Sunteți pe pagina 1din 12

TRANSPORTATION

PLANNING
HANDBOOK

INSTITUTE OF
TRANSPORTATION ENGINEERS

peak-hour factor may be divided into the hourly passenger


flow to obtain the peak flow rate. A peak-hour factor close
to 1.0 may indicate system overloading and reveal the po
tential need for increased service.

12.8.6

Loading variations

It is highly improbable that all vehicles in a stream of buses


or streetcars or in a train will be loaded equally. In mixed
traffic situations, vehicles often bunch in groups of two to
five. Vehicles at the front of a bunch are usually loaded to
capacity, but those toward the back often do not carry a
full load. In trains, those cars farthest removed from the
platform's access points are less crowded than the average
load for the entire train. Often, dwell times at a stop vary
depending upon schedule variations.
To allow for variations in dwell times, the R factor is
introduced into capacity calculations. It can be obtained
from field observations and may range from .50 to .85 for
buses and from .85 to 1.00 for rapid transit.
Because all cars on a rail transit are not loaded equally,
an "unequal" loading factor is sometimes used. This factor
is likely to be in the range of . 70 to .90. It will exceed .90
only under conditions of extreme overcrowding. A meas
ured loading variation close to 1 may actually indicate sys
tem overload, because some passengers are being forced to
wait for a later vehicle.

12.9

Level of service

The concept of level of service for transit is far more com


plex than for highways. It includes such factors as coverage
of major residential and activity areas, comfort, speed, and
reliability (i.e., on-time performance). Convenient sched
ules, comfortable vehicles, and frequent, fast, and reliable
service contribute to level of service. Speed is influenced
not only by the number of riders using a transit line, but to
an even greater extent by stop frequency and dwell times,
traffic interferences, and right-of-way design.
Productive capacity, the product of passenger capacity
and speed, is an important measure of mass transit system
efficiency. It is important in that it distinguishes between
equal passenger throughputs achieved at different speeds.
Thus, express bus service normally has a higher productive
capacity than local bus service; similarly, a commuter rail
line operating at 40 mph is twice as "productive" as an
urban rail transit line carrying the same number of people
at 20 mph. In general, "productive capacity" will be influ
enced by the type of technology (rail versus bus), the
method of operation (private right-of-way versus shared),
and the spacing of stops.
Two aspects of level of service are important from a ca
pacity perspective: the number of passengers per vehicle and
the number of vehicles per hour. Capacity-related level of
service criteria should reflect both. Figure 12.10 illustrates
this two-dimensional nature of urban transit capacity.
It can be seen that it is possible to operate many transit
vehicles, each carrying few passengers. From a roadway ca
pacity perspective, the number of vehicles could be at or
near capacity, even if they run nearly empty.
430

Transportation Planning Handbook

Max. People/vehicle
Crush load

Max. design load (peak)


Maximum vehicles/
channel/hour

QJ

>

C.
QJ

Many vehicles,
few passengers

B
A

Vehicles per Hour (level of service)

Figure I 2.10 Nature of Urban Transit Level of Service and Capacity

A few vehicles could operate, each overcrowded. This


represents a poor level of service from a passenger comfort
(user) perspective, Long waiting times would also detract
from user convenience.
Finally the domain of peak-period operations commonly
involves a large number of vehicles each heavily loaded.

12.10

Vehicle capacities and


loading criteria

The total passengers carried varies depending on bus or


railcar capacity and the trade-off between seated capacity
and standees. The largest number of seats and lowest num
ber of standees should occur on longer suburban bus routes
or on commuter raii routes where higher levels of comfort
are essential. See Chapter 5, Mass Transit Systems for typi
cal transit vehicle types, dimensions, and passenger capac
ities. A typical 40-foot urban transit bus can normally seat
53 passengers and can carry up to 32 additional standees.
Similarly, a 60-foot articulated bus can carry 69 passengers
and 41 standees. An 8-car train of 75-foot rail transit cars
normally can seat about 500 and carry a "crush load" of
about 1,800 to 2,000 people. Doorways on buses range
from 22 to 30 inches each, while doors on rail vehicles typi
cally average 50 inches each.
Table 12.24 gives suggested "passenger" levels of service
for a conventional 40-foot bus, based on 53 passengers per
bus and 340 gross square feet per vehicle. These approxi
mate comfort-related levels of service are from the perspec
tive of passengers on the vehicle rather than the number of
vehicles in a given channel. They are based on local bus
operations where short trips at relatively low speed allow
standees. Express bus service on expressways and busways
should not allow standees; hence, their scheduling should
be guided by level of service C.
Suggested passenger levels of service for urban rail tran
sit vehicles are given in Table 12.25. LOS D, which allows
up to two persons per seat and a minimum 5.0 square feet.

TABLE 12.24

12.11

Passenger Loading Standards and Levels of Service


for Bus Transit Vehicles (5-Seat, 340-Sq.-Ft. Bus)
Peak-Hour
Level of Service
A
B
C
D
E (Max. scheduled load)
F (Crush load)

Passengers

Approx.
Sq. Ft./Pass.

Pass./Seat
(Approx.)

0 to 26
27 to 40
41 to 53
54 to 66
67 to 80
81 to 85

13.1 or more
l 3.0 to 8.5
8.4 to 6.4
6.3 to 5.2
5.1 to 4.3
<4.3

.00 to .50
.51 to .75
.76 to 1.00
1.01 to 1.25
1.26 to 1.50
1.51 to 1.60

SOURCE: TRANSPORTATION RESEARCH BOARD, Highway Capacity


Manual, 1985 (Washington, D.C.: TRB, 1986), Table 12-5.

By far the largest percentage of transit passengers in the


United States are served by transit buses operating in
mixed traffic on conventional city streets. Although the
theoretical capacity of a lane assigned exclusively to buses
has been shown to be about 1,400 vehicles, that is, two
thirds of the capacity of a 100 percent passenger vehicle
flow, this value is of little practical significance because it
assumes continuous flow without stops or pickups to re
ceive passengers. Bus flow per lane per hour is much lower
when passenger stops are taken into account.

12.11 .1
TABLE 12.25
Passenger Loading Standards and Levels of Service
for Urban Rail Transit Vehicles
Peak-Hour
Level of Service

Approx.
Sq. Ft./Pass.

Approx.
Pass./Seat

A
B
C
D
E-1
E-2 (Maximum scheduled load)
F (Crush load)

15.4 or more
15.2-10.0
9.9-7.5
6.6-5.0
4.9-4.0
3.9-3.3
3.2-2.61

.00- .65
.66-1.00
1.01-1.50
1.51-2.00
2.01-2.50
2.51-3.00
3.01-3.80

1
The maximum crush load can be realized in a single car, but not in every
car on the train.
Note: Fifty percent standees reflects a load factor of 1.5 passengers per seat.

SOURCE: TRANSPORTATION RESEARCH BOARD,

Manual, 1985 Table 12-6.

Highway Capacity

per person, provides a reasonable balance between operat


ing economy and passenger comfort. It is consistent with
the use of 5.4 square feet per passenger suggested by
Pushkarev et al. as a realistic passenger capacity for rapid
transit lines.6
Level of service E is synonymous with "schedule design
capacity" assuming a reasonable number of standees. It
represents the upper limit for scheduling purposes. These
maximum scheduled loads are normally 65 to 75 percent
of the crush loads.
Level of service F defines "crush load" conditions in
which standees and other passengers are subject to unrea
sonable discomfort. Such loads are unacceptable to passen
gers. Although LOS F represents the theoretically offered
capacity, it cannot be sustained on every vehicle for any
given period, and it exceeds the maximum utilized capac
ity. Moreover, it is not reasonable to assume that passen
gers will be equally distributed among all cars of all trains.
Therefore, level F should not be used for transit capacity
calculations. Note, however, that when the maximum
schedule design loads are used, some transit units will op
erate at LOS F.
6Pushkarev, B.S., Zupan, J.M., Cumella, R., Urban Rail in America: A
Regional Plan Association Book, Indiana University Press, Bloomington,
IN (1982).

Bus transit

Operating experience

Bus service volumes are dependent upon a number of fac


tors, including the frequency of service, equipment type,
seating capacity, bus layout, boarding points, and peak
hour factor. Observed bus passenger volumes on U.S. and
Canadian roadways are shown in Table 12.26.
1. The highest bus volumes, more than 800 buses per hour,
are found on the approach to the Midtown Bus Terminal
in New York City. The highest freeway bus volumes per
lane per hour (about 700) operate nonstop westbound
along 1-495 on the approach to the Lincoln Tunnel.
2. Peak flows are much less when buses make intermediate
stops. Flows as high as 200 buses per hour have been
reported when (a) buses operate in platoons, (b) two or
more lanes in one direction enable buses to pass each
other, and/or (c) passengers can also enter through more
than one door.
3. Downtown streets generally carry bus volumes of 80 to
100 buses per hour where there are two to three boarding
positions per stop and passenger boarding is not concen
trated at a single stop. This frequency corresponds to
about 5,000 to 7,500 passengers per hour, depending
upon bus loading factors.
4. Table 12.27 gives peak-hour bus flows at major bus ter
minals in North America. At peak-hour flows, most com
muter bus terminals average about 8 buses per berth per
hour.

12.11 .2 Bus capacity on arterial streets


Suggested arterial street bus capacity ranges based on ac
tual operating experience are given in Table 12.28. This
table gives representative service volumes for downtown
streets and arterial streets leading to the city center for
each level of service. Where stops are not heavily patron
ized, as along outlying arterial streets, volumes could be
increased by about 25 percent.
The values for LOS F, forced flow conditions, should
not be used for planning or design. They are merely given
for comparative purposes. The service volumes shown in
Table 12.28 may be used for planning purposes. The values
assume that key boarding points are sufficiently dispersed
to achieve these bus loads. They suggest maximum person
flow rates up to 7,500 people per hour per lane on down
town streets and 10,000 people per hour per lane on
arterial streets. Corresponding maximum values for seated
Capacity in Transportation Planning

431

TABLE 12.26
Observed Peak-Hour Bus Passenger Volumes on Urban Roadways-U.S. and Canada, 1972-1990
(ranked by buses per hour)
Buses/Hour
In Peak
Direction

City and Location

Passengers/Hour
In Peak
Direction

Passengers
/Bus

Tunnels and Bridges


New York-Lincoln Tunnel
San Francisco-Oakland Bay Bridge (before BART)
San Francisco-Oakland Bay Bridge (after BART)
New York-Brooklyn Battery Tunnel
Philadelphia-Ben Franklin Bridge
New York-George Washington Bridge
Washington, D.C.-Memorial Bridge (before Metro)

830
360
200
150
137
108
100

32,700
14,900
8,740
7,500
5,065
4,245
4,020

39
41
44
50
37
39
40

Freeways
New Jersey-1-495
New York-Gowanus Expressway
Washington, D.C.-Shirley Highway
New York-Long Island Expressway
Chicago-N. Lake Shore Drive
Philadelphia-Schuylkill Expressway
San Francisco-Golden Gate Bridge
Los Angeles-San Bernadina Freeway

700
106
200
90
80
78
80
70

28,000
5,300
10,000
3,560
4,000
2,800
3,750
3,500

40
50
50
40
50
36
47
50

Bus-Only streets
Portland, OR (Mall)-5th Avenue (dual lanes)*
Portland, OR (Mall)-4th Avenue (dual lanes)*
Chicago-State Street Transitway-south bound
Chicago-State Street Transitway-north bound

167
175
130
155

8,300
8,500
7,200
8,600

50
50
60
60

200
200-220
170
160
130
155
143

10,000
10,000
8,500
8,000
6,500
9,900
8,300

50
50
50
50
42
58
58

116
90
80
67

3,580
3,800
4,000
2,807

31
42
50
42

City Streets-Some Bypassing Possible


Ottawa-Carleton Transport
New York-Madison Avenue (dual lanes)*
New York-Hillside Avenue*
Washington, D.C.-14th Street Bridge (before Metro)*
Washington, D.C.-K Street N.W. (before Metro)
San Francisco-Market Street (before BART and MUNI)*
Philadelphia-Market Street
City streets-one lane
Chicago-Michigan Ave.
Chicago-Washington St.
Rochester-Main St.
Atlanta-Peachtree St.
*Buses use more than one lane.

SOURCE: H. S. LEVINSON, and California Department of Transportation, New York City Department of Trans
portation, Urban Mass Transit Administration, Texas Transportation Institute.
TABLE 12.27
Peak-Hour Bus Flows at Selected Major Terminals, U.S. and Canada
Peak Hour Buses
(one way)

City and Terminal


New York-Midtown'
San Francisco-Trans-bay (before BART)
San Francisco-Trans-bay (after BART)
Toronto-Eglinton2
Chicago-Jefferson Park
Toronto-W ilson2.3
New York-George Washington Bridge
Chicago-95th-Ryan
Philadelphia-69th St.
Washington, D.C.-SW
Cincinnati-Dixie
Chicago-69th-Ryan
Pittsburgh-McKeesport
Denver-Denver Mall
3 4

730
350-400
250
250
140
136
108
106
90
80
48
40
30
NA

Loading Berths
184
37
37
13
14
18
43
22
10
10
6
4
7
NA

Buses per
Berth/Hour
4.0
9.5-10.8
6.6
19.2
10.0
7.6
2.5
4.8
9.0
8.0
8.0
10.0
4.3
15.0

NA - Not available.
'Includes 26 intercity buses before expansion. Since expansion, bus volumes have grown to about 900.
2Free transfer to subway
3
Includes LRT
'Before Yonge St. extension
SOURCE: H ERBERT S. LEVINSON, ET AL., Bus Use of Highways, Stale of the Art, National Cooperative Highway Research
Program Report 143, Transp. Research Board, National Research Council 1973; Texas Transportation Institute and Her
bert S. Levinson, Lockwood Transit Center, Conceptual Planning and Design, 1985.

432

Transportation Planning Handbook

TABLE 12.28

Dwell times, D, can be computed as follows:

Suggested Bus Service Volumes for Planning Purposes


(hourly flow rates, based on 50 seats per bus)
Level of Service
(street)
Arterial streets
A
B
C
D
E
CBD streets
A
B
C
D
E

o Boarding only, one-way passenger flow

Level of Service (passengers)


A
B
E
D
C

Buses/Hour
25 or less
26-45
46-80
81-105
106-135

625
1,125
2,000
2,625
3,375

940
1,690
3,000
3,940
5,060

1,250
2,250
4,000
5,250
6,750

1,560 1,875
2,810 3,375
5,000 6,000
6,560 7,875
8,440 10,125

20 or less
21-40
41-60
61-80
81-100

500
1,000
1,500
2,000
2,500

750
1,500
2,250
3,000
3,750

1,000
2,000
3,000
4,000
5,000

1,250
2,500
3,750
5,000
6,250

1,500
3,000
4,500
6,000
7,500

SOURCE: TRANSPORTATION REsEARCH BOARD, Highway Capacity


Manual, 1985 (Washington, D.C.: TRB, 1986), Table 12-12, p. 12-14.

D=bB

(12.14)

o Alighting only, one-way passenger flow


D=aA

(12.15)

o Two-way flow through single door


D =(aA +bB) y

(12.16)

where
Y = 1.2 for major passenger transfers, 1.0 otherwise

a = alighting coefficient, in seconds per passenger

passenger flow rates are 5,000 and 6,750 people, respec


tively. Exclusive use of articulated buses would increase
these values 15 to 20 percent. A peak-hour factor of .80
would result in a maximum one-day hourly volume of
about 6,000 persons on downtown streets and 8,000 per
sons on arterial streets.

12.11.3

Bus capacity at transit stops

The flow rates shown in Table 12.28 are useful for planning
purposes. However, they may overstate capacities where
most passengers board or alight buses at a single point. For
greater precision, it is essential to determine the minimum
headway that can be provided at the maximum load point
and then compute the transit units per hour and, in turn,
the passenger capacity.
The following set of formulas should be used for esti
mating on-street capacity of a transit stop, berth, or route.
They apply to bus transit, but they also can be used for
light rail transit with appropriate adjustments for car and
train length, door configuration, and fare collection meth
ods. They assume that the transit lane or stop area would
be used exclusively for transit use. They are keyed to the
busiest stop along the route in terms of passenger inter
change.

A = number of alighting passengers per vehicle


b = boarding coefficient, in seconds per passenger
B = number of boarding passengers per vehicle
D = dwell time, in seconds

Table 12.29 gives the capacity of buses per hour for a sin
gle loading position for various passenger dwell times per
stop, clearance times between buses, and green-per-cycle ra
tios. It assumes a value of R = 0.833 in formula (12.14).
For example, if 3 passengers unload through the front
door at 2 seconds per passenger, and 8 passengers load
through the front door at 3 seconds per passenger, the total
dwell time will be 30 seconds. Entering Table 12.29 with
this dwell time, a clearance time of 15 seconds and a G/C
ratio of 6.5 results in an estimated capacity of 54 buses per
hour.
Suggested guidelines for the parameters in (12.12),
(12.13), (12.14 ), and (12.15) as well as for the efficiency of
multiple stopping positions are given in Table 12.30.
o The boarding and alighting coefficients apply to buses
with single channel doors.
o The clearance times apply to SO-passenger (40-foot)
buses. For articulated buses, they should be increased
by 3 to 5 seconds.

12.11.3.1 Buses per stop. The number of transit vehi


cles per hour that a transit stop can accommodate can be
estimated as follows:

TABLE 12.29
Capacity of Bus Stops at Signalized Intersections

c 3, 600 R
Cb= (g/ )

[D + tc]

Clearance Time (sec.)

(12.13)

where
R = reduction factor to compensate for dwell times and

arrival variations

Cb = vehicles per hour


g = green time per cycle, in seconds

C = cycle length, in seconds

D = dwell time, in seconds

tc = clearance time, in seconds

15

10
Signal g/C>
Dwell Time (sec.)
15
45
60
90
120
150

.5

1.0
.5
(buses per hour)

1.0

86
46
38
28
22
18

120
54
42
30
23
18

67
40
33
25
20
17

100
50
40
28
22
18

SOURCE: TRANSPORTATION RESEARCH BOARD, Highway Capacity


Manual, 1985 (Washington, D.C.: TRB, 1986), Table 12-9, p. 12-12 and
Table 12-19, p. 12-21.
Capacity in Transportation Planning

433

TABLE 12.30
Suggested Capacity Guidelines for Application (Bus Transit)

I. Boarding time per passenger-pay full fare on bus


2.6 seconds single fare
3.0 seconds exact fare(general value)
3.5 seconds exact fare-standees on bus
2. Alighting time per passenger
1.7-2.0 seconds-use 2 seconds
3. Heavy two-way passenger flows-through a single door
1.2 (2A + 3B)
where: A= alighting passengers/bus
B = boarding passengers/bus
4. Clearance between successive buses
l0" -Absolute minimum
15" -Desirable minimum
20"-Minimum for operations on high speed roadways
5. Levels of service

LOS
A
B
C
D
E ( Capacity)
E (Capacity)

3600R

0.400
0.500
0.667
0.750
0.833
l.000

1440
1800
2400
2700
3000
3600

Proportion
ofE
(LOS factor)

Approximate
Probability
of Queues
Forming behind
Bus Stop(%)

.4
.6
.8
.9
1.0

2.0
2.5
10.0
20.0
30.0
50.0

6. Berth efficiency (cumulative effective berths)


Berths

On-Line Stations
Factor

Off-Line Stations
Factor

I
2
3
4
5

1.00
l.75
2.25
2.45
2.50

1.00
0.85
2.60
3.25
3.75

in entering or leaving an available berth by buses in adja


cent berths. Thus, in applying (12.13) and (12.14) to esti
mate bus volumes for additional berths, it is necessary to
increase the value of tc. Similarly, where routes terminate tc
must include the layover time if buses layover in a regular
berth.
If the number of buses to be accommodated along a
street exceeds the capacity of the busiest stops, routes may
be separated into two groups of about equal hourly vol
umes with different sets of stops provided for each group of
routes. If buses can pass each other along the street, the
total street capacity for carrying transit buses approaches
the sum of the capacities of the stops for each group.
12.11.3.2 Passenger capacity of a berth. The capacity
of a berth in persons served per hour can be estimated from
the previous formula by trial and error. This involves assum
ing a dwell time at major stops, estimating the transit units
per hour per stop, and, based on the service times per pas
senger, estimating the boarding, alighting, or interchanging
passengers per bus.
A more direct approach is to estimate the maximum
passengers per berth, Q, as follows (boarding conditions
govern):
Q

(g/c) 3600 BR
Bb(g/c)+tc

Nb, the number of effective berths needed to serve J pas


sengers, is found as follows:
(12.19)

SOURCE: Highway Capacity Manual, Special Report 209 Transporta


tion Research Board, National Research Council, 1985.

o The position efficiency factors apply to bus operations.


(It is reasonable to assume that they can also apply to
light rail transit.)
An R value of .833 corresponds to LOS E. R values of
.667 to .750 should be used for downtown streets. They
correspond to LOS C and D and result in probabilities of .2
to .3 that queues will develop. An R value of .5 should be
used for bus stops along arterial streets. It results in level of
service B, with queues resulting about 2.5% of the time.
Thus, the capacity of an outlying bus stop (per berth) can
be obtained as follows:
Cb =

(g/c) 1,800

[!D + tc J

(12.17)

where
Cb is the frequency of buses per berth for LOS B
This formula results in a low probability of a bus having
to queue to enter the stop and block other traffic. It is de
sirable for design purposes, but it is too stringent for down
town streets.
Multiple-berth stops do not have the same capacity per
berth as single-berth stops, because buses may be delayed
434

Transportation Planning Handbook

(12.18)

Hence
J[tc +Bb(g/c)J

=
Nb (g/c) 3600 BR

(12.20)

Design of a bus terminal or transit center involves infor


mation of the number of buses that will use the center,
route and service patterns, and operating practices. Ideally,
a maximum of two distinct services should be provided per
loading_ position.
While berth space requirements can be estimated by the
preceding formulas, it is more realistic to set berth criteria
based on operating standards and practices. Schedule re
cover and driver relief times must be considered. When
service is infrequent, departures should be based on
"clock" headways. Similarly, the need for "timed transfer"
connections may increase dwell times.
Current operating experience suggests 8 to 10 buses per
berth per hour for commuter buses and 1 to 2 buses per berth
per hour for intercity buses. Such guidelines provide a firs
approximation and should be adjusted to reflect specific op
erating and scheduling needs.
12.11.3.3 Passenger capacity of a route. The passenger capacity of a berth and the number of berths determine
the passenger flow of people at the major boarding point.
These riders may constitute all or a portion of the riders at

the maximum load point or section, depending upon pas


senger distribution along the bus line.
The number of passengers passing the maximum load
point, P, can be estimated as follows:
(g /c) (3 600 R N b S )
p =_ _ _ _ _ _ _ _ _ _
[Bb (g/c) + tc]

(12.21)

where S is the number of passengers per transit unit, that


is, passengers per bus.
This capacity can be estimated as a function of the pro
portion of the total passenger volume at the maximum load
sections boarding at the busiest stop x, as follows:
(g /c) (3 600 N b R )
p =- _ _ _ _ _ _ _ _ _
xb(g/c)+tc !S

Where LRT trains operate on streets under manual con


trol, capacity estimates can be obtained by adapting the
bus transit equations to account for differing vehicle sizes,
train lengths, and clearance requirements.
When cars share a lane with buses, bus capacities will be
reduced. These adjustments can be made by estimating the
amount of green time per cycle preempted by cars and reduc
ing the green time for buses accordingly. See Table 12.31.

12.11.4

A review of transit operating practices, and the various ca


pacity formulas suggests that bus capacities can be in
creased by

(12.22)

o Increasing the number of major boarding points in the


central area or at other major activity centers
o Reducing passenger service times by using wider
doors, allowing prepayment of fares and separating
loading and unloading activities
o Using longer units o( r where feasible higher load fac
tors)to reduce the number of units needed and the
clearance times between successive vehicles

where x is the fraction of passengers passing maximum


load point boarding at heaviest stop.
Finally, the maximum number of passengers, P b, that
can be carried at the maximum load point for each effec
tie berth, can be estimated by dividing P by Nb:
P
3 600 R
Pb=-=-----Nb xb (g/c) + tc/S

(12.23)

In these relationships, g /c=1.00 for off-street busways and


terminal approaches. For on-street operations, g/c is site
specific.
Typical bus service volumes at maximum load points,
based on 2.5 effective berths and derived from these for
mulas, are shown in Table 12.31. For 50 percent standees
and a peak-hour factor of .8, system service volumes are
o Maximum capacity, dispersed loading conditions
5,000 to 6,000 persons per hour
o Maximum capacity, typical CBD loading conditions,
45 to 60 seconds /stop, 3,300 to 5,000 persons per hour

Increasing capacities

In extreme cases, buses o( r trains)cannot be unloaded


or loaded at certain stops as rapidly as passengers accumu
late o( r before the next unit arrives). Thus, the headway
that theoretically would be adequate for the demand vol
ume as measured at the maximum load point cannot be
delivered as line throughput. Such situations can be allevi
ated by changing vehicle or stop configuration, using col
lectors to load rear doors, or having prepaid areas.

12.12

Light rail transit

"Light rail" transit has evinced considerable interest in


Canada and the United States in recent years. The term light

TABLE 12.31
Typical Arterial Street Bus Service Volumes at Maximum Load Point (Service Level E)
Seated Passengers, 50 Persons/Bus
Condition
20% board at busiest stop
25% board at busiest stop
30% board at busiest stop
40% board at busiest stop
50% board at busiest stop
Level of Service E (Based on
Current Operating Experience)

Approximate
Dwell Time
at Busiest Stop

Flow Rate

30
38
45
60
75

50% Standees, 75 Persons/Bus

Hourly Volume
(PHF- .80)

Approximate
Dwell Time
at Busiest Stop

Flow Rate

Hourly Volume
(PHF- .80)

6250
5560
5000
4170
3560

5000
4450
4000
3340
2850

45
56
68
90
112

7500
6270
5770
4690
3950

6000
5020
4620
3750
3160

5000

4000

7500

6000

Assumptions:
I. 15" clearance between buses
2. Ratio of clearance time (sec)/to (pass/bus)- .3 for seated and .2 for standees
3. Service time per passenger- 3 seconds
4. g/C Ratio - 0.5
5. Peak hour factor - 0.80
6. All buses stop at busiest stop
7. 2.5 effective berths
8. R factor - 0.8333

Capacity in Transportation Planning

435

rail refers not so much to the strength of the rail structure


(which may be identical to a full rail transit structure), but
rather to the typically lighter cars, shorter trains, and meth
ods of operation. In many cases, light rail transit operates on
surface routes with occasional grade crossings and shared
use of the roadway with automobiles. Power supply is nor
mally overhead rather than through a "third rail."

12.12.1

Operating characteristics

The observed volumes in the United States and Canada


generally reflect passenger demands and scheduling poli
cies rather than maximum possible flows. Streetcars oper
ated at 30-second headways on-street (Pittsburgh) and
in-tunnel (Philadelphia) following World War II; peak-hour
peak-direction flows approximated 9000 passengers per
hour. Table 12.32 provides operating data for street car and
light rail facilities in North American and world cities.
Several European LRT systems report peak flows of
15,000 to 18,000 persons per hour. Philadelphia's Market
Street subway has carried 140 cars per hour with a mini
mum headway of 23 sec, serving as many as 12,000 passen
gers per hour. Boston's Tremont-Boylston LRT subways
traditionally scheduled 60 trains and 250 cars per hour. The
MBTA estimated the capacity of 15,000 persons per hour in
1971 when inbound peak flows approximated 12,000 per
sons. In the 1920s, LRT trains in Boston carried up to
25,000 people per hour (241 cars running on two tracks).

12.12.2 Line capacity


The main-line passenger carrying capacity of a light rail
facility will be determined by car and train sizes, safe oper
ating headways, and passenger loading standards. Modern,
three-car light rail transit trains providing 5 square feet per
seated and standing passenger (about 15-inch-radius circle)
will accommodate about 13,000 passengers per hour per
track if 2-minute headways can be maintained safely.
The practical capacity of a light rail line will, however,
usually be determined by the loading and unloading facili
ties at the heaviest load point, not by its main-line capacity.
If the most heavily used lQi,lding point is an on-street inter
sectional stop, then that point will normally determine the
line's capacity with a boarding rate of up to 2.5 to 3 sec
onds per passenger, plus an appropriate clearance time, for
determining practical capacity. The loading rate can be
changed significantly by the number of doors available for
loading, platform levels, and fare collection processes.
LRT trains usually are limited to a maximum of three
cars where on-street operation is involved since: (I) longer
trains could not operate on city streets without simulta
neously occupying more than the space between adjacent
cross streets when traversing short blocks, (2) long trains
could not clear at-grade intersections rapidly, and (3) long
trains need long platforms lengths at stations.
Minimum hadways for light rail systems will depend
on train length, platform design (high versus low), fare

TABLE 12.32
Peak-Hour Passenger Volumes on Street Car and Light Rail Systems in United States, Canada, and the World. (Peak Direction)

City

Location

On street
Pittsburgh
Pittsburgh
San Francisco

Smithfield St. Bridge


Smithfield St.
Market Street
(before subway)
Toronto
Queen St. East
East Edge CBD
Hong Kong
K6ln
Deutzer Brucke
Basel
Post
Melbourne
Princess Bridge
In tunnel or off street
Philadelphia
Market St.
'3ostoq
Green Line
(Boyiston St.)
Philadelphia
Market Street
Juniper St. Station
Boston
Arlington Station
Koln
Neumarket
San Francisco Market Street
Cleveland
Shaker Hts.
Boston
Green Line
Boston
(Lechmere)
Newark
City Subway
Edmonton
LRT Line
San Diego
LRT

Year

Trains
per
Hour

Cars
per
Hour

1949
1976
1977

120
51
68

120
51
68

30
71
53

46.5
46.5
46.0

9,000 1
3,800
4,900

1978
NA
NA
NA
NA

66
NA
32
70

66
96
NA
70
89

55
38
113
51
40

46.5
29.2
101.7

1956
1985

133
45

133
45

1978
NA
NA
NA
1983
1976
1978
1985
1978
1978
1981

73

73
20
41
48
62
60 1
48
12
30
24
6

NA
30 1
16
12
30
12
3

Headway
Seconds

Length
of Car
or Train

Passenger/Hour
in Peak
Direction

Passengers/
Train

Equipmen

75 1
74
72

75 1
74
92

PCC
PCC
PCC

46.5

4,200
7,975
6,500
5,726
4,394

64
83
203
82
49

64
NA
203
NA
NA

PCC
NA
NA
NA
NA

27
50

46.0
71.0

9,000
10,600

67
125

67
235

PCC
Boeing LR:

180
180
88
75
NA
120 1
225
300
120
300
1,200

46.0
46.0

3,700
3,840
6,722
10,000
6,340
4,400
1,500
1,600
1,500
2,100
600

51
192
164
208
7
73
31
135
50
87
200

51
NA
NA
NA
NA
143
911
135
50
174
400

Boeing LA
NA
NA
NA
Boeing
PCC
PCC
Boeing
PCC
DUWAG
DUW.\G

101.7
70.0
50.0
46.5
71.0
46.5
77.0
151.0

Passenger/
Car

1 Estimated.
'Length varies.

SOURCES: Vukan, Vuchic; Light Rail Transit System, A Definition & Evaluation, 1972; Highway Capacity Manual, 1985 Special Report 209. T.
portation Research Board, National Research Council, 1986; Massachusetts Bay Transportation Authority, Hamburger, Wolfgang S. and Quinby, Hen:. Chapter 7, "Urban Transit" in Transportation and Traffic Engin eering Handbook, S econd Edition, Institute of Transportation Engineers, Prentice H
Englewood Cliffs, New Jersey, 1982; Vukan, Vuhic, Urban Public Transportation Systems and Technology, Prentice Hall, Englewood Cliffs, New J
1981.

436

Transportation Planning Handbook

TABLE 12.33
Typical Ranges in Capacity-Light Rail Systems
Passenger Level of Ser vice
D

Type of Operation

Units
per Hour

5.0 Sq. Ft.


per Person

Maximum Schedule
Design Loads
3.3 Sq. Ft.
per Person

Street cars
(single 46-50-ft.
unit on-street)

60
90

4,300
6,400

6,800
10,200

25
30
35

7,800
9,350
11,050

12,400
14,900
17,000

LRT-off-street
(three 75-ft.
car units)

three-car train per block, clearing on a 60-second cycle,


could result in 180 units per hour.
This condition, however, cannot always be achieved in
practice because of varying arrival and dwell times. With
one unit every other cycle, capacity could be as high as 30
trains or 90 units per hour.
Typical ranges in capacities, assuming a 15 percent re
duction for unequal car loading and/or schedule variations,
are shown in Table 12.33.

12.13

SOURCE: Estimates by H. S. Levinson.

collection methods (prepayment versus pay on train), and


headway controls (manual versus block signals). Under
manual operations, 80 to 100 single-unit cars per track per
hour can be accommodated. When trains run under block
signal controls, as is common with rapid transit systems,
120-second headways are common, although shorter head
ways could be realized.
For LRT operations, the number of transit vehicles per
city block should not exceed the length of the block. Good
operations result when there is only one unit per block
which can clear within a single signal cycle. For example, a

Heavy rail transit

"Heavy rail" transit systems are designed to carry large


numbers of people quickly and dependably. Long trains op
erating on exclusive rights-of-way have door and seating ar
rngements that are designed to expedite passenger loading
and alighting. High-level platforms also expedite passenger
flow.

12.13.1

Operating experience

Tables 12.34 and 12.35 show peak-hour rail transit rider


ship for U.S., Canadian, and overseas transit systems. The
wide range of peak-hour passengers carried reflects ranges
in the number, length, frequency of trains operated, accept
able levels of passenger crowding, and track and junction
constraints.

TABLE 12.34
Observed Peak-Hour Volumes on Rapid Transit Systems-U.S. and Canada

City and Year


ew York City

Line/Location
1982

1960

Toronto
ontreal
Chicago

1978
1974
1960
1976
1984
1978

Philadelphia
Boston

1976
1977-78
1985
1977-78
1985
1977
San Francisco

ashington
.\tlanta
Oeveland

1980
1980
1976
1960

IND E, F, 53rd St. Tunnel


IND A, D, 8th Ave Express
!RT 4, 5, Lexington Ave. Exp.
PATH-World Trade Center 1
IND E, F, 53rd St. Tunnel
IND A, D, 8th Ave. Express
!RT 4, 5, Lexington Ave. Exp.
!RT 2, 3, 7th Ave. Express
Yonge St.
Yonge St.
Yonge St.
N Line
Milwaukee
Lake-Ryan
North-South
Lake-Ryan
North-South
North Broad (2 tracks)
Red Line
Red Line
Orange Line
Orange Line
BART-Transbay
BART-Mission
Blue-Orange
East Line
West Side
West Side

Trains
per
Hour

Cars
per
Hour

26
21
25
38
32
30
31
24
30
28
28
23
17
19
15
21
20
23
17
17
13
5
ll
10
20

208
210
250
266
320
300
310
240
210
168
224
207
136
152
120
168
160
126
68
68
52
60
98
85
120
36
52
80

14
20

Headway
Seconds
128
159
157
98
112
120
116
150
i20
129
129
157
212
189
240
Ill
180
157
212
190
277
150

321

360
180
600
258
180

Approximate
Car Length
Feet
(Rounded)
75
60, 75
50
50
60
60
50
50
75
75
57
56
50
50
50
50
50
67
70
70
55
55
75
75
75
75
50, 70
50

Persons/Hour in
Peak Direction
(Maximum
Load Section)

Passengers
per Train
(Rounded)

Passengers
per Car
(Rounded)

54,500
43,500
38,100
25,500
61,400
62,000
44,500
36,800
32,000
36,000
32,200
28,200
12,400
12,300
11,400
16,500
14,000
10,600
13,000
13,000
8,400
9,000
8,000
6,500
13,000
4,250
5,400
6,200

2,100
2,070
1,520
670
1,920
2,070
1,430
1,530
1,060
1,290
1,260
940
730
647
760
790
700
460
460
460
650
600
730
650
650
710
390
360

260
210
150
100
190
210
140
150
150
210
140
140
90
80
95
100
90
85
140
190
160
150
80
75
110
120
105
80

ultiple track terminal.


SOURCE: Adapted from Transportation Research Board, Highway Capacity Manual, 1985 (Washington, D.C.: TRB, 1986), Table 12-13, p. 12-15.

Capacity in Transportation Planning

437

Highway Capacity Manual. The "crush load" capacities are


shown for comparative purposes, but they should not be
used in determining capacity or in system design.

12.13.3 Commuter railroads


Path of front
of train

Most commuter rail lines operate 20 or fewer trains per


track during the peak hour. The number of passengers per
train may exceed 1,000 if double-decker coaches are used.
Paris RER commuter service-which operates along the
lines of rapid transit-carries over 1,700 passengers per
peak-hour train.
Passenger volumes of 15,000 per person per track are
common. Table 12.37 contains peak-hour data on com
muter rail for world cities. Commuter rail operations on
"segregated" rights-of-way have minimum headways of 3
minutes. When urban commuter trains share tracks with
intercity freight and passenger trains, headways of 10 to 12
minutes are more realistic.

Rate of change of slope


represents acceleration

"'
(.)

.;
0

12.14
Constant slope represents
balancing speed

Pedestrian facilities8

Concentrated pedestrian movement occurs at public


events; in and near transit terminals, high-rise buildings,
department stores, theaters, stadia, and parking garages;
and at other major traffic generators. Pedestrian safety,
trip patterns, and convenience are necessary considerations
in all multimodal traffic and transportation studies. Table
12.38 presents some of the higher pedestrian volumes ob
served in several major urban centers.
The concentration of pedestrian activity at street cor
ners and crosswalks makes them critical traffic links for
both sidewalk and street networks. An overloaded corner
or crosswalk not only affects pedestrian convenience but
can also affect vehicular turning movements and intersec
tion capacity.

Time
Figure 12.11 Distance-Time Diagram for Rapid Transit Train
Headway and Frequency Analysis

The precise values for these equations will vary among


individual transit properties depending on the type of
equipment used and their operating and scheduling policy.
Values should be discounted by 10 to 15 percent to reflect
unequal loading within cars.
Illustrative rail rapid transit capacities are summarized
in Table 12.36 based on 30 trains per track per hour; 50and 75-foot cars and 6-, 8-, and IO-car trains; and a 15
percent reduction for unequal car loading. The levels of
service shown are consistent with those set forth in the

8This section was prepared by Leon Goodman of the Port Authority of New
York and New Jersey.

TABLE 12.36
Rail Rapid Transit Capacities (Include 15% Reduction for Unequal Passenger Distribution)
Passengers per Hour
Cars/Train

Cars/Hour

180

240

10

300

Passengers per Seat (approximate)

Car/Length
(feet)

Approximate
Seats/Train

1.00

1.50

2.00

2.50

3.00

3.50

50
75
50
75
50
75 1

300
450
400
600
500
750

7,650
11,500
10,200
15,300
12,750
19,100

11,500
17,200
15,300
23,000
19,000
28,700

15,300
23,000
20,400
30,600
25,500
38,200

19,100
28,700
25,500
38,250
31,900
47,800

22,950
34,500
30,600
45,900
38,250
57,300

26,800
40,250
35,700
53,550
44,600
66,850

10.0

6.7

5.0

4.0

3.3

2.6

E-1

E-2

Sq. Ft. per Passenger:


Passenger Level of Service (U.S. and Canada Conditions)
Comments:

Maximum
schedule
loads

F
Not
attainable
on a train
basis

This condition does not exist in the United States.


SOURCE: Adapted from Transportation Research Board, Highway Capacity Manual, 1985 (Washington, D.C.: TRB, 1986), Table 12-15, p. 12-17.

Capacity in Transportation Planning

439

TABLE 12.37
Observed Peak-Hour Volumes on Commuter Railroads, 1978

Location

Trains/
Hour/Track

Headway
(seconds)

Length of Trains
(feet)

Passengers/
Hour/Track
(peak direction)

Passengers/
Train

Auber-Chiitelet
Borough Market Jct.
Laim-Donnersbergbriike
Paris Nord
Jamaica
Harvard Division
Sud-Nord
Park Avenue Tunnel
Zurich-Oerlikon
Penn Station

18
30
22
17
15
II
21
20 1
15
17

200
120
164
212
240
327
171
100
240
212

718
515- 798
442- 663
604- 732
1085
340- 935
656-1066
255- 970
1168
NA

31,150
28,520
21,650
17,420
15,550
14,370
12,600
12,000 1
10,815
9,070

1730
950
985
1025
1035
1300
600
600
720
530

City
Paris (RER)
London
Miinchen
Paris (SNCF)
New York (LIRR)
Chicago (CNW)
Bruselles
New York (Metro)
Zurich
New York (NJT)

'Estimated from total count of three parallel tracks.


SOURCE: HOMBURGER, WOLFGANG S. AND QUINBY, HENRY D., Chapter 7, "Urban Transit" in Transportation and Traffic Engineering Handbook, 2nd
Edition, Prentice Hall, Englewood Cliffs, New Jersey/Hub-Bound Travel (New York: Metropolitan Transportation Council, 1982).

12.14.1

Pedestrian flow analysis

The concept of pedestrian flow and the methods to calcu


late pedestrian capacities for walkways, corridors, and
other basically "linear" situations are covered extensively
in the 1985 Highway Capacity Manual. The following ma
terial provides a framework for planning. Application pro
cedures and sample calculations are available in the

Highway Capacity Manual.


The principles of pedestrian flow analysis are similar to
those used for vehicular flow. The fundamental relation
ships among speed, volume, and density are similar. As the
volume and density of a pedestrian stream increases from
free flow to more crowded conditions, speed and ease of
movement decrease. When pedestrian density exceeds a
critical level, volume and speed become erratic and rapidly
decline.
The level of service concept, first used to define relative
degrees of convenience on highways, is also applicable to
pedestrian facilities. With this concept, such convenience
factors as the ability to select walking speeds, bypass slower
pedestrians, and avoid conflicts with others are related to
pedestrian density and volume.
Pedestrian analysis uses some familiar traffic terms, as
well as others not used elsewhere. The following listing de
fines the major terms used.
Pedestrian speed: the average pedestrian walking speed,
generally expressed in units of feet per second.
Pedestrian flow rate: the number of pedestrians passing a
point per unit time, expressed as pedestrians per 15 minutes
or pedestrians per minute; "point" refers to a perpendicular
line of sight across the width of a walkway.
Unit width flow: the average flow of pedestrians per unit
of effective walkway width, expressed as pedestrians per
minute per foot.
Platoon: a number of pedestrians walking together in a
group, usually involuntarily, because of signal control and
other factors.
Pedestrian density: the average number of pedestrians per
unit of area within a walking or queuing area, expressed in
terms of square feet per pedestrian; this is the inverse of
440

Transportation Planning Handbook

density, but it is a more practical unit for the analysis of


pedestrian facilities (see Figure 12.12).

12.14.2

Density, speed, and flow


relationships

The relationship between density, speed, and flow for


pedestrians is of the same form as for vehicular traffic
streams; that is,
Flow = Speed x Density
v=SxD

( 12.26)

where flow is expressed as pedestrians per minute per foot,


speed is expressed as feet per minute, and density is ex
pressed as pedestrians per square foot.
The flow variable used in this expression is the "unit
width flow" defined earlier. An alternative and more use
ful expression can be developed using the reciprocal of
density, or space, as follows:

Speed

Flow= Space
v=-

(12.27)

The basic relationships between speed, flow and den


sity, as recorded by several researchers, are illustrated in
Figure 12.13.
Research on pedestrian movements has demonstrated
that a speed-density relationship exists which is analogous
to the vehicle speed-density relationship on a free-flowing
roadway. A maximum flow rate of about 25 pedestrians
per minute per foot of walkway takes place when there are
5 to 9 square feet available per person. All movement effec
tively ceases at about 2.5 square feet per person.
As density decreases from these high levels, comfort and
convenience increase, and the walker's ability to maintain
a selected walking speed is increased. For these reasons,
density has been selected as the measure of effectiveness
for pedestrian facilities. For convenience, the reciprocal of
density-space-is used as the criterion. Criteria
for pedestrian service levels are given in Table 12.39 and

S-ar putea să vă placă și