Housing Evacuation of Mixed Abilities Occupants in Highrise Buildings by dffhrtcv3

VIEWS: 9 PAGES: 101

									Ser
 TH1      National Research       Conseil national
  ~427    Council Canada          de recherches Canada
no. 706
          Institute for           lnstitut de
          Research in             recherche en
          Construction            construction




          IUC-CIWC
          Housing Evacuation of Mixed Abilities
          Occupants in Highrise Buildings




          by Guylene Proulx, John C. Latour, John W. McLaurin, Joelle Pineau,
          Lana E. Hoffman and Chantal Laroche




          Internal Report No. 706


          Date of issue: August 1995




          This is an internal report of the lnstitute for Research in Construction. Although not
          intended for general distribution, it may be cited as a reference in other publications.
                                           TABLE OF CONTENTS

                                                                                                                          ...
LIST OF FIGURES ........................................................................................................ m
                                                                                                                          ...
LIST OF TABLES ...................................................................................................   .....
                                                                                                                     :

ACKNOWLEDGEMENTS .............................................................................................
                                                                                                             v
EXECUTIVE SUMMARY ............................................................................................ vi
1.0 INTRODUCTION .................................................................................................... 1
2.0 OBJECTIVES .......................................................................................................... 2
3.0 GENERAL METHODOLOGY ................................................................................ 3
4.0 METHOD AND RESULTS FOR EACH BUILDING .............................................. 7
          4.1. Building A .Montreal................................................................................... 7
                 Occupants................................................................................................. 9
                 Procedure ................................................................................................. 9
                 Results .Building A .................................................................................. 9
                        Behaviour of Occupants .............................................................. 10
                        Occupants' Use of the Staircases and Exits.................................. 12
                        Evacuation Timing .................................................................... 13
                        Time Com~arisons             According to Gender ..................................... 15
                                                                   .
                        Time Comparisons According to Age Categories......................... 15
                        Time Comparisons According to Limitation ................................ 17
                                                                   .
                        Speed of Occupants on the Stairs ................................................ 17
                 Questionnaires ................................................................................... 19
                 Alarm Audibility Measurement................................................................ 20
                 Conclusions ..................................................................................... 22
          4.2 Building B .Calgary .................................................................................... 24
                Occupants.......................................................................................... 25
                Procedure ........................................................................................ 25
                            . .
                Results . Bulldlng B ................................................................................ 26
                        Behaviour of Occupants .............................................................. 27
                        Evacuation Timing ...................................................................... 28
                        Time Comparisons According to Gender ..................................... 31
                        Time Comparisons According to Age .......................................... 31
                        =me Comparisons According to Limitation ................................ 33
                        Speed of Occupants on the Stairs ................................................ 33
                Questionnaires ........................................................................................ 35
                Alarm Audibility Measurement................................................................ 36
                Conclusions ....................................................................................... 39
          4.2 Building C .Gloucester ............................................................................... 40
                Occupants............................................................................................... 42
                Procedure ............................................................................................... 42
                                     .   .
                    Results .Building C ................................................................................ 42
                           Behaviour of Occupants ..............................................................43
                           Evacuation Timing ......................................................................         44
                           Time Comparisons According to Gender ..................................... 47
                           Time Comparisons According to Age ..........................................48
                           Time Comparisons According to Limitation ................................49
                           Speed of Occupants on the Stairs ................................................ 49
                    Questionnaires ........................................................................................ 51
                    A a m Audibility Measurement................................................................ 52
                     lr
                    Conclusions ........................................................................................ 55
5.0 JOINT RESULTS .ALL THREE HIGHRISE BUILDINGS .................................. 56
          5.1 Comparisons of Mean Times to Start. Exit and Move .................................. 57
                       Time to Start............................................................................... 57
                       Time to Exit ............................................................................. 58
                       Time to Move ........................................................................... 58
          5.2 Speed Comparisons According to Building .................................................. 59
          5.3 Time and Speed Comparisons Regrouping the Three Buildings.................... 59
                       Time and Speed Comparisons According to Gender .................... 59
                       Time and Speed Comparisons According to Age .........................60
                       Time and Speed Comparisons According to Limitation................ 61
                       Composition of Evacuation Groups ............................................. 63
                       A a m Audibility.......................................................................... 65
                        lr
6.0 JOINT RESULTS: HIGHRISE AND MIDRISE BUILDINGS.............................. 66
7.0 GENERAL CONCLUSIONS .................................................................................                 68
8.0 RECOMMENDATIONS ........................................................................................              71
9.0 REFERENCES ...................................................................................................        73
ANNEX 1: MEMO DISTRIBUTED TO OCCUPANTS .............................................. 75
ANNEX 2: POST-EVACUATION QUESTIONNAIRE...............................................                                     79
ANNEX 3: GUIDE TO MANAGERS ..........................................................................                     85
                                           LIST OF FIGURES

Figure 1: Floor Plan .Montreal Building....................................................................... 7
Figure 2: Elevation .Montreal Building .....................................................................8
Figure 4: Audibility Design Window for a 55 year old man
           with mild to severe hearing loss . Bedroom of Unit 1519 ............................. 21
Figure 5: Audibility Design Window for a 55 year old man
           with normal hearing . A-side Corridor. 15th floor ........................................ 22
Figure 6: Typical Floor Plan . Calgary Building .............................................. ........... 24
Figure 7: Elevation . Calgary Building ........................................................................ 25
Figure 8: Audibility Design Window for a 55 year old man
           with normal hearing . Bedroom of Unit 508 ................................................. 37
Figure 9: Audibility Design Window for a 55 year old man
           with normal hearing . A-side Conidor, 5th floor .......................................... 38
Figure 10: Audibility Design Window for a 55 year old man
           with normal hearing . A-side Staircase, 5th floor .......................................... 38
Figure 11: Floor Plan .Gloucester Building ................................................................. 40
Figure 12: Elevation .Gloucester Building ................................................................... 41
Figure 13: Audibility Design Window for a 55 year old man
           with normal hearing . Bedroom of Unit 819 ................................................. 54
Figure 14: Audibility Design Window for a 55 year old man
           with mild to severe hearing loss .Corridor in Front of Unit 1219 ................. 55


                                            LIST OF TABLES

Table 1:      Occupants Appearing in Corridor but not Evacuating.................................. 10
Table 2:      Occupants Exiting onto their Balconies ........................................................ 11
Table 3:      Occupants Egress Route According to their Apartment Location ................. 12
Table 4:      Evacuation Timing at 1 min Intervals .Full Duration of the Alarm ............... 13
Table 5:      Evacuation Timing at 15 s Intervals .First 5:00 min of the Alarm Only ........14
Table 6:      Time Comparisons According to Gender ..................................................... 15
Table 7:      Number of Occupants in Each Age Category .............................................. 16
Table 8:      Time Comparisons According to Age .......................................................... 16
Table 9:      Time Comparisons According to Limitation................................................. 17
Table 10:     Gender Comparisons wh11e Descending One Floor and Speed in Stairs......... 17
Table 11:     Speed in Stairs of Occupants According to Age Category............................ 18
Table 12:     Speed in Stairs Comparisons According to Age Category ............................ 18
Table 13:     Speed on Stairs Comparisons According to Limitation ................................. 18
Table 14:     Speed on Stairs Comparisons According to Groups ..................................... 19
Table 15:     Background Noise Levels and Detectsound Analysis of Alarm Audibility .....20
Table 16:     Occupants Exiting onto their Balconies ...................................................... 27
Table 17:     Occupants' Egress Route According to their Apartment Location................ 28
Table 18:     Evacuation Timing at 1min Intervals .Full Duration of the Alarm ...............29
Table 19:     Evacuation Timing at 15 s Intervals .First 5:00 min ..................................... 30
Table 20:    Time to Move vs. Starting Floor .................................................................. 31
Table 21:    Time Comparisons According to Gender ..................................................... 31
Table 22:    Number of Occupants in Each Age Category ............................................... 32
Table 23:    Time Comparisons According to Age - Building B ....................................... 32
Table 24:    %me Comparisons According to Limitation ................................................. 33
Table 25:    Gender Comparisons while Descending One Floor and Speed in Stairs......... 34
Table 26:    Speed of Occupants According to Age Category ......................................... 34
Table 27:    Speed on Stairs Comparisons According to Age Category ........................... 34
Table 28:    Speed on Stairs Comparisons According to Limitation................................. 35
Table 29:    Background Noise Levels and Detectsound Analysis of Alarm Audibility .....36
Table 30:    Occupants' Egress Route According to their Apartment Location................ 44
Table 31 :   Evacuation Timing at 1 min Intervals - Full Duration of the Alarm ...............45
Table 32:    Evacuation Timing at 15 s Intervals - Fist 5:00 min ..................................... 46
Table 33:    Time to Move vs. Starfing Floor .................................................................. 47
Table 34:    Time Comparisons According to Gender ..................................................... 47
Table 35:    Number of Occupants in Each Age Category ............................................... 48
Table 36:    Time Comparisons According to Age - Building C ....................................... 48
Table 37:    Time Comparisons According to Limitation ................................................. 49
Table 38:    Gender Comparisons while Descending One Floor and Speed in Stairs......... 50
Table 39:    Speed of Occupants According to Age Category ......................................... 50
Table 40:    Speed on Stairs Comparisons According to Age Category ........................... 50
Table 4 1:   Speed on Stairs Comparisons According to Limitation ................................. 51
Table 42:    Activities Undertaken After Hearing the Alarm ............................................ 52
Table 43:    Background Noise Levels and Detectsound Analysis of Alarm AudibiLity .....53
Table 44:    Mean T i e s to Start. Exit and Move - All Buildings .................................... 57
Table 45:    Comparison of Mean Times to Start. Exit and Move - All Buildings............. 57
Table 46:    Comparison of Mean Time Exit - Full Duration of the Alarm - Calgary and
             Gloucester Buildings Combined ................................................................... 60
Table 47:    Mean Speeds for each Age Group - Al Buildings Combined ........................ 61
                                                           l
Table 48:    Comparison of Mean Speeds for each Age Group - All Buildings Combined 61
Table 49:    Time Comparison According to Limitation - Full Duration of the Alarm ......62
Table 50:    Time Comparison According to Limitation -
             Occupants with Times to Start S 5:M) min ................................................... 62
Table 5 1:   Distribution of Evacuees - All Buildings....................................................... 63
Table 52:    Composition of Evacuation Groups - All Buildings ...................................... 64
Table 53:    Mean Times to Evacuate - Occupants Alone vs. Occupants in Groups -
             Al Three Buildings ...................................................................................... 65
              l
Table 54:    Mean Times to Start, Exit and Move - All Buildings .................................... 66
ACKNOWLEDGEMENTS

         Many people participated in this research project. We would f i s t like to thank the
occupants of the buildings studied. Most of them participated willingly in the evacuation
drill, not only providing us with interesting data but also useful insights, either directly or
in answering the post-evacuation questionnaire. We are also grateful to the managers and
the boards of directors of the three buildings; they were patient and helpful in providing
information on their buildings and their occupants, as well as allowing us full access to the
premises to install the equipment

        The participation of the firefighters was essential in ensuring the success of this
study. We are grateful to all the firefighters who willingly took part in the drills; they did a
great job:

Montreal Fr Protection Service (Service de la prbvention des incendies):
          ie
      Surintendant Ronald Dubeau, Chef de section Jean Renaud, Benoit L'Ecuyer

Calgary Fire Department:
       District Chief D. Dickenson, Division Chief E. Prince, Supervisor of Inspectors at
       the Prevention Bureau Sandy Mackenzie, Amateur Television Unit

Gloucester F i e Department:
      Captain Jim Atwill, " B Platoon, Station #1

       Many people at the National Research Council of Canada helped in installing the
equipment, distributing questionnaires, acting as observers or editing the manuscript. We
would especially like to thank Roch Monette and Mike Denham.

     This work was jointly funded by the National Research Council of Canada
(NRCC) and the Canada Mortgage and Housing Corporation (CMHC).
EXECUTIVE SUMMARY

        A joint research project was undertaken by the National Research Council of
Canada (NRCC) and Canada Mortgage and Housing Corporation (CMHC) to study
evacuation drills in highrise apartment buildings with mixed abilities occupants. A similar
project had been carried out in 1993-94 on midrise apartment buildings. This second
study collected real data on time and movement during evacuation drills in three highrise
buildings. The buildings were located in Montreal, Calgary and Gloucester, Ont. These
different locations reflected provincial variations in fue safety procedures. Each f i e drill
was organized, supervised and carried out with the full participation of local fire
departments or local fire protection services.

        Before evacuation drills in two of the three buildings, NRCC staff met with the
                                                             - .
residents identified as having some limitations to discuss their knowledge of the building
evacuation procedures and their ability, or inability, to participate in the drill. The initial
goal of those meetings was to better prepare for the drii and to inform the fire department
of problems that could occur during the exercise. These meetings, however, became
information sessions to reassure occupants about the drills. Many older residents and
people with disabilities wondered if the firefighters would cany them down the stairs or
down a ladder. Those occupants then received further explanation a s to how the
firefighters would perform the horizontal evacuation. One week before each drill, all
occupants were informed in writing of the evacuation drill. The message, distributed to
each door, did not specify the exact day or time of the evacuation drill, however, it
highlighted the importance of the exercise and described the participation of NRCC,
CMHC and the local fire department. It specified that researchers would observe the drill
and that video cameras, positioned in hallways and staircases, would record the drill.

        The Montreal and Calgary drills took place on weekdays between 6:30 P.M.and
7:00 P.M. in order to have the largest possible number of occupants at home, yet before
bedtime for most people. These two evacuation drills were carried out during the summer
and early fall of 1994; for each evacuation, the weather was sunny and warm. The
Gloucester drill was held on a Saturday morning in December 1994, between 10:30 a.m.
and 11:OO a.m. This drill allowed the study of occupants' evacuation behaviour under
winter conditions. During each drill, data were gathered concerning the following: the
time to respond to the alarm, the direction of movement, the time for the occupants
(including disabled persons and elderly) to reach a safe area, the overall evacuation
behaviour of disabled persons as well as other occupants, and the time to evacuate the
entire building. Video cameras recorded the location, time and frequency of movements,
providing data used in various statistical calculations. During the evacuation drill, the
research team remained outside the building in such a way as to not interfere with either
the building occupants or the fuefighters. In addition, data on the audibility of the building
  m
f alarm, at different locations in the building, was recorded during each drill.

        Firefighters arriving at the scene followed their usual procedures and participated
in the evacuation of occupants. People with limited mobility were moved by fuefighters
                                             vii

horizontally to their balconies or to ground level using the elevator. Mobile occupants,
still in the building, were asked to leave by the nearest exit. When the firefightersjudged
that the situation was satisfactory, the Captain gave the "all clear", the alarms were reset
and occupants went back to their apartments.

       After the evacuation drills, a questionnaire was distributed to each apartment
Occupants were asked if they had heard the fire alarm and the P.A. messages; intewiewees
were also asked to list the chronology of their actions and describe their evacuation
movements.

        A series of statistical analyses was conducted on the data gathered using the video
cameras. The times to start the evacuation, to move to safety and to totally evacuate the
building were studied. An analysis of variance was used to compare buildings and to
assess the impact of gender, age and physical limitation. The speed of movement on stairs
was also analyzed.

        The results show that the time to start the evacuation is dependent on the ability of
occupants to hear the alarm and on the required preparation time to evacuate. When
evacuating in winter, the time to start was much longer, since occupants were getting
dressed before leaving their apartments. The time to move to safety appeared similar for
buildings of comparable architectural design, while the overall evacuation time was
dependent on the time at which occupants started to evacuate. In the buildings studied,
most occupants reported in the questionnaire hearing the fire alarm in their apartments,
even though alarm audibilitv measurements showed that alarm sound levels were
inadequate in most apartments. Questionnaires showed that the time delay before leaving
their apartments was used by the occupants to get dressed, gather valuables, find children
or find pets. In buildings where a longer time to start was observed, there was,
consequently, a longer time to totally evacuate the buildings.

        The characteristics of gender, age and limitations presented little impact on the
timing and movement during evacuations. Some older occupants tended t move slower
                                                                             o
on starts than other adults. Such occupants, however, did not impede the evacuation of
others since the hallways and staircases were never crowded and faster occupants were
able to overtake slower ones. Occupants with serious mobility limitations also did not
impede the evacuation of others since they usually stayed in their apartments to be rescued
by firefighters.

        Occupants in all three buildings showed a comparable mean speed on the stairs,
travelling, on average, at 0.95 m/s, 1.05 m/s and 1.07 m/s on non-crowded stairs.
Gloucester occupants were slightly slower than the others, which could be attributed to
the fact that they were wearing winter clothing. Gender did not play a role in the
differences in speed while evacuating the buildings. Those over 65 travelled significantly
slower than younger people during the evacuations. Children aged 2 to 5 were also
significantly slower; going down the stairs one step at a time, holding the handrail at a
height over their shoulders.
                                            viii

        Many occupants tended to evacuate in groups. The majority of these groups were
couples. Children, for the most part, evacuated in groups which included an adult. These
group formations likely delayed the speed of movement of the whole group because
members tended to assume the speed of the slowest person. In Calgary, however,
occupants travelling in groups had faster speeds. Occupants tended to use the staircase
closest to their apartments. None of the staircases were crowded during the drills,
however, in each building, occupants travelled slightly faster in the staircase with the
lowest density.

        Recommendations are made regarding the audibility of fue alarms. Such alarms
should be tested to make sure that everv occupant can hear the alarm from everv area of
the building. Compensatory alarm sys&ms should be provided for occupants with hearing
limitations. All occupants with some limitation should be informed of the emergency
                                                                                 -    .
procedures and whatis expected of them in case of a fire emergency. Building
management should be familiar with the functioning of the alarm, and should ensure that
the alarm system configuration respects regulations. Training should be provided through
annual evacuations drills, to ensure that every occupant recognizes the fire alarm, knows
the different means or egress, and understands the general fire evacuation procedures.
These measures could improve fire safety for residents of highrise apartment buildings
with mixed abilities occupants.

        The results from the highrise building evacuations were similar to results obtained
for the four midrise apartment buildings. All evacuation start times were not significantly
different, except for two of the midrise buildings where the alarm was inaudible, and for
the highrise evacuated during the winter. Times to Move, as can be expected, were
shorter for the midrise buildings, even though occupants' speeds in the midrise were
slower than in the highrise. This can possibly be explained by the fact that occupants of
the midrise appeared to entertain closer relationships with their neighbours and were much
more likely to stop or slow down to speak to other occupants. This is believed to have
caused them to travel significantly slower than highrise building occupants.
1.0 INTRODUCTION

       The 1991 National Census of Population in Canada showed that 4.2 million
Canadians reported some level of disability [I]. This number represents 15.5% of the total
population. Among all people with disabilities, 93.7% are living in private households
while 6.3% are living in special institutions. Within the group of people with disabilities
between the ages of 15 and 64,97.9% are living in private households.

       Canada is also experiencing an increase in the number of people over theage of
65. The proportion of such elderly people has risen from 10.7% of the total population in
1986 to 11.8% in 1992. Over one-third of these elderly people are living in private
households.

        Clearly most disabled people and many people over the age of 65 are living in
typical residential buildings in Canada. Consequently, it could be expected that
approximately 20% of the residents in regular housing are people with some sort of
limitation on movement, perception or cognitive capacity.

       In the last decade, there has been great concern with regard to providing
accessibility for most people to a large range of buildings. The housing sector was one
area where the need for accessibility was acknowledged and where a serious attempt to
provide accessibility has been observed. Several agencies in Canada, such as the Canada
Mortgage and Housing Corporation (CMHC), have strongly encouraged the creation of
accessible housing throughout the country. More and more apartment blocks now contain
accessible entrances or ramps. Most buildings have accessible elevators and, occasionally,
a few units are specially designed to accommodate occupants with disabilities.

        Due to economics and the growing number of people with limitations, it would be
unrealistic, in the near future, to expect the construction of only new specialized buildings
where people with disabilities can reside. Furthermore, most people with disabilities and
the elderly are determined to stay autonomous and live as long as possible in standard
types of housing. It is not surprising, therefore, that more and more housing projects
contain occupants with mixed abilities. Mixed abilities occupancy implies that residents
are a mix of adults, children, family groups, elderly people and people presenting some
level of disability or limitation.

         While accessibility is offered in more housing projects, the question of egressibity
is now becoming a subject of concern. Egressibility can be defined as the possibility of
leaving a building, or reaching an area of safety, in case of an emergency such as a fr.
                                                                                       ie
The egressibility concept does not mean that every occupant should egress in the same
fashion or through the same route; rather, it intends to promote equivalent opportunity of
life safety for everyone [2].

       At the National Research Council of Canada (NRCC), researchers in the National
Fire Laboratory of the Institute for Research in Construction (IRC) are developing a F i e
Risk Evaluation and Cost Assessment Model (FiRECAM"), which is a tool to assess the
relative fire safety of building designs for all people. To account for possible increases in
evacuation times as a result of the presence of people with disabilities and elderly people,
data are needed on such users' evacuation timing and movement It has previously been
assumed that the presence of people with mixed abilities will increase egress time,
however, no quantitative data exist to determine the overall change in fue safety resulting
from evacuations involving such user groups.

        The increasing tendency to fmd occupants with mixed abilities in standard housing
raises new issues regarding life safety. This is the reason that NRCC and CMHC
undertook this joint research project to study evacuation drills in highrise apartment
buildings with mixed abilities occupants.


2.0 OBJECTIVES

        The general objective of this project was to collect real data on evacuation times
and the movement of occupants in highrise apartment buildings. Furthermore, it was
important to observe the impact of occupants with mixed abilities on the overall building
evacuation. This study will also lead to the development of guidelines for improving fire
safety procedures in apartment buildings.

       It is important to understand that the goal of the research was not only to study the
evacuation of disabled persons, but to gather data on the time and the manner in which
occupants with mixed abilities left their apartment buildings. The research, thus, aims at
improving fue safety for all occupants of a building.

        This is the second joint research project between NRCC and CMHC to study
housing evacuations with mixed abilities occupants. A first project, canied out in 1993-
94, studied evacuation M l s in four midrise apartment buildings. The results of this first
study can be found in several reports and papers [3,4,5,6]. This first study led to a
number of findings which needed to be compared to results from the evacuation of
highrise buildings to determine if these results could be generalized to apartment buildings
of larger dimensions. The fust study suggested that the building fue a l m is not always
heard inside apartments [S,61. Another study also suggested that many older occupants
cannot hear f i e alarms [7]. Consequently, measurements of alarm audibility were taken in
the highrise buildings. It was also decided that one of the evacuation drills should be
carried out during the winter to study the effects of winter conditions on the egressibility
and evacuation times of occupants.

        The results of highrise building evacuations will be compared to those of midrise
buildings to identify differences in the timing and in the occupants' behaviours. The
benefits of the study for the occupants, and for society in general, are the identification of
problems associated with evacuations to help develop design and architectural solutions
that increase safety for all occupants. More realistic expectations regarding, for example,
the time required for the evacuation of such buildings, may enable the f i e departments to
be better prepared and potentially to save lives.

       Findings are used to draft f i e safety guidelines for managers and occupants of
apartment buildings which are presented in Annex 3 of this report. These guidelines
provide general information that can be tailored to specific buildings according to their
occupants and fire safety features.

        Furthermore, the results of this research will no doubt be included, alongwith
results of other projects, in updates to f i e safety regulations and in the National Building
Code of Canada. The results will also be used in the computerized F i e Risk-Cost
Assessment Model (FiRECAM'), currently being developed by NRCC's National Fire
Laboratory.


3.0 GENERAL METHODOLOGY

        The research involved collecting real data on the evacuation times and movement
of occupants in three highrise apartment buildings identified by CMHC. The buildings met
several requirements. Fist, the buildings were over seven storeys in height, qualifying
them as highrise apartment buildings. Secondly, the buildings were of mixed occupancy
meaning that there were adults, children, seniors and people with limitations living in the
apartments. As well, the buildings chosen were located in different cities to reflect local
fire department variations in their evacuation procedures. Researchers at NRCC discussed
the study with the building management in each case and obtained their permission to
observe an evacuation drill.

       Information, such as the type of alarm system, the location of the alarms, staircase
and corridor measurements, the number of exits and demographic information about the
occupants, was obtained on-site.

        In order to prepare occupants for the drill, a brief memo was sent to each
apartment one week before the exercise took place. The memo provided information
about the purpose of the research, the presence of video cameras during the drill, and the
name and telephone number of the principal researcher. The occupants were not made
aware of the exact time and date that the drill would take place. As well, whenever
possible, the researchers met with those occupants who had limitations in order to discuss
the study and determine their ability to participate. These meetings provided information
for the fmfighters and eased occupants' concerns about the drill.

        On the day of the evacuation drill, video-cameras were positioned in the corridors
and staircases to capture the movement of the evacuees inside the buildings. Video-
cameras were also set up outside to record the occupants exiting from the building or
moving onto their balconies. The NRCC researchers remained outside during the
evacuation in order to not disrupt the movement or influence the behaviour of evacuees.
        The evacuation drills were performed in 1994. The first two evacuations occurred
during the summer while the third building evacuation was in December to study the
effects of winter conditions on occupants' behaviour. The evacuation of the fust two
buildings took place between 18:30 and 19:OO on weekday evenings in order to have a
  -
lawe number of residents at home. The winter evacuation took olace on a Saturdav    -
morning to have good natural lighting for the evacuees and the participating firefighters.
In all buildings, the evacuation began when one of the occupants received a call from the
principal researcher, who was standing outside the building, asking that occupant to pull
the fue alarm.

         Members of the local fue department were contacted orior to the drill to discuss
the evacuation procedure and the 1oEation of the hypotheticalfire to determine a strategy
for evacuating occupants. The firefighters' role was to assist occupants with limitations
                                                                      above and the floor
and to ensweethat occupants on the fire floor, the two floors diirec~ly
immediately below were evacuating the building. The drill ended when the fuefighters
judged that all the occupants had reached a safe location. The fire chief would then give
the "all clear", the alarm was turned off and occupants were allowed to return to their
apartments.

        Immediately after the drill took place, the occupants received a questionnaire.
This post-evacuation questionnaire gathered information regarding the ability of occupants
to hear the alarm, the types of activities they engaged in while preparing to evacuate, their
age, gender and limitations, and their evacuation route.

         The videotapes were carefully viewed and put into a transcript The use of video
allowed for precise recording of the timing of each occupant's actions. The transcript
identifies the unit number from which each occupant started to evacuate, and the exact
time to leave that unit as well as the speed of movement in the corridor and staircase and
the general behaviour during the evacuation. These data were then analyzed using
Microsoft Excel, Version 5.0 and SPSS, Version 6.01. Microsoft Excel is a spreadsheet
application that was used to input the transcript data, calculate time differences and to
conduct other basic calculations. SPSS is a comprehensive software package for
statistical data analysis. This program was used to determine if significant differences
existed between occupants of different genders, ages, and levels of ability when examining
their times to start the evacuation, exit the building, and move from start to finish.

        The evacuees were placed in one of 7 age categories. Category 1 defined children
0-2 years old who had to be carried down the stairs. Category 2 was for children 3-5
years old. These children may have been able to walk down the stairs but often needed
assistance from an adult. Category 3 consisted of children 6-12 years old. These children
had no difficulty travelling without assistance. Category 4 was for teenagers 13-19 years
old. Teenagers may have travelled without their parents and had no problems with the
stairs. Category 5 was for adults between the ages of 20 and 39. These adults travelled
quickly and without problems. Category 6 was for older adults between the ages of 40
and 64. These adults may have been a bit slower than the younger adults, but they still
had a relatively easy time with the stairs. Category 7 was for seniors, ages 65 and up.
These people often experienced some difficulty with the stairs due to their age or other
limitations.

        Limitations are those factors which may hinder an occupant's ability to travel with
ease, particularly in the staircase. In this study, people with limitations included: slow
elderly people; people carrying children, pets, or objects; toddlers and the adults helping
them down the stairs; as well as people using a walker to move. The presence of a
limitation was judged by the researchers from the videotapes. Only those limitations
which were visible and obvious were noted. The Dresence of hearing ~roblems.
                                                                     - A
                                                                                     heart
problems, respiratory problems and the like could not be inferred from the videotapes,
although they may have had an effect on occupants' movement times and behaviour.

       Many people were seen evacuating in groups of two or more. When a number of
people travelled in a group, the speed of the group was generally that of its slowest
member. The presence of others when evacuating could affect the speed of evacuees and
an ANOVA test was completed to see if this was the case in the evacuation of the
Montreal building.

        These group differences were tested using the Analysis of Variance (ANOVA).
The ANOVA compares the scores of groups and determines their similarity by producing
a probability level (p-level). For the purposes of this research, a p-level less than p=0.05
indicates that a significant difference exists between the two groups. This means that
there is less than a 5% probability that the results occurred by chance. There is a 95%
confidence level that the differences between the two groups are real differences that
would occur again if the experiment was repeated.

        The three buildings were also compared to one another to see if significant
differences could be found between the three sets of occupants. Further analyses were
conducted to compare these results with findings from the previous study on midrise
apartment evacuation.

        The use of videotapes made it possible to calculate the speed of the occupants
travelling down the stairs. For each person, it was possible to determine the time at which
they fust appeared in the staircase and the time at which they reached the fust floor
landing. These times were used to calculate the total time each person travelled in the
staircase. It was also possible to determine how many flights of stairs the person had
travelled down. These two calculations were used to determine the average time for a
person to descend one flight of stairs. Measurements of the staircase were also taken at
the time of the drill to determine the distance in metres that occupants travelled. This
made it possible to calculate a subject's speed in metres per second. These two speed
calculations, mean time per floor and speed in d s were used to determine if any speed
differences existed between occupants of different genders, ages and limitations.
        In collaboration with Dr. Chantal Laroche, Department of Audiology and Speech
Language Pathology of the University of Ottawa, the sound of the fire alarm was
measured during the drill. The first series of measurements was carried out one hour
before the drill to measure the ambient sound. The measurements were taken using a
Type 1 microphone ( C i s MK 224) plugged into a digital audio tape recorder (TEAC
DAP-20). Different samples of approximately 15 to 20 s duration were recorded in
different rooms of different apartments. Samples were also taken in the public areas such
as the lobby and on different levels of the conidors and the staircases. In each apartment
visited, sound samples of the ambient noise were taken in the living room, kitchen,
bedroom and bathroom. In the apartments, the sound measurements were taken from the
centre of the room with the bedroom door open and closed, the window open and closed
and the air conditioning on and off. During the drill, a second series of measurements was
taken in the same rooms. The rooms were selected in order to give a good representation
of the different apartment locations in the buildings. Some rooms were close to an alarm
bell while others were far from it.

        Data gathered were analyzed using a Toshiba T-5200 computer equipped with a
digital-to-analog board and two specialized software packages. The first software
package, dBFEl'E, gives the 113 octave band levels (80-8000 Hz) and the overall level in
dBA of each sample. The second software package, DetectsoundTM, was used to analyze
the audibility and recognition of the alarm for different background noises recorded in the
different apartments. The audibility was assessed taking into account the hearing status of
a group of 5 5 year old men with normal hearing according to age and a group of 55 year
old men with a mild to severe hearing loss. This age group gives an interesting picture of
the alarm audibility and is representative of the occupants in the buildings.

        Detectsoundm is a software package developed by the University of Montreal
Groupe d'Acoustique to evaluate and help in the selection of warning sounds [8,9]. The
soh-are analyzes the audibility and recoinition of the alarm considering the background
noise, the hearing status of the receivers and, if appropriate, the use of hearing protectors.
Different algorithms, based on scientific literature and experimental studies, are included in
the software. They allow comparisons to be made between: a) the spectral content of an
alarm measured at a specific location, and b) the audibility "design window" calculated
according to the background characteristics and the hearing status of the receivers. The
hearing status of the receiver refers to the hearing thresholds and the frequency selectivity
of the ear. Hearing thresholds in noise are computed for each 113 octave band and 10 to
25 dB are added to these values in order to make sure that the alarm will not only be
detected but also attract attention and be recognized. The design window goes from 125
to 3000 Hz in such a way as to take into account that spectral elements over 3000 Hz can
be missed by people suffering from different kinds of hearing losses, for example, older
people or workers exposed to noise.
4.0 METHOD AND RESULTS FOR EACH BUILDING

       Three highrise apartment buildings were studied in this research. Although these
buildings contain similar features, it is important first to look at each building individually.

                  -
4.1. Building A Montreal

        The fust building studied was located in Montreal, Quebec. It was constructed of
concrete and completed in 1974 under Regulation 1900, Ville de Montreal. The building
was 14 floors high and housed 244 apartments. The 16 apartments on the first floor were
two storeys high, causing the second floor to have no common comdor. The remaining
floors, 3 through 15 (without a Floor 13). were similarly designed with 19 apartments per
level. There were two elevators located in the middle of the corridor, and two staircases,
one at each end. Each apartment had a balcony, except for those on the fust floor which
had a patio at ground level. The floor plan of the building is shown in Figure 1 while
Figure 2 shows the elevation of the building.

        Occupants had three exits to choose from during the evacuation. Those using
Staircase A had two options for exiting: they could enter the fust floor and exit through
the front entrance or they could exit into a comdor that led to a shopping mall behind the
building. Those exiting from Staircase B also had the option of entering the first floor or
they could continue going down the stairs and exit into the parking garage.




                                      Entree Principale


                                                 -
                        Figure 1: Floor Plan Montreal Building
                                             8
         Escalier A                                                       Escalier   B




                                              -
                        Figure 2: Elevation Montreal Building

        The closest fire hall was located 1 km from the building. Two alarm bells, one 15
cm and the other 25 cm in diameter, were located in the conidor recessed into the wall on
every floor, covering an area per floor of 2500 mZ. On each floor, a Public Address (P.A.)
 .                                                                               .
svstem with two soeakers over the two exit doors was also installed. The P.A. svstem
was only used to play music and there was no plan to use it to give instructions during a
fire evacuation. Smoke detectors and heat detectors were installed throughout the
building and each staircase was vented to the outside at the top floor level. The landings
in the staircases were not large enough to be areas of refuge for those occupants with
mobility limitations, nor were there any fire doors in the corridors to create separate
compartments.

        Two security guards were permanently on location; one pauolled the building or
guarded the front door, while the other worked in a small office inside the parking garage.
The fire annunciator control panel was located by this latter office. In the eventuality of
the activation of the f r alarm, the guard on patrol had to go to manually recall the two
                       ie
elevators to the ground floor. This guard would also investigate the fire, help evacuate
occupants from around the affected area and fight the fire if necessary. Meanwhile, the
other guard would call the fire department and wait to direct and inform firefighters. The
building had not previously experienced an organized fire drill, although it had had several
false alarms in the past few years.
Occupants

        At the time of the evacuation, approximately 500 people lived in the building.
About half, or 250, of these occuoants were seniors. There were also four wheelchair
users, although no apartments were specifically designed to accommodate their needs.
The names and apartment numbers of 5 - - who would need assistance in case of an
                                         people
emergency were on a list. This list was posted in the security guard's office in the
basement Approximately 10 children and 10 multi-cultural families resided in the
building.

Procedure

         Four days before the evacuation drill took place, the occupants of the building
received a memo informing them about the drill. The building manager did not allow the
researchers to talk to the occupants before the drill. In previous evacuations, the
researcher had met with elderly and disabled occupants in order to discuss the exercise, to
ease their fears about the evacuation and to explain to them the procedure they should
follow. The building manager, in this case, refused these meetings for fear of upsetting the
occupants. It should also be noted that the building manual given to residents stated that,
in the case of a fire not in their own unit, occupants are to evacuate onto their balconies
and wait until the situation is under control. Only those who live in the apartment where
the fire has started are instructed to use the staircases to evacuate.

        There were 26 video cameras positioned in every second corridor and in both
staircases. Another four cameras were set up outside facing the back and front of the
building to record movement on the balconies. The cameras were activated shortly before
the alarm began and were not turned off until the alarm ended.

        The local fire department was not permitted by the firefighters' union to participate
in the drill. F e prevention officers of the prevention branch, however, were on the scene
               ir
to assist in the exercise. A fue scenario was discussed with them and it was decided that
the hypothetical fire would occur on the fourth floor near Staircase A. The fire prevention
officers' role was to assist those people on the fire-list and to make sure that occupants on
the fire floor, the two floors above and the floor underneath were evacuating the building.

       The questionnaires were distributed after the drill to all apartments and the
respondents were asked to place the completed questionnaires in a box near the front
entrance.

        -
Results Building A

        The evacuation drill took place on August 18, 1994 on a warm, 23OC, summer
evening. A woman in her 50's who lived on the 4th floor was the designated "alarm
puller" for the building. After receiving a call from the principal researcher, who was
standing outside the building, she emerged from her apartment at 18:54:26 and proceeded
to pull the fire alarm. This started the evacuation at 18:54:42. She travelled down
Staircase A and exited into the mall corridor at 18:55:07. Her total time to move from the
location of the alarm to an exit was 00:25 s.

        The alarm system shut itself off unexpectedly at 19:00:00. After a few minutes of
confusion, it was decided to end the drill and the occupants were permitted to return to
their apartments. The alarm system had been equipped years ago, with a system to silence
the alarm automatically after 5 min. The building management and the fire department
were not aware of the presence of this device. Automatic silencers for fue alarms are not
permitted in this municipality.

        The fire prevention officers, wearing their official uniform, entered the building as
planned 4 3 0 min after the beginning of the alarm and started by consulting the control
panel located in the basement garage. They hardly had time to reach Floors 3.4 and 5
when the alarm shut itself off. When the alarm stopped, they returned to the first floor and
exited the building.

Behaviour of Occupants

        Only 107, or 21.4%, of the 500 occupants were seen throughout the evacuation
drill. Many of these occupants exhibited a variety of behaviours such as appearing but not
evacuating, entering the building from outside, checking the safety of others, returning to
places where they had previously appeared and exiting onto their balconies.

        There were 22 occupants who chose not to evacuate but who were seen looking
into the corridor from their apartments, often talking to neighbows, before returning
inside. Table 1 shows the data for these occupants. On average, these people spent about
00:46 s displaying this type of behaviour.

           Table 1: Occupants Appearing in Corridor but not Evacuating




        There were also six people who entered the building during the exercise. Four of
these people came in from the mall entrance. Of these four, one occupant turned around
and went back into the mall upon hearing the alarm. Another went up the stairs to the
second floor landing and seems to have entered the top floor of a double level apartment
The other two entered the fust floor and exited through the front entrance. However, one
of these occupants returned inside, started to go up Staircase A, came back down, and
exited again through the front entrance. The remaining two occupants were seen in
Staircase B coming up from the parking garage. They entered the fust floor and also
exited through the front entrance. Overall, only one person who entered the building
during the exercise did not exit the building when hearing the fire alarm.
                                                                    -
        Three occupants from the 15th floor worked for the building administration. They
                        the
took re~~onsibilitifor safety of the other occupants, although, helping occupants
during an evacuation is not part of their duty or included in the evacuation procedure.
One of these occupants travelled down staircase A and the other two travelled down
Staircase B. At most floors they stopped, opened the door to the corridor and gave
instructions to other occupants starting to evacuate. The worker in Staircase A had
reached the 5th floor by the time the alarm ended and the other two had reached Floors 6
and 7. After the alarmknded, they continued to travel down the stairs and eventually
exited through the front entrance of the building.

        An unusual behaviour displayed was the tendency for a few of the occupants to
return to places where they have previously appeared. For example, one occupant
appeared from his apartment on the 1lth floor, began to descend Staircase A, reached the
8th floor and then returned up the staircase to the 1lth floor landing, probably intending to
return to his apartment. At this point, however, he joined other o c c u p a n ~ began to
                                                                               and
descend again, eventually reaching an exit

        Another occupant emerged in Staircase B from the 10th floor. He descended two
floors, entered the 8th floor, crossed the building to Staircase A, descended one floor, and
proceeded to enter the 7th floor. Here, he tried the elevator but was unsuccessful so he
returned to Staircase A. At this point he joined other occupants and eventually reached an
exit

        There were two other occupants who displayed similar behaviour. These
occupants were two children who appeared in Staircase A on the 5th floor before the
alarm began. At the sound of the alarm, they began to go up the staircase to Floor 6 but
then returned to Floor 5 and continued going down until they reached an exit.

       There were also about 47 occupants who appeared on their balconies during the
evacuation. It was impossible to judge from the videotapes the gender or age of the
occupants. At times, the only indication of a subject's appearance was the slight flickering
of movement on the balcony. Table 2 shows the mean time of f i t appearance on their
balconies for these occupants and the number of occupants who retumed inside their
apartments.

                   Table 2: Occupants Exiting onto their Balconies




       On average, the occupants appeared on their balconies 2:25 min after the alarm
began. Most occupants appeared alone but there were also six groups consisting of two
people each and one group of three people. The group of three, however, was not
included in calculating the time of f m t appearance since they were already on their
balcony when the alarm began. As well, many of the occupants, approximately 16, did not
remain on their balconies for the duration of the alarm and were seen re-entering their
apartments before the end of the alarm.

         Interestingly, there were four cases where the occupants, seen on the balconies,
may have also made an appearance in the comdors. For example, an occupant on the
I lth floor appeared in the conidor 00:22 s after the alarm began but returned inside the
apartment 00:34 s later. At 1:15 min a person, likely the same as the one appearing in the
comdor, was seen on the balcony of that apartment. Also on the 1lth floor, occ'upants
were seen on their balcony 00:06 seconds after the alarm began. They were then seen re-
entering their apartment. After the alarm had been ringing for about 3:35 min, a person
appeared in the conidor. Again, this is liiely to be the same person. There are two other
cases similar to these examples.

         There were seven people who had started to evacuate but were unable to reach an
exit by the end of the alarm. Three of these people were part of the building
administration mentioned earlier. Although they started to evacuate fairly quickly after the
start of the alarm, they were slow to move down the stairs since they stopped to talk to
others and often entered corridors. After the alarm ended they continued down the stairs
and eventually reached the front exit.

        The remaining four occupants were either from the upper floors of the building or
                                          l
they had started their evacuation later. Al of these occupants continued to travel to an
exit even after the alarm had ended. The average time it took for all seven occupants to
reach an exit was 1:37 rnin after the end of the alarm.

Occupants' Use of the Staircases and Exits

       Of the occupants seen evacuating the building by the two staircases, 55% (n=16)
used Staircase A, and 45% (n=13) used Staircase B. Table 3 shows that most people
tended to use the staircase closest to their apartment. For instance, 85.7% of those
evacuees seen leaving apartments closest to-staircase A also chose this strairway as their
means to exit the building. Only 14.3% of those from the A-side apartments used
Staircase B to evacuate. o f the-people seen evacuating the B-~idea~artments,    100% used
Staircase B to exit.

     Table 3: Occupants Egress Route According to their Apartment Location
              . ...............................
                 ..
          ...K.................................;........ .................    .;.. .....................:
                                                                               ;                            .....................   ................
          ...
           :..    armen.c,:t6$ : uSeii6f'Staif&&s'A:i: iiaseiof,~staf*wei'ei'g:i
                             .    F
               A-side apartments                                             85.7%                                         14.3%
               B-side apartments                                             0.0%                                         100.0%

       The exit doors were also used with differing frequencies. The front door to the
building was most popular and was used by 58% of the evacuees. The second most-
popular exit was on the first floor in Staircase A; a corridor leading to a shopping mall was
used to exit 25% of the evacuees. In Staircase B, evacuees could continue down into the
parking garage to exit; only 17% of evacuees used this exit



         The tables below presents the subjects' evacuation times. Table 4 displays the
times at 1 min intervals and Table 5 displays the times at 15 s intervals. Three different
evacuation times are calculated for this analysis: a) Time to Start, b) Time to Exit, and
c) Time to Move. The Time to Start represents the amount of time elapsed between the
start of the alarm and the moment the person initially leaves his or her apartment. The
Time to Exit represents the amount of time elapsed between the start of the alarm and the
moment when the uerson reaches an exit. Finally. the Time to Move is the difference
                                                   <.

between these two-times. It provides an indication of the time taken by occupants to
travel from start to finish, regardless of when and where they started their evacuation. All
of these times are calculatedonly for occupants who use the staircases. Those who
appear but do not leave the building (e.g., balcony users, people entering the building from
outside, people re-entering their apartments, etc.) are not used to calculate these times.

         In calculating the Time to Start, 31 occupants were used. For the Time to Exit,
the same occupants were used, including those occupants who reached an exit after the
alarm had shut off. Three of the occupants, however, were eliminated from the Time to
Exit calculations since they were members of the building administration and their T i e s
to Exit were distorted by the fact that they stopped at every floor. The Time to Move
included all the occupants used in the Time to Exit but it was also possible to make a few
additions to this group. Five occupants who appeared before or shortly after (less than 5
S) the alarm began, including the "alarm puller", were added, for a total of 33 occupants.

                                                            -
     Table 4: Evacuation Timing a t 1min Intervals Full Duration of the Alarm




                                   --                               ~       -

6:Ol-7:OO            0           0.0%          2        I    7.1%       1       0          0.0%
   TOTAL      I     31      I   100
                                 0.%    I     28        1   100
                                                             0.%        1       33   I   100.0%
                                             14


                                                    -
  Table 5: Evacuation Timing at 15 s Intervals First 5:00 min of the Alarm Only




         The alarm in Montreal lasted only 5:18 min. Many occupants, however, started to
evacuate within the fust four minutes of the a l m . The average time for the occupants to
begin evacuating was 1:30 min after the start of the alarm. This statistic seems impressive
since 64% of the evacuees started within the fust 1:15 min, as shown in Table 5. This
result, however, is not a reliable representation of the average Time to Start because late
starters did not take part in the drill. From previous evacuation studies, it is known that
some occupants start to leave only when told to do so by the firefighters or when they
hear the fire trucks. This situation did not occur during this drill because of the short
duration of the alarm and the non-participation of the local fue department

       The Time to Exit shows more variation throughout the duration of the alarm. The
average Time to Exit was 3 0 7 min. The first occupants to reach an exit did so at 0 5 3 s.
The last person reached an exit 1:20 min after the end of the alarm.
       The Time to Move for occupants varies from 0:25 s up to 4:22 min. The average
Time to Move was 1:43 min. These times were not affected by any crowding in the
comdors or staircases. As mentioned earlier, however, some of the occupants stopped at
various floors, started back up the staircase, crossed comdors, and displayed other similar
behaviours, which may lead to longer Times to Move. As well, some of the occupants
simply had a greater distance to travel.

Time Comparisons According to Gender

       The ANOVA test was used to determine if any differences exist between males
and females for evacuation times. The results of this test as well as the mean times for
each gender are presented in Table 6.

                  Table 6: Time Comparisons According to Gender




        Overall, the male occupants experienced slightly faster Times to Start and Exit
than the female occupants. These results, however, are not statistically significant as seen
by the obtained p-levels of p=0.55 for the Time to Start and p=0.70 for the Time to Exit.
For the Time to Move, females were faster than males but this result was also not
significant The slower average T i e to Move for males may be accounted for by the fact
that several of the male occupants returned to places where they had previously travelled,
crossed comdors and displayed other similar behaviour, which increased their travel time.

Time Comparisons According to Age Categories

       Ages of the evacuees were estimated by the researchers when watching the
videotapes. Table 7 presents a frequency table for the number of occupants in each age
category. Table 8 presents the results of the ANOVA test comparing different age
categories for the Times to Start, Exit and Move.
                                             16

                Table 7: Number of Occupants in Each Age Category




                     Table 8: Time Comparisons According to Age




Note: Age categories 1,2,4, and 7 had fewer than 5 subjects, therefore these groups
       were omitted from the analysis since such a small sample would not produce valid
       results.

       A comparison of all age groups for the Times to Start, Exit, and Move reveals no
significant difference. This means that the people in different age groups did not have
different evacuation times. Many of the age groups, such as seniors and children,
however, are under-represented in this data. The only age groups with more than five
occupants were Groups 3 (6-12 year olds), 5 (20-39 year olds) and 6 (40-64 year olds).
All three groups were compared but no significant difference was found. The groups were
then compared two at a time. These comparisons also failed to produce any significant
result.

        Overall, it would seem that there are no significant differences in the T i e s to
Start, E i ,and Move among different age groups in this building. It is important,
        xt
however, to note that many of the groups lacked a sufficient number of occupants to be
compared to other groups.
Time Comparisons According to Limitation

        The occupants' Time to Start, Exit, or Move may be affected by the presence of a
limitation. The occupants with limitations were compared to the occupants without
limitations to determine if differences in evacuation times exist between the two groups.
The results of these comparisons are presented in Table 9.

                       Table 9: Time Comparisons According to Limitation
            ........
            ......                      .. ....... .. . .
                       . . . ..,,. . ,,,................ .. . . .   ... . . . . . .                   .. .   ..   .:. :.. . .. .   ..........
            . ..
                                                            ;$;.;p-y&uc$j:
            :z:.Me~iTi;m@;i+l& t ~ w f i:,:.p~o i t a t i b n
            . ...        i .~ i.
                            ...         ; .. ~ f m
                                             .

            Time to Start                               2:05 (N=5)                    1:23 (N=26)            p=O. 18
            Time to Exit                                4:00 (N=5)                    2 5 5 0\1=23)          ~3.20
            Time to Move                                1:48 (N=6)                    1:42 (N=27)            p=0.85

        In this building, the occupants with limitations were slower than the occupants
without limitations to Start, Exit and Move; however, none of these differences were
significant. These results may be affected by the small number of occupants in the
limitation group and by the short duration of the alarm.

Speed of Occupants on the Stairs

        Twenty-nine occupants were used in the speed calculations. Occupants who
exited from the fust floor were omitted from the sample since they did not use the stairs to
evacuate. The mean time to descend one floor was 11.3 s. The mean speed of the
occupants was 1.07 mls.

        The mean times per floor for each gender and the speed in d s are compared in
Table 10. Males show a slightly faster time than females, but this difference was not
significant. Overall, just as demonstrated earlier, gender appears to have no effect on
evacuation timing.

  Table 10: Gender Comparisons while Descending One Floor and Speed i n Stairs




        The same age categories as used earlier are represented in Table 11. The table
shows the mean time per floor and the mean speed in m/s for each age category. There is
no speed for babies aged 0 to 2 years old since they would be carried down the stairs by
other people. No occupants of age Categories 2 (small children) or 7 (seniors), used the
stairs during the exercise. Only one teenager used the stairs.
                                                          18

         Table 11: Speed in Stairs of Occupants According to Age Category




        An ANOVA was conducted on the data to determine if any significant differences
in speed existed among the age categories. Table 12 presents the results of the ANOVA
for the speed in d s . Initially all age groups were compared against one another and then
only groups with five or more occupants were used to produce statistically valid results.

         Table 12: Speed in Stairs Comparisons According to Age Category




        Overall, it would seem that there are no significant differences in average travelling
speeds among the age categories. Just as before, however, many of the age groups do not
have a sufficient number of occupants to allow for valid comparisons. In particular, this
sample of occupants' speeds lacks a sufficient number of subjects who would be expected
to travel at dower speeds, such as seniors and younger children.

         It is expected that the presence of a limitation would affect an occupant's speed on
the stairs. Table 13 shows the results of the ANOVA test comparing the mean time per
floor and the mean speed in d s , for occupants with or without limitations.

          Table 13: Speed on Stairs Comparisons According to Limitation
          ................................................
         ...............................................
                                                      :Fldss< : ,:i;M&;$S P&d iti;:mk.i
                 ...................
          .............................................:::............
           .................................. . . . . . .
           . . . ............................................
         ..:::..................................... . . ,
               ................                               i        ,



          Limitation                         12.71 (N=3)              0.88 W=3)
          No Limitation                      11.10 (N=26)             1.09 (N=26)
         ,Result (p-value)                     ~4.6260                   ~3.40
       Although the occupants with liiitations were slightly slower than the occupants
without limitations, these differences were not significant. This result may be affected by
the small number of occupants in the limitation group. Perhaps a larger sample size would
produce sigmcant differences.

        The occupants travelling in groups were generally slightly slower than those
travelling alone, the difference, however, is not significant. The results are shown in Table
14.

                 Table 14: Speed on Stairs Comparisons According to Groups
           . . . . .,.. .
          ..:.:::~~.:. .. .... ..... . . .: . ...~.~~ .
         .~.. .....:.::,.,::..:.::.::::.:.:~:. ....::...:~..:.:.:.. ::: .. .. ........ .
                ....                  ...........            ~                             . .. .. .. .   . . .. ..   . . .   . .... . . .. ..,, .,,.,.,., ,,   .. ..... ... .
                       ... ...
         ...... ,,.~............ ..~. ....... ... ..
                   . ... . . .               .
                                              .
         ...................................... ..~. .
                                                     . .~         ~ ~ s ~ : ' ~ ~ & ' . g;~j& @ ~fs ,(: s ~ ~ d ,:i ,~ ? & & ;
                                                                                          : ,~ ~ g l ~) : :
          With Group                                                                 12.29 (N=19)                                 1.00 (N=19)
          Alone                                                                       9.34 (N=10)                                 1.18 (N=lO)
         .Result (p-value)                                                               p=O. 15                                     ~4.25


Questionnaires

       Fifty-eight questionnaires were returned, which represents 24% of the units in the
building. Of these 58 people, 43 stated that they were present for the drill and 41
mentioned that they had heard the fire alarm.

         A few occupants, 11.6%, found that the sound of the alarm was not loud enough
in their apartments. Although their apartments were located on different floors, these
people all lived in apartments at the far ends of the corridor, near the A or B staircases.
Only 4.7% of the people found that the sound of the alarm was not loud enough in the
staircases. None of these people, however, stated that they had used the stairs to
evacuate.

        Most of the people, 83.7%. were in their apartments at the time of the drill. In
response to the sound of the alarm, the occupants listed several activities that they
engaged in before starting their evacuation. For instance, looking out into the corridor
was frequently listed, accounting for 22.8% of responses. Exiting onto a balcony
accounted for 21.7% of responses and leaving the building accounted for another 13.0%
of responses. These were the most frequently listed activities.

        The average time people stated that they took before starting their evacuation was
2:18 min after the alarm sounded, which is a good assessment on their part of the actual
time observed of 1:30 min. Only 21.0% of people revealed that they always leave the
building at the sound of the alarm. Thirty people, or 70.0%, said that they don't always
leave, which is in accordance with the procedure explained in the tenants' manual.

       The average age of the respondents was around 60 years. The respondents were.
mainly female, 65.1%. with males respondents being, 32.5%.
Alarm Audibility Measurement

       Table 15 presents the measurements in dBA, of the background noise and of the
background noise with the alarm operating for different locations in the building. The
analysis was limited by the small number of samples that could be taken during the
shortened drill. The alarm should have a minimum of three spectral elements in order to
meet audibility requirements. The two last columns in Table 15 presents the number of
spectral components reaching the audibility design window. Tthe numbers in parentheses
represent the number of spectral components that exceeded the audibility design window,
which confirms the intense sound pressure in certain locations.

 Table 15: Background Noise Levels and Detectsound Analysis of Alarm Audibility




                                     Frequency (Hz)
             Figure 3: Audibility Design Window For a 55 year old man
                                          -
                    with normal hearing Bedroom of Unit 1519
         Figures 3 to 5 illustrate some of the results obtained with DetectsoundTM thefor
fire alarm analysis of   the Montreal building. In these figures, the x-axis represents the
frequency content in hertz (Hz). The y-axis gives the sound pressure level (dBSPL).
Measurements of the spectral content of the background noise is represented by the
continuous horizontal line and the alarm level by the vertical lines for each 113 octave
band. For the buildings studied, the alarm is often lost in the background noise making it
difficult to extract its exact spectral content. The diagonal hatched area represents the
audibility design window into which the alarm should have a minimum of three spectral
elements in order to meet the requirements of audibility. The spectral elements that reach
the audibility design window are represented with a thicker vertical line, as shown in the
figures.




                                       Frequency (Hz1
              Figure 4: Audibility Design Window for a 55 year old man
                                                   -
               with mild to severe hearing loss Bedroom of Unit 1519

        In the living room of Unit 1519, only one spectral component reached the
audibility design window for a 55 year old man with normal hearing, as shown in Figure 3.
The spectral element reaching the design window might be loud enough for normal
hearing subjects but would be missed by people with mild to severe hearing loss (from 1 to
8 kHz) as shown in Figure 4. It should be noted that in Figures 3 and 4, the background
noise level and the alarm level are the same, the only difference is the hearing status of the
receivers. The hearing status of the receivers changes the shape of the design window
which becomes an important factor in an alarm audibility analysis Without paying
attention and being awake, it is difficult for an occupant to hear the fire alarm under such
circumstances.
                                      Frequency (Hz)
              Figure 5: Audibility Design Window for a 55 year old man
                                          -
                  with normal hearing A-side Corridor, 15th floor

        Figure 5 shows that the alarm met the audibility requirements in the conidor with
three spectral elements reaching the audibility design window. There are however,
thirteen spectral elements which exceed the design window and a series of high levellhigh
frequency elements. These two factors c o n f m the intense sound pressure at that
location.

Conclusions

        The evacuation drill in this highrise apartment building was unusual in many
respects. The alarm rang only for a short period of time, 5:18 min, before it unexpectedly
shut itself off. The building management was apparently not aware of the fact that a timer
shutting off the alarm after 5 minutes had been installed. Such timers have never been
permitted by the local jurisdiction. The occupants seemed to be confused about what they
should do when the alarm stopped after such a short period.

         During the drill, many occupants appeared in the comdors and then returned inside
their apartments. There were also several occupants who exited onto their balconies, in
accordance with the building Tenant's Manual which instructs occupants that, in the case
of fire, they should wait for help on their balconies instead of attempting to leave. This
may explain why many occupants did not leave their apartments. The low participation
rate of the occupants can also be explained by the fact that there had never been a fire drill
in this buildig and that the researchers involved in this study were not allowed to talk to
the occupants prior to the drill.

       The occupants who did evacuate, tended to travel in a direction that was familiar
to them. Most occupants used the staircase closest to their apartments and 58% of the
evacuees used the front entrance to leave the building.

        The average time for the occupants to start evacuating was 1:30 min after the start
of the alarm. This is an unusually fast T i e to Start. The short duration of the alarm
probably prevented occupants who were late starters from participating in the drill and,
therefore, they could not be taken into account in the results. The average Time to Exit
the building was 3:07 min. Again, the exclusion of late starters causes this to be a fairly
quick time. The average Time to Move from start to finish was 1:43 min. The absence of
crowding in the staircases and the poor representation of many age groups, such as small
children and seniors, probably accounts for this quick Time to Move.

        The results of the statistical analyses show that there were no significant
differences between occupants of different genders, ages, and levels of ability when
comparing their times to start, exit and move. These results, however, are probably
affected by the small sample sizes and the short duration of the alarm. Occupants who are
expected to be slower, such as seniors and those with limitations, may not have started to
evacuate within the first 5 3 8 min of the alarm. This under-representation of late starters
affects the results obtained from these comparisons. The time for occupants to descend
one flight of stairs and the speed at which they travelled also did not yield any significant
result when the groups mentioned above were compared. The mean speed while going
down the stairs was 1.07 m/s.

        The responses from the questionnaires revealed that most occupants found the
                                                                        people
alarm to be loud enough in their apartments and in the staircases. Most . . tended to
look in the conidor, exit onto the; balcony or leave the building upon hearing the alarm.
This is consistent with the observed behaviour of the occupants. Many people also
admitted that they don't leave their apartments every time the alarm sounds.

        The analysis of the alarm audibility shows that the alarm was not audible from
inside any of the three apartments measured. Unfortunately, due to the short period of the
alarm sounding, it was impossible to take more samples. The sound pressure in the
staircase and the corridors appeared, in contrast, to be overpowering.
                  -
4.2 Building B Calgary

        The second building studied was located in Calgary, Alberta. It was a concrete
building built in 1984 under the 1981 Alberta Building Code. It was 14 floors high and
housed 117 apartments. The fust floor contained offices and meeting rooms; no
apartments were located on this floor. The remaining floors had 8 or 10 apartments:
Apartment Numbers 1 and 11 on Floors 3 through 15 were double level apartments and
there was no Floor Number 13. Each floor had two staircases located at opposite ends of
the corridor and two elevators located close to Staircase A. Each apartment had a
balcony facing either the front or back of the building with the exception of the double
level apartments whose balconies were on the sides of the building. Figure 6 shows the
floor plan of the building and Figure 7 shows its elevation.


                                           Parking
                             Rear
                             Exit
                               m                          /-P.A.    Speaker




                               Entrance


                                                      -
                      Figure 6: Typical Floor Plan Calgary Building

       There were four possible exits for occupants to leave the building during the drill.
Occupants using Staircase A had the choice of exiting the staircase through a door at
ground level, or they could enter the ground floor lobby and exit through the front or back
entrances. Staircase B led only to an outside door exit at the bottom of the stairs.

        The alarm system was directly connected to the local Fire Department which was
located 0.7 km away. Two alarm bells were located on every floor. A P.A. system also
existed but there was no plan to use it during the drill to issue instructions. Smoke and
heat detectors were also found throughout the building. The elevators were set to
automatically return to the fust floor upon activation of the fue alarm. The control panel
was located in the lobby, but it did not contain a frre list or a master key for the building.
In the past, false fire alarms had been frequent, with as many as 37 occurring in one year
(1990). For the first nine months of 1994, there had been only 3 false alarms.

                                  Stair "A"                        Stair "B"
                                moor     Elevators                        Floor
                                                                           15
                                                                           14
                                                                           12
                                                                           11
                                                                           10
                                                                            9
                                                                            8
                                                                            7
                                                                            6
                                                                            5
                                                                            4
                                                                            3
                                                                                Exit "B"
                                                                            1   +
                                               L
                                            Main
                                          Entrance
                                               -
                        Figure 7: Elevation Calgary Building

Occupants

        At the time of the evacuation, approximately 165 people lived in the building.
Over 20% of the occupants were seniors, 17 people had limitations and about 15 families
lived in the building. Apartment Number 8 on every floor was specifically designed to be
accessible to people with limitations.

Procedure

       Four days before the evacuation drill took place, the occupants of the building
received a memo informing them about the exercise. The researchers also met with the
older occupants and those with limitations to discuss the procedure they should follow.
About 20 people were met prior to the drill.

        There were 27 video cameras positioned in every corridor and in both staircases,
giving a complete overview of the occupants' movement. As well, a camera was located
in the lobby and two cameras were positioned outside to capture balcony movement on
the front and back sides of the building. The cameras were turned on slightly before the
alarm began and were not turned off until the alarm ended.
        The Calgary Fire Department agreed to participate in the drill. A strategy was
discussed with the firefighters prior to the drill. The hypothetical fire was staged on the
fourth floor near the A exit. The firefighters were to move throughout the building
evacuating occupants with mobility impairments from the 4th, 5th and 6th floors to the
ground level using the elevator. Upon receiving the alert call at the fire station,
firefighters in full uniform, who were waiting only a few streets from the apartment
                                      -
building. headed towards the building in three fire trucks, with the sirens activated. The
        ",
Fire Department also used this opportunity to test their new Amateur TV unit which
would assist them in real fire situations. Two local TV networks, CBC and CFCN had
been invited by the Fue Department to shoot the exercise, they remained outside the
building so that they would not interfere with the evacuation.

       After the drill questionnaires were distributed to all apartments and the
respondents were asked to place the completed questionnaires in a box near the front
entrance.

         -
Results Building B

       The evacuation drill took place on September 15, 1994, on a warm, 25OC, autumn
evening. A woman who lived on the 5th floor was the designated "alarm puller" for the
building. After receiving a phone call from the researchers, she proceeded to pull the
alarm. This started the evacuation at 18:30:21. The "alarm puller" re-entered her
apartment at 18:30:33, went onto her balcony and did not evacuate the building.

        The alarm ended at 18:47:51, 17:30 min after the start, when the fire chief gave the
 al
"l clear" signal, indicating that the evacuees had reached an area of safety. The
occupants were then allowed to return to their apartments.

       The alarm systems activated as anticipated, with the exception of the elevators.
The elevators were supposed to automatically return to the f i s t floor at the sound of the
alarm but this did not occur. In fact, several of the occupants used the elevators to
evacuate.

        There were several firefighters in full uniform, as well as three fue trucks, on the
scene, to participate in the evacuation. The first fuefighters were seen entering the
building at 18:36:45,6:24 min after the alarm began. As planned, they knocked on doors
and checked on the occupants of the 4th, Sth, and 6th floors. They were also seen using
the elevators and staircases to travel from floor to floor. On the 6th floor, the firefighters
used the elevator to evacuate an elderly lady with a walker, since she was unable to use
the stairs. The firefighters were also seen testing their Amateur TV unit on the 4th and
5th floors, throughout the drill.
Behaviour of Occupants

       Only 48, or 29%, of the 165 occupants were seen throughout the evacuation drill.
Several of these occupants displayed a variety of behaviours, such as appearing in the
comdor but not evacuating, entering the building during the alarm, or exiting onto their
balconies.

         There were three occupants who appeared in the comdor during the evacuation
but who did not evacuate. The f i s t was the "alarm puller" from the 5th floor, who
returned to her apartment 00:12 s after the alarm began. A woman on the 1lth floor
appeared in the comdor 00:39 s after the alarm began but returned to her apartment 00:04
s later. Finally, a man on the 5th floor appeared at 00:44 s and stayed in the comdor for
about 1:03 min before re-entering his apartment. On average, these people spent 00:26 s
outside their apartments displaying these behaviours.

        Only one person appeared to enter the building during the alarm. It can be
assumed that he was in the elevator when the evacuation began, since he emerged from
the elevator on the 9th floor 00:Ol s after the start of the alarm. He entered his apartment
00: 11 s later and did not proceed to evacuate.

       There were about 12 occupants who appeared on their balconies during the
evacuation, as shown in Table 16. It was impossible to judge the gender or age of the
occupants from the videotapes. In some instances, the only indication of a subject's
appearance was the slight flickering of movement on the balcony.

                   Table 16: Occupants Exiting onto their Balconies




        On average, the occupants appeared on their balconies 2:24 min after the alarm
began. Most occupants appeared alone, but one group composed of two people was seen
on a third floor balcony. As well, most occupants remained on their balconies throughout
the drill, with the exception of a person on the 8th floor who re-entered his or her
apartment.

       There was one case where a person may have made two appearances, one in the
comdor and one on the balcony: a woman from the 1lth floor who made a brief
appearance in the corridor is likely to be the same person as the one who appeared on the
balcony of the same apartment 03:55 min after the alarm started.

       Of the occupants seen evacuating the building, 51.5% (n=17) used Staircase A,
33.3% (n=ll) used Staircase B, and 15.2% (n=5) used the elevator. Table 17 reveals that
most people tended to use the staircase closest to their apartment. For those people living
in apartments closest to Staircase A, 81.3% chose this as their means to exit the building.
Only 6.3% of those in A-side apartments used Staircase B, and 12.5% used the elevators.
Interestingly, of the occupants living closest to Staircase B, 58.8% used these stairs to
evacuate, 23.5% used Staircase A, and 17.6% used the elevators. This means that 41.1%
of the occupants in B-side apartments chase to move towards the elevator, which is
probably their familiar route to exit, and then they used Staircase A to leave, rather than
using the closest staircase.

    Table 17: Occupants' Egress Route According to their Apartment Location




         The exits doors were also used with differing frequencies. Occupants travelling in
Staircase A had the option of entering the frst floor lobby and exiting through the front or
back door, or they could continue down 5 steps and exit on the side of the building. There
were 10 people in Staircase A who chose to go to the side exit, representing 30.3% of the
total evacuees. The other seven people using Staircase A chose to exit through the lobby
and front door. No one chose to exit through the back door. The five people travelling in
the elevators also exited through the front entrance in the lobby. This made a total of
36.4% evacuees exiting through the front door. In Staircase B, the evacuees could only
continue going down the stairs to the side exit. In total, 11, or 33.3%, of the occupants
left the building through this exit

Evacuation Timing

         The tables below present the occupants' evacuation times. Table 18 displays the
times at 1 min intervals for the full duration of the alarm and Table 19 displays the times at
15 s intervals for the first 5:00 min only. Three different evacuation times are calculated
for this analysis: a) Time to Start, b) Time to Exit, and c) T i e to Move. The Time to
Start represents the amount of time elapsed between the start of the alarm and the moment
the person initially appears from their apartment. The Time to Exit represents the amount
of time elapsed between the start of the alarm and the moment when the person reaches an
exit. The Time to Move is the difference between these two times. It provides an
indication of the time taken by occupants to travel from start to finish, regardless of when
and where they started their evacuation. All of these times are calculated only for
occupants who use the staircases. Those who appear but do not leave the building (e.g.,
balcony users, people entering the building from outside, people re-entering their
apartments, etc.) are not used to calculate these times.
                                             29

                                                         -
    Table 18: Evacuation Timing a t 1 min Intervals Full Duration of the Alarm




       In calculating the Times to Start, Exit and Move in Table 18.33 occupants are
used. This table covers the full evacuation time for the drill and includes a l occupants
                                                                             l
seen evacuating the building. Table 19 presents data only from the first 5:00 rnin of the
alarm so occupants numbers vary for Times to Start, Exit and Move.

        As seen in Table 18,30 of the 33 people seen evacuating had a T i e to Start
within the first 7:00 min after the commencement of the alarm. The first occupants to
start were a group of two adults on the 7th floor, 0:15 s after the alarm. The last
occupants seen evacuating were a group of three, two adults and a small child, who left
their 10th floor apartment 10:10 min after the start of the alarm. The average T i e to
Start for all occupants was 2:48 min. Table 19 shows that the majority of the evacuees,
63.6%, presented Times to Start which are fairly evenly distributed across the fust 2:00
rnin of the alann. After this point, the various Times to Start are more spread out.

        The occupants' Times to Exit are quite varied. The times range from 1:02 to
13:10 min, with an average Time to Exit of 4:34 min. Most evacuees, 87.9%, reached an
exit within the fust 9:00 min after the alarm sounded. By the end of the first 5:00 min, as
shown in Table 19,63.6% of the 33 evacuees had reached an exit. The Time to Start, the
distance to travel, travelling with children, the age of the evacuee, and other such factors
are likely to have affected the Time to Exit
                                           30

                                                             -
           Table 19: Evacuation Timing at 15 s Intervals First 9 0 0 min




         Almost all of the evacuees, 97%. had a Time to Move under 4:00 min. The
average Time to Move was 1:46 min. The longest Time to Move was 7 5 6 min; it was the
time it took a senior women using a walker to evacuate from the 6th floor by the elevator,
with the help of a fuefighter. Table 20 shows the Times to Move according to the floor
where occupants started. Two occupants from the 2nd floor were the fastest, with a Time
to Move of 0:13 s. The longest Time to Move, excluding the elderly lady mentioned
above, was 3:18 min for an occupant on the 14th floor. Four other occupants who had no
apparent limitations took the elevator to evacuate. The use of the elevator explains the
quick Time to Move of four occupants: two from the 12th floor with 0059 and 1:32, one
from the 11th floor with 1:01, and one occupant on the 15th floor with 2:16.
                                            31

                      Table 20: Time to Move vs. Starting Floor




Time Comoarisons Accordino to Gender

        The ANOVA test was used to determine whether any differences exist between
males and females for evacuation times. The results of this test, as well as the mean times
for each gender, are presented in Table 21.

                  Table 21: Time Comparisons According to Gender




        The male occupants of the building have faster Times to Start, Exit and Move than
do the females; however, none of these differences are significant. It seems that gender
did not play a significant role on the evacuation timing.

Time Comoarisons According to Aoe

       Ages of the evacuees were estimated by the researchers when watching the
videotapes. The evacuees were then placed in one of the seven age categories based on
this estimate. Table 22 presents a frequency table for the number of occupants in each age
category. Table 23 presents the results of the ANOVA test comparing different age
categories for the Times to Start, Exit and Move.

               Table 22: Number of Occupants in Each Age Category




             Table 23: Time Comparisons According to Age - Building B




                     I                 I
                     * = a significant result
Note: Age categories 1,2,3,4 and 7 had less then 5 subjects, so these groups were
       omitted from the analysis since these small numbers would not produce valid
       results.

        The ANOVA test initially included all of the age categories. It reveals significant
differences for Time to Start, Exit and Move amongst all groups. The remaining analyses
were conducted only with groups containing more than five occupants. This included only
Groups 5 and 6, adults and older adults. The results for Times to Start, Exit and Move
were not significant This means that there were no significant differences in evacuation
times between the two groups. This is not surprising since the two age categories should
have fairly similar characteristics.
       Overall, although the results suggest that some differences exist between age
categories, it is important to keep in mind that many of the age groups lack a sufficient
number of occupants to be able to make valid comparisons between them. I sufficient
                                                                             f
sample sizes were obtained, one could be more confident that the results are reliable.

Time Comparisons According to Limitation

        An occupant's Time to Start, Exit, or Move may be affected by the presence of a
limitation. The occupants with limitations were compared to the occupants without
limitations to determine if differences in evacuation times exist between the two groups.
The results of these comparisons are presented in Table 24.

                    Table 24: Time Comparisons According to Limitation
            . ......... ,.,~ .
            .~....... :
            ........~.. : .
              .. .
            .~ .............
             .......... ...
              .... ............ . . .
            ::::::I:
                    ......... .
                     .
               Time to Start
                                  .~.... :
                                      .        .............. . . . ..: ::,.. ,, .... .
                      ~~~~~nm.~;i~~~;~~.~.:~l:IILimi,ta
                                 ,.N,&:u~tati&~!. aIue::i
                                           .    ,
                                              ;:+
                                                . . : : ....... ..... . .. .. . .. .

                                                               5:34 (N=9)
                                                                                          .. .
                                                                                          .
                                                                                                    . .. .
                                                                                                 .. . . .. .. . . .

                                                                                              1:46 (N=24)
                                                                                                                           ' ....:,..:...~
                                                                                                                      . . .. .

                                                                                                                                  *p=O.OO
                                                                                                                                         .....:.,.:..:..... :.::...::
                                                                                                                                          .             .



               Time to Exit                                    8:18 (N=9)                     3:lO (N=24)                         *p=O.OO
               Time to Move                                    2:44 (N=9)                     1:24 (N=24)                         *p=O.OO
            * = a significant result
         In this building, the occupants with limitations were significantly slower than
occupants without limitations to Start, Exit and Move. The mean Time to Start for
occupants with limitations was 5:34 min while those without limitations had a mean Time
to Start of 1:46 min. This difference was significant at p=O.OO. Occupants with
limitations also reached an exit a lot later than occupants without limitations. The mean
times were 8:18 min and 3:10 min, respectively. This difference was also significant at
p=O.OO. The Time to Move showed that occupants without limitations moved twice as
fast as those without limitations. This difference was significant at p=O.OO.

Speed of Occuoants on the Stairs

        Twenty-eight occupants were used in the speed calculations. Occupants who went
on their balconies or used the elevator to evacuate were omitted from the sample since
they did not use the stairs during the drill. The mean time to descend one floor was 10.18
s. The mean speed of all the occupants was 1.05 m/s.

        The mean times per floor for each gender and the speed in m/s are compared in
Table 25. Females show a slightly faster time than males, but this difference was not
significant. Overall, however, just as demonstrated earlier, gender appears to have no
effect on evacuation timing.
                                              34


  Table 25: Gender Comparisons while Descending One Floor and Speed in Stairs




       The same age categories as used earlier were used to compare the speeds.
Table 26 shows the mean time per floor and the mean speed in m/s for each age category.

              Table 26: Speed of Occupants According to Age Category




       An ANOVA was conducted on the data to determine if any significant differences
in speed existed between the age categories. Table 27 presents the comparison for the
Speed in mls. Initially, all age groups are compared against one another and then only
groups with five or more occupants are used to produce statistically valid results.

         Table 27: Speed on Stairs Comparisons According to Age Category

                     Ages Compared (          p-value   I   Ss (NS)
                         1,2,5,6.7        (   *p=O.OI   I      No
                             5.6          (   *p=O.O2   1     Yes
                   * = a significant result
        The comparison of all age groups' average travelling speeds yields significant
                                                                          -- -
results. This means that there are differences in travelling speeds among people of the
                                                           -*

different age categories. It is important, however, to keep in mind that many of these age
groups do not contain five or more occupants. Only Groups 5 and 6, adults and older
adults, meet this requirement. When these two groups are compared, the speeds of the
occupants in each group are significantly different at p=0.02. The group of adults aged
20-39 was significantly faster in the stairs than the group of older adults aged 40-64.
        It is expected that the presence of a limitation would affect an occupant's speed on
the stairs. Table 28 shows the results of the ANOVA test comparing the mean speed in
m/s for occupants with and without limitations.

           Table 28: Speed on Stairs Comparisons According to Limitation
                    ......................
                   ........... ; :............: ::..::.. :.. .....: .:
                    ............... : : . .. .
                    .        ;
                        ......
                                                                  ,;:M$$n:;S":
                                                                           .sed        :::,   -. v ~ , ~ ~ l j
                     Limitation                                          0.61 (N=8)           p1.00
                     No Limitation                                       1.22 (N=20)

        The occupants with limitations were significantly slower than the occupants
without limitations when travelling down the stairs. The mean time per floor for
occupants with limitations was 15.45 s and the mean time for those without limitations
was 8.07 s. The speed comparisons show that those with limitations travelled at a speed
of 0.59 m/s while those without limitations travelled at nearly twice this speed, 1.14 m/s.
Both of these differences are significant at p1.00. Overall, the results show that occupants
with limitations are slower than occupants without limitations when travelling down the
stairs.

        The speed on the stairs of occupants alone and the speed of occupants travelling in
groups were also calculated and compared. The occupants without limitations travelling
alone had an average speed of .94 mls while the occupants travelling in groups of two or
more had an average speed of 1.12 m/s. One of the groups, formed of two older adults
and a small child of approximately 4 years old, was particularly slow, travelling at a speed
of .53 m/s. During the evacuation of the Calgary building, however, people travellimg in
groups had a tendency to move faster than people travelling alone but the difference was
not ~ i g ~ c a(p=0.22).
                nt

Questionnaires

        Fifty-three questionnaires were returned to the researchers, representing 45.3% of
the units in the building. Of these 53 people, 26 stated that they were present for the drill
and 25 stated that they had heard the alarm.

        Only 7.7% of the people had difficulty hearing the alarm in their apartments. One
of these people lived in an apartment designated for people with limitations. The other
person lived on the 15th floor, but the location of the apartment is not known. As well,
3.8% of the people found that the sound of the alarm was not loud enough in the corridor.

         Almost all of the respondents, 84.6%, were in their apartments at the time of the
alarm, mostly in their living rooms. In response to the sound of the alarm, the occupants
listed several activities in which they were engaged. For instance, leaving the building was
a frequently listed activity that accounted for 22.4% of responses. Gathering valuables
accounted for 14.9% of responses and exiting onto a balcony accounted for 13.4% of the
responses. These were the most frequently listed activities.
       According to the respondents, the average time they began to evacuate was 3:35
min after the start of the alarm, which is an overestimation of the observed T i e to Start
of 2:48 min. Only 30.8% of the people said that they used their balconies to evacuate,
7.7% stated that they tried the elevator, and 53.8%used the stairs. More than half of the
respondents, 53.8%,revealed that they do not always leave the building when they hear
the sound of the fire alarm.

      The average age of the respondents was around 40 years old. The respondents
were mainly female, 65.496, although some males did answer the questionnaire as well,
30.8%.

Alarm Audibility Measurement

       Table 29 presents the measurements, in dBA, of the background noise and of the
background noise with the alarm for different locations. The alarm should have a
minimum of three spectral elements in order to meet audibility requirements. The two last
columns in Table 29 presents the number of spectral components reaching the audibility
design window. Tthe numbers in parentheses represent the number of spectral
components that exceeded the audibility design window, which confirms the intense sound
pressure in certain locations.

 Table 29: Background Noise Levels and Detectsound Analysis of Alarm Audibility
       Figures 8 to 10 illustrate some of the results obtained with DetectsoundTM for the
fr alarm analysis of the Calgary building. In these figures, the x-axis represents the
 ie
frequency content in hertz (Hz). The y-axis gives the sound pressure level (dBSPL).
Measurements of the spectral content of the background noise are represented by the
horizontal continuous lime and the alarm level by the vertical lines for each 113 octave
band. For the buildings studied, the alarm is often lost in the background noise making it
difficult to extract its exact spectral content. The diagonal hatched area represents the
audibility design window into which the alarm should have a minimum of three spectral
elements in order to meet the requirements of audibility. The spectral elements reaching
the audibility design window are represented by a thicker line, as shown in the figures.




                                    Frequency (Hz)
             Figure 8: Audibility Design Window for a 55 year old man
                                             -
                     with normal hearing Bedroom of Unit 508

       In the living room of Unit 508,no spectral component reached the audibility
design window for a 55 year old man with normal hearing, as shown in Figure 8. The
alann was completely lost in the background noise, except for a few 113 octave bands
which did not reach the design window. Without paying attention and being awake, it is
almost impossible for an occupant to hear the fue alarm under such circumstances.
                    Frequency (Hz)
Figure 9: Audibility Design Window for a 55 year old man
                         -
     with normal hearing A-side Corridor, 5th floor




                     Frequency (Hz)
Figure 10: Audibility Design Window for a 55 year old man
     with normal hearing - A-side Staircase, 5th floor
        Figures 9 and 10 show, that the alarm met the audibility requirements in the
comdor and staircase with eight and seven spectral elements, respectively, reaching the
audibility design window. There are, however, seven (in the comdor) and two (in the
staircase) spectral elements exceeding the design window and a series of high levellhigh
frequency elements, confuming the intense sound pressure at these location.

Conclusions

        The evacuation drill in this highrise building lasted 17:30 min. The fire safety
systems activated as planned, with the exception of the elevators. The two elevators were
supposed to be automatically recalled to the first floor at the sound of the alarm, however,
they continued to function until the fuefighters recalled them manually. The elevators'
accessibility allowed a few occupants to use them to evacuate. Most of the occupants
who appeared during the drill evacuated using the stairs, but a few people appeared on
their balconies as well. The low participation rate for this building may be affected by the
vacancy of several apartments and by the fact that many occupants were out of the
building at the time of the drill.

        The occupants who did evacuate, tended to travel in familiar directions. Most
occupants used the staircase closest to their apartments but, in several cases, the
occupants who lived in the B-side apartments used Staircase A. This is probably due to
the fact that the elevators are closer to this staircase and this is a familiar route. The front
entrance and the two side exits were used almost equally as a means to leave the building.
The back entrance was not used.

        The average time for the occupants to start evacuating was 2:48 min after the start
of the alarm. The average Time to Exit the building was 4:34 min and the average Time
to Move, from start to finish, was 1:46 min. The lack of crowding in the staircases
contributed to this fairly quick Time to Move.

         The results of the statistical analyses provide some insight into the differences
between occupants of different genders, ages and levels of ability when comparing their
Times to Start, Exit and Move. There were no significant differences between male and
female occupants, although the males tended to have faster evacuation times. Another
finding was that when all the age groups were compared to one another, there were
significant differences. Many of the age groups, on which these findings are based,
however, had very small sample sizes. Therefore, these results must be interpreted with
caution. The differences in times for occupants with and without limitations were also
significant This means that those with limitations are, in fact, slower than those without
The occupants had a mean time of 10.18 s to descend one flight of stairs and their mean
travelling speed was 1.05 m/s. When these speed calculations were used to compare the
groups mentioned above, similar results were found. For instance, gender still had no
effect, but the significant differences between age groups and occupants' levels of ability
were observed once again.
          The responses from the questionnaires revealed that most occupants found the
   alarm to be loud enough in their apartments and in the staircases. Nevertheless, the
   analysis of the alarm shows that the alarm is not readily audible inside any of the
   apartments visited, although the alarm appeared overpowering in the comdor.

           In response to the alarm, occupants mentioned in the questionnaire that they
   tended to gather their valuables, leave the building, or exit onto their balconies. This is
   consistent with the observed behaviour of the occupants. Many people also admitted that
   they don't always leave their apartments at the sound of the alarm. This could be due to
   the frequent occurrence of false alarms in this building.



                    -
   4.2 Building C Gloucester

          The thud evacuation was canied out under winter conditions. It was very
   important to measure the time to get prepared to evacuate and the time to move with
   winter clothing. The building studied for evacuation under winter conditions was located
   in Gloucester, Ontario. It was a concrete building built in 1979, under the 1975 Ontario
   Building Code. It was 12 storeys high and housed 213 apartments. The first floor
   contained 15 apartments and a laundry room. The remaining floors each contained
   18 apartments. The basement was divided into two large areas: a meeting room and a
   room for physical conditioning with a hot-tub. Sprinklers had been installed in the
   basement




                                                Main
                                              Entrance
                                                     7'
Exit                                                                           Exit
                                                                               4




                                     P.A. Speaker
                                    OAlarrn



                                                 -
                         Figure 11: Floor Plan Gloucester Building
        There were two staircases located at opposite ends of the corridor and two
elevators were located closer to Staircase B. Only apartments located on the four comers
of the building had balconies. Figure 11 shows the floor plan of the building and
Figure 12 shows an elevation plan.


                      A                                                B
                  . Stairs                         Elevators         Stairs Floor
            Floor
             12                                                              12
              11                                                             11
             10
              9
              8                                                                8
              7                                                                7
              6                                                                6
              5                                                                5
              4                                                                4
              3                                                                3
   Exit "A" 2                                                                      Exit
                                                                               1   +
                                                      .
                                                      L
                                                     Main
                                                   Entrance

                                               -
                       Figure 12: Elevation Gloucester Building

        There were three possible exits for occupants to leave the building during the drill.
Occupants using either Staircase A or B had the choice of exiting the staircases through a
door at ground level, or they could enter the ground floor and exit through the lobby &d
front entrance.

        The alarm system was directly connected to a security company. The local F i e
Department was located approximately 1 km away. Five 15 cm alarm bells were located
on every floor. A P.A. system had also been installed, with four speakers on every floor,
plus an additional speaker on every second level of the staircases for fue department use.
The staircases contained vents at the top floor level. Exit doors at the bottom of the
staircases opened automatically with the activation of the alarm. Smoke and heat
detectors were found throughout the building. The elevators had to be recalled manually
to the ground floor by the two superintendents who lived in the building. The control
panel was located in the lobby and contained a fire list of occupants who would need
assistance in case of an emergency. In the past, there had been few false alarms, generally
only 2 or 3 per year.
Occupants

       At the time of the evacuation, approximately 325 people lived in the building. As
many as 33 apartments were vacant on the day of the drill. Close to 20% of the occupants
were seniors and about 15 families lived in the building. According to the fue list, there
were 16 people with limitations residing in the building.

Procedure

         Four days before the evacuation drill took place, occupants of the building
received a memo informing them about the exercise. The researchers also met with older
occupants and people with limitations to discuss the procedure they should follow during
the drill. About 14 people were met prior to the drill.

        There were 30 video cameras used to record the evacuation, one in every corridor,
in the hall and at each level in both staircases to give a complete overview of the
occupants' movements. The cameras were turned on shortly before the alarm began and
were not turned off until the alarm ended.

        The Gloucester F i e Department agreed to participate in the drill. A strategy was
discussed with the fuefighters prior to the drill. The hypothetical fue was staged on the
fourth floor near Exit B. The firefighters were to move throughout the building
evacuating occupants with mobility impairments from the 3rd, 4th and 5th floors to the
ground level, using the elevator. Upon receiving the alert call at the fue station, the
information was dispatched to the fuefighters who where waiting a few streets from the
building. They headed towards the building in two fue trucks, with sirens on and a rescue
unit aboard, 4 min after the call.

       Questionnaires were distributed after the drill to all apartments and the
respondents were asked to place the completed questionnaires in a box near the front
entrance.

        -
Results Building C

        The evacuation drill took place on December 17, 1994, during the morning so that
there would be daylight. Perfect winter conditions were observed, with heavy snow
coming down and a temperature of -l°C. One of the superintendents was the designated
"alarm puller" for the building. He proceeded to pull the alarm on the fourth floor. This
started the evacuation at 10:41:07. The alarm systems activated as anticipated. The exit
doors at the bottom of the stairs opened automatically. A superintendent recalled the two
elevators to the ground floor.

        The alarm ended at 11:00:59, 1952 min after the start, when the fue chief gave the
"all clear" signal to indicate that the evacuees had reached an area of safety. The P.A.
system was used only to inform the occupants that the drill was over. The occupants were
then allowed to return to their apartments.

        Several firefighters in full uniform arrived on the scene in two f ~ trucks and a
                                                                               e
rescue unit truck. The fust firefighters were seen entering the building at 10:50:15,8:38
min after the alarm began. As planned, they knocked on doors and checked on the
occupants of the 4th, 5th, and 6th floors. They were also seen using the elevators and
staircases to travel from floor to floor.

Behaviour of Occupants

        A total of 95, or 29%, of the estimated 325 occupants were seen throughout the
evacuation drill. Several of these occupants displayed a variety of behaviours such as
appearing in the corridor but not evacuating, entering the building during the alarm or
returning to their apartment while the alarm was still ringing.

         Six occupants appeared in the corridor during the evacuation, but did not
evacuate. Three of them were senior women with limitations. The fust one had a look in
the 8th floor corridor 5:01 min after the start of the alarm,stayed for 1 rnin re-entered her
apartment, and looked again 18:50 rnin after the start of the alarm for only 25 s. The
second woman was located on the 3rd floor, looked in the corridor for 25 s, 10:08 rnin
after the beginning. On the 7th floor, 10:30 rnin after the start of the alarm, the third
woman looked into the corridor for 1:52 min. The other three people who were seen
looking into the corridor were men, aged 30.50 and 60 years old and none of them had a
limitation. One of them, situated on the 4th floor, was seen 9:20 rnin after the start for 5
s. A second one was seen for 10 s on the 7th floor, 10:57 rnin after the alarm started
ringing Finally, 12:15 rnin after the start, on the 8th floor, a man looked into the corridor
for 2 s.

        Three people appeared to he entering the building during the alarm. One couple
coming back from the grocery store entered the building briefly to drop off their purchases
but evacuated immediately after. Another man who had left his apartment 1:22 rnin after
the start of the driil re-entered the building and reached his unit 17:37 rnin into the drill,
before the alarm stopped ringing.

       Only units situated in the comers of the building have balconies: Units 11, 12, 18,
and 19. None of the occupants were seen going to their balconies during the drill. The
small number of balconies, as well as the winm conditions, probably prevented the
occurrence of this behaviour.

        Of the occupants seen evacuating the building, 46.3% (n=44) used Staircase A,
37.9% (n=36) used Staircase B, 9.5% (n=9) were occupants of the fust floor who did not
use the stairs and, finally, 6.3% (n=6) did not evacuate. Table 30 reveals that most people
tended to use the staircase located closer to their apartments. For people living in
apartments closest to Staircase B, all but four chose this staircase as their means to exit
the building. The other four people living in a B-side apartment used Staircase A. Of the
occupants living closest to Staircase A, two used Staircase B but all others used Staircase
A. Two people from the A-side tried using the elevators and 5 people from the B-side did
the same. Table 30 indicates the percentage of people who attempted to use the elevators.
After trying unsuccessfully to use the elevators, all of them proceeded to exit by the
staircases.

    Table 30: Occupants' Egress Route According to their Apartment Location




        The exit doors were also used with differing frequencies. Occupants using
Staircase A or B had the option of entering the fust floor and going to the lobby to exit
through the front door or they could also continue straight ahead to the exit on the side of
the building. There were 39 people in Staircase A who chose to go to the side exit,
representing 44.3% of the total evacuees. The other 10 people in Staircase A chose to
exit through the lobby and front door and 11 people from Staircase B did the same, which
means that 23.9% of evacuees used the front door. Finally, 28 people in Staircase B,
31.8% of evacuees, exited through the door on that side.



         The tables below present the occupants' evacuation times. Table 31 displays the
times at 1 min intervals for the full duration of the alarm and Table 32 displays the times at
15 s i n k ~ a l for the first 5:00 min only. Three different evacuation times are calculated
                 s
for this analysis: a) Time to Start, b) Time to Exit, and c) Time to Move. The T i e to
Start represents the amount of time elapsed between the start of the alarm and the moment
the person initially appears from his or her apartment. The Time to Exit represents the
                                                     l
amount of time elapsed between the start of the am and the moment that the person
reaches an exit. The T i e to Move is the difference between these two times. It provides
an indication of the time taken by occupants to travel from start to finish, regardless of
when and where they started their evacuation.

        In calculating the Times to Start in Table 31,95 occupants are used but only 94 of
them are included in the Time to Start statistics (the 95th being the alarm puller who
evacuated immediately), 87 of them are included in the Time to Exit statistics and only 86
have a known Time to Move. This table covers the full evacuation time for the drill and
includes all occupants seen evacuating the building. Table 32 presents data only from the
fust 5:00 min of the alarm so occupant numbers vary for T i e to Start, Exit and Move.
                                              45


                                                          -
    Table 31: Evacuation Timing a t 1min Intervals Full Duration of the Alarm




         As seen in Table 31,71.3% of people seen evacuating had a Time to Start within
the first 6:00 min after the commencement of the alarm and 87.2% started within 11 min.
The fust occupants to start were a group of three: a woman with two young girls (aged 4
and 5) from the 8th floor, 0:26 s after the alarm. The last occupant seen evacuating left
his apartment on the 4th floor, 18:34 min after the start of the alarm. His wife or
roommate had left the unit only 45 s after the start of the alarm. The average Time to
Start for all occupants was 5.19 min but Times to Start range from 26 s to 18:34 min.

        The occupants' Times to Exit are quite varied. The times range from 1:35 to
20: 11 min (excluding the alarm puller who had left the building after 22 s), with an
average Time to Exit of 6:14 min. Most evacuees, 84.1%, reached an exit within the first
11:OO min after the alarm. By the end of the first 7:00 min, as shown in Table 32,71.6%
of the 88 evacuees had reached an exit The Time to Start, the distance to travel,
travelling with children, the age of the evacuee, and other similar factors are likely to have
affected the Time to Exit. The fust person to exit was the alarm puller, 22 s after the start
of the alarm, followed, at 1:35, by a resident of the 2nd floor. The last person to exit, at
20:11, was a man in his forties, with no apparent limitation, from the 12th floor.

                                                                -
            Table 32: Evacuation Timing at 15 s Intervals First 5 0 0 min




         Almost all of the evacuees, 98.8%, had a Time to Move under 5:00 min. The
 average Time to Move was 1:19 min. The longest Time to Move was 6% min; it was the
time it took a young man with no apparent limitation to exit, however, he went down the
stairs, turned back up and then went down again before evacuating the building, which
explains why it took him so long. He was excluded from the speed calculations as his
Time to Move down the stairs was not representative of his travelling speed and of the
actual time it would take him to evacuate. The longest Time to Move, excluding the man
just mentioned, was 4 3 0 min for a senior from the 10th floor who had an age-related
limitation. A senior woman from the 1st floor, who also had an age-related limitation, was
the fastest with a Time to Move of 0 2 1 s. The eight occupants with the shortest Times to
Move were all from the 1st floor. Table 33 shows the various Times to Move, according
to the floor where the occupant started. There is a correlation between the Time to Move
and the starting floor (r = 0.63, p=O.OO). Occupants starting their evacuation on higher
floors had longer Times to Move than those starting on lower floors.

                       Table 33: Time to Move vs. Starling Floor




Time Comparisons Accordine to Gender

       The ANOVA test was used to determine if any differences exist between males
and females for evacuation times. The results of this test, as well as the mean times for
each gender, are presented in Table 34.

                  Table 34: Time Comparisons According to Gender




        The female occupants of the building had faster Times to Start and Exit than the
males, while the male occupants had faster Times to Move than the females. None of
these differences, however, were significant. It seems that gender did not play a
significantrole in the evacuation timing.
Time Comparisons According to Ace

         Ages of the evacuees were estimated by the researchers when watching the
videotapes. The evacuees were then placed in one of the seven age categories, based on
this estimate. Table 35 presents a frequency table for the number of occupants in each age
category. Table 36 presents the results of the ANOVA test comparing different age
categories for the T i e s to Start, Exit and Move.

               Table 35: Number of Occupants in Each Age Category




                                                                -
            Table 36: Time Comparisons According to Age Building C




                    * = a significant result
Note: Age categories 1.2 and 3 had less then 5 subjects, therefore these groups were
      omitted from the analysis since such a small sample would not produce valid
      results.
       The ANOVA test initially included all of the age categories. It revealed significant
differences for Times to Start among all groups. The remaining analyses were conducted
only with groups containing more than five occupants. This included Groups 4 , 5 , 6 and
7: teenagers, adults, older adults and seniors. The results for Times to Start, Exit and
Move were not significant except when comparing the Times to Start and Exit of
teenagers and seniors (Groups 4 and 7) and the same times for adults and seniors (Groups
5 and 7). The seniors took significantly longer to start their evacuation and to exit the
building than the teenagers (aged 13-19) and the younger adults (aged 20-39).

Time Comparisons According to Limitation

        An occupant's Time to Start, Exit, or Move may be affected by the presence of a
limitation. The occupants with limitations were compared to the occupants without
limitations to determine if diferences in evacuation times exist between the two groups.
The results of these comparisons are presented in Table 37.

                       Table 37: Time Comparisons According to Limitation
             .....................................................................
             ...........                             .........................................................                              .:...... u... : ::... *.. . :,A:: .. ::...!..............: ...:. ....?..
                                                                                                                                                          ,         ........ :      ::               :     .: : ::
            .~~;i;il;i~e~~;~mesi:,:.:i12~;:~~~1mItafiOn:
            ............. .
             .:.~ . . . .
               ..                                                                              .,pJ@;Lrrm,&fioni:
                                                                                                              ii+p;-&ueGt
            ............ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . .           . . ..............
                                                                                                                                                                .
                    Time to Start                                                   6:25 (N=28) 4:51 (N=66)      pd.14
                    Time to Exit                                                    7:42 (N=25) 5:40 (N=63)      p=0.07
                    Time to Move                                                    1:25 (N=24) 1:17 (N=62)      p=0.59

         In this building, the occupants with limitations were slower than the occupants
without limitations to Start, Exit and Move, however, the differences were not significant
The mean T i e to Start for occupants with limitations was 6:25 min, while those without
limitations had a mean Time to Start of 4:51 min. Occupants with limitations reached an
exit later than occupants without limitations; the mean times were 7:42 rnin and 5:40 min
respectively. The average Time to Move was similar for people with a limitation and
those without, 1:25 min and 1:17 min, respectively.

Speed of Occuvants on the Stairs

        Seventy-six occupants were used in the speed calculations. Occupants who
appeared in the conidor but did not leave were omitted from the sample since they did not
use the stairs during the drill. The mean time to descend one floor was 12.14 s. The mean
speed on the strairs of all the occupants was 0.95 m/s.

       The mean times per floor for each gender and the speed in m/s are compared in
Table 38. Males show a significantly faster time than Females. Overall, unlike what was
demonstrated in previous building evacuations, gender appears to have had an effect on
the speed in the stairs. There was a higher proportion of female evacuees who were
seniors, compared to the male evacuees. This may explain why the mean speed per floor
was significantly greater for females since most of them were seniors.
  Table 38: Gender Comparisons while Descending One Floor and Speed in Stairs




                * = a significant result
       The same age categories as used earlier were considered to compare speeds.
Table 39 shows the mean time per floor and the mean speed in m/s for each age category.

             Table 39: Speed of Occupants According to Age Category




            Table 40: Speed on Stairs Comparisons According to Age Category




                         4.5-67            I   *p=O.OO I   Yes
                          4.5.6            I   *n=O.01 1   Yes
                          5,6,7                *p=O.Oo     Yes
                           4,5                  p=O. 10    Yes
                           4,6                 *p=O.O2     Yes
                           4.7             I   *p=O.OO (   Yes
                            5-6            I   *p=O.O2 I   Yes
                                        I      *p=O.OO I
                            5,7
                            6.7         1
                    * = a signifcant result
                                               *D=O.OO 1
                                                           Yes
                                                           Yes      I
       An ANOVA was conducted on the data to determine if any significant differences
in speed existed between the age categories. Table 40 presents the comparison for the
Speed in mls. Initially, all age groups are compared against one another and then, only
groups with five or more occupants are used to produce statistically valid results.

        The comparison of all age groups' average travelling speeds yields significant
results. This means that there are differences in travelling speeds between people of the
different age categories. It is important, however, to keep in mind that only Groups 4.5,
6 and 7 (teenagers, adults, older adults and seniors) had five or more subjects. When
these four groups are compared, the difference in speeds is significant at pd.00. Two
groups were found to be significantly different from each other: older adults and seniors,
Groups 6 and 7, were significantly slower than the teenagers and younger adults, Groups
6 and 7.

        It is expected that the presence of a limitation would affect an occupant's speed on
the stairs. Table 41 shows the results of the ANOVA test comparing the mean speed in
m/s for occupants with or without limitations.

          Table 41: Speed on Stairs Comparisons According to Limitation



                  [ NO Limitation I 1.09 (N=55)             I           1
                     * = a significant result
        The occupants with limitations were significantly slower than the occupants
without limitations when travelling down the stairs. The speed comparison shows that
occupants with limitations travelled at an average speed of 0.57 mls and the mean speed
for those without limitations was almost twice that speed, 1.07 mls. The difference
between the two groups' respective speeds is significant at p 10.00. Overall, the results
show that occupants with limitations are slower than occupants without limitations when
travelling down the stairs.

       The speeds on the stairs of occupants alone and in groups were also calculated and
compared. No significant difference was found (p=0.10). The occupants travelling alone
had an average speed of 1.00 mls, while the occupants travelling in groups of two or more
had an average speed of 0.87 mls.

Questionnaires

       Eighty-two questionnaires were returned to the researchers, representing 45.6% of
the occupied units in the building. Of these 82 people, 49 stated that they were present for
the M 1 and all of those 49 stated that they had heard the building fire alarm.

        As many as 20% of the respondents to that question (9 people), however, had
difficulty hearing the alarm from their apartments. These nine people were distributed
throughout the building and it is not possible, from the responses, to identify specific
zones in the building where the alarm was not loud enough. As well, 6.3% or 2 people
                                                                                    l
found that the sound of the alarm was not loud enough in the staircases. People of a l age
groups perceived the alarm as not being loud enough.

                  l
         Almost al of the people, 96.2%. were in their apartments at the time of the alarm,
mostly in their living rooms. In response to the sound of the alarm, the occupants listed
several activities that they engaged in as shown in Table 42. For instance, 35 of the
49 respondents listed getting dressed as one of their first three activities. Leaving the
building, looking in the corridor and gathering valuables were other frequent choices of
activities.

              Table 42: Activities Undertaken After Hearing the Alarm




         According to respondents, the average time required for them to begin to evacuate
was 4 2 4 min after the start of the alarm, which is an underestimation of 55 s from the
observed Time to Start (5:19). Only one of the respondents tried using the elevators to
evacuate, but then proceeded to use the stairs, just l i e 36 other respondents who used the
stairs to evacuate. Twelve of the respondents said that they did not use either the stairs or
the elevators which leads to the conclusion that they did not exit the building since none of
them had balconies on which they could exit. An important number of the respondents,
31.3%. revealed that they do not always leave the building went they hear the sound of the
alarm.

       The average age of the respondents was around 45 years old. A majority of
respondents were females, 53.1%, and 46.9% were males.

Alarm Audibility Measurement

       Table 43 presents the measures in dBA of the background noise and of the
background noise with the alarm for different locations. The alarm should have a
minimum of three spectral elements in order to meet audibility requirements. The two last
columns in Table 43 presents the number of spectral components reaching the audibility
design window. The numbers in parentheses represent the number of spectral components
that exceeded the audibility design window, which confirms the intense sound pressure in
certain locations.

Table 43: Background Noise Levels and Detectsound Analysis of Alarm Audibility
       The sound pressure was acceptable in Units 601,819 and in the living rooms of
Units 1010 and 1209. Only the livinx rooms of Units 910 and 1209 had an acce~table
sound pressure for occup&ts over 55, or with mild hearing loss. The sound of &e alarm
                       l
was over-powering for al the corridor and staircase measurements.

        Figures 13 and 14 illustrate some of the results obtained with Detectsoundm for
the fue alarm analysis of the Gloucester building. In these figures, the x-axis represents
the frequency content in Hz. The y-axis gives the sound pressure level (dBSPL).
Measurements of the spectral content of the background noise are represented by the
horizontal continuous line and the alarm level by the vertical lines for each 113 octave
band. For the buildings studied, the alarm is often lost in the background noise making it
difficult to extract its exact spectral content. The diagonal hatched area represents the
audibility design window into which the alarm should have a minimum of three spectral
elements in order to meet the requirements of audibility. The spectral elements within the
audibility design window are represented by thi~ker     lines, as shown in the figures.




                                     Frequency (Hz1
             Figure 13: Audibility Design Window for a 55 year old man
                                             -
                     with normal hearing Bedroom of Unit 819

        In the living room of Unit 8 19, four spectral components are within the audibility
design window for a 55 year old man with normal hearing, as shown in Figure 13. The
four spectral elements reaching the design window might be loud enough for normal
hearing subjects but three of them would be missed by people with mild to severe hearing
loss (from 1 to 8 kHz) as shown in Table 43.
                                     Frequency (Hz)
            Figure 14: Audibility Design Window for a 55 year old man
                                               -
           with mild to severe hearing loss Corridor in Front of Unit 1219

        Figure 14 shows that the alarm does not meet the audibility requirements in the
corridor with only two spectral elements within the audibility design window. There are
however, thirteen spectral elements exceeding the design window and a series of high
level/high frequency elements, confiiing the intense sound pressure at that location.

Conclusions

        The evacuation d i l in this highrise building lasted 19:52 min. Most of the
                          rl
occupants who appeared during the drill evacuated using the stairs. There was a good
participation rate for this building, even though the winter conditions could have deterred
a few.

        The occupants who did evacuate, tended to travel in familiar directions. Most
occupants used the staircase located closest to their apartments, hut in 7.6% of cases,
occupants living on one side used the opposite staircase. This is probably due to the fact
that the elevators are closer to one staircase and it is a familiar route. The front entrance
was used less frequently than the two side exits, possibly because the staircases led
directly to the side doors, which opened automatically when the f r alarm was activated.
                                                                      ie

        The average time for the occupants to start evacuating was 5: 19 min after the start
of the alarm. The average Time to Exit the building was 6:14 min and the average T i e
to Move from start to finish was 1:19 min. The lack of crowding in the staircases
contributed to this fairly quick Time to Move.

       The results of the statistical analyses provide some insight into the differences
between occupants of different genders, ages and levels of ability when comparing their
Times to Start, Exit and Move. There were no significant differences in evacuation times
between male and female occupants, although females tended to Start and Exit earlier
while the males tended to Move faster. In fact, men had a significantly faster average
speed of evacuation.

        Another finding was that when all the age groups were compared to one another,
they were significantly different in some respects. Many of the age groups, however, had
very small sample sizes and could not be included in some of the comparisons. For this
reason, some of the results must be interpreted with caution. When considering only the
groups containing more than five people, it was found that the evacuation times were not
significantly different between groups, with the exception of the Start and Exit Times
which were significantly slower for seniors than for young adults and teenagers. The
average speed of seniors was significantly slower than that of the teenagers, younger
adults and older adults; older adults had a significantly slower mean speed than teenagers
and younger adults.

        The differences in times for occupants with or without limitations were also
significant. This means that those with limitations were slower than those with no
limitations. The mean travelling speed for all occupants was 0.95 mls.

       The responses from the questionnaires revealed that most occupants found the
alarm to be loud enough in their aDartments, in the corridors and in the staircases.
Nevertheless, an analysis of the al&n showed that the alarm is not readily audible inside
some of the apartments visited, although, the alarm appears to be overpowering in
corridors and staircases.

       In response to the alarm, respondents mentioned in the questionnaire that they
tended to get dressed, gather their valuables andlor leave the building. This is consistent
with the actual observed behaviour of the occupants. Many people also admitted that they
don't always leave their apartment at the sound of the alarm.


                         -
5.0 JOINT RESULTS ALL THREE HIGHRISE BUILDINGS

         The analyses presented in this section compare the three buildings using only the
first 5 min of the evacuation, since the alarm in Montreal only lasted 5:18 min. Further
analyses are performed using Calgary and Gloucester data, because only these two
buildings had a complete evacuation with local fr department participation.
                                                  ie
5.1 Comparisons of Mean Times to S a t Exit and Move
                                  tr,

        The mean Times to Start, Exit and Move for occupants with Times to Start under
5:00 in any of the three buildings are presented in Table 44. Table 44 also shows the mean
times for the full duration of the alarm for the Calgary and Gloucester occupants. Table
45 shows the results of the ANOVA tests that compare the Times to Start, Exit and Move
for the three buildings.

                       Table 4 . Mean Times to S a t Exit and Move All Buildings
                              4                 tr,                                                                                                                -
          . ..... .
     ................... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...............................................................................................
                                                                                                                                                 . . . . . . .
        .......................
                ~:::~:mm
                     . .-
                     ~:!:;c::
      ...............
     ......................
     :........... . ...:   ;;-~~!~a:ii$~rs~~&nutes;~;ii~~~;;j~ ~
                                                ~j~~>;~.
       .:... .......................................................................................
      .......................... .............. ...
      ..................................................
                                                   ::
     ............................... :..... .........................                .....~.I....... . . ~ . . .
                                                                                                                   .. . . . . . .
                                                                              . .. . . . . . . . . . . .............................. .................................
                                                                                                                                                                                          .....................................
                                                                                                                                             ................. : : :......:~ $.. fi ;:.?c::
                                                                                                                                          ~::..~                           .....
                                                                                                                                                                                                                                  &~&~&~$~
     .....................
      ...................... ,,(fi,fl:s)ig.f<iil<::
          . .
      ............
      .. . ..... . ....                                      :Montreal!;
                                               ..............................      G1@ucester
                                                                       ,,;:C&tlg&ry{;;
                                                                                    . .              :Gpwresf:"
                                                                                            iiCCdgarJr,i,
            Time to S a t
                     tr                                               1:30                             1:37                             211                             2:48                              5:19
            Time to Exit                                              3:07                             3:Ol                             3:36                            4:34                              6: 14
            Time to Move                                              1:43                             1:25                             1:26                            1:46                              1:19

   Table 45: Comparison of Mean Times to S a t Exit and Move All Buildings
                                          tr,                                                                                                                                            -




* = a significant result
Time to Start

         The Time to Start represents the amount of time elapsed between the start of the
alarm and the moment an occupant first appeared from hidher apartment When looking
at the full duration of the alarm for all the buildings, the mean Time to Start was 1:30 min
for the Montreal building, 2:48 min in Calgary and 5:19 min in Gloucester. This difference
is significant at p=O.OO. Thii fast mean Time to Start for Montreal is probably caused by
                                    hi
the short duration of the alarm. T i allowed only the quicker occupants to evacuate,
consequently affecting the mean. The late staters are clearly under-represented in the
Montreal data. The much longer Time to Start observed in Gloucester was expected,
since it was the only evacuation to be conducted during the winter. Occupants generally
required a longer time to get ready, mainly because they had to put on winter clothing
before starting to evacuate.

       The mean Time to Start of the Montreal and Calgary buildings are similar when
only the first 5:00 min are considered. The mean Time to Start in Montreal remains the
same at 1:30 min but the mean for Calgary decreases to 1:37 min. The obtained plevel of
p=0.73 indicates that when comparing the occupants of these two buildings over a similar
time span, no significant difference is noted. This implies that the occupants of the two
buildings are quite similar in their reaction to the alarm. Even when considering only the
fvst five minutes following the activation of the alarm, residents of the Gloucester building
were significantly slower than Montreal and Calgary evacuees.

Time to Exit

        The Time to Exit represents the amount of time elapsed between the start of the
alarm and the moment that the person reaches an exit When considering the full duration
of the alarm, the mean Time to Exit was 3:07 min in Montreal, 4:34 min in Calgary and
6:14 in Gloucester. The slower mean Time to Exit in Calgary and Gloucester can
probably be accounted for by the fact that there was a greater number of late starters.
Occupants who take a longer time to leave their apartments will obviously not reach an
exit until later into the alarm. The difference between the three means is significant at the
pd.00 level. The diference, however, is not significant between the Montreal and
Calgary results. The reason this difference is not significant is probably due to the fact
that some of the occupants in Montreal, who were used in this calculation, reached an exit
after the alarm had shut itself off, making the times longer and more comparable to the
Calgary data. Once again, since Gloucester residents started, in general, later than
residents of Montreal and Calgary, it is not surprising that their Times to Exit were longer,
and significantly so when compared to Montreal results.

         When only occupants with Times to Start under 5:00 min are considered, the mean
Times to Exit for Montreal, Calgary and Gloucester become 3:07 min, 3:01 min and 3:36
rnin, respectively. There is no significant difference between these results, which indicates
that occupants who start at similar times also exit at similar times, for all three buildings.
Even though for the first five minutes, Montreal residents had significantly faster Start
Times than Gloucester residents, the difference between their T i e to Exit is not
significant. This could imply that Gloucester residents have faster Times to Move than
Montreal residents, which is discussed in the next section of this report.

Time to Move

         The T i e to Move is the difference between the T i e to Start and the Time to
Exit. It provides an indication of how long it takes occupants to travel from start to finish,
regardless of when and where they started their evacuation. In Montreal, the mean Time
to Move for the full duration of the alarm was 1:43 rnin, while the mean in Calgary was
 1:46 min and the mean in Gloucester was 1:19. It is of interest to note that both the
Montreal and the Calgary buildings are 15 storeys high, while the Gloucester building is
12 storeys high, possibly explaining why the mean Time to Move is shoaer in the
Gloucester building. These three means, however, are quite similar and do not yield any
significant diference. The occupants take about the same amount of time to move from
start to finish, regardless of which building they live in, or of the time at which they leave
their apartment.
        When late starters are eliminated, the mean Time to Move for Montreal remains
the same at 1:43, but the mean for Calgary decreases to 1:25 min while the mean for
Gloucester increases to 1:26. The decrease in Calgary might be an indication that the late
starters were slower movers in general and took a longer time, not only to leave their
apartments, but also to travel through the building, while the results from the Gloucester
building seem to indicate the contrary, though the change is so small that it is inconclusive.
There is still, however, no significant difference among of the buildings, indicating that, as
a whole, occupants still take about the same amount of time to move from start to finish
for buildings of approximately the same height

5.2 Speed Comparisons According to Building

        The mean time for occupants to travel from floor to floor and their mean speed in
m/s were calculated for each building. These times and speeds were then compared to
determine if any differences exist between the buildings. The mean time to descend one
floor was 11.27 s in Montreal, 10.18 s i n Calgary and 12.14 s in Gloucester. These means
do not differ significantly since the p-value obtained when comparing the three buildings is
p=0.30. The mean speed of occupants in Montreal was 1.07 d s , the mean speed in
Calgary was 1.05 m/s and the mean speed in Gloucester was 0.95 m/s. Overall, it appears
that there is little difference in evacuation speeds between the Montreal and Calgary
buildings, but that people in the Gloucester building were travelling at a slightly slower
speed (p=0.22). Since the Gloucester evacuation took place in the winter, it is not
surprising that people had a tendency to travel slower, their winter clothing affecting their
mobility and speed.

5.3 Time and Speed Comparisons Regrouping the Three Buildings

        To adjust for the problem of small sample sizes, the occupants of the different
buildings can be combined in order to place a larger number of people in a particular
group, as long as the different buildings are not significantly different from one another,
                                -
see Table 45. The three buildings cannot be combined for the T i e to Start over the fl  ul
duration of the alarm as they are significantly different from one another in that respect
When considering only the first 5 minutes, the Montreal and Gloucester buildings cannot
be combined because they are significantly different, however, the Montreal a n d ~ a l ~ a r y
building do not show a significant difference which means that they can be combined; it is
the same for Calgary and Gloucester. Similarly, Montreal and Gloucester buildings cannot
be combined for the T i e to Exit over the full duration of the alarm as they significantly
differ on that aspect, but either of these two buildings can be paired with the Calgary
building for this case.

Time and Speed Comparisons According to Gender

        No significant difference was found between men and women for their evacuation
times when combining the buildings to provide a larger sample size. The only significant
difference found was between their mean speeds, when combining the Montreal and the
Gloucester buildings (p=0.01). It was shown that men (mean speed = 1.07 d s ) were
faster than women (mean speed = 0.90 d s ) . When all three buildings were combined,
however, the difference was not judged to be significant (p=0.06).

Time and Speed Comparisons According to Age

        Many age-based comparisons involving seniors resulted in a significant difference.
When comparing age groups over the full duration of the alarm, for the combination of the
Calgary and the Gloucester buildings, it appears that seniors had a later mean Time to Exit
than teenagers, younger adults and older adults. These results are shown in Table 46. No
analysis was completed for the three younger age groups because of an insufficient
number of subjects in those groups, which would not give valid comparisons. The mean
Times to Move of the various age groups were not significantly different (p=0.66). The
                                                                             -
mean Times to Move could not be studied for anv combination of buildings as the
buildings were all significantly different from each other in that respect

                                                -
    Table 46: Comparison of Mean Time Exit Full Duration of the Alarm Calgary    -
                        and Gloucester Buildings Combined

                     Age Groups       1      Mean Time to Exit           I


                 I      6,7           I             *p=0.04              1
                  * = a significant result
         All three buildings can be combined for comparisons when considering only the
first five minutes following the start of the alarm. When comparing the different age
-     ..                      -                 -
u o u ~ s with the three buildings combined. no significant difference is found. When
          .
comparing the age groups of people from only the Montreal and Calgary buildings,
                                  -
however. the Time to Move is significantly different between age groups (p=0.03). More
                                                                  - - -
specifically, it appears that young adults had a significantly shorter mean Time to Move
than children (p=0.04) and older adults (p=0.01). Only these three groups could be
compared since other groups contained less than five subjects and would not produce
valid comparisons.

        For the three buildings combined, all age groups were compared for their mean
speeds, with the exception of babies and small children, since these groups contain less
than five subjects. Significant differences between age groups' mean speeds were found
as shown in Table 48. When excluding the babies and small children, it was shown that
generally, the younger the age group, the faster its respective mean speed (see Table 47).
                                            61

                                                        -
        Table 47: Mean Speeds for each Age Group AU Buildings Combined




        Table 48 shows the results of the comparison tests performed between the
different age groups. It was found that seniors were significantly slower than all other
groups. Older adults were also significantly slower than younger adults and children. It
may appear surprising to some that older adults are significantly different from younger
adults, while not being ~ i g n ~ c a n tdifferent from teenagers even though the absolute
                                         ly
mean speed difference is more important between older adults and teenagers than between
older and younger adults. This could be due to the fact that there was only a small
number of teenagers, therefore, it is not certain, at 95%, that the same difference would be
observed again if another sample were to be taken. The important number of young adults
available for analysis provides a better assurance that there is a 95% certainty that ts
                                                                                       hi
mean speed difference would re-appear if the experiment were repeated.

                                                                      -
     Table 48: Comparison of Mean Speeds for each Age Group All Buildings
                                  Combined

                                                      Age Group                          1



  * = a significant result
Time and Speed Comparisons According to Limitation

       The buildings were combined to obtain a bigger sample of people with limitations,
which allows for a better comparison between people with a limitation and those without,
for such variables as the mean Times to Start, Exit and Move, as well as the mean speeds
when travelling down the stairs. Table 49 presents the results for the full duration of the
alarm, while Table 50 presents the results for the occupants with Times to Start under
5 0 min.
 :0

                                                           -
 Table 49: Time Comparison According to Limitation Full Duration of the Alarm



                      Al 3 bldgs
                       l           I                  I                  I               I




* = a significant result
                 Table 50: Time Comparison According to Limitation       -
                       Occupants with Times to Start 15:00 min




* = a significant result
        The results in Table 49 indicate that people with limitations have slower Times to
Start, Exit and Move, as well as having slower travelling speeds than people without
limitations. The difference in speed is significant below p=0.05, for any building
combination. This implies that people with limitations take longer during an evacuation to
go down the stairs than people without limitations. People with limitations were also
shown to have a significantly later mean Time to Exit, as well as a significantly longer
Time to Move, when combining the results of the Calgary buildings with either the
Montreal or Gloucester building.

         Table 49, however, combines data that is not obtained over similar lengths of time,
which is why results from only those people starting within the fust 5 minutes were also
analyzed in order to use Calgary and Gloucester data that was comparable to the Montreal
data. Results for those analyses are shown in Table 50. The occupants with limitations
are still slower than the occupants without limitations in every respect, but many of these
results are not significant. The Time to Exit of people with limitations, however, was
significantly longer than that of people without limitations (pd.03).

       Overall, it would seem that the longer the evacuation drill lasted, the more likely it
was to obtain significant differences between occupants with or without limitations.
Occupants with limitations may take longer to start evacuating and may well be under-
represented in shorter drills.

Composition of Evacuation Groups

        During evacuation drills, many people travelled with others, rather than alone.
Most often, the groups were formed of people living together who evacuated at the same
time, but some groups also formed in the corridors or in the staircases as people moved
towards an exit The videotapes revealed that many of the occupants in these buildings
prefer to travel in groups. In this study, a group was defined as two or more people who
travel together for the majority of the time that they spend evacuating. People who mainly
travel alone, but join a group for a short period of time, are not considered to be part of a
group. Table 51 presents the number of occupants who travel alone and the number of
those who travel in groups, while Table 52 presents the composition of the evacuation
groups.

                                Table 51: Distribution of Evacuees All Buildings                                 -
          ......                  .:.
      .............. . . . . . . ...:..::::. . . .    ...........................               . . .           .     .    .     .     .     .
                                                                                .....................................................................
           B"ming,iij2ii/j
     ............... . .
      ............
     .........
             .                              ,Occup&:Ec::Exi*ngFq&f$Ups
                     i:s;Occupan~:;Exrti.ng:iA,oneli,jj::
                                     ................ . .
                                                   .
          Montreal              10                        26
          Calgary               11                        16
          Gloucester            64                        31
          Total                 85                        73
          Percentage          54.8%                     46.2%
        In the Montreal and Calgary buildings, a majority of people travelled in groups. In
the Gloucester building, however, over two-thirds of people travelled alone. When the
data from the three buildings are combined, the total number of occupants travelling in
groups of two or more people is 46.2%,while 54.8% of evacuees travelled alone.

                                                                -
             Table 52: Composition of Evacuation Groups A11 Buildings




        In all three buildings, the most frequent type of group consisted of 2 adults. The
other groups were mainly composed of a mix of children and adults. Only two groups of
children travelling without adults were identified: one composed of two 16 year olds and
one composed of a teenager and a younger child.

       In most cases, the groups of adults consisted of people of similar ages and abilities.
These groups usually travelled fairly quickly down the stairs. This fast average pace for
groups could be due to the fact that many seniors did not travel in groups when
evacuating.

        The groups with adults and children displayed a variety of combinations. In many
instances, there were older children in the group. These children tended to travel very
quickly down the stairs and were often seen several steps ahead of the others. Younger
children tended to hold the handrail or their parent's hand, and moved very slowly down
the stairs. Only two children had to be carried down the stairs. One was in Calgary,
where a young mother carried her baby to an exit. The other was in Gloucester, where a
neighbour carried down a small girl. In a few of the cases, it appeared as if older adults,
perhaps grandparents, accompanied the children. Table 53 shows the mean times and
speeds for people travelling in groups and for people travelling alone.
                                          -
 Table 53: Mean Times to Evacuate Occupants Alone vs. Occupants in Groups                      -
                             All Three Buildings




         Occupants travelling in groups in all three buildings tended to start evacuating
earlier than others, they had an earlier mean Time to Exit, and in both Montreal and
Calgary, they had a shorter mean Time to Move. In Gloucester, people travelling in
groups had a longer mean Time to Move. When comparing people travelling in groups
and those travelling alone, only one significant difference was found: when considering
the full duration of the alarm and combining buildings in Calgary and Gloucester, people
travelling in groups had a significantly shorter T i e to Exit than people travelling alone.

        People in groups tended to travel at the speed of the slowest member. For groups
that included small children or people with limitations, the pace might be slower than it
would be if occupants had travelled alone. It still appears, however, that people prefer this
slower pace to travelling alone. People travelling in groups in Montreal and Gloucester
had slower evacuation speeds than those travelling alone, but the opposite situation was
observed in Calgary.

Alarm Audibility

        The alarm audibility was not shown to have an important effect on the T i e to
Start of occupants when comparing the Calgary and Gloucester buildings. Occupants in
Calgary had a much sooner %me to Start (2:48 compared to 5:19 in Gloucester),
however, the alarm audibility measurements in that building showed that the alarm level
was not acceptable in any of the units tested. Ln Gloucester, the alarm level was
satisfactory in approximately half of the units tested.

        There is some contradiction in the alarm audibility measurements and occupants'
perception of the alarm as reported in the questionnaires. Even though the alarm level in
apartments was shown to be inadequate by the measurements, in Calgary, all but one
occupant heard the alarm, whiie all the occupants of the Gloucester building heard the
alarm. Respondents did not say, however, if they could hear the alarm from their unit, or
if they heard it when travelling in the conidors and stairs, which would not have been
surprising since the alann was measured to be too loud in all tested corridors and
staircases of both buildings. When asked to evaluate the audibility of the alarm, 8% of
Calgary occupants reported it was not loud enough and 20% of Gloucester occupants had
a similar opinion.

        It is obvious that a factor, other than the alarm audibility, affected the Time to
Start of occupants since, even though the alarm was louder in Gloucester, Calgary
occupants started evacuating earlier. The most plausible explanation is that people in
Gloucester had to get dressed before evacuating, which, delayed their T i e s to Start,
since the evacuation in Gloucester took place during the winter, unlike the one in Calgary.


6.0 JOINT RESULTS: HIGHRISE AND MIDRISE BUILDINGS

         A first study collecting real data on time and movement during evacuation drills in
four buildings was undertaken by NRCC and CMHC in 1993-94, to study evacuation
drills in midrise apartment buildings with mixed abilities occupants. The results of the
second study on highrise building evacuations were detailed in this report In this section,
the results of both studies are compared in an attempt to identify similarities and
differences in the evacuation behaviour of occupants in midrise and highrise buildings.

         The four midrise buildings of the first study were situated in Ottawa, Montreal,
                                                              -
North York (Toronto) and Vancouver. The midrise buildings were com~osed 6 to     of
7 storeys, with elevators, and the drills were conducted in the early evenkg during the
Summer of 1993. In all other respects, the evacuations were conducted similarly to the
drills in the highrise buildings.

       Table 54 shows the mean evacuation times for occupants of both highrise and
midrise buildings. It also includes the mean speed while travelling in the stairs and the
average time needed to go down one floor in each of the buildings.

                                                                   -
              Table 54: Mean Times to Start, Exit and Move All Buildings

 Type of        Lourtiun of                  Mean 'limes               M a n Time       Mean
 Building        Rullding          Start        Exit       hfove       per Floor        speed
                                   minx        mins        min:u           Y             mls
Highrise       Monueal             (1:30)*     (3 07)'      1.43         1 1 27     I    1.07




* Times for a shorthen evacuation of 5: 18
** Estimated dimensions of stairs
***   Unable to determine srair dimensions

       Two of the midrise buildings, Montreal and Toronto, had Times to Start and Exit
much longer than the other buildings. This was explained by the fact that some residents
could not hear the building fire alarm from their apartments and only started evacuating
when the firefighters knocked on their doors. As a results, their Times to Start were much
later than could be expected and, consequently, the mean Time to Exit in these two
buildings was also much longer than anticipated. It took about 25 min to totally evacuate
these two buildings. The Time to Move of these two buildings, however, was not affected
by the delay to start the evacuation.

        As can be expected, on average, the midrise buildings were shown to have
significantly shorter Times to Move (p=O.OO). The Time to Move for the Gloucester
highrise was not shown to be significantly different from the Montreal, Toronto and
Vancouver midrises.

       Residents of highrise buildings appeared to travel significantly faster than residents
of midrise buildings (~4.00).T i can be explained by the fact that in the midrise
                                 hs
buildings, a higher proportion of elderly residents participated in the drills, and elderly
occupants tend to have slower travelling speeds.

        From the tapes of the evacuations, observed behaviours showed that these slower
times could be explained by another reason. Occupants of the midrise buildings were seen
stopping frequently whiie going down the stairs, slowing down to chat with their fellow
evacuees. People in midrise buildings seemed to know their neighbours and delayed their
evacuation as they stopped to talk. This type of behaviour was seldom observed in the
highrise buildings.

       Results of the midrise evacuations showed that there was a problem with the fire
alarm audibility. For the Montreal, Toronto and Vancouver midrise buildings, between
17% and 25% of the occupants mentioned in the post-evacuation questionnaire that the
alarm was not loud enough. This problem not observed in the Ottawa midrise where
alarm bells had been installed in every unit

        Alarm sound level measurements were taken in each of the highrise buildings on
the day of the evacuation. Measurements confirmed that, in the three highnse buildings,
the alarm had an acceptable number of spectral components for a normal hearing person in
only two of the 18 bedrooms tested. Furthermore, the alarm was not audible inside any of
the units tested in either the Montreal or Calgary buildings, although the sound was
generally over-powering in the corridors and staircases. In Gloucester, the alarm level
was appropriate in some of the apartment living rooms, but was always too loud in the
staircases. The sound levels were never judged appropriate in any of the units for
someone with mild to severe hearing impairment, except for the living rooms in two of the
Gloucester units.
7.0 GENERAL CONCLUSIONS

        The results presented in this study reveal a number of interesting findings. When
comparing evacuation times of the three highriie buildings, Montreal results must be
considered with caution. In Montreal, the fire alarm s t o ~ w dunexwctedlv onlv minutes
after the beginning of the drill, which disrupted the evac&ion andprovidkd incomplete
sample of occupants' behaviour. This unexpected event, however, brought to light the
fact that no occupant started the evacuation after the alarm sound had stopped. It shows
that occupants starting to evacuate are affected by the continuous sounding of the alarm.
As was observed in the Montreal building, people will not start to evacuate if the alarm
stops before they have had a chance to initiate their evacuation. This finding shows how
important it is that, during an emergency, for the fire alarm to ring continuously to inform
the occupants that the problem is not yet cleared and that the emergency procedure is in
place.

       When considering only the fust 5 min of the three drills, the evacuations in
Montreal and Calgary yield very similar results in terms of evacuation times. Gloucester
occupants are behind in some respects because it took them longer to start and exit. The
longer Time to Start in Gloucester can be explained by the fact that, due to the winter
conditions, occupants took longer to get ready, putting on appropriate clothing before
proceeding to evacuate the building. The Time to Move for the Gloucester building,
however, is shorter than in the two other buildings. One explanation is the fact that the
Gloucester building has only 12 storeys, compared to 14 for the other two buildings.

        Results of the statistical analyses showed that gender had very little influence on
the timing of the evacuation for any of the buildings. Age was not shown to have an effect
on evacuation times with one exception: in Calgary, young adults started evacuating and
exited the building significantly faster than seniors. There were too few occupants with
limitations who participated in Montreal to draw any conclusions on their overall impact
Occupants with imitations, however, had a significant effect on evacuation times in
Calgary while there was no such impact on the Gloucester evacuation times.

         Occupants of the three buildings had average travelling speeds that were not
significantly different Gloucester occupants were slightly slower, with an average speed
of 0.95 mls, compared to 1.07 m/s and 1.05 m/s for Montreal and Calgary, respectively.
This can be explained, in part, by the fact that the Gloucester occupants were wearing
winter clothing. The evacuation in Montreal, however, lasted less than 6 min, resulting in
an incomplete subject sample, since slower occupants were not likely to have had the time
to start evacuating before the end of the alarm. There were differences between the
speeds of males and females but none of these differences were significant enough to
support any conclusion on the role played by gender in the speed of an evacuation. In
general, no crowding was observed in the stairs, therefore, occupants were able to travel
at their own pace, their speeds were unaffected by the presence of other evacuees.
        Looking at all three buildings, seniors, and to a lesser degree older adults, were
significantly slower than younger occupants when evacuating. Many of the occupants
with limitations were older occupants, and it was also found that occupants with
limitations were significantly slower than other occupants. Analysis showed that both
older occupants and those with limitations had slower travelling speeds while going down
the stairs.

         Occupants travelling in groups in all three buildings tended to start evacuating
earlier than others. People in groups, however, tend to travel at the speed of the slowest
member, therefore, in groups that included small children or people with limitations, the
pace was slower than it would have been if occupants travelled alone. People travelling in
groups in Montreal or Gloucester were slower than those travelling alone, but in Calgary,
the opposite situation was observed since most groups were made up of two adults. In
Calgary and Montreal, a majority of occupants travelled in groups, however, in
Gloucester, two-thirds of the occupants travelled alone. A majority of occupants used the
staircase closer to their apartments to exit. In all three buildings, speeds in both staircases
were similar.

        The Gloucester evacuation, canied out during the winter, yielded results that are
different from summer evacuation findings. The average Time to Start the evacuation was
much longer during the winter evacuation. This result is possibly due to the longer
preparation time needed to get dressed before the occupants left their units. Most
occupants were wearing boots, coat, mitts and hat when they left their units. These winter
clothes slightly affected the speed of occupants as they travelled down the stairs. In
Gloucester, occupants took an average of 12.14 s per floor, compare to 10.18 s in Calgary
and 11.27 s in Montreal.

        This study also shows that more information and training is necessary to make
people understand that an immediate response is important when the fire alarm sounds, for
both their safety and the safety of others. Education should be provided on what course
of action is expected from occupants when the fire alarm sounds. Occupants with
limitations should receive specific information on what action they are expected to take.
The local fire department can be of great help in providing specific information on the
evacuation procedures they would like to see implemented.

        Results from the highrise building evacuations were similar to results obtained for
the midrise buildings in many respects. Evacuation Times to Start were not significantly
different, except for the two midrise buildings where the alarm audibility was poor and for
the highrise winter evacuation. Times to Move were generally shorter in the midrise
buildings due to the shorter distance to travel. The main difference between midrise and
highrise evacuations was in the occupants' speeds. In the midrise buildings occupants
moved at an average speed of 0.63 mls, while in the highrises, the average speed was 0.99
m/s. The slower evacuation speed of midrise occupants can possibly be explained by the
fact that occupants of midrises appeared to know their neighbours much better and were
more likely to stop or slow down to speak with other occupants during their descent This
caused the midrise occupants to wave1 significantly slower than highrise occupants who
generally did not show this type of behaviour.

        Even though the measurements of the alarm sound level showed a low or non-
existent audibility in most apartments, most respondents to the questionnaire reported that
the alarm was loud enough. This does not imply, necessarily, that they could hear the
alarm from inside their apartments. Occupants could have become aware of the M 1
through exterior noises such as the arriving fire trucks, movement in the corridors, or
neighbours knocking on their doors. The results of the questionnaire might not be truly
re~resentative the situation since it is ~ossible occu~ants
               of                                   that         who did not hear the
a h also did not participate in the evachation drill.

         Building management should make sure that every occupant can hear the fue alarm
from every location in the building. One way to ensure that the alarm is audible inside
units is to install alarm bells in each apartment This strategy is better than trying to
increase the sound of the alarm bells already installed in corridors and stairs, since the
sound of these alarms is usually over powering in these locations. An alarm which is too
loud becomes counter-productive because occupants cannot discuss a plan of action or
exchange information about the situation, which is essential in a real fue emergency. As
shown in a previous study, the addition of alarm bells set at a lower sound level in each
apartment appears to be the appropriate strategy to ensure that every occupant is aware of
the situation. The overall results for the three high-rise apartment buildings in this study
show that, in many rooms of different units, the fue alarm was barely audible, while in the
corridors, the sound pressure was very high.
8.0 RECOMMENDATIONS

1   I t is imoerative to make sure that e v e n occuoant in a residential buildine can
    hear the fire alarm from e v e n area of the build in^. An evaluation of the
    audibility of the fire alarm system should be made once the building is occupied.
    This assessment should be made with the residents of the building. An easy way to
    assess if everyone can hear the fm alarm is to perform an evacuation drill. After the
    drill, a short questionnaire can ask occupants if they heard the alarm, or the building
    manager can go to every door and ask the occupants if they heard the alarri~.

2   To imorove the audibilitv of the fire alarm, the number of fire alarm bells
    should be increased rather than increasing the sound of existing alarm bells. In
    most buildings, the sound of each alarm-bell is loud enough. In fact, increasing the
    sound of each alarm-bell would not be a viable solution because the resulting sound
    levels would be too high for people directly exposed to them and could cause
    hearing impairment for evacuees. A too-loud alarm would also prevent exchange of
    information between people, which is often essential. To improve the audibility of
    the fue alarm, the best way is to increase the number of alarm-bells, ideally locating
    an alarm-bell in each apartment Researchers Sultan & Halliwell [lo] at NRCC have
    developed guidelines for proper location of alarm bells in apartment buildings.

3   The number of false alarms should be limited. False alarms tend to reduce
    occupants' participation in evacuations since the occupants consider the alarm as an
    unreliable source bf information to indicate an emergeicy situation. It is important
    to reduce the number of false alarms to rebuild faith of occupants about the meaning
    of the alarm.

4   In buildings where a oublic announcement svstem is installed. it should be
    used to give evacuation instructions to the occuoants. It is, however, necessary
    to test the system to make sure that evervone can hear and understand the content of
    the messages from all areas of the building.

5   Occuoants with hearing disabilities should have a com~ensatow         alarm svstem
    linked to the central fire alarm. People with hearing limitation may not use their
    hearing aids at all time. It is essential that these occu~ants provided with an
                                                                 be
    acceptable "alarm system" to make them aware of ankmergeniy situation at the
    same time as other occupants.

6   Occuoants with limitations should met with the building manaeement to
    discuss the evacuation aroeedure. People with limitations are often very worried
    about the idea of an emergency and have very little knowledge of what would be
    expected of them. A specific procedure should be planned by the building
    management and be explained to the occupants. The local fue department can help
    in providing specific instructions according to their own procedures.
7       Saecific training should be undertaken with occuaants of non-canadian
        cultural backeround. Problems arose during evacuations because some occupants
        -             - -   --

        did not understand the meaning of the fire alarm. Some were also frightened by the
        firefighters and refused to vacate their apartments. It is necessary to provide an
        information session for these residents in their own language, if necessary.

8       Building occuaants should be aware of the alanned evacuation arocedures for
        their buildings. The evacuation procedures should be posted in the elevator lobby,
        the staircases and other specific areas, such as the laundry room or the underground
        garage. It would be worthwhile to provide a copy of the evacuation procedures to
        every resident when they sign their lease, for example.

9       Buildine occuoants should aractice the alanned evacuation orocedures. Once a
        year, it would be useful to prform an evacuation drill to make sure that occupants
        &cognize the sound of the kre alarm, experience different means of egress and apply
        any specikk procedures for their building.

10      Building manaeement should be familiar with the functionine of the alarm
        .
        -         The configuration of some alarm systems is not standard and building
        management should be aware of the specific features of their alarm system. Some of
        these features might not be justified any longer and the alarm system confguration
        should be revised to ensure that the system will work adequately in an emergency
        situation.

    1   The use of the balcony as an emergencv exit should be defined clearlv in
        buildine emereencv orocedures. It must be emphasized in building emergency
        procedures that balconies can be used as a place of refuge, unless the person is in the
        apartement of fire. It must be stressed, that the door to the balcony must remained
        closed to reduce the risk of smoke, heat and flames propagation and to limit the air
        pressure fluctuation within the building.
9.0 REFERENCES

1 Statistics Canada, 1  9 Catalogue
                           ,
                       9 1 92-305E, Statistics
  Canada Ottawa. 1992.

2 E. JUILLET, J. PAULS, "Recent social and technical developments influencing the
  life safety of people with disabilities: the North American scene", Proceedinps of the
  IAPS 12th International Conference, 11-14 July, Greece, 1992.

3 G. PROULX, 1. C. LATOUR, 1. W. MACLAURIN, "Housing Evacuation of Mixed
  Abilities Occupants," Internal Report, IRC-IR-661, Institute for Research in
  Construction. National Research Council of Canada, Ottawa, Ontario, 1994.

4 G. PROULX, C. McQUEEN, "Evacuation timing in apartment buildings," Internal
  Report, IRC-IR-660, Institute for Research in Construction. National Research
  Council of Canada, Ottawa, Ontario, 1994.

5 G. PROULX, "The time delay to start evacuating upon hearing a fm alarm,"
  Proceedings of the Human Factors and Ereonomics Society 38th Annual Meeting,
  Nashville TN, 24-28 October 1994, Vol. 2, HFES, Santa Monica, CA, 1994.

6 G. PROULX, C. LAROCHE, J. C. LATOUR, "Audibility problems with fue alanns in
  apartment buildings," Proceedinos of the Human Factors and Er_ponomicsSocietv 39th
  Annual Meeting, San Diego, CA, 9- 13 October 1995, in press, HFES, Santa Monica,
  CA, 1995.

7 R. W. HUEY, D. S. BUCKLEY, N. D. LERNER, "Audible Performance of Smoke
  Alarm Sounds", hoceedinps of the Human Factors and Ergonomics Society 38th
  AnnualMeetinp, Monica, CA, 1994, pp. 147-151.
                 Vol. 1, HFES, Santa

8 C. LAROCHE, R. &TU, H. TRAN QUOC, R. LAROCQUE, "'Detectsound': a
  user-friendly software for the analysis of warning sounds in noisy workplaces",
  Proceedinos of the 6th International Congress on Noise as a Public Health Problem,
  Nice, France, July 1993, Vol. 2, pp. 169-172

9 C. LAROCHE, H. TRAN QUOC, R. HETU, S. MCDUFF, "'Detectsound': A
  computerized model for predicting the detectability of warning signals in noisy
  workplaces", Applied Acoustics, Vol. 32, 1991, pp. 193-214

10 M. A. SULTAN, R. E. HALLIWELL, "Optimum location for fm alarms in apartment
   buildings", in Fire Technolog, Vol. 26, No. 4, 1990, pp. 342-356.
                  75


ANNEX 1: MEMO DISTRIBUTED TO OCCUPANTS
               Exercice d'evacuation
*** Un exercice d'evacuation aura lieu samedi matin***

  Votre irnmeuble a ete selectionnee pour faire partie d'une etude concernant
les procedures d'evacuation en cas d'incendie. Nous savons qu'un exercice
d'evacuation cause bien des inconvenients mais rappelez vous qu'il s'agit de
votre securite et de celle des autres qui est en jeu. Participer a un exercice
permet de rnieux connaitre les procedures de securite et aide a reagir
correcternent en cas d'urgence. Nous vous demendons votre
collaboration, lorsque I'alarme d'incendie se ferra entendre il est
essentiel de s'habiller et de quitter le bstiment.

   Dans le but de realiser cette etude, des cameras-video vont filmer I'exercice.
Ces cameras vont Gtre placees dans certains corridors, des cages d'escaliers
et a I'exterieur. Tous les films video seront tenus confidentiels.

  Apres I'exercice on vous demandera de rernplir un questionnaire. Vos
norns et adresse ne seront pas mentionnes. Notez que vous pouvez refuser
de rernplir le questionnaire sans fournir d'explication.

   Les resultats de I'etude serviront a ameliorer les procedures de securite
dans les tours a apparternents. Cette etude est realisee par le Conseil
national de recherche du Canada, ainsi que la Societe canadienne
d'hypotheque et de logernents et le Services d'incendie de Gloucester.

  Si vous avez des questions n'hesitez pas a cornrnuniquer avec Mme Proulx
au 993-9634 ou vous pouvez ecrire a:

                       Dr. Guylene Proulx
                       Laboratoire national d'incendie
                       Conseil national de recherche du Canada
                       Ottawa, Ontario K1A OR6
                 EVACUATION DRILL
**An evacuation drill will be carried out Saturday morning **

  Your apartment building has been selected to be part of a
research project on fire evacuation drills. We know that an
evacuation drill is disruptive but we need to carry this out for your
safety and for that of others. Drills and training are the best way to
know how to deal with an emergency situation. We are asking for
your participation; it is imperative for everyone to get dressed
and to leave the building when the fire alarm will go off.

  For this drill, the evacuation will be filmed with video cameras
located in different corridors and staircases. After the evacuation
drill, you will be asked whether you are willing to fill out a short
questionnaire on your experiences during the drill. Your name and
address will be kept confidential.

 This research project is carried out by the National Research
Council of Canada, in collaboration with the Canada Mortgage and
Housing Corporation, and the Gloucester Fire Department.

  If you have any questions, please call the principal investigator,
Dr. Guylene Proulx, at 993-9634, or you can write:

                     Dr. Guylene Proulx
                     National Fire Laboratory
                     Institute for Research in Construction
                     National Research Council of Canada
                     Ottawa, Ontario, KIA OR6
                  79

ANNEX 2: POST-EVACUATION QUESTIONNAIRE
                                                          En franpais sur la page suivante


                      Residents of
                  Questionnaire on the evacuation drill
1. Were you here during the evacuation drill this evening?
   a a Yes                       t
                 No. if not ~ uthis auestionnaire in the box bv the main entrana.

2. Did you hear the fire alarm?     a aYes           No

3. If you heard the fire alarm, in your o inion, was the sound of the alarm
      in your apartment     a   too loud   b   loud enough        a   not loud enough
      in the corridor           too loud   a   IU
                                               O,     enough      a   not IOU,    enough
      in the staircase      a   too loud   a   loud enough        a   not loud enough

4. Where were you when you heard the fire alarm?
   a   in apartment;                 a     Outside apartment;
    a                                        a
    a
    a
           kitchen
           living room
           bedroom,    a   asleep?
                                             gh  on balcony
                                                       corridor
                                                    n elevator
    a      bathroom                          a      in lobby
                                                    in laundry room
      On which floor?

5. What did you do as first, second and third action when you heard the fire
   alarm? write 1, 2, 3 in boxes;
               a
        :: i ~ a
        i : ~~ a
               -
                                                    seek information from others
                                                    gather valuables (purse, wallet, etc.)
      -   find a parent                             look through the window
      U        U
          find ~ e t                                move to the balconv,
      a
      -        a
               -
          get dressed                               wait for help
               U
          instruct others                           leave the buildina
                                                                     "
      a
      -        a
          call to get information                   other
      Ugo into the corridor to have a look
6. How long did you take before leaving your apartment?                          minutes

7. Did you try to use the elevator to evacuate?         a a Yes          No
                                                      En fran~ais la page suivante
                                                                sur


8. Did you use a staircase to evacuate?    a Yes a No
   If you used the staircase, which one was it?
   Staircase A, Ogilvie St.a   Staircase B, parking      a
   While using the staircase to evacuate did you have any problems such as;
      a opening the door               a entering the staircase
      a going down the steps           0other
      a finding the handrail           a no problem
9 Have you taken part in an evacuation in this building before? a Yes # No

10.Do you always leave the building when you hear an alarm? #Yes     a No
11.Do you suffer from      limitations that could prevent you from evacuating a
              [I
   building? II Yes

   If Yes, explain the kind of limitation you have:




12.Your age               sex:   a a
                                   F     M
   Age and sex of others who are living with you:       Age -     UFDM
   Age -      UFPM
                 A               g e a F n M Age-                  a   F   QM

Thank you for your help.

              Return this questionnaire by Monday.
              Put it in the box by the main entrance

Dr. Guylene Proulx, Researcher,
National Fire Laboratory
National Research Council of Canada
Ottawa, Ontario KIA OR6 Phone: (613) 993-9634
                         Residents du
                 Questionnaire sur I'exercice d'evacuation
1. Avez-vous articipe a I'exercice d'evacuation de votre immeuble ce matin?
        Oui    6   Non, si non mettre le ouestionnaire dans la boite dans le hall d'entde.

2. Avez-vous entendu I'alarme d'incendie?       II
                                               [; Oui    a    Non

3. Si oui, a votre avis, I'alarme d'incendie etal;
         dans votre appartement       a  trop forte     a  assez forte   a    pas assez forte
            dans le corridor          0    trop forte                    0    pas assez forte
            dans les escaliers        a    trop forte                    a              fe
                                                                              pas assez o,

4. Oh etiez-vous lorsque I'alarme s'est mise a sonner?
                                            a
       i
       Dans votre appartement;
             dans la cuisine
             dans le salon
                                                A I'exterieur de votre appartement;


                                                        ~~k:leb~~~~dor
       a dans la chambre, a endormi?                    dans I'ascenseur
       a     dans la salle de bain                      dans le hall d'entrbe

       Sur quel etage?
                                                   a    dans la salle de lavage


5. Qu'est-ce que vous avez fait comme premiere, deuxieme et troisieme action
   apres avoir entendu I'alarme? ~ c r i v e z 2, 3 dans les carres;
                                             1,

                                                   Q chercher sacoche, porte-monnaie, etc.
      ce
      rhr
      ::
       ;           les enfants                     a    regarder par la fenetre
   a    chercher un parent                         a    aller sur le balcon
   I[I1 chercher un chat, un chien, etc.                attendre de I'aide
       s'habiller                                  a    demander aux autres quoi faire
   a   diriger les autres                          a
                                                  quitter le bitiment
   a   telephoner pour avoir de I'information IC] autres
   a   regarder dans le corridor ce qui se passe

6. Combien de temps avez-vous pris avant de quitter votre appartement? -minutes

7. Avez-vous essay6 d'utiliser I'ascenseur;     IC] Oui    a   Non
8 Avez-vous utilis6 les escaliers?        Oui     Non
   Si oui, quel escalier avez-vous utilise?
   C6t6 A, rue Ogilvie      C6te B, stationnement a
   -                                     -
   En utilisant les escaliers avez-vous rencontre des problemes du genre;
   Udifficult6 ouvrir la ~ o r t e                                         -
                                         Udifficult6 a entrer dans la caae d'escalier
   a   difficult6 a trouver la rampe    0   autres choses
   C]I difficult6 a descendre les marches      #   pas de probleme

9. Aviez-vous deja participe a des exercices d'evacuation dans ce bfitiment
   avant ce matin?     Oui
   Si, Oui combien de fois,
                              a   Non
                                        la date de la derniere fois,

10.Est-ce que vous quittez toujours le bfitiment quand vous entendez I'alarme?
   aOui     a Non
11.Est-ce que vous avez des limitations hysiques qui pourraient vous
   empecher d16vacuerun bfitiment?      8   Oui     Non
  Si Oui, expliquer le genre de limitations que vous avez:


12.Votre fige               Sexe: 0 Q  F     M
   L'fige et le sexe des autres qui etaient avec vous au moment de I'exercice:
   Age -Sexe:      UF UM                Age        ~exe:      F   a    M
   Age -~ e x e :  UF UM                Age        Sexe:  U F ~ M
       Deposer le questionnaire complete avant lundi
    dans la bolte rouge a cet effet dans le hall d'entree
Merci de votre collaboration,

                       Dr. Guylene Proulx, Chercheure
                       Laboratoire national d'incendie
                       Conseil national de recherche du Canada
                       Ottawa, Ontario KIA OR6       Tel.: 993-9634
            85


ANNEX 3: GUIDE TO MANAGERS
                                Guide to Managers


                   Fire Safety for All Occupants
            in Midrise and Highrise Apartment Buildings


        In Canada, most fires and fire deaths take place in residential buildings. Over the
years, a large range of fire safety equipment has been developed and installed in buildings.
Still, Canada and the United States have the worst fire records amongst developed
countries. As a way to limit the number of fires and the fire injuries and deaths, it is
essential to invest more effort in fue prevention through training and education. Fire
prevention is important because it is the best way to ensure that occupants know what
they are supposed to do and whdt actions can be dangerous during a fire. It is the building
manager's responsibility to inform the occupants about fire safety and the evacuation
procedures for the building and the responsibility of every occupant to learn them.

       This guide has been developed to help you supply fue safety information to
occupants of your building. Each building has different fue safety equipment, so it is
important that you tailor the information to include features specific to your building.

  Supply all residents with a brochure entitled "Evacuation Procedure" which is enclosed
  at the end of this Guide. The brochure explains what occupants should do in case of a
  fire. You should ensure that the brochure is given to each new tenant moving into the
  building.

  Complete the "Evacuation Procedure" brochure by providing explanations on the fire
  safety equipment in place in your building. For example, specify if appropriate, when
  the building fire-alarm activates:
          * their is a two-stage alarm, the intermittent ring means that you should
             prepare to evacuate and you should leave your apartment at the full ring,
          * the elevators will go automatically back to the first floor;
          * the fire doors with magnetized-holders will close or open automatically
             (specify the locations);
          * the vent or pressurization system will start in the building or the staircase,
          * sprinklers will activate in the fii area only,
          * information on action to take will be given over the P.A. system,
          * the roof is (or is not) accessible in case of fue,
          * the areas of refuge are located (specify the locations),
          * etc.
Make sure that the "Evacuation procedure" is acceptable to occupants who have
limitations. For example, make sure that residents who have hearing impairments have
strobe-alm lights in their apartment and that wheelchair users have access to their
balconies or another refuge area.

Put together a "Fire-List" to keep a record of all occupants who may need assistance in
case of an emergency. This Guide provides you with a method to identify the
occupants who should be listed and how to present the information on the Fi-List.

Special effort should be made to inform occupants of other cultural backgrounds about
the fue evacuation procedure.

Once a year, organize a "Fi Safety Information Session". Your local F i e Department
can help you organize this information session; they can usually provide written
instructions and videos to animate the session.

Once a year, a few days after the Fire Safety Information Session, cany out an
evacuation drill. Carrying out a drill helps to ensure that occupants can hear and
recognize the fire alarm from every apartment in the building. It also helps occupants
know where the closest staircase is and it gives them some knowledge of the staircase
organization and the time it takes to reach ground level, as well as where the exit doors
lead. Simple training, such as an evacuation drill once a year, will help occupants
decide what to do, if they are ever faced with a fire in the building.
Steps to carry out an evacuation drill:
        * Inform your local Fire Department about the drill.
        * To start the drill: Activate the fue-alarm at a pull-bar station.
        * Use the P.A. system to give out information.
        * Evaluate the participation of occupants.
        * A drill should last between 20 and 30 min to give enough time to occupants
           to reach ground level.
        * After the drill, ask the occupants if they heard the fire alarm; understood the
           messages from the P.A. system; and knew what to do. To get this
           information you can use a short questionnaire or ask occupants in person.
       * Meet with the occupants who did not participate and explain to them the
           importance of practising the evacuation procedure.
                                      The Fire-List

       The building manager should prepare a Fire-List. It is a list of the occupants that
may require help during a fire emergency. This list should be given to the local fire
department and be enclosed in the fire alarm panel in a way that firefighters arriving at the
building will immediately have access to the list.

       In determining who should be included on the Fire-List, the manager has to
consider many different characteristics of each occupant
                            to
       Abilitv to res~ond alarm: can the occupant hear the fire alarm and understand
       its meaning?
                  pet
       Abilitv to - out of bed and eet dressed: is the occupant able to get out of bed
       without help, transfer himselfmerself into a wheelchair or put shoes on and stand
       up without help?
       Abilitv to move to an area of refuge: is the occupant able to reach an area of
       refuge which implies horizontal movement to reach either a balcony, an elevator
       lobby, or a stair landing?
       Abilitv to move on stairs to exit the building: is the occupant able to leave the
       apartment, open the staircase door and go down a certain number of floors?

        Occupants who meet any one of these criteria should be included on the Fire-List
The occupants with limitations who live with someone else or those who have visiting
support staff should also be included on the list. The list should contain specific
information: the apartment number, the name of the occupant, the telephone number, and
the limitation of the person, as well as any special care that should be considered by the
rescuers. It is important to know that, although firefighters and police officers receive
training on how to rescue people, they do not necessarily know all the specific problems of
people with different types of limitations. Information should be provided on the Fie-List
and the occupant with the limitation should also be able to inform the rescuers as to how
to handle them.

      The names appearing on the Fire-List should be listed according to the apartment
number, from the 1st floor to the upper floors.

       The Fire-List should be updated every six months. During an emergency, each
second counts and f~refightershave no time in such lie-threatening situations to try to
rescue people who no longer live in the building, while new occupants with limitations are
ignored because they were not included on the Fire-List. It is important to indicate on the
document the date at which the Fire-List was last updated.
                                  SAMPLE
                               Fire-List
                    Occupants who need assistance
                       in case of an emergency
                               Four Seasons Place
                                1111 Main street
                                 Beautiful City

Superintendent: Mr. Bob Lachance, #201




Date:     15 March 1995
                                Evacuation Procedures

        In any residential building, the activation of the fire alarm should never be ignored.
Residents of the building should take specific actions according to the safety features of
the building, the fm conditions and th& physical capacity to move around. A fue is an
emergency situation that can develop in unexpected ways, which is why there is no perfect
plan or absolute evacuation procedure to follow. Occupants have to use their judgement
and evaluate the situation to decide what is best to do. Here are a few tips to help make
appropriate decisions.

IF YOU DISCOVER A FIRE:
      Leave the area immediately
      Close the door to prevent the fire and smoke from spreading
      Activate the fue alarm at the pull-bar station located near the staircase
      Do not use the elevator
      Use the closest staircase to leave the building
      Once in a safe area, call 91 1 or the fire department

IF YOU HEAR THE BUILDING FIRE ALARM
     Immediately prepare to leave the building, get dressed and take your door keys.
     Go to the door and feel it with your hands for heat.
      3 If the door or the knob is hot, do not open it. Call 91 1 or the fire department
         and explain your situation. Use wet rags and towels to seal the cracks around
        your door to prevent the smoke from entering. Go onto your balcony, if you
        have one, close the balcony door behind you and wait to be rescued.
     3 If the door is cool, open it a crack bracing your shoulder against it to prevent




       -
        the pressure created by heat and smoke from knocking you over.
     Have a look in the comdor. If the corridor is clear of smoke. Leave your
     apartment and lock the door.
        If you see smoke in the comdor, return to your apartment. Call 911 or the fire
        department and explain your situation. Use wet rags and towels to seal the
        cracks around your door to prevent the smoke from entering. Go onto your
        balcony, if you have one, close the balcony door behind you and wait to be
        rescued.
     Do not use the elevator.
     Move to the closest staircase and make sure it is clear of smoke before starting
     your descent.
        If you encounter smoke in the staircase, leave the staircase immediately and
        move to another staircase. You can also go back to your apartment or enter
        any floor and knock on doors to ask other people to let you into their
        apartments.
     Once you have reached ground level, go to the designated meeting point to let
     others know that you are out safely.
     Do not re-enter the building until the fire department gives the "all clear".
                                             91


Some Basic Fire Knowledge

        It is important to remember that smoke is very dangerous, in fact, most fire deaths
are due to smoke inhalation. If there is smoke in an area, avoid entering, but if you have
no other choice, crawl low since the smoke tends to rise; it also helps if you c& protect
your face with a wet cloth. Always close doors behind you to stop the smoke movement.
If there is light smoke in an area, opening a door or a window may create a chimney effect
that will increase the quantity of smoke. If you need to open a window to breathe, open it
only a crack first because by opening a window you will bring more air into the room
which may draw more smoke into the room. Ideally, go to the balcony and close the door
behind you. If you don't have a balcony, stay at the window and hang out a sheet to
signal your location to the fire department

         You should never try to use an elevator to evacuate a building when a fire is
suspected. The elevator shaft will act as a chimney, drawing a lot of smoke into the
elevator shaft and cabins. During a fue, the electricity may be shut off, which could stop
the elevator. If you are in an elevator at that time, you will be trapped. Some elevator
control panels are heat-sensitive, which means that the buttons to call floors on the panel
are triggered by the pressure of a fmger combined with static and the warmth of the skin.
In case of fire, the warm smoke may trigger all the buttons of the control panel which will
make the elevator stop at every floor including the fire floor. If the elevator stops at the
fire floor and the door opens, it is very likely that the smoke in the area will prevent the
door from closing again. This is because some elevator door closing mechanisms work
with an optical beam on both sides of the door which, when interrupted by a passing
person, prevent the door from closing. The smoke from a fire can be thick and dark
enough to interrupt the beam so that the door will not close and a person in the elevator
will be trapped on the fire floor,

        The roof of most buildings, whether residential or commercial, is usually not
accessible. Even when there is a door to the roof, this door is usually locked. Since
smoke tends to rise in a staircase, it can be very dangerous to try to escape by going up
the stairs.

        You should talk openly with other members of your family about what you would
do in case of a fire. Decide in advance who will be responsible for the children or pets and
which route you will use to go outside. Fix a meeting point outside - a specific location
where you will meet in case you are separated. Practice your evacuation route with all
your family members. It is essential to know where the closest staircase is, to have a feel
for the staircase organization in case you have to move in the dark. Measure how long it
takes to make your way down, leave through the exit door to see where it leads to. Do
the same thing with another staircase of the building, to make sure you have an alternate
way out, in case it is impossible to use the closest staircase.
Occunants who have limitations

         When the fire alarm activates, every occupant who is physically capable to reach
safety at ground level should use the staircase and fire exits, if the fire conditions permit,
to exit the building. Occupants who have limitations and cannot, without harm, move to
the outside, should stay in their unit or go to an area of refuge. If you have limitations,
make sure that your name is on the Fie-List so that for the firefighters know that you
need help when they arrive at the building.

       Occupants who are moving to an area of refuge should use their judgement to
decide which refuge is safest according to the location of the fire. A balcony, for example,
can become a dangerous area if the fire is in the apartment directly underneath as f i e and
smoke tend to rise and can engulf the balcony above. The areas of refuge available
depend on the building fire safety provisions. In some cases, an area of refuge can be a
balcony, a comdor away from the f i e location separated by fue doors, an elevator lobby
separated by fire doors from a corridor, a staircase landing, if it is large enough to
accommodate people waiting to be rescued, or another apartment away from the fire.

        It is important to understand that people who cannot move to ground level should
move horizontally to an area away from the fire and the smoke and wait to be rescued if
necessary. It may be best to stay in your apartment if the f i e is not on your floor or the
floor beneath you.

        Occupants should stay in the refuge area or their unit as long as the fire alarm is
ringing or until they receive the "all clear" from the fire department Rescuers can take a
long time to reach you, so you should be patient. They will f i s t rescue the people most at
risk, usually the ones closest to the fire. Consequently, it can take a long time, sometimes
hours, before they arrive to rescue you if you are in no immediate danger. Make sure the
rescuers know where you are by informing 911 or your local Fire Department

								
To top