between 76/F and 77/F

2. Staircase between 71/F and 72/F

3. Staircase between 67/F and 68/F

4. Staircase between 62/F and 63/F

5. Staircase between 58/F and 59/F

6. Staircase between 53/F (Refuge Floor) and 55/F

Number of evacuees in the measuring staircases will be recorded at each 30 seconds

interval in order to analyze the variation of population density with time in the

evacuation routes.

29

6. RESULTS

6.1. Empirical Calculation

The hydraulic model developed by Nelson is a calculation tool to evaluate the egress

time of occupants, while the geometry of different egress components are briefly

required in the calculation process. Hence, some assumptions were made on the missing

information to facilitate the calculation.

According to the building layout plan, the width of corridor and staircase are given in

1200mm and 1100mm respectively. For the width of exit door, 850mm is assumed to

comply with the minimum width requirement in local regulation. Apart from that, the

staircase in configuration with 7.0 inch riser and 11.0 inch tread is assumed as the

geometry of staircase in the selected building. All assumptions made are fulfill the

minimum width of tread and maximum height of riser in the local requirement.

Two scenarios, five occupants and twelve occupants inside each flat, are calculated

under total evacuation strategy. The movement velocities, population density and

effective width of different egress components are considered in the calculation process.

For the case with five occupants inside each, around 620 seconds (10 minutes) (Refer

to Appendix A1) is required to evacuate all occupants, while the case with twelve

occupants inside each flat will require around 1239 seconds (20.6 minutes). (Refer to

Appendix A2)

30

6.2. Computer Simulation

6.2.1. Case 1

For the building with five occupants inside each flat, there are 690 evacuees involved

in the evacuation between the roof and refuge floor under total evacuation strategy. 831

seconds (13.85 minutes) is required to evacuate all occupants from their original

locations to refuge floor. (Refer to Appendix B1)

Figure 6.2.1.1. Total Cumulative Persons Arriving Refuge Floor

Based on the simulation result, the entire area along the staircases is occupied by

evacuees during the whole evacuation process. Although there is same number of

evacuees using each staircase, the resultant time required for two staircases have a

significant difference. From Figure 6.2.1.2. and Figure 6.2.1.3., shorter evacuation time

as 665 seconds are able to transport all evacuees to the refuge floor for staircase 1 when

comparing with staircase 2. Staircase 2 needed 830 seconds to achieve the same target

which involving 165 seconds difference between two staircases.

0

100

200

300

400

500

600

700

800

5

30

55

80

10

5

13

01

55

18

0

20

52

30

25

5

28

0

30

53

30

35

53

80

40

54

30

45

54

80

50

5

53

05

55

58

06

05

63

06

55

68

07

05

73

0

75

57

80

80

5

83

0

31

Figure 6.2.1.2. Cumulative Persons Arriving Refuge Floor via Staircase 1

Figure 6.2.1.3. Cumulative Persons Arriving Refuge Floor via Staircase 2

0

50

100

150

200

250

300

350

400

52

54

56

58

51

05

12

51

45

16

51

85

20

52

25

24

52

65

28

53

05

32

53

45

36

53

85

40

54

25

44

54

65

48

55

05

52

55

45

56

55

85

60

56

25

64

56

65

0

50

100

150

200

250

300

350

400

5

30

55

80

10

5

13

01

55

18

0

20

52

30

25

5

28

0

30

53

30

35

53

80

40

54

30

45

54

80

50

5

53

05

55

58

06

05

63

06

55

68

07

05

73

0

75

57

80

80

5

83

0

32

Figure 6.2.1.4. Population Density Recorded at Measuring Points (Staircase 1)

Figure 6.2.1.5. Population Density Recorded at Measuring Points (Staircase 2)

The population density is varies with the time since number of evacuees is each part of

the staircases will be totally different at each time interval. In Staircase 1, the highest

population density reached 3.18 persons/ square meter (m2) in the staircase connecting

62/F and 63/F at 120s. Comparing the overall population density between two cases,

0

0.5

1

1.5

2

2.5

3

3.5

30

60

90

12

0

15

0

18

0

21

0

24

0

27

0

30

0

33

0

36

0

39

0

42

0

45

0

48

0

51

0

54

0

57

0

60

0

63

0

66

0

69

0

72

0

75

0

78

0

81

0

84

0

Den

sity

(P

erso

ns/

sq.m

)

Time (s)

77/F -> 76/F 72/F -> 71/F 68/F -> 67/F

63/F -> 62/F 59/F -> 58/F 55/F -> Refuge

0

0.5

1

1.5

2

2.5

3

3.5

30 60 90 120 150 180 210 240 270 300 330 360 390 420 450 480 510 540 570 600 630

Den

sity

(P

erso

ns/

sq.

m)

Time (s)

77/F -> 76/F 72/F -> 71/F 68/F -> 67/F

63/F -> 62/F 59/F -> 58/F 55/F -> Refuge

33

the results are quite similar with the range between 2 and 2.5 persons/ m2. Based on the

theoretical equation, the movement velocity is kept at 0.36 m/s in most of the time of

evacuation process.

After t=360s, 314 occupants have evacuated from the building and reached the refuge

floor, 173 and 141 of them are evacuated from Staircase 1 and Staricase 2 respectively,

which is around 45% of the occupants inside the buildings.

6.2.2. Case 2

Case 2 is the worst situation among four case studies, which involving 1656 occupants

living between refuge floor and the roof. The longest evacuation time as 1548s (25.8

minutes) is required to evacuate all involved residence. (Refer to Appendix B2)

Figure 6.2.2.1. Total Cumulative Persons Arriving Refuge Floor

Similar to Case 1, the entire staircase is packed with a crowd of evacuee, but situation

of Case 2 is much worse than other cases. In Figure 6.2.2.2, population density is the

most significant issue in this case. The population density began to increase from a low

0

200

400

600

800

1000

1200

1400

1600

1800

5

50

95

14

0

18

5

23

0

27

5

32

0

36

5

41

0

45

5

50

0

54

5

59

0

63

5

68

0

72

5

77

0

81

5

86

0

90

5

95

0

99

5

10

40

10

85

11

30

11

75

12

20

12

65

13

10

13

55

14

00

14

45

14

90

15

35

34

level to around 3.8 persons/sq. m within first two minutes after the initiation of

evacuation. In the first seven minutes, the population density in upper floors (62/F-77/F)

maintained in a high level. For the later part in the evacuation period, the entire staircase

involved in the evacuation zone kept at 2.5 persons/sq. m in average. This situation

happened for both Staircase 1 and Staircase 2. Based on the SFPE Handbook, no any

movement can take place if the population density exceeds 3.8 persons/ sq. m until the

crowd has passed from the area to reduce the population density along the evacuation

path. The value even reached 4.3 persons/ sq. m which already exceed the allowed

upper limit for movement. In the animation of result generated by the software, all

occupants squeezed into the staircase and only a slightly movement can be achieved.

Figure 6.2.2.2. Population Density Recorded at Measuring Points (Staircase 1)

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

30

90

15

0

21

0

27

0

33

0

39

0

45

0

51

0

57

0

63

0

69

0

75

0

81

0

87

0

93

0

99

0

10

50

11

10

11

70

12

30

12

90

13

50

14

10

14

70

15

30

Den

sity

(P

erso

ns/

sq.m

)

Time (s)

77/F -> 76/F 72/F -> 71/F 68/F -> 67/F

63/F -> 62/F 59/F -> 58/F 55/F -> Refuge

35

Figure 6.2.2.3. Population Density Recorded at Measuring Points (Staircase 2)

The evacuation flow between Staircase 1 and Staircase 2 are similar with highest

population level occurred at the first seven minutes. The evacuation flow completed at

1548s and 1515s for Staircase 1 and Staircase 2 respectively. In evacuating a large

number of occupants, it is not a significant difference of the evacuation time required

for two staircases compared to case 1. The reason is due to the high population density

in the staircases. In this case, although there was various occupant types in the building,

each of evacuee moved with the limited speed in the crowd instead of their own

movement speed, the effect of speed variance is less significant to the result. Hence, the

total evacuation time required for two identical staircase would be similar in the

situation of same population loading in each staircase.

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

30

90

15

0

21

0

27

0

33

0

39

0

45

0

51

0

57

0

63

0

69

0

75

0

81

0

87

0

93

0

99

0

10

50

11

10

11

70

12

30

12

90

13

50

14

10

14

70

Den

sity

(P

erso

ns/

sq.

m)

Time (s)

77/F -> 76/F 72/F -> 71/F 68/F -> 67/F

63/F -> 62/F 59/F -> 58/F 55/F -> Refuge

36

Figure 6.2.2.4. Cumulative Persons Arriving Refuge Floor via. Staircase 1

Figure 6.2.2.5. Cumulative Persons Arriving Refuge Floor via. Staircase 2

0

100

200

300

400

500

600

700

800

900

5

55

10

5

15

5

20

5

25

5

30

5

35

5

40

5

45

5

50

5

55

5

60

5

65

5

70

5

75

5

80

5

85

5

90

5

95

5

10

05

10

55

11

05

11

55

12

05

12

55

13

05

13

55

14

05

14

55

15

05

0

100

200

300

400

500

600

700

800

900

5

55

10

5

15

5

20

5

25

5

30

5

35

5

40

5

45

5

50

5

55

5

60

5

65

5

70

5

75

5

80

5

85

5

90

5

95

5

10

05

10

55

11

05

11

55

12

05

12

55

13

05

13

55

14

05

14

55

15

05

37

In the evacuation flow, it is obvious that the population density would reach a high level,

while it is interested for the density recorded in staircase connected between the lowest

floor of the zone (55/F) and refuge floor (53/F) kept at a relatively lower level in range

of 2 – 2.5 persons/sq. m. during the whole period of time.

As mentioned before, it is the worst scenario within four case studies. At 360s, only

366 occupants successfully reached the refuge floor which implies that 1290 evacuees

still stayed inside the staircase. It is the low efficiency to evacuate 22% of occupants in

the building within six minutes. Additionally, the population density of upper floor was

still reaching the upper limit value at 360s. There was the high probability for the fire

fighters faced the counter-flow problem as they were moving upstairs to the fire scene

for rescue. In the simulation result, it shows that no any extra spaces available to

accommodate the fire-fighting teams in an overcrowded staircase, which might cause

the delay in providing the rescue to the injuries.

6.2.3. Case 3

Case 3 has the least number of population as 120 occupants involved in the evacuation

and thus requires the least time which about 679s (11.3 minutes) for the completion of

phased evacuation. (Refer to Appendix B3) In Staircase 1 and Staircase 2, it took 625s

and 679s to finish the evacuation process respectively.

38

Figure 6.2.3.1. Cumulative Persons Arriving Refuge Floor

In the evacuation flow, majority of evacuees were evacuating in small groups of twelve

to fourteen people. In each staircase, four groups of evacuees were found from the

animation result. In Figure 6.2.3.2, which expressed in statistical format, it clearly

shows that four significant slants are plotted at t = 305s, 365s, 410s and 545s, while

there are only slightly increases or even a flat line in the remaining period of time. It

reveals that the majority of evacuation flow is divided into four batches of evacuee. The

slightly increase of value in the graph means that there were one or two evacuees who

travelling in a faster pace did not follow the movement in groups. They were able to

evacuate with their own speed for overtaking other evacuees.

0

20

40

60

80

100

120

140

52

54

56

58

51

05

12

51

45

16

51

85

20

52

25

24

52

65

28

53

05

32

53

45

36

53

85

40

54

25

44

54

65

48

55

05

52

55

45

56

55

85

60

56

25

64

56

65

39

Figure 6.2.3.2. Population Density Recorded at Measuring Points (Staircase 1)

Since there was not a continuous evacuation flow along the staircase, the measured

population density was zero if no any evacuee passing through the measuring points.

When the groups of evacuees approached to the measuring points, the population

density increased gradually. The maximum population density recorded is about 2.15

persons/sq. m when the evacuees reached the measuring points. The peak of population

density only kept as a short period of time since the group of people could able to travel

a floor within 30 seconds. After the group of people passed through the measuring

points, value of population density gradually decreased to zero. The same situations

happened in every measuring points of staircases.

0

0.5

1

1.5

2

2.5

30 60 90 120 150 180 210 240 270 300 330 360 390 420 450 480 510 540 570 600 630 660 690

Des

nsi

ty (

Per

son

s/sq

. m

)

Time (s)

77/F -> 76/F 72/F -> 71/F 68/F -> 67/F

63/F -> 62/F 59/F -> 58/F 55/F -> Refuge

40

Figure 6.2.3.3. Population Density Recorded at Measuring Points (Staircase 2)

At t = 360s, only 10 persons had evacuated from the building through each staircase.

Upon the number of evacuees finished the evacuation, it seems that phased

evacuation is not an efficient approach as less than a half of occupants could reach the

refuge floor before the arrival of fire fighters. Actually, the “inefficient evacuation” is

due to the long travelling distance but not the population density. It is obvious that it

require a long period of time for occupants travelling around 200m from topmost floor

to refuge floor. If the fire happened in lower level of the zone, the occupants lived

between 55/F and 58/F are required to evacuate, only 171s (2.85 minutes) is required

to complete the phased evacuation for the same population loading. This value

indicates that the duration of phased evacuation is highly dependent on the origin of

fire floor, but not the population density.

0

0.5

1

1.5

2

2.5

30 60 90 120 150 180 210 240 270 300 330 360 390 420 450 480 510 540 570 600 630

Den

sity

(P

erso

ns/

sq.

m)

Time (s)

77/F -> 76/F 72/F -> 71/F 68/F -> 67/F

63/F -> 62/F 59/F -> 58/F 55/F -> Refuge

41

6.2.4. Case 4

712s (11.9 minutes) is required to evacuate 288 occupants from alerted zone in phased

evacuation approach. (Refer to Appendix B4) There is slightly time difference between

the evacuation time of two staircases, Staircase 1 needed 690s only while Staircase 2

needed 712s.

Figure 6.2.4.1. Cumulative Persons Arriving Refuge Floor

In this case, although occupants evacuated from four floors only, the evacuation flow

did not separated into different groups of evacuee. In contrast, merging flow between

floors occurred in this case, people from each floor combined into a continuous

evacuation flow in the staircase. Figure 6.2.4.1. shows that an ongoing flow of evacuees

arrived the refuge floor is recorded for the consecutive periods of time.

In the previous case, population density of the measuring points dropped to zero

between two batches of evacuees since there was the separation of evacuation flow. In

this case, a slightly drop of density was recorded instead of dropped to zero for the

majority of evacuation flow. The highest population density as 3 persons/sq. m.

occurred at the topmost floor during the initial stage of evacuation. But in view of the

0

50

100

150

200

250

300

350

5

30

55

80

10

5

13

0

15

5

18

0

20

5

23

0

25

5

28

0

30

5

33

0

35

5

38

0

40

5

43

0

45

5

48

0

50

5

53

0

55

5

58

0

60

5

63

0

65

5

68

0

70

5

42

average density in the remaining period of time, the density kept in the range between

1.5 – 2.5 persons/sq. m.. In this range of density, evacuees were still able to move in

the speed around 0.4m/s to 0.62m/s by using correlation between density and movement

speed in SFPE handbook.

Figure 6.2.4.2. Population Density Recorded at Measuring Points (Staircase 1)

Figure 6.2.4.3. Population Density Recorded at Measuring Points (Staircase 2)

0

0.5

1

1.5

2

2.5

3

30 60 90 120 150 180 210 240 270 300 330 360 390 420 450 480 510 540 570 600 630 660 690 720

Den

sity

(P

erso

ns/

sq.m

)

Time (s)77/F -> 76/F 72/F -> 71/F 68/F -> 67/F

63/F -> 62/F 59/F -> 58/F 55/F -> Refuge

0

0.5

1

1.5

2

2.5

3

3.5

30 60 90 120 150 180 210 240 270 300 330 360 390 420 450 480 510 540 570 600 630 660 690

Den

sity

(P

erso

ns/

sq.

m)

Time (s)

77/F -> 76/F 72/F -> 71/F 68/F -> 67/F

63/F -> 62/F 59/F -> 58/F 55/F -> Refuge

43

Owing to the long travelling distance between alerted zone and refuge floor, around

210s is required for the first evacuee travelled to refuge floor. Thus, only 15 evacuees

successfully reached the refuge floor at t = 360s.

44

7. RESULT ANALYSIS

7.1. Comparison between Numerical Calculation and Simulex Result

Comparing the results obtained from two methods, it found that the travelling time

simulated by Simulex are longer than calculated result from equations in both occupant

population. For 5 occupants inside each flat, 831s (13.85 minutes) is required for last

evacuee reaching the nearest refuge floor in Simulex simulation while only 620.2s (10.3

minutes) is required for empirical calculation. For 12 occupants inside each flat, similar

situation is happen. 1548s (25.8 minutes) is found from computer simulation while only

1239s (20.7 minutes) is needed in empirical calculation.

Although the time taken for travelling between the occupants’ origin location and the

main door of a flat is one of factors contributing the difference, other reasons such as

the density play more important role in the variation of results. In computer simulation,

all occupants are assumed to stay inside the bedrooms as the origin which attain the

longest distance away from the unit door. Once simulation begins, time taken to travel

across the unit door is also recorded; while empirical equation method analyzes the time

of evacuation at the unit door rather than randomly distributed in the flat. In the

simulation result, it reveals that occupants take 10 -15s to evacuate from point of origin

to flat door. Hence, comparing to the 200 – 300s difference between the obtained results,

the relative short time span for travelling inside the flat might not contribute the major

variation of result.

In fact, the deviation of movement speed and population density cause more important

influences to the difference. In computer simulation, the movement velocity of

evacuees depends on the occupant type selected for the program including to proportion

of child and elderly. Different occupant types have its movement velocity. Besides,

evacuees were not evenly distributed along the evacuation path, especially in staircase,

45

from computer simulation. Some groups of evacuee with higher population density

were formed along the staircase rather than even distribution with an average

population density. It might be caused by the relatively slow movement velocity of

elderly or child. In the database set by Simulex, the average movement velocity of an

elderly and child is only around 60% of the velocity for an adult. Hence, an evacuee

with lower movement speed lived in lower storey will affect the evacuation flow with

higher movement speed from upper floor, while overtaking situation seldom occur in

the evacuation process from the simulation result. Thus, several groups of crowd with

each contains more than twenty occupants were induced, the movement speed

significantly reduced around 0.4 m/s, which used 18s to travel a floor.

In empirical calculation, although different parts of evacuation route would have

different population density, the population density would remain the same in a single

egress component. Hence, it is assumed that evacuation flow is even distributed along

the with a constant population density along the entire staircase. Besides, the movement

speed of all occupants are the same based on the specific flow and population density

of a particular egress component. In contrast, variance of population density and

movement velocity exist in real situation or even computer simulation which induce the

blockage problem. As a result, the time required for travelling one floor is 10.4s and

13.4s for 5 and 12 occupants inside each flat respectively, while at least 20s is required

from the computer simulation result.

46

7.2. Comparison between Total Evacuation and Phased Evacuation Strategy

In studying the feasibility or appropriateness of an evacuation strategy, the total

evacuation time definitely is an important factor, however, there are also other

considerations to determine the quality of evacuation process, such as population

density along egress components or existence of counter-flow problem. Hence, the

discussion and analysis are made on different scenarios to identify the performance of

phased evacuation strategy in a building.

7.2.1. 5 Occupants inside each flat

By comparing both scenarios, even the population involved in the evacuation for case

1 is 5.75 times to the population loading in Case 2, the difference of resulted time

required for two cases is not quite significant. The reason might due to the population

density of the evacuees along the staircase.

Indeed, the population density of Case 1 kept in the range between 2 – 2.5 persons/sq.m.

in most of the time, while the population density of each groups of evacuees was around

2.15 persons/sq.m. for Case 2. Since population density is the dominant factor to

determine the overall evacuation time required in different situation, it is reasonable

that a similar population density of crowd will result in a similar evacuation time in

theory.

Upon the value of population density, it seems that overall population density of the

movement of crowd are similar for both cases, however, the duration of that population

density is also a concern. For total evacuation which involving more number of

evacuees, the population density level usually keep a certain period of time especially

for the lower storeys within the evacuation zone. The lower of the floor level, more

evacuees will pass through the staircase on that floor due to the cumulation of

47

evacuation flow. For example, the evacuation flow ended at t = 270s in 67/F, but the

end of flow in 58/F had been prolonged to t = 570s even it is assumed with the same

initiation time of evacuation at all floors. This situation might come to the counter-

flow problem against the upward movement of fire fighters.

In the concept of total evacuation approach, all floors between the refuge floor and

roof are required to carry out the evacuation procedure in this case study, regardless

of the location of fire scene in the high zone of the residential building. Therefore, it

might have the chance facing the counter-flow problem once fire happens between

55/F and 62/F. Local fire services department pledges the fire team can respond to the

building fire calls within 6 minutes. However, simulation shows that around 170 – 200

of evacuees were still having their evacuation after the arrival of fire fighters.

On the other hand, counter-flow problem is not likely to occur by adopting counter-

flow approach. In the same scenario as fire happened at topmost floors, all evacuees

left 76/F within three minutes. Hence, if fire fighters takes longer than three minutes

to arrive the fire scene, staircases near the fire floor will be free of evacuees. Fire

fighters can easily reach the fire scene without facing any blockage.

48

7.2.2. 12 Occupants inside Each Flat

The evacuation time can be significantly reduced to 712s by adopting phased

evacuation, while it took 1548s for the evacuation time in original approach. Thus,

phased evacuation can carry out its function in shortening the evacuation time under

high population density scenario.

In the first half of total evacuation period, the population density at upper floors kept

at high level which in range of 3.5 – 4 persons/m2, the value even rose to 4.3

persons/m2 for a short instant when evacuation flow passing through the measuring

points. The high density induced by large evacuees involved into the evacuation is the

dominant reason for long evacuation time. Even after the arrival of fire fighters, the

staircases of upper floors still packed with crowd of evacuees. If fire fighters need to

move upstairs to the fire scene via staircases, they might face the difficulties as no

extra free spaces are available to accommodate the contra flow of rescue team, while

some of them carries the bulk equipment.

The situation can be alleviated in phased evacuation. The highest population density

is only 3 persons/m2 which measured at the initial stage of evacuation. In the

remaining time, the density had not exceed 2.5 persons/m2 anymore. At t = 360s, the

evacuation flow already left the alerted zone. As a result, counter-flow problem is no

longer to happen when rescue teams moving upwards to the fire scene.

49

8. DISCUSSION

In the literature, there are lots of parameters contribute the effects on the effectiveness

and efficiency of an evacuation including population density, movement speed, human

behaviors or counter-flow problems. Every parameter plays an important role to the

result. The comparison between empirical calculation and computer simulation is

already a simple example. Based on the same population loading and building geometry,

the differences between two approached are mainly due to the variations of speed of

occupant types and the merging flow. As mentioned before, computer simulation

software can consider the movement speed of individual. Apart from that, twisting or

overtaking are the human behaviors commonly found in a merging flow. Only these

behaviors already result in the delay of evacuation time. However, neither the empirical

calculation nor the computer simulation have considered all the human behavior during

the evacuation due to the lack of statistical information about the effect of behavior to

the evacuation nowadays. Nevertheless, it is proved that the inevitable human behavior

during emergency situation will prolong the total evacuation time required for the

building.

Comparing total evacuation and phased evacuation, the time required for total

evacuation is usually longer than that required for phased evacuation. Moreover, the

population density is the major concern in an evacuation. Past experience shows us that

the over-crowded situation can result in a more serious consequence than the actual

causes such as fire or hot smoke. Even evacuees might not directly exposed to the fire,

panic or non-adaptive behavior during emergency case will cause peoples stampede. In

a crowd environment of staircase, evacuees might be trampled to injury or death. As

the concern on human behavior is getting higher and higher, many papers found that

crowded environment is the root of non-adaptive behavior.

50

Apart from the uncertainty of human behavior, the fatigue of evacuees will contribute

to the further extension of evacuation time. In the selected supertall residential building,

the longest distance required for the evacuees travelling from topmost floor to refuge

floor is 192m. Experiment already found that the depreciation of movement speed will

getting more significant for each 100m travelling distance. The result is based on an

evacuation of a healthy person, while the depreciation rate is twice of a healthy person

for people with lower physical ability.

In Hong Kong or even in the world, ageing problem is a hot social issue nowadays. It

is estimated that the ageing problem will getting more and more serious. For an elderly,

it is reasonable that the physical ability is lower than a middle age person in general.

Meanwhile, there will be more and more elderly in the society in future. Hence,

Spearpoint & MacLennan (2012) already realized the relevance of data which

conducted few decades ago with today or future, since the population profile in terms

of age and gender changed in past decade, and also predicted to change in next decades.

Take Hong Kong as an example, the percentage of elderly aged 65 years or above to all

age group has increased from 8.31% in 1991 to 21% in 2014. Recently, the population

policy forecasts that the percentage will further increase to 32% in 2041. [11,23] A

modelling conducted in New Zealand suggested that the total evacuation time will

extend up to 20% when comparing the data for New Zealand scenario in 2031 and

nowadays. (Spearpoint & MacLennan, 2012) Therefore, the extension of required

evacuation time in future should be considered nowadays since buildings designed in

this moment will serve for the next 50 years as the services life.

As the longer and longer evacuation time due to human behavior or ageing problem,

solution be proposed to minimize these effects acting on the evacuation time. Having a

51

wider staircase ostensibly can accommodate more evacuees to speed up the evacuation

rate. However, Helbing et al. (2000) found that there is no any significant improvement

can be attained by widening the egress routes since it only resulted in more evacuees

congested in the staircase and thus no alleviation of the population density.

Comparing between different cases between total evacuation and phased evacuation in

Figure 8.1, phased evacuation brings the notable advantages on both reducing the total

evacuation and population density along staircase. To the viewpoint of evacuees, they

are able to travel with a faster pace without facing much merging situation and

congestion due to high population. In the viewpoint of safety, less unwanted non-

adaptive behavior is likely to appear since no crowd with high density is found in

phased evacuation.

Strategy Scenario No. of people Time Required

Case 1 Total Evacuation 5 occupants / flat 690 13:51

Case 2 Total Evacuation 12 occupants / flat 1656 25:48

Case 3 Phased Evacuation 5 occupants / flat 120 11:19

Case 4 Phased Evacuation 12 occupants / flat 288 11:52

Figure 8.1. Summary of Results in Different Cases

Based on the facts mentioned above, it is feasible to adopt the phased evacuation and

defense-in-place strategy instead of total evacuation in a residential building. Actually,

this is not a new evacuation strategy for Hong Kong. In local commercial and hotel

buildings with sprinkler provisions, phased evacuation is already adopted in these

premises. [9]

Higher safety level for an occupants in unaffected area can be experienced if they stay

inside their own flats rather than taking the evacuation procedure. It is because

52

occupants can keep in clam when stayed in a familiar places rather than congested in

the crowded staircases. Moreover, occupants can take the preventive measures such as

tightly close the door and seal the cracks or bottom gap between the door and floor. In

contrast, if occupants in unaffected area are required to initiate the evacuation process,

it will cause a serious impact to the population density in staircase, thus, a series of

human behavior will also be induced as the chain effect. However, the prerequisite for

adopting phased evacuation is the adequate fire safety provisions and non-combustible

construction of the building.

It is a controversial issue about the fire safety provisions of residential building in Hong

Kong after every tragedy happened with severe casualties. Citizens or commentators

usually blamed insufficient fire safety provisions for the casualties of accident.

Compared with the safety standard of other Asian cities including Shanghai and Taiwan,

automatic sprinkler system is already one of the essential elements required in the

standard.

It is widely accepted that the hazard level of supertall buildings is higher than others

since no any fire fighters can directly approach the fire scene by ladder to carry out fire

fighting and rescue operation at the ultimate high level. Fire fighters can only rely on

the staircase and fireman’ lift as the means of access to the fire scene. In the current

evacuation approach, simulation results shows that fire fighters in upward movement

will highly probably face the counter-flow of evacuees moving downwards. Heliovaara

et al. (2012) and Hofinger et el. (2014) stated that jams and collisions are induced by

the counter-flow since evacuate might need to adjust their moving direction or rotating

their bodies to minimize the collisions. Both evacuation and fire fighters flow of

movement will be affected. For the fire growth, the continuous fresh air supply at high

53

elevation height and large window pane can facilitate the combustion and growth of

fire. The later the start of fire fighting operation, more time will be available for the fire

growth.

In order to minimize the obstruction to the flow of both evacuees and fire fighters,

phased evacuation is a feasible and better strategy to supertall residential building with

the provision of automatic sprinkler system. Only the occupants in fire floor, two floors

above and one floor below will be alerted by the fire bell rather than all floors within

the zone will be alerted in phased evacuation. It can significantly reduce the number of

evacuees involved and thus alleviate the crowd flow situation in staircase. For the

remaining occupants inside the building, they will not be alerted by the alarm. Since

they might not realize the fire happens in other part of their building like the defense-

in-place approach, automatic sprinkler system is proposed to be installed in the public

corridor of each floor and entire staircases in order to prevent the spread of fire through

staircases. Developers might argued that the new system will cost much money on

purchasing and installing. In fact, the installation of sprinkler system do not require the

change of building geometry. It will only cost less than $10,000 for installing the

sprinkler system on a floor. [13] Once installed the sprinkler system in a residential

building, phased evacuation can be implemented instead of the original total evacuation

approach.

In order to carry out the phased evacuation in a good order, it might require a higher

level of fire safety management to the building. Fire drill and well-trained staff can help

to achieve the successful evacuation procedure. In Hong Kong, supertall residential

building are usually classified as luxurious building with well-trained property

management staff. Hence, the proposed evacuation strategy and fire safety provisions

54

for high-rise residential building can be implemented to super-tall residential building

as the pilot scheme since these kinds of building usually with a higher risk level.

9. CONCLUSION

Although there are many studied related to the impact of human behavior and physical

ability on the required evacuation time, few of them can be quantified in the calculation

or simulation result. Nevertheless, numerous research studies agreed that human

behavior will extend the required evacuation time, while the severity of impact of

human behavior has a close relationship with the crowdedness of egress routes. Based

on the findings, case studies reveal that the inefficiency and potential problems caused

by total evacuation strategy which currently adopted in residential building. Upon the

problems found, phased evacuation and defense-in-place are proposed as the alternative

strategy to reduce the population density along the egress routes. Along with the new

strategy, automatic sprinkler system is also proposed as one of the fire safety provisions

should be installed in residential buildings. It is not only help achieving the prerequisite

for phased evacuation, but also help to enhance the fire safety level of buildings as

general purpose. Moreover, the new strategy is recommended to be used in the new

building design in order to cater the ageing problem in society in future, since higher

proportion of elderly involved in evacuation will face the physical challenge and panic

during the travelling long distance in emergency situation. The new strategy can plays

an important role to alleviate the severity of human behavior and thus the extension

time required to complete the evacuation process.

55

10. REFERENCE

1. BS 7974-6:2004 , “The application of fire safety engineering principles to fire

safety design of buildings – Part 6: Human factors: Life safety strategies –

Occupant evacuation, behaviour and condition”, British Standards Institution

2. Building Code 2008, Department of Buildings, New York City

3. C.L. Chow and W.K. Chow (2009), “Fire safety aspects of refuge floors in

supertall buildings with computational fluid dynamics”, Journal of Civil

Engineering and Management, Vol. 15, Issue 3, pp. 225-236

4. C.L. Chow and W.K. Chow (2010), “Heat release rate of accidental fire in a

supertall building residential flat”, Building and Environment, Vol. 45, Issue 7, pp.

1632-1640

5. Code of Practice for Fire Safety in Buildings 2011, Buildings Department, Hong

Kong SAR

6. Code of Practice for Minimum Fire Service Installations and Equipment and

Inspection, Testing and Maintenance of Installations and Equipment 2012, Fire

Services Department, Hong Kong SAR

7. Dirk Helbing, Illes Farkas & Tamas Vicsek (2000), “Simulating dynamical

features of escape panic”, Nature 407, pp. 487-490

8. Simo Heliovaara, Timo Korhonen, Simo Hostikka & Harri Ehtamo (2012),

“Counterflow model for agent-based simulation of crowd dynamics”, Building

and Environment, Vol. 48, pp. 89-100

9. Egress: Simulex User Guide (2014), Integrated Environmental Solutions Limited

10. Enrico Ronchi and Daniel Nilsson (2013), “Assessment of Total Evacuation

Systems for Tall Buildings”, The Fire Protection Research Foundation, January

2013

11. FSD Circular Letter No.4/ 1996, Fire Services Department, HKSAR

56

12. Gesine Hofinger, Robert Zinke & Laura Kunzer (2014), “Human factors in

evacuation simulation, planning and guidance”, Transportation Research Procedia,

Volume 2, pp. 603-611

13. Helbing, D & Monlar, P. (1995), “Social force model for pedestrian dynamics”,

Physical Review E, Vol. 51, Issue 5, pp. 4282-4286

14. H L Mu, J H Wang, Z L Mao, S M Lo & Q S Wang (2013), “Pre-evacuation human

reactions in fires: An attribution analysis considering psychological process”,

Procedia Engineering, Volume 52, pp. 290-296

15. Home Fire Sprinkler Cost Assessment – 2013, NFPA

16. Hong Kong 1991 Population Census, Census and Statistics Department, Hong

Kong

17. Jan Klerks (2011), “Tall Buildings in Numbers”, International Journal on Tall

Buildings and Habitat, Issue II, pp. 46-47

18. Jeongin Koo, Byung-In Kim and Yong Seog Kim (2014), “Estimating the effects

of mental disorientation and physical fatigue in a semi-panic evacuation”, Expert

Systems with Applications, Vol. 41, Issue 5, pp. 2379-2390

19. Kayvan Aghabayk, Omid Ejtemai, Majid Sarvi & Amir Sobhani (2014),

“Understanding pedestrian crowd merging behavior”, Transportation Research

Procedia, Vol. 2, pp. 768-773

20. K.F. Wong (2004), “Study on reliability of manual call points in residential

buildings”, International Journal on Engineering Performance-Based Fire Codes,

Vol. 6, No. 4, pp. 344-352

21. K.W. Lau and W.K. Chow (2014), “A Study on the Fire Safety Issues for Large

Window Openings in Supertall Residential Buildings in Hong Kong”, 3rd

International High Performance Buildings Conference

22. Michael Spearpoint & Hamish A. MacLennan (2012), “The effect of an ageing

and less fit population on the ability of people to egress buildings”, Safety

Science, Vol. 20, pp. 1675-1684

57

23. Mingchun Lao and Kelvin H.L. Wong (2006), “Evacuation Strategy for Super

Highrise Building”, Proceedings of 5th Annual Seminar on Tall Building

Construction and Maintenance

24. Ming Lo, S. and Will, B.F. (1997), “A view to the requirement of designated

refuge floors in high-rise buildings in Hong Kong”, Fire Safety Science, Vol. 5,

pp. 737-745

25. M.Y. Ng (2009), “Airport evacuation on comparing SIMULEX against

buildingEXODUS”, International Journal on Engineering Performance-Based Fire

Code, Number 1, pp.14-22

26. N. Cai and W.K. Chow (2010), “Numerical studies on evacuation for supertall

commercial buildings”, Journal of Applied Fire Science, Vol. 20, No. 2, pp. 119-

133

27. Peter A. Thompson & Eric W. Marchant (1995), “A computer model for the

evacuation of large building populations”, Fire Safety Journal, Vol. 24, Issue 2, pp.

131-148

28. Population by age group and sex 2014, Population Estimates, Census and

Statistics Department, HKSAR

29. Sales Brochure Part 1 – Part 3, The Harbourside, Hang Lung Properties

30. The SFPE Handbook of Fire Protection Engineering, National Fire Protection

Association (NFPA) and Society of Fire Protection Engineers (SFPE), 3rd edition

31. The SFPE Handbook of Fire Protection Engineering, National Fire Protection

Association (NFPA) and Society of Fire Protection Engineers (SFPE), 4th edition

32. The Skyscraper Center, The Global Tall Building Database of the CTBUH

33. UAE Fire and Life Safety Code of Practice, 2011 Edition, United Arab Emirates

34. W. Bukowski (2007), “Emergency Egress Strategies for buildings”, NIST

Building and Research Laboratory

58

35. W.K. Chow (2012), “Concerns on estimating heat release rate of design fires in

fire engineering approach”, International Journal on Engineering Performance-

Based Fire Codes, Vol. 11, No. 1, pp. 11-19

36. Xiaoshan Pan, Charles S. Han, Ken Dauber & Kincho H. Law (2006), “Human

and social behavior in computational modeling and analysis of egress”,

Automation in Construction, Volume 15, Issue 4, pp. 448-461

37. Year 2014 Statistic Report, Fire Services Department, Hong Kong SAR

59

APPENDIX A1 – Empirical Calculation Result for Case 1

Density (D) = Number of persons per unit area

= (5x3)/(14.2x1.2)

= 0.88 person per m2

By equation S = k – akD ,

= 1.4 – ( 0.266 x 1.4 x 0.88 )

= 1.07 m/s

By equation Fs = (1 – aD)kD ,

= (1 – (0.266 x 0.88)) x 1.4 x 0.88

= 0.94 persons/s/m of effective width

While Fs is less than the maximum specific flow (Fsm), Fs will be used for the

calculation.

Effective width (We) = Actual width – boundary layer width

Effective width of corridor = 1200 – ( 2 x 200 )

= 800mm (0.8m)

Effective width of exit door = 850 – ( 2 x 150 )

= 550mm (0.55m)

60

Calculated flow of corridor:

By equation Fc = Fs x We

= 0.94 x 0.8

= 0.752 persons/s

Since there is a difference of the effective width between corridor and exit door, in

other words, the exit route become narrower. Hence, the situation of transitions need

to be considered in this case to calculate the specific flow after the transition point.

Fs(door) = ( Fs(corridor) x We(corridor) ) / We(door)

= 0.94 x 0.8 / 0.55

= 1.37 persons /s/m of effective width

The calculated specific flow is 1.37 which exceeds 1.3 as the maximum value of

specific flow, hence, 1.3 will be the value used for following calculation rather than

1.37 persons /s/m of effective width.

Calculated flow of exit door:

By equation Fc = Fs x We

= 1.3 x 0.55

= 0.715 persons/s

Although the calculated flow for corridor is 0.752 persons/s (45 persons /min), which

is slightly larger than the calculated flow for exit door as 0.715 persons/s (43

persons/min), a slight congestion may occur is implied from the values, the number of

evacuees in each floor is only 15 persons toward each staircase based on assumption.

61

Hence, the congestion is not likely to occur even the narrower situation exists.

For the evacuation in staircase:

Effective width of staircase = 1100 – (2 x 150)

= 800mm (0.8m)

Fs(staircase) = ( Fs(door) x We(door) ) / We(staircase)

= 1.37 x 0.55 / 0.8

= 0.94 persons /s/m of effective width

By equation Fs = (1 – aD)kD ,

Population density is 1.4 persons/m2 where Fs is 0.94 persons /s/m of effective width,

a is the constant with value of 0.266 and k is 1.08 based on the staircase

configuration.

By equation S = k – akD ,

The movement speed is 0.67 m/s when D is 1.4 persons/m2.

For the floor to floor height of building is 3m, conversion factor is used to transfer the

vertical travel distance into relevant line of travel distance. The conversion factor for

the staircase upon the assumed configuration is 1.66. Besides, there is a landing when

travel a floor, the horizontal distance for the landing is around 2m.

Hence, the floor to floor travel distance is ((3 x 1.66) + 2) = 6.98m.

The time required for travel one floor is 6.98/0.67 = 10.4s.

62

Evacuation flow:

In the staircase, it can carry 10 evacuees each time. However, there are 15 evacuees

from each floor assumed in this case. For the calculated flow of staircase is 0.94 x 0.8

= 0.752 persons/s. Hence, 5 evacuees might need to wait 6.65s for entering the

staircase for evacuation.

The time required to travel from corridor to exit door:

= 14.2 / 1.07

= 13.3 s

At 13.3s, evacuees start passing through the exit door towards staircases. 10 persons

can travel in staircase while 5 persons waiting in queue to enter the staircases area.

After further 10.4s (23.7s), the first evacuee in staircase reach the next floor. At

30.35s, 5 persons reach the next floor, hence, the last people queuing at the staircase

entrance can start evacuate downstairs.

Time (s) Status

30.35 All persons have evacuated the 79/F

43.75 The end of the flow reaches the 77/F

60.8 All persons have evacuated the 77/F

71.2 The end of the flow reaches the 76/F

88.25 All persons have evacuated the 76/F

98.65 The end of the flow reaches the 75/F

115.7 All persons have evacuated the 75/F

126.1 The end of the flow reaches the 73/F

143.15 All persons have evacuated the 73/F

153.55 The end of the flow reaches the 72/F

170.6 All persons have evacuated the 72/F

181 The end of the flow reaches the 71/F

63

198.05 All persons have evacuated the 71/F

208.45 The end of the flow reaches the 70/F

225.5 All persons have evacuated the 70/F

235.9 The end of the flow reaches the 69/F

252.95 All persons have evacuated the 69/F

263.35 The end of the flow reaches the 68/F

280.4 All persons have evacuated the 68/F

290.8 The end of the flow reaches the 67/F

307.85 All persons have evacuated the 67/F

318.25 The end of the flow reaches the 66/F

335.3 All persons have evacuated the 66/F

345.7 The end of the flow reaches the 65/F

362.75 All persons have evacuated the 65/F

373.15 The end of the flow reaches the 63/F

390.2 All persons have evacuated the 63/F

400.6 The end of the flow reaches the 62/F

417.65 All persons have evacuated the 62/F

428.05 The end of the flow reaches the 61/F

445.1 All persons have evacuated the 61/F

455.5 The end of the flow reaches the 60/F

472.55 All persons have evacuated the 60/F

482.95 The end of the flow reaches the 59/F

500 All persons have evacuated the 59/F

510.4 The end of the flow reaches the 58/F

527.45 All persons have evacuated the 58/F

537.85 The end of the flow reaches the 57/F

554.9 All persons have evacuated the 57/F

565.3 The end of the flow reaches the 56/F

582.35 All persons have evacuated the 56/F

592.75 The end of the flow reaches the 55/F

609.8 All persons have evacuated the 55/F

620.2 The end of the flow reaches the 53/F

In total evacuation, the travel time for all occupants to refuge floor is around 620s

(10min).

64

APPENDIX A2 – Empirical Calculation Result for Case 2

Density (D) = Number of persons per unit area

= (12x3)/(14.2x1.2)

= 2.113 person per m2

By equation S = k – akD ,

= 1.4 – ( 0.266 x 1.4 x 2.113 )

= 0.613 m/s

By equation Fs = (1 – aD)kD ,

= (1 – (0.266 x 2.113)) x 1.4 x 2.113

= 1.3 persons/s/m of effective width

While Fs just reach the value of maximum specific flow (Fsm), Fs will be continuously

used for the calculation.

Effective width (We) = Actual width – boundary layer width

Effective width of corridor = 1200 – ( 2 x 200 )

= 800mm (0.8m)

Effective width of exit door = 850 – ( 2 x 150 )

= 550mm (0.55m)

65

Calculated flow of corridor:

By equation Fc = Fs x We

= 1.3 x 0.8

= 1.04 persons/s

Since there is a difference of the effective width between corridor and exit door, in

other words, the exit route become narrower. Hence, the situation of transitions need

to be considered in this case to calculate the specific flow after the transition point.

Fs(door) = ( Fs(corridor) x We(corridor) ) / We(door)

= 1.3 x 0.8 / 0.55

= 1.89 persons /s/m of effective width

The calculated specific flow is 1.89 which exceeds 1.3 as the maximum value of

specific flow, hence, 1.3 will be the value used for following calculation rather than

1.89 persons /s/m of effective width.

Calculated flow of exit door:

By equation Fc = Fs x We

= 1.3 x 0.55

= 0.715 persons/s

Although the calculated flow for corridor is 1.04 persons/s (62 persons /min), which is

slightly larger than the calculated flow for exit door as 0.715 persons/s (43

persons/min). From the above values, it implies that 19 occupants are required to

queue up near the exit door entrance for evacuation. However, the average number of

66

evacuees of each floor is assumed to be 36 occupants for each staircase, which below

the maximum flow capacity of both corridor and exit door. Hence, queuing will not

happen due to the narrowing of the evacuation route.

For the evacuation in staircase:

Effective width of staircase = 1100 – (2 x 150)

= 800mm (0.8m)

Fs(staircase) = ( Fs(door) x We(door) ) / We(staircase)

= 1.89 x 0.55 / 0.8

= 1.3 persons /s/m of effective width

Based on the SFPE handbook, stair riser and treads with 7.0 inch wide and 10 inch

high respectively can provide the maximum specific flow of 1.01 persons/s/m of

effective width (Fsm). While the specific flow calculated from above is 1.3 persons

/s/m of effective width, which exceed the limitation of the staircase, hence, the

maximum specific flow of 1.01 persons/s/m of effective width will be used for the

calculation.

By equation Fs = (1 – aD)kD ,

Population density is 1.8 persons/m2 where Fs is 1.01 persons /s/m of effective width,

a is the constant with value of 0.266 and k is 1.08 based on the staircase

configuration.

By equation S = k – akD ,

The movement speed is 0.52 m/s when D is 1.8 persons/m2.

67

For the floor to floor height of building is 3m, conversion factor is used to transfer the

vertical travel distance into relevant line of travel distance. The conversion factor for

the staircase upon the assumed configuration is 1.66. Besides, there is a landing when

travel a floor, the horizontal distance for the landing is around 2m.

Hence, the floor to floor travel distance is ((3 x 1.66) + 2) = 6.98m.

The time required for travel one floor is 6.98/0.52 = 13.4s.

Evacuation flow:

Based on the density and flow of staircase, each of it can carry 13 evacuees. In this

case, 36 evacuees is assumed to evacuate towards each staircase. While the calculated

flow of staircase is 1.01 x 0.8 = 0.808 persons/s, it means that the 36th evacuee may

need to wait around 28.5s for entering the staircase.

The time required to travel from corridor to exit door:

= 14.2 / 0.613

= 23.2 s

At 23.2s, evacuees start passing through the exit door towards staircases. 10 persons

can travel in staircase while 26 persons waiting in queue to enter the staircases area.

At further 13.4s (36.6s), the first evacuee in staircase reach the next floor. At 65.1s,

the last person queuing at staircase entrance can start evacuate downstairs.

68

Time (s) Status

65.07 All persons have evacuated the 79/F

78.47 The end of the flow reaches the 77/F

120.33 All persons have evacuated the 77/F

133.73 The end of the flow reaches the 76/F

175.60 All persons have evacuated the 76/F

189.00 The end of the flow reaches the 75/F

230.86 All persons have evacuated the 75/F

244.26 The end of the flow reaches the 73/F

286.13 All persons have evacuated the 73/F

299.53 The end of the flow reaches the 72/F

341.39 All persons have evacuated the 72/F

354.79 The end of the flow reaches the 71/F

396.66 All persons have evacuated the 71/F

410.06 The end of the flow reaches the 70/F

451.92 All persons have evacuated the 70/F

465.32 The end of the flow reaches the 69/F

507.19 All persons have evacuated the 69/F

520.59 The end of the flow reaches the 68/F

562.45 All persons have evacuated the 68/F

575.85 The end of the flow reaches the 67/F

617.72 All persons have evacuated the 67/F

631.12 The end of the flow reaches the 66/F

672.98 All persons have evacuated the 66/F

686.38 The end of the flow reaches the 65/F

728.25 All persons have evacuated the 65/F

741.65 The end of the flow reaches the 63/F

783.51 All persons have evacuated the 63/F

796.91 The end of the flow reaches the 62/F

838.78 All persons have evacuated the 62/F

852.18 The end of the flow reaches the 61/F

894.05 All persons have evacuated the 61/F

907.45 The end of the flow reaches the 60/F

949.31 All persons have evacuated the 60/F

962.71 The end of the flow reaches the 59/F

1004.58 All persons have evacuated the 59/F

1017.98 The end of the flow reaches the 58/F

69

1059.84 All persons have evacuated the 58/F

1073.24 The end of the flow reaches the 57/F

1115.11 All persons have evacuated the 57/F

1128.51 The end of the flow reaches the 56/F

1170.37 All persons have evacuated the 56/F

1183.77 The end of the flow reaches the 55/F

1225.64 All persons have evacuated the 55/F

1239.04 The end of the flow reaches the 53/F

In total evacuation, the travel time is around 1239s (20.6 min).

70

APPENDIX B1 – SIMULEX RESULT OF CASE 1 (Excerpt)

Number of Floors = 24

Number of Staircases = 2

Number of Exits = 2

Number of Links = 48

Number of People = 690

-----------------------------------------------

All people reached the exit in 13:51.9.

Number of people through all exits over 5-second periods

Time(s) N (People) Time(s) N (People)

5 0

10 0

15 0

20 0

25 0

30 0

35 0

40 3

45 0

50 3

55 0

60 4

65 6

70 5

75 5

80 6

85 6

90 5

95 3

100 4

105 5

110 6

115 4

120 5

125 2

130 3

135 5

140 5

145 7

150 6

155 6

160 5

165 5

170 7

175 4

180 4

185 4

190 7

195 7

200 3

205 3

210 5

215 7

220 5

225 6

230 8

235 4

240 4

245 3

250 5

71

Time(s) N (People)

255 7

260 5

265 8

270 4

275 5

280 6

285 7

290 5

295 5

300 6

305 4

310 5

315 5

320 2

325 3

330 4

335 6

340 6

345 7

350 4

355 5

360 5

365 6

370 4

375 5

380 2

385 4

390 5

395 5

400 6

405 5

410 4

415 4

420 3

425 4

430 5

435 6

Time(s) N (People)

440 6

445 6

450 4

455 4

460 7

465 4

470 6

475 5

480 7

485 4

490 6

495 5

500 6

505 5

510 3

515 6

520 6

525 7

530 4

535 6

540 5

545 5

550 3

555 7

560 5

565 4

570 6

575 5

580 5

585 5

590 3

595 7

600 5

605 6

610 7

615 4

620 2

72

Time(s) N (People)

625 7

630 7

635 5

640 4

645 3

650 5

655 3

660 6

665 5

670 5

675 2

680 3

685 3

690 2

695 2

700 0

705 1

710 3

715 3

720 2

725 3

730 3

Time(s) N (People)

735 2

740 3

745 0

750 0

755 0

760 0

765 1

770 4

775 3

780 2

785 3

790 3

795 2

800 0

805 3

810 1

815 3

820 3

825 2

830 2

835 2

73

APPENDIX B2 – SIMULEX RESULT OF CASE 2

Number of Floors = 24

Number of Staircases = 2

Number of Exits = 2

Number of Links = 48

Number of People = 1656

-----------------------------------------------

All people reached the exit in 25:48.2.

Number of people through all exits over 5-second periods

Time(s) N (People)

5 0

10 0

15 0

20 0

25 0

30 0

35 0

40 2

45 3

50 2

55 2

60 3

65 7

70 6

75 6

80 6

85 6

90 4

95 8

100 7

105 6

110 6

115 5

120 6

125 7

Time(s) N (People)

130 6

135 5

140 6

145 5

150 5

155 7

160 7

165 6

170 5

175 4

180 7

185 6

190 5

195 6

200 6

205 4

210 7

215 7

220 7

225 6

230 5

235 6

240 8

245 4

250 7

74

Time(s) N (People)

255 5

260 7

265 6

270 5

275 8

280 5

285 5

290 6

295 4

300 4

305 8

310 7

315 5

320 7

325 3

330 6

335 6

340 4

345 7

350 7

355 5

360 7

365 4

370 8

375 5

380 5

385 6

390 7

395 7

400 6

405 7

410 3

415 4

420 7

425 6

430 6

435 4

Time(s) N (People)

440 7

445 7

450 5

455 7

460 7

465 8

470 4

475 5

480 6

485 6

490 7

495 3

500 7

505 5

510 7

515 6

520 5

525 7

530 5

535 5

540 5

545 7

550 7

555 2

560 8

565 7

570 2

575 8

580 2

585 7

590 6

595 2

600 9

605 6

610 5

615 6

620 7

75

Time(s) N (People)

625 5

630 6

635 6

640 4

645 7

650 6

655 6

660 4

665 6

670 7

675 5

680 4

685 3

690 6

695 6

700 8

705 6

710 6

715 6

720 7

725 4

730 5

735 4

740 7

745 8

750 2

755 7

760 5

765 4

770 2

775 7

780 5

785 5

790 3

795 6

800 5

805 8

Time(s) N (People)

810 6

815 6

820 5

825 7

830 6

835 6

840 5

845 3

850 7

855 5

860 7

865 6

870 7

875 6

880 6

885 5

890 3

895 9

900 7

905 6

910 5

915 4

920 5

925 8

930 7

935 5

940 7

945 5

950 6

955 5

960 7

965 4

970 2

975 8

980 6

985 3

990 4

76

Time(s) N (People)

995 6

1000 4

1005 2

1010 3

1015 3

1020 3

1025 6

1030 8

1035 4

1040 2

1045 9

1050 7

1055 6

1060 5

1065 6

1070 7

1075 5

1080 6

1085 6

1090 6

1095 4

1100 6

1105 6

1110 5

1115 5

1120 6

1125 6

1130 6

1135 7

1140 6

1145 5

1150 4

1155 7

1160 5

1165 5

1170 8

1175 5

Time(s) N (People)

1180 5

1185 5

1190 6

1195 6

1200 4

1205 6

1210 4

1215 4

1220 8

1225 4

1230 8

1235 4

1240 6

1245 5

1250 6

1255 6

1260 7

1265 5

1270 4

1275 8

1280 7

1285 4

1290 7

1295 5

1300 7

1305 3

1310 5

1315 6

1320 6

1325 3

1330 5

1335 3

1340 4

1345 9

1350 4

1355 5

1360 5

77

Time(s) N (People)

1365 7

1370 3

1375 7

1380 7

1385 5

1390 5

1395 3

1400 6

1405 7

1410 5

1415 4

1420 3

1425 2

1430 5

1435 6

1440 7

1445 7

1450 4

1455 4

1460 6

1465 7

1470 5

1475 6

1480 5

1485 6

1490 7

1495 2

1500 8

1505 5

1510 6

1515 4

1520 4

1525 2

1530 3

1535 3

1540 2

1545 3

Time(s) N (People)

1550 2

78

APPENDIX B3 – SIMULEX RESULT OF CASE 3

Number of Floors = 24

Number of Staircases = 2

Number of Exits = 2

Number of Links = 48

Number of People = 120

-----------------------------------------------

All people reached the exit in 11:19.5.

Number of people through all exits over 5-second periods

Time(s) N (People) Time(s) N (People)

5 0

10 0

15 0

20 0

25 0

30 0

35 0

40 0

45 0

50 0

55 0

60 0

65 0

70 0

75 0

80 0

85 0

90 0

95 0

100 0

105 0

110 0

115 0

120 0

125 0

130 0

135 0

140 0

145 0

150 0

155 0

160 0

165 0

170 0

175 0

180 0

185 0

190 0

195 0

200 0

205 0

210 0

215 0

220 0

225 0

230 0

235 0

240 1

245 0

250 0

79

Time(s) N (People)

255 0

260 0

265 1

270 0

275 0

280 0

285 0

290 0

295 0

300 0

305 2

310 4

315 8

320 3

325 1

330 0

335 0

340 0

345 0

350 0

355 0

360 0

365 3

370 2

375 1

380 6

385 6

390 4

395 3

400 0

405 0

410 2

415 3

420 3

425 3

430 3

435 2

Time(s) N (People)

440 0

445 0

450 0

455 0

460 0

465 0

470 0

475 0

480 0

485 0

490 0

495 0

500 0

505 0

510 0

515 0

520 0

525 0

530 0

535 1

540 3

545 0

550 3

555 4

560 5

565 4

570 7

575 3

580 5

585 3

590 2

595 4

600 3

605 2

610 3

615 2

620 2

80

Time(s) N (People)

625 1

630 0

635 0

640 0

645 0

650 0

655 0

660 0

665 0

670 0

675 0

680 2

81

APPENDIX B4 – SIMULEX RESULT OF CASE 4

Number of Floors = 24

Number of Staircases = 2

Number of Exits = 2

Number of Links = 48

Number of People = 288

-----------------------------------------------

All people reached the exit in 11:52.9.

Number of people through all exits over 5-second periods

Time(s) N (People) Time(s) N (People)

5 0

10 0

15 0

20 0

25 0

30 0

35 0

40 0

45 0

50 0

55 0

60 0

65 0

70 0

75 0

80 0

85 0

90 0

95 0

100 0

105 0

110 0

115 0

120 0

125 0

130 0

135 0

140 0

145 0

150 0

155 0

160 0

165 0

170 0

175 0

180 0

185 0

190 0

195 0

200 0

205 0

210 1

215 1

220 4

225 0

230 0

235 0

240 0

245 0

250 0

82

Time(s) N (People)

255 0

260 0

265 0

270 0

275 0

280 0

285 0

290 0

295 0

300 5

305 0

310 0

315 0

320 0

325 0

330 0

335 0

340 0

345 0

350 0

355 4

360 0

365 0

370 0

375 4

380 3

385 3

390 6

395 6

400 7

405 5

410 0

415 0

420 0

425 0

430 0

435 0

Time(s) N (People)

440 0

445 0

450 1

455 3

460 4

465 3

470 1

475 0

480 0

485 0

490 3

495 5

500 7

505 7

510 4

515 3

520 2

525 7

530 7

535 5

540 8

545 3

550 7

555 8

560 6

565 3

570 2

575 3

580 3

585 4

590 2

595 8

600 4

605 2

610 6

615 6

620 7

83

Time(s) N (People)

625 5

630 6

635 6

640 4

645 6

650 5

655 5

660 8

665 6

670 7

675 5

680 5

685 6

690 7

695 3

700 1

705 3

710 4

715 1

84