State Rail of New South Wales appointed PB to undertake a passenger
flow analysis of Sydney’s
Town Hall station in 2000. The aim of the study was to determine
the current and future operating conditions of the station and develop
and test station redevelopment concepts designed to meet future
passenger demands. The challenge of the study was to develop a model
that is robust enough to represent the complex circulation patterns
in the station but practical for implementation in a constrained
time frame and budget.
Town Hall Station
Town Hall station was designed around 1916 based on turn-of-the-century
concepts. It opened in 1932 to coincide with the opening of the
Sydney Harbour Bridge. The structural design and pedestrian infrastructure
enabling movement between levels reflect this period. Latter modifications
were fit within the constraints of the original design.
Town Hall station is now the largest and busiest station on the
CityRail network with more than 120,000 passengers entering or exiting
the station each day. Large volumes of passengers also transfer
between platforms within the paid area of the station. By 2016,
the station’s passenger movements are expected to be at least
20 percent higher than current levels.
The station consists of three levels. A concourse level just below
the street has a central paid area surrounded by a free circulation
area, and it provides fare control and ticket purchasing facilities.
Each of the two platform levels below contains two platforms and
three tracks. The central track on each level is accessible from
only one platform and there is no direct movement between platforms
on the same level. Figures 1 and 2 show the concourse level and
one of the two platform levels with analysis zones.
Figure 1: Town Hall Station Complex |
Figure 2: Town Hall Station Upper Platform Level |
The free circulation area of the concourse level is becoming an
increasingly important part of Sydney’s underground pedestrian
network. Its retail concessions constrain pedestrian flows while
attracting additional non-rail related pedestrians.
Town Hall station is already experiencing high levels of congestion
and probably could not sustain significant future growth in its
current form. With the addition of planned rail expansions and general
patronage growth, pressure on Town Hall station will increase, as
will demand for increased interchange between rail lines. If this
growth is to be accommodated, then Town Hall station will need to
be improved to increase both its capacity and efficiency.
Approach and Model Structure
The approach applies a spreadsheet model that is relatively straightforward
in its design, but expansive in order to model the complex pattern
of movement and spaces in the station. Thus, whereas “micro-simulation”
models have not yet been adequately developed to model complex pedestrian
movement, the model applied presents a practical and informative
solution for a complex problem without becoming conceptually convoluted.
The passenger flow model consists of a series of worksheets that
estimate the number of people who would pass through each element
or zone and the number who would pause in the zones and for how
long. Each of the worksheets in the series is described below.
Pedestrian Volumes by Origin and Destination. The
station’s sixteen possible origins and destinations comprise
six platforms and ten external access/egress points. The pedestrian
volume worksheet presents existing or future forecast volumes between
any combination of the 16 origin/destination (O/D) points and any
other, including pedestrians who transfer from platform to platform
and those who enter the free area of the concourse but do not enter
the station. Allowance is also made for those who may transfer to
a different train on the same platform or enter and leave the concourse
by the same door, as a person might do when visiting one of the
shops on the concourse.
The base year pedestrian volume data is derived from passenger surveys
and counts. The future travel demand volumes are forecasts supplied
by the State Rail Authority with estimates of appropriate distribution
between access points. The data is input into the model in the form
of AM and PM peak five-minute origin destination matrices.
Routing Assignment. This table includes an assignment
by percentage of people travelling between each of the 16 origins
and destinations (256 combinations) to any of the 171 elements or
zones (resulting in 43,776 assignment cells). Due to the change
in direction of two escalators and gates at one entrance from exit-only
to entrance-only from the morning to the evening, different assignments
are needed for each period. Additional routing assignment tables
are required to analyse any proposed physical changes to the station.
Walk Volumes. This table calculates the pedestrian
volumes passing through each zone by multiplying the origin and
destination volumes with the percentage assignment for each zone.
Walk Time. This table includes the approximate
time in seconds to walk through each analysis zone. Different walk
times through a zone representing different paths can be associated
to each origin and destination pair. The three typical choices are:
- The full length of the zone
- Half the length, either as an average for people who end their
walk in the zone or cut through it diagonally
- A cross measurement, which may be used for particular routes
across some zones.
Walk time is calculated based on distance in meters divided by
an assumed walking speed of 1.2 m/sec (4.3 kph or 2.6 mph).
Table 1: Element Design Capacity
 |
Table 2: Fruin Level of Service Definitions
 |
Table 3: Levels of Service Rangers for Station Elements
 |
Dwell Percent. This table indicates the percentage
of pedestrians passing through a particular zone who dwell within
that zone to either wait for a train, purchase a ticket, make a
purchase, or for other purposes. No dwell time is assumed on stairs,
escalators or fare control barriers.
Dwell Time. This table includes an average time
in seconds that pedestrians who dwell in a zone spend there. On
platforms this time is related to train headways, but it is less
than half the average headway since commuters on the Sydney system
time their arrival for trains to some extent. Appropriate times
are also assigned for ticketing, browsing or other dwell activities.
A function representing crowding at the entrance to the circulation
element was added to all circulation element queuing zones. The
dwell time for the zones leading to the circulation elements, at
the base of escalators and stairs, is based on a function related
to the capacity of the element. When the circulation element tends
towards capacity, the dwell time in the zone leading to the circulation
element increases.
Time Space Demand. The demand for walk time-space
is calculated for each analysis zone by multiplying pedestrian volumes
in each zone by the walk time required, and an assumed design standard
of 1.4 m2 (14 square feet) per person. The demand for dwell time-space
is calculated by multiplying pedestrian volumes in each zone by
the dwell percent, the average dwell time and an assumed dwell space
of 0.65 m2 (7 square feet) per person. The two are totalled for
a combined time-space demand in each zone.
Design Capacities of Operational Elements. Table
1 shows the design capacities for the key pedestrian circulation
elements of the station. The design capacity of the ticket barrier
gates, ticket windows and ticket vending machines is based on the
capacity observed during the surveys.
Level of Service. The operating condition of each
zone is assessed using levels of service (LOS) as developed by John
Fruin (Pedestrian Planning & Design, 1974). Design capacity
for all elements is generally considered LOS C. Descriptions of
the pedestrian levels of service are given in Table 2.
The LOS in each walk zone is calculated by comparing the ratio of
time-space demand to time space supply with a look-up table. The
LOS for stairs is based the volume per effective meter width compared
to the appropriate look up table. The LOS for fare control barriers
and escalators is based on a demand to capacity ratio, where a ratio
of 1.0 equals the threshold of LOS C/D.
The definitions of level of service measures applied in the Town
Hall models were adapted from Fruin’s Austroads Guide to Engineering
Practice: Part 13 Pedestrians (1995). For various station elements,
the flow rates indicated in these sources were converted to volume
to capacity ratios, based on design capacities. The LOS ranges for
various elements are shown in Table 3. The LOS for each station
zone was presented in colour on plans of each station level, making
areas of congestion graphically apparent.
Conclusion
The passenger flow model, while extensive in its calculations, provided
a relatively quick analysis of a very complex condition. The results
of the analysis identified key problem areas and helped evaluate
alternate plans for modifications to make the station work better
and to add capacity. |
Mark Walker is a Senior Professional Associate
specializing in transit planning and design, pedestrian circulation,
and transit oriented development. With PB for 16 years, he is currently
analyzing passenger circulation for the extension of the Number
7 subway line to Manhattan’s west side and studying the complex
interaction of pedestrians, sidewalk vending, and curbside loading
in New York’s Chinatown neighborhood. Mark is also a Ph.D.
candidate and Adjunct Professor of Planning at Columbia University
in New York City.
|
