-------
Center traffic to reach peaks at the same time. The average traffic
volume generated by the 'shopping center on these four peak days was 30
percent higher than the average for all days surveyed.
As indicated by the data collected at the Oakbrook Shopping Center,
identification of the maximum 8-hour and 1-hour traffic periods
for a proposed shopping center is not a simple process. However, the
following guidelines should be used to determine when these periods are
most likely to occur.
1. The peak periods of total traffic are highly dependent
upon the ratio of through traffic to shopping center
traffic. Typically, this ratio is approximately 2:1
for suburban regional centers.-'-" With this ratio, fluc-
tuations in through traffic have twice the impact that
the same percentage fluctuations in shopping center
traffic would have.
2. The most likely periods of maximum traffic are:
a. Preschool sale days in August and September
or spring sale days (pre-Easter) when (1)
through traffic is generally at its peak (as
high as 120 percent of the average day on a
Friday), and (2) shopping center traffic is
close to its design level (approximately 140
percent of the average day).
b. Pre-Christmas Saturdays when through traffic
(in an area that does not contain intense
commercial development) is approximately 80
percent of the annual average daily volumes
and shopping center traffic is approximately
180 to 200 percent of the average day.
c. Post-Christmas (day after) exchange and sale
days when through traffic on a Friday may be
100 to 110 percent of the average annual day
and shopping center traffic is approximately
150 to 160 percent of the average day.
3. The maximum 8-hour period in the vicinity of a
regional shopping center (through traffic plus shop-
ping center traffic) typically occurs between 11:00
a.m. and 7:00 p.m. on weekdays and between 10:00 a.m.
and 6:00 p.m. on Saturdays and Sundays. The peak
65
-------
1-hour of traffic occurs simultaneously with the
peak period of through traffic (usually 4:00 p.m. to
5:00 p.m.).
4. The following set of calculations illustrates a sug-
gested methodology for analyzing the maximum traffic
flows on any critical section of roadway:
Given; Projections of through traffic
indicate that a section of roadway adja-
cent to the site will carry annual aver-
age daily traffic (AADT) volume of 10,000
vehicles. Traffic counts indicate the
evening peak hour is 10 percent of AADT
and the 8-hour period (11: a.m. to
7:00 p.m.) is 50 percent of AADT.
Information from the State Highway Depart-
ment indicates December traffic flows are
90 percent of AADT, April and September
flows are 110 percent of AADT, Fridays are
110 percent and Saturdays are 90 percent
of the average day.
Trip generation and direction of approach
analyses indicate 4,000 shopping center
vehicles utilize the section of road on an
average day.
Find ;
a. Total 8-hour and 1-hour traffic
volumes on a Friday preschool sale day
in August:
• 24-hour maximum through traffic = 10,000
x 1.10 (month factor) x 1.10 (day of
week factor) = 10,000 (1.21) = 12,100
vehicles.
• 8-hour maximum through traffic = 12,100
x 0.50 (8-hour factor) = 6,050 vehicles.
• 1-hour maximum through traffic = 12,100
x 0.10 (1-hour factor) = 1,210 vehicles.
e 24-hour maximum shopping center traffic =
4,000 x 1.40 (sale day factor) = 5,600
vehicles.
66
-------
• 8-hour maximum shopping center traffic =
5,600x 0.70 (8-hour factor) = 3,930
vehicles.
• 1-hour maximum shopping center traffic =
5,600 x 0.12 (1-hour factor) = 672
vehicles.
• Total maximum 24-hour volume = 12,100 +
5,600 = 17,700 vehicles.
• Total maximum 8-hour volume = 6,050 +
3,930 = 9,980 vehicles.
• Total maximum 1-hour volume = 1,210 +
672 = 1,882 vehicles.
b. Total traffic volumes on a pre-Christmas Saturday:
• 24-hour maximum through traffic = 10,000
x 0.90 x 0.90 = 8,100 vehicles.
• 8-hour maximum through traffic = 8,100 x
0.50 = 4,050 vehicles.
• 1-hour maximum through traffic = 8,100 x
0.10 = 810 vehicles.
• 24-hour maximum shopping center
traffic = 4,000 x 1.80 = 7,200
vehicles.
• 8-hour maximum shopping center
traffic = 7,200 x 0.80 = 5,760
vehicles.
• 1-hour maximum shopping center
traffic = 7,200 x 0.12 = 864
vehicles.
o Total 24-hour maximum traffic =
8,100 + 7,200 = 15,300 vehicles.
» Total 8-hour maximum traffic =
4,050 + 5,760 = 9,810 vehicles.
9 Total 1-hour maximum traffic =
810 + 864 = 1,674.
67
-------
c. Total traffic on a post-Christmas Friday:
• 24-hour maximum through traffic =
10,000 x 0.90 x 1.10 = 10,000
vehicles.
• 8-hour maximum through traffic =
10,000 x 0.50 = 5,000 vehicles.
• 1-hour maximum through traffic =
10,000 x 0.10 = 1,000 vehicles.
• 24-hour maximum shopping center
traffic = 4,000 x 1.50 = 6,000
vehicles.
• 8-hour maximum shopping center
traffic = 6,000 x 0.80 = 4,800
vehicles.
« 1-hour maximum shopping center
traffic = 6,000 x 0.12 = 720
vehicles.
• Total 24-hour maximum traffic =
10,000 + 6,000 = 16,000 vehicles.
• Total 8-hour maximum traffic =
5,000 + 4,800 = 9,800 vehicles.
o Total 1-hour maximum traffic =
1,000 + 720 = 1,720 vehicles.
For this section of roadway, the maximum 24-hour, 8-
hour, and 1-hour volumes are most likely to occur on
the special preschool or spring sale days on a Friday
in April or August.
Level of Service - The term "level of service" is a broadly defined
measure of traffic operating conditions at intersections and continuous
segments of roadways.
Of critical concern to the estimation of vehicle emissions at inter-
sections is the determination of the vehicular operating modes (i.e.,
approach speeds, deceleration, queueing, and acceleration modes).
68
-------
Although a level of service condition at an intersection cannot directly
be used to predict the vehicle operating modes (i.e., a given level of
service does not correspond to a single condition of approach speeds,
deceleration, etc.), the level of service function provides a framework
for estimating these parameters. Therefore, it is necessary to under-
stand the process for determining intersection capacity, service
volumes, and level of service at signalized intersections as part of
the process for estimating operating modes. The level of service
terminology and procedures are outlined in more detail in Appendix G.
The process for estimating operating modes is presented in the following
section of this report.
Methodology for Estimating Vehicular Operating Modes at Intersections
As stated previously, the critical elements of traffic operating con-
ditions at intersections are the vehicular operating modes. The esti-
mation of the vehicular approach and departure speeds, acceleration and
deceleration rates, and queuing characteristics is essential for the
determination of carbon monoxide emissions.
•Approach and Departure Speeds - For this study, the approach and
departure speeds were defined as the vehicle speeds experienced outside
of the influence zone of the signalized intersection, (i.e., the free-
flow speeds prior to deceleration and after acceleration at the
signal). The average approach and departure speeds observed at the
intersection of Illinois Route 83 and 22nd Street are presented in
Table 14. Except on the south leg, where approach and departure
speeds are close to constant, there is a discernable pattern of lower
speeds during the morning (7:00 a.m. to 9:00 a.m.) and evening
(4:00 p.m. to 6:00 p.m.) rush hours. Midday speeds are approximately
5 to 10 miles per hour below the existing speed limits (50 miles per
hour on Illinois Route 83 and 40 miles per hour on 22nd Street),
while peak-hour speeds are 15 to 20 miles per hour below the existing
speed limits. Figures 26 and 27 illustrate the relationship between
69
-------
Table 14. OBSERVED APPROACH AND DEPARTURE SPEEDS
Hour ending
8
9
10
11
12
13
14
15
16
17
18
19
Average:
Approach speeds
North leg
southbound
39
36
39
41
38
38
41
38
35
32
32
38
37.3
South leg
northbound
48
48
49
50
50
50
51
49
50
49
49
50
49.6
East leg
we s tb ound
38
39
38
37
35
33
34
32
32
24
26
37
33.3
West leg
eastbound
32
33
41
45
42
39
37
40
42
38
42
44
39.4
Hour ending
8
9
10
11
12
13
14
15
16
17
18
19
Average:
Departure speeds
South leg
southbound
42
44
44
45
45
43
43
44
41
43
43
43
43.3
North leg
northbound
35
37
37
38
36
36
40
37
39
35
40
39
37.4
West leg
westbound
40
42
39
38
36
35
39
36
37
33
33
36
36.4
East leg
eastbound
32
31
36
36
35
35
34
36
35
34
35
37
34.6
70
-------
o o u o
u •-; iJ u
-• -> J -I
r 5 I i
*- >- i- t-
cc cr -> 3
O Og o
2 Z b) «
3
rH
O
CO
3
CO
w
T3
0)
8UJ
„ Z
X
uj
o.
O
IIMW) Q33dS 3TOIH3A 30WU3AV
71
-------
o o
ill!
< * < £
o ct o i
a < cc <
a. a a. a.
a m a. u
< Q < Q
O O O O
Ul Ul
J 5 3
O
(U
0)
a,
CO
cu
4J
Vi
rt
a,
cu
-o
a,
a.
to
cd
M
0)
CM
Q)
I
•H
(HdVNI 033dS 313IH9A 30VU3AW
72
-------
the average observed approach and departure speeds versus the
average volumes per lane on the intersection legs. Assuming the
single-lane capacity for each leg is constant, a pattern similar
to the relationships between volume and average speeds for rural
higtways indicated in Chapter 3 of reference 4 would be indicated.
The lower overall speeds (5 to 10 miles per hour) on the east and north
legs is most likely caused by disruption from adjacent signalized
intersections. The east leg approach speeds appear to be more signif-
icantly impacted by the higher volumes per lane. This is most likely
due to excessive congestion at the intersection of Illinois Route 83
and 22nd Street, itself, during the peak hours. The posted speed
limit appears to provide the upper boundary for all the approach and
departure speeds as no average single observed speed of more than
5 miles per hour over the posted speed limit was recorded on either
roadway during the study period.
The reasonable estimation of approach and departure speeds for inter-
sections in the vicinity of a proposed indirect source is difficult.
Approaching and departing flow conditions can vary from uninterrupted
(free-flow) conditions to interrupted conditions for numerous reasons.
Speed limits may vary from 55 miles per hour to below 25 miles per hour.
The limitations on vehicular acceleration and deceleration rates also
influence the speeds that can be obtained between consecutive signal-
ized intersections. The influence of decelerating and accelerating
vehicles may be experienced at distances of up to 400 feet upstream
or doxmstream from a signalized intersection. Thus, midblock unin-
terrupted flow conditions (average speeds > 30 miles per hour) are
unlikely to exist if signal spacing is less than 800 feet. Where
signal spacings of 800 feet or greater exist, the approaching and
departing speeds can be approximated by uninterrupted rural conditions,
and the curves relating volume to average speeds as found in Figures
3.41-3.43 of reference 4 should be applicable. However, for road-
ways in the vicinity of an indirect source, the posted speed limit-
should be used in place of the average highway speed.
73
-------
For signal spacings of 800 feet or less, the relationship of V/C to
average speeds for urban and suburban arterials indicated by Figure
10.3 of reference 4 can be used. However, the V/C ratio shown in
the graph reflects the midblock V/C and not the typical controlling
intersection V/C condition.
The speeds found from the various curves presented in reference 4 are
only approximations. Results should be compared with any available
recent studies. Wherever possible, local conditions should be analyzed
and appropriate adjustments made. Field measurements of existing
conditions in the vicinity of a proposed regional shopping center or
at locations having conditions equivalent to future proposed condi-
tions would provide the best measure.
Acceleration and Deceleration Rates - The average acceleration and
deceleration rates observed at the intersection by hour of the day are
shown in Table 15. Peak-hour congestion does not appear to have any
significant effect on these events. The average observed rates for
this study agree well with typical roadway deceleration and accelera-
tion performances recorded in several other traffic engineering
studies. For the purpose of emission estimation, it is suggested
that deceleration rates of between 2.5 and 3.0 mph/sec and acceleration
rates of 2.25 to 2.75 mph/sec be used. More applicable data should
be substituted if available.
Queuing Characteristics - The third critical element of the traffic
operating conditions at signalized intersections is queuing. The
number of vehicles waiting and the average waiting time per queued
vehicle are important parameters for the estimation of emission in-
tensities at intersections because of the high density emission rates
of idling vehicles.
74
-------
Table 15. OBSERVED ACCELERATION AND DECELERATION RATES
Hour ending
8
9
10
11
12
13
14
15
16
17
18
19
Average :
Average deceleration rates (mph/sec)
North leg
southbound
2.5
2.4
2.5
3.1
2.4
2.5
-
2.6
2.6
2.2
2.2
2.0
2.45
South leg
northbound
2.7
2.7
3.0
3.1
2.8
2.9
-
3.0
2.4
3.1
2.9
2.8
2.85
East leg
westbound
3.0
2.6
3.7
3.4
3.1
2.9
-
2.5
2.4
2.1
2.0
3.6
2.85
West leg
eastbound
2.0
2.9
2.8
2.8
2.2
3.1
-
2.9
2.8
3.6
3.2
3.1
2.85
Hour ending
8
9
10
11
12
13
14
15
16
17
18
19
Average:
Average acceleration rates (mph/sec)
South leg
southbound
3.0
2.7
2.7
3.1
2.4
2.4
-
2.4
2.1
2.0
2.2
2.2
2.47
North leg
northbound
3.0
3.1
2.1
2.8
2.2
2.8
-
2.8
3.0
2.5
2.5
1±1
2.70
East leg
eastbound
2.2
2.5
2.8
2.4
2.4
1.9
-
2.2
2.4
1.8
1.9
2.2
2.24
West leg
westbound
2.2
2.5
2.7
2.4
1.9
2.2
.
2.4
2.4
2.7
2.3
lil
2.45
75
-------
The hourly averages of the maximum queue lengths (in vehicles) observed
per signal cycle at the intersection of Illinois Route 83 and 22nd Street
are presented in Table 16. The data indicate an increase in the number
of queued vehicles per cycle as volumes on the approach increase.
Several mathematical models for predicting queue lengths have been
developed. As part of the analysis of queue formation, two different
models were analyzed. The first model investigated is frequently
used in operations research problems, and it is also used by traffic
engineers for the design of toll booth or drive-in bank storage areas.
The model, or the "Toll Booth Formula," assumes a Poisson distribution
of arrivals and an exponential distribution of service time. The
average number of vehicles in the queue is given as:
N =
where: N = expected number of vehicles waiting.
V = expected number of arrivals per unit time (equivalent
to approach demand volume) .
C = expected number of units departing per unit time
(equivalent approach capacity).
The time unit of the volume and capacity variables has no effect on
the number of expected vehicles in the queue as long as the volume
demand to capacity ratio is constant over each unit of time. Using
the observed hourly data on approach volumes, and the hourly capacity
calculations (based on the level of service calculations presented in
Appendices G and H) , the expected number of waiting vehicles was cal-
culated. This calculated queue length was then plotted against the
observed values. These plots are shown in Figure 28. The graphs
indicate the Toll Booth Formula almost consistently underestimates the
queue length by a factor of two or greater. Some inconsistency is
due to the Toll Booth Formula averaging the queue over both the green
and red signal phases. Corrections for this discrepancy would increase
76
-------
Table 16. AVERAGE OBSERVED VEHICLES QUEUED PER SIGNAL CYCLE
Hour
of
day
8
9
10
11
12
13
14
15
16
17
18
19
Number of vehicles queued
North leg
Left-
turn lane
(1)
1.0
5.5
2.0
3.0
1.0
2.0
1.3
3.4
3.0
2.8
2.8
1.0
Through
lanes
(2)
4.0
8.0
7.2
6.3
7.5
10.5
8.3
11.8
16.8
23.3
28.4
11.5
South leg
Left-
turn lane
(1)
6.0
3.0
3.0
3.0
2.0
2.3
1.5
1.5
2.9
3.0
3.0
2.0
Through
lanes
(2)
19.9
13.3
9.0
5.3
7.8
8.4
8.9
10.4
10.9
12.7
10.6
6.5
East leg
Through
lanes
(2)
4.4
5.9
6.6
8.3
12.2
20.6
16.8
10,6
16.5
33.4
33.1
15.0
West leg
Through
lanes
(2)
18.4
24.5
13.2
11.0
16.5
20.8
32.6
1.3.5
13.0
17.2
16.4
10.9
the calculated values by only a factor of between 1.10 and 1.70. This
would not improve the performance of this formula significantly.
The second model tested was a modification of the "Red Time Formula"
which is typically used by traffic engineers to calculate storage
length requirements for intersection approaches. This modified
formula is given by:
N =
(V) (l-G'/GO
CPH
where N = expected number of vehicles queued per signal cycle.
V = expected number of arrivals per hour (hourly volume
demand).
G'/C'
CPH = number of signal cycles per hour.
(2)
green time provided the approach as percent of total
cycle time ( (1-G/C/) is the red time plus amber time).
77
-------
NoRfH APPROACH
5 10
OeSERVED QUEUE LENGTH IN VEHICLES
SOUTH APPROACH
10-v>
CALCULATED QUEUE LEWiTH IN VSHICLEJ
NOTE : TOLL BOOTH FORMULA
Figure 28. Calculated versus observed queue length using tollbooth
formula
78
-------
EAST APPROACH
5 050-f-
2
0 020- •
O
5 o 10 - • * . „' .
5
10 JO 30
OOSIKVID QUEUE LENGTH IN VEHICLE!
WEST APPROACH
I 10-
o
8 5-
ra 20 30
OBSERVED QUEUE LENGTH IN VEHICLES
MOTE . TOLL OOOTH FORMULA
N.-J*L
CIC V|
Figure 28 (continued).
Calculated versus observed queue length using
tollbooth formula
79
-------
Using the data on hourly volumes and average cycle length observed
at the intersection, and'the calculated G'/C' ratios (see Appendix G
for a discussion of provided G//C/ ratio calculations), the queue
lengths predicted by the Red Time Formula were also plotted against
the observed values. The results are displayed in Figure 29. The
graphs indicate that although the Red Time Formula in most cases
slightly over predicts the queue length, it estimates the queue length
much better than the Toll Booth Formula.
Further refinements of the simple Red Time Formula can easily be made.
A factor to account for a Poisson distribution of arrivals can be
used along with a desired confidence level to determine more accurately
the expected number of arrivals. Also, the (l-C'/C') factor can be
modified to account for first vehicle and overall queue delay times
which result from delayed driver reactions to the green indication.
Several additional studies using these refinements should be consulted
for more detailed information. ' '
Another important aspect of queue formation is the average waiting
time per vehicle in the queue. Table 17 summarizes the waiting times
of vehicles queued at the intersection of Route 83 and 22nd Street.
Table 17. VEHICLE WAITING TIMES AND ESTIMATED RED TIME
Minimum time (seconds)
Maximum time (seconds)
Average time (seconds)
Standard deviation
Average red plus amber
time (estimated)
North leg
Left
turn
2.0
121.0
36.9
28.6
144
Through
3.0
180.0
27.2
27.5
104
South leg
Left
turn
1.0
84.0
36.9
22.6
144
Through
2.0
138.0
30.9
23.0
104
East leg
All
move-
ments
2.0
191.0
47.8
40.6
88
West leg
All
move-
ments
2.0
128.0
40.5
28.3
104
80
-------
NORTH APf-ROACH
M Of
g
;
X
I JOO
I
10 0 30.0 30.0
OSSEBVtO QUEUE LENGTH IH VEHICLE!
40.0 500
SOUTH APPROACH
50.0 -r
200- - *
. 10.0 50.0 300
OBSERVED QUEUE LENGTH IN VEHICLES
•WO 50.0
NOTE: RED TIME FORMULA
N _ Hi^i/ci)
CPH
Figure 29. Calculated versus observed queue length using red time
formula
81
-------
EAST APPHOACH
MO T
100 20 0 30 0
O6SERVED QUEUE LENGTH IN VEHICLES
400 500
WEST APPROACH
3
w
O 100- •
a
20M)
10 0 20.0 30.0
OBSERVED QUEUE LEKGTH IN VEHICLE*
NOTE: RED TIME FORMULA.
N = -^
CPU
Figure 29 (continued). Calculated versus observed queue length using
red time formula
82
-------
An analysis of queue time versus queue position was inconclusive. The
major east and west through movements experienced significantly more
delay than the north and south through movements. The estimated
average red plus amber times (cycle time minus green time) for the
approaches is also given in Table 17. The average waiting time per
vehicle varies from 25 to 55 percent of the estimated red plus amber
time for the approaches.
The estimation of emission intensities at intersections requires a
detailed evaluation of all vehicle operating modes. The approach
speeds, being a function of the midblock V/C ratio, are sensitive to
the capacity of the roadways. Roadway width and surging characteristics
are key determinants of the capacity of a roadway. Increased roadway
widths (number of lanes) can significantly reduce the V/C ratio and
thus allow for higher average speeds. However, for most urban and
suburban conditions, increases beyond level of service C will not
generally yield significantly increased average speeds. In many cases,
the posted speed limit will restrict the maximum speeds.
The acceleration, deceleration, and queuing characteristics are related
to the intersection control. The total number of vehicles having to
stop at a signalized intersection depends on the arrival pattern of the
traffic and the G//C/ ratio provided by the signal control. Assuming
an uncontrolled (random) arrival rate, the number of vehicles queued
during an hour is approximately equal to the volume demand times the
hourly percent cycle time given to red and amber time (or V x (1-G//C/).
As the green time per cycle for an approach increases, the number of
vehicles queued decreases. However, the green time given to one approach
is usually limited by the green time required by an opposing leg. In-
creasing the G//C/ ratio will thus have a limited effect in reducing
the queuing at the intersection as a whole, although queue lengths
could be "traded" among approaches.
83
-------
A more promising measure to reduce queuing is to modify the vehicle
arrival pattern. This can be done by progressive signal systems. A
progressive signal system times the release of traffic at signal-con-
trolled intersections so that vehicles arrive at succeeding intersec-
tions during the green phase, provided they travel at the proper speed.
Such a coordinated signal system will maximize the use of green time
for the direction of travel being progressed, and it will reduce the
vehicle start-up and queue delays. The successive intersections must
be free of queues when the progressed vehicles arrive.
To determine the effect of signal progression on queuing at an inter-
section, the designer must provide an estimate of the percent of total
approach traffic which has been progressed. The percent of vehicles
progressed is essentially the progression bandwidth divided by the
cycle length, and the bandwidth is the width (seconds) of the "window"
of green in a given direction which allows the vehicles to proceed
unopposed through the succeeding intersection. This information is
generally available from the analysis of the traffic distribution at a
regional center. This percent/100 can be designated as P, the pro-
gression factor. The through traffic adjacent to a shopping center
will often be progressed since it represents the majority of traffic
flow, and the value of P may vary from 0.40 to 0.70. Assuming this
platooned traffic proceeds through the intersection with only minor
disruptions due to turning movements, the emissions generated by the
progressed traffic can be estimated by the cruise emission strength.
The volume of progressed traffic is simply V = P (V total) and the
speed is the design progression speed. Techniques to model the effects
of a progression on emissions have been developed.
The remaining nonprogressed traffic (V-V ) should be analyzed according
to the following procedures developed for determining the operating cha-
racteristics of vehicles at intersections.
84
-------
1, Determine maximum 8-hour and 1-hour volumes by approach
according to procedures outlined in Section IV.
2. Determine the approach speed by calculating the midblock
V/C ratio and using the appropriate chart relating V/C
to average speeds,, Capacity should be based on analysis
of a continuous roadway section using factors for road-
way width and trucks.
3. Calculate the provided G' /C> ratio for each approach ac-
cording to the procedure described in Appendix G. The
total (G + Y)/C ratio for all approaches should be
< 1,00 at level of service D for worst conditions.
(Y represents amber time).
40 The number of vehicles decelerating, accelerating and
idling during a given hour (by approach) is calculated
by NIJQ = V x (1-G'/C')« The average number of vehicles
queued per cycle length is given by the Red Time Formula.
5, The volume of vehicles passing through the intersection
during an hour (at a constant speed equal to approach
speeds) is given by V-N^.
The cycle length for a traffic actuated signal varies with demand simi-
larly to the G/C ratio. Typical cycle lengths can vary from 60 seconds
to 180 seconds for arterial roads. At Oakbrook Shopping Center, the
cycle length at Route 83 and 22nd Street intersection varied from 140 to
210 seconds throughout the day. Table 18 indicates average cycle lengths
observed at Oakbrook Shopping Center by hour of the day.
The determination of cycle length (existing or proposed) is not a diffi-
cult process. The estimation of the existing cycle length required only
a brief field observation of the intersection during peak conditions,
In predicting future conditions, the designer usually selects the optimal
cycle lengths based on the approach speeds and the phasing necessary to
accommodate traffic at the desired level of service. Several methods for
determining optimal cycle lengths for fixed time or actuated signals are
available. \
• ' \
Cycle lengths for fixed-time signals are limited between 60 and 120
seconds due to equipment restrictions. Good estimates would be 60-,
85
-------
Table 18. OBSERVED CYCLE LENGTH - ROUTE 83 AND
22ND STREET INTERSECTION
Date
6/13/74
6/13/74
6/13/74
6/13/74
6/13/74
6/13/74
6/17/74
6/17/74
6/17/74
6/17/74
6/17/74
6/17/74
Day
Thursday
Thursday
Thursday
Thursday
Thursday
Thursday
Monday
Monday
Monday
Monday
Monday
Monday
Hour of
day
8
9
10
11
12
13
14
15
16
17
18
19
Average cycle
length (sees)
174
170
133
147
152
160
191
181
177
179
174
166
Cycles/
hour
20.7
21.2
27.1
24.5
*
23.7
22.1
18.8
19.9
20.3
20.1
20.7
21.7
75-, or 90-second cycle lengths for fixed time, depending upon the
progressive system (if any). Fixed-time signals are usually found only
in urban conditions where the signal coordination is essential.
Actuated signals are usually found in suburban and rural conditions,
where most signals are isolated, due to the great expense to coordinate
actuated signals. Cycle lengths for actuated equipment are highly
variable. However, cycle lengths of 90 or 120 seconds would be good
estimates. In all cases, local data should take precedence over these
estimates.
86
-------
SECTION V
MODEL DEVELOPMENT AND EVALUATION
This section describes a technique for estimating emission profiles at
intersections. Estimated profiles are then used as inputs to the HIWAY
model for comparison of calculated and observed carbon monoxide con-
centrations. A discussion of area source evaluation follows this. The
section concludes with the presentation of a methodology for estimating
concentrations near intersections.
ESTIMATION OF EMISSION PROFILES FROM QUEUING VEHICLES
A signalized intersection resolves conflicts between opposing streams
of traffic by alternately "blocking" and then allowing free passage of
vehicles on intersecting approaches. This feature, plus the relatively
fine detail in emission strength variations which must be known to
assess nearby concentrations adequately, imply that the widely used
"average route speed" method of computing emission strengths is not
suitable in this case. A detailed knowledge must be provided of the
emission variations from a vehicle undergoing mode changes from cruise
through deceleration to idle, and acceleration back to cruise. The
recentl;
detail.
9
recently developed "Modal Analysis Model" supplies this emission
The Modal Analysis Model computes total emissions of hydrocarbons, car-
bon monoxide, and oxides of nitrogen from a user-specified vehicle mix
through 1971. Any desired driving sequence falling within the range of
model applicability may be used. The model was modified for this study
87
-------
to calculate emissions at equal length intervals for a given "approach"
cruise speed, rate of deceleration, rate of acceleration, and "departure"
speed for an "average" vehicle representing a 1974 low altitude mix. A
1974 mix was obtained by calculating emissions for a 1971 mix using
Table 3.1.2.7 of AP-4210 and multiplying the results by the ratio of 1974
to 1971 test results. For carbon monoxide this ratio is 39/47. The
length interval was chosen as 8 meters, since this is the average distance
occupied by queued vehicles (front bumper to front bumper) used in the pro-
18
posed interim guidelines. Figure 30 shows the carbon monoxide emission .
profile calculated by the modified Modal Analysis Model for one vehicle
decelerating from 35 mph/sec at -2.75 mph/sec and accelerating from 0 back
to 35 mph at 2.50 mph/sec with no idle time. These values of speed, decel-
eration, and acceleration are representative of the values observed at
Oakbrook as described in Section IV. The units on the abscissa are meters
upstream (negative values) and downstream (positive values) from the inter-
section stop line which, for the single vehicle depicted here, is approxi-
mately the location of the front bumper when the vehicle is stopped. The
vehicle moves from left to right. The ordinate units are grams/8 meters -
time is not included here, but it will be later in the discussion of queu-
ing and signal cycle lengths. All that is of interest at this point is
the mass of CO emitted in each 8 meter length interval by a stopping and
starting (but not idling) vehicle.
The modeling of traffic flow requires simplifying assumptions to keep the
analysis tractable, and an idealized model of the behavior of queuing ve-
hicles was used in calculating emission profiles at intersections. All
vehicles which stop are assumed to decelerate at a constant rate from a
constant cruise speed. They queue up with each vehicle occupying an
8 meter interval, and then they accelerate at a constant rate back to
cruise speed. The emission profile is calculated by adding the emissions
from each vehicle in each interval according to the vehicle's speed and .
mode in the interval. It consists essentially of adding up 10 of the
curves shown in Figure 30, with each successive curve displaced -8 meters
from the previous curve; i.e., 8 meters upstream, and then subtracting out
the cruise emission component. Figure 31 shows the excess emission profile
88