Module 3 Handout 1
Module 3 Handout 1
Traffic Stream
Models
Outline
Learning Outcomes
1. Explain the speed-flow-density relationship;
2. Understand the fundamental diagram of traffic flow; and
3. Apply the knowledge in calculus in exploring Greenshield’s and Greenberg model
INTRODUCTION
While the preceding definitions and relationships provide the basis for the measurement
and calculation of traffic stream parameters, it is also essential to understand the interaction of
the individual macroscopic measures in order to fully analyze the operational performance of
traffic streams. The models that describe these interactions are discussed in the following
sections, and it will be shown that Eq. 1 serves the important function of linking specific models
of traffic into a consistent, generalized model.
𝑞 = 𝑢𝑘 Equation 1
where
Based on the definitions presented, a simple identity provides the basic relationship
among traffic flow, speed (space-mean), and density (denoting space-mean speed, us as
simply u for notational convenience).
where
𝑘2
𝑞 = 𝑢𝑓 (𝑘 − 𝑘 ) Equation 3
𝑗
The general form of Eq. 3 is shown in Fig. 2. Note in this figure that the maximum flow
rate, qcap, represents the highest rate of traffic flow that the highway is capable of handling. This
is referred to as the traffic flow at capacity, or simply the capacity of the roadway. The traffic
density that corresponds to this capacity flow rate is kcap, and the corresponding speed is ucap.
Equations for qcap, kcap, and ucap can be derived by differentiating Eq. 3, because at maximum flow
𝑑𝑞 2𝑘
= 𝑢𝑓 (1 − 𝑘 ) = 0 Equation 4
𝑑𝑘 𝑗
𝑘𝑗
𝑘𝑐𝑎𝑝 = Equation 5
2
𝑘 𝑢𝑓
𝑢𝑐𝑎𝑝 = 𝑢𝑓 (1 − 2𝑘𝑗 ) = Equation 6
𝑗 2
𝑢𝑓 𝑘𝑗
𝑞𝑐𝑎𝑝 = 𝑢𝑐𝑎𝑝 𝑘𝑐𝑎𝑝 = Equation 7
4
3.3 SPEED-FLOW MODEL
Again returning to the linear speed–density model (Eq. 2), a corresponding speed-flow
model can be developed by rearranging Eq. 2 to
𝑢
𝑘 = 𝑘𝑗 (1 − 𝑢 ) Equation 8
𝑓
𝑢2
𝑞 = 𝑘𝑗 (𝑢 − 𝑢 ) Equation 9
𝑓
The speed-flow model defined by Eq. 9 again gives a parabolic function, as shown in Fig. 3. Note
that Fig. 3 shows that two speeds are possible for flows, q, up to the highway’s capacity, qcap
(this follows from the two densities possible for given flows as shown in Fig. 2). It is desirable, for
any given flow, to keep the average space-mean speed on the upper portion of the speed-flow
curve (above ucap). When speeds drop below ucap, traffic is in a highly congested and unstable
condition.
Data have been collected that tend to confirm the argument postulated above but there
is some controversy regarding the exact shape of the curve. A similar argument can be postulated
for the general relationship between the space mean speed and the flow. When the flow is very
low, there is little interaction between individual vehicles. Drivers are therefore free to travel at
the maximum possible speed. The absolute maximum speed is obtained as the flow tends to 0,
and it is known as the mean free speed (uf). The magnitude of the mean free speed depends on
the physical characteristics of the highway. Continuous increase in flow will result in a continuous
decrease in speed. A point will be reached, however, when the further addition of vehicles will
result in the reduction of the actual number of vehicles that pass a point on the highway (that is,
reduction of flow). This results in congestion, and eventually both the speed and the flow become
0. Figure 4c shows this general relationship. Figure 4b shows the direct relationship between
speed and density.
From Eq. 1, space mean speed is flow divided by density which makes the slopes of lines
0B, 0C, and 0E in Figure 4a represent the space mean speeds at densities kb, kc, and ke,
respectively. The slope of line 0A is the speed as the density tends to 0 and little interaction exists
between vehicles. The slope of this line is therefore the mean free speed (u f); it is the maximum
speed that can be attained on the highway. The slope of line 0E is the space mean speed for
maximum flow. This maximum flow is the capacity of the highway. Thus, it can be seen that it is
desirable for highways to operate at densities not greater than that required for maximum flow.
Macroscopic Approach
The macroscopic approach considers traffic streams and develops algorithms that relate
the flow to the density and space mean speeds. The two most commonly used macroscopic
models are the Greenshields and Greenberg models.
Corresponding relationships for flow and density and for flow and speed can be developed.
Since q = usk, substituting q>us for k in Eq. 10 gives
𝑢
𝑢𝑠2 = 𝑢𝑓 𝑢𝑠 − 𝑘𝑓 𝑞 Equation 11
𝑗
𝑢
𝑞 = 𝑢𝑓 𝑘 − 𝑘𝑓 𝑘 2 Equation 12
𝑗
Eqs. 11 and 12 indicate that if a linear relationship in the form of Eq. 10 is assumed for speed and
density, then parabolic relationships are obtained between flow and density and between flow
and speed. The shape of the curve shown in Figure 4a will therefore be a parabola. Also, Eqs. 11
and 12 can be used to deter- mine the corresponding speed and the corresponding density for
maximum flow.
Consider Eq. 11:
𝑢𝑓
𝑢𝑠2 = 𝑢𝑓 𝑢𝑠 − 𝑞
𝑘𝑗
𝑢𝑓 𝑑𝑞
2𝑢𝑠 = 𝑢𝑓 −
𝑘𝑗 𝑑𝑢𝑠
that is
𝑑𝑞 𝑘𝑗 𝑘𝑗 𝑘𝑗
= 𝑢𝑓 − 2𝑢𝑠 = 𝑘𝑗 − 2𝑢𝑠
𝑑𝑢𝑠 𝑢𝑓 𝑢𝑓 𝑢𝑓
𝑑𝑞 𝑘 𝑢𝑓
=0 𝑘𝑗 = 2𝑢𝑠 𝑢 𝑗 𝑢0 = Equation 13
𝑑𝑢𝑠 𝑓 2
Thus, the space mean speed uo at which the volume is maximum is equal to half the free mean
speed.
𝑢𝑓 2
𝑞 = 𝑢𝑓 𝑘 − 𝑘
𝑘𝑗
𝑑𝑞 𝑢𝑓
= 𝑢𝑓 − 2𝑘
𝑑𝑘 𝑘𝑗
𝑑𝑞 𝑢𝑓 𝑘𝑗
𝑑𝑘
=0 𝑢𝑓 = 2𝑘 𝑘 2
= 𝑘0 Equation 14
𝑗
Thus, at the maximum flow, the density ko is half the jam density. The maximum flow for the
Greenshields relationship can therefore be obtained from Eqs. 1, 10, and 11, as shown in Eq. 15:
𝑘𝑗 𝑢𝑓
𝑞𝑚𝑎𝑥 = Equation 15
4
𝑘𝑗
𝑢𝑠 = 𝑐 𝑙𝑛 Equation 16
𝑘
𝑘𝑗
𝑢𝑠 𝑘 = 𝑐𝑘 𝑙𝑛 ; 𝑢𝑠 𝑘 = 𝑞
𝑘
𝑑𝑞 𝑘𝑗
= 𝑐 𝑙𝑛 − 𝑐
𝑑𝑘 𝑘
𝑑𝑞
=0
𝑑𝑘
then,
𝑘
𝑙𝑛 𝑘𝑗 = 1 Equation 17
0
𝑢0 = 𝑐
Use of these macroscopic models depends on whether they satisfy the boundary criteria
of the fundamental diagram of traffic flow at the region that describes the traffic conditions. For
example, the Greenshields model satisfies the boundary conditions when the density k is
approaching zero as well as when the density is approaching the jam density kj. The Greenshields
model can therefore be used for light or dense traffic. The Greenberg model, on the other hand,
satisfies the boundary conditions when the density is approaching the jam density but it does not
satisfy the boundary conditions when k is approaching zero. The Greenberg model is therefore
useful only for dense traffic conditions.
The traffic models discussed thus far can be used to determine specific characteristics,
such as the speed and density at which maximum flow occurs, and the jam density of a facility.
This usually involves collecting appropriate data on the particular facility of interest and fitting
the data points obtained to a suitable model. The most common method of approach is
regression analysis. This is done by minimizing the squares of the differences between the
observed and expected values of a dependent variable. When the dependent variable is linearly
related to the independent variable, the process is known as linear regression analysis. When the
relationship is with two or more independent variables, the process is known as multiple linear
regression analysis.
𝑦 = 𝑎𝑥 + 𝑏 Equation 18
1 𝑏
𝑎 = 𝑛 ∑𝑛𝑖=1 𝑦𝑖 − 𝑛 ∑𝑛𝑖=1 𝑥𝑖 = 𝑦̅ − 𝑏𝑥̅ Equation 19
1
∑𝑛 𝑛 𝑛
𝑖=1 𝑥𝑖 𝑦𝑖 −𝑛(∑𝑖=1 𝑥𝑖 )(∑𝑖=1 𝑦𝑖 )
𝑏= 1 2 Equation 20
∑𝑛 2 𝑛
𝑖=1 𝑥𝑖 −𝑛(∑𝑖=1 𝑥𝑖 )
where
∑𝑛 (𝑌 −𝑦̅)2
𝑅2 = ∑𝑖=1 𝑖
𝑛 (𝑦 −𝑦
̅)2
Equation 21
𝑖=1 𝑖
where Yi is the value of the dependent variable as computed from the regression equations.
The closer R2 is to 1, the better the regression fits.
A section of highway is known to have a free-flow speed of 55 mi/h and a capacity of 3300
veh/h. In a given hour, 2100 vehicles were counted at a specified point along this highway section.
If the linear speed–density relationship shown in Eq. 2 applies, what would you estimate the
space-mean speed of these 2100 vehicles to be?
SOLUTION
The jam density is first determined from Eq. 7 as
4𝑞𝑐𝑎𝑝
𝑘𝑗 =
𝑢𝑓
4(3300)
𝑘𝑗 = = 240 𝑣𝑒ℎ/𝑚𝑖
55
𝑢2 𝑘𝑗 2
𝑞 = 𝑘𝑗 (𝑢 − ) → 𝑢 − 𝑘𝑗 𝑢 + 𝑞 = 0
𝑢𝑓 𝑢𝑓
Substituting,
240 2
𝑢 − 240𝑢 + 2100 = 0
55
which gives u = 44.08 mi/h or 10.92 mi/h. Both of these speeds are feasible, as shown in
Fig. 3.
EXAMPLE 2. TRAFFIC FLOW MODEL
Free-flow speed of vehicles on a road section is 88 km/h. Jammed density is 228 veh/km.
Estimate the traffic density on the road if traffic flow is at a maximum level. Estimate also the
possible maximum flow for the road section and the average speed of vehicles at that maximum
traffic volume.
SOLUTION
Density when traffic flow reaches the maximum volume, use Eq.5
𝑘𝑗 228
𝑘𝑐𝑎𝑝 = = = 114 𝑣𝑒ℎ/𝑘𝑚
2 2
𝑘𝑗 𝑢𝑓 228(88) 𝑣𝑒ℎ
𝑞𝑚𝑎𝑥 = = = 5016
4 4 ℎ𝑟
𝑢𝑓 88
𝑢0 = = = 44 𝑘𝑚/ℎ𝑟
2 2
A highway section has an average spacing of 25ft under jam conditions and a free-flow
speed of 55mph. Assuming that the relationship between speed and density is linear, determine
the jam density, the maximum flow, the density at maximum flow, and the speed at maximum
flow.
SOLUTION
1 1 𝑣𝑒ℎ
𝑘𝑗 = = = 212.7~213
𝑠 25 𝑚𝑖
5280
𝑘𝑗 𝑢𝑓 213(55) 𝑣𝑒ℎ
𝑞𝑚𝑎𝑥 = = = 2929
4 4 ℎ𝑟
At maximum flow:
𝑘𝑗 213 𝑣𝑒ℎ
𝑘𝑐𝑎𝑝 = = = 106.5~107
2 2 𝑚𝑖
At maximum flow:
𝑢𝑓 55 𝑚𝑖
𝑢0 = = = 27.5
2 2 ℎ𝑟
A road has capacity 4000 veh/hr, and a free-flow speed of 50 mi/hr. If the density is 100
veh/mi, what is the speed and flow?
SOLUTION
𝑘𝑗 𝑢𝑓
𝑞𝑚𝑎𝑥 =
4
(50)𝑘𝑗
4000 =
4
𝑣𝑒ℎ
𝑘𝑗 = 320
𝑚𝑖
Speed-density relationship,
𝑘 100
𝑢 = 𝑢𝑓 (1 − ) = 50 (1 − ) = 34.37 𝑚𝑖/ℎ𝑟
𝑘𝑗 320
Flow-density relationship,
𝑘2 1002 𝑣𝑒ℎ
𝑞 = 𝑢𝑓 (𝑘 − ) = 50 (100 − ) = 3437.5~3438
𝑘𝑗 320 ℎ𝑟
EXAMPLE 5. FITTING SPEED AND DENSITY DATA TO THE GREENSHIELD’S MODEL
Use the data shown in Table 1 (columns 1 and 2) to demonstrate the use of the method
of regression analysis in fitting speed and density data to the macro- scopic models.
SOLUTION
𝑢𝑓
𝑢𝑠 = 𝑢𝑓 − 𝑘
𝑘𝑗
∑ 𝑦𝑖 = 404.8
∑ 𝑥𝑖 = 892
𝑥̅ = 63.71 𝑦̅ = 28.91
∑ 𝑥𝑖 𝑦𝑖 = 20,619.8
∑ 𝑥𝑖2 = 66,628
Table 2 Computations
𝑎 = 28.91 − 63.71𝑏
892(404.8)
20,619.8 −
𝑏= 14 = −0.528
(892)2
66,628 − 14
then,
Since a = 62.55 and b = - 0.53, then uf = 62.55 mi/h, uf /kj = 0.53, and so kj = 118.02 veh/mi,
and us = 62.55 - 0.53k.
2
∑𝑛𝑖=1(𝑌𝑖 − 𝑦̅)2
𝑅 = 𝑛 = 0.95
∑𝑖=1(𝑦𝑖 − 𝑦̅)2
Using the above estimated values for uf and kj, determine the maximum flow from Eq. 15 as
𝑘𝑗 𝑢𝑓 118.02(62.55)
𝑞𝑚𝑎𝑥 = = = 1846 𝑣𝑒ℎ/ℎ𝑟
4 4
Using Eq. 13 we also obtain the velocity at which flow is maximum,
𝑢𝑓 62.55
𝑢0 = = = 31.3𝑚𝑖/ℎ𝑟
2 2
𝑘𝑗 118.02
𝑘0 = = = 60 𝑣𝑒ℎ/ℎ𝑟
2 2
EXAMPLE #6. FITTING SPEED AND DENSITY DATA TO THE GREENBERG MODEL
The data in Table 2 can also be fitted into the Greenberg model shown in Eq. 16:
𝑘𝑗
𝑢𝑠 = 𝑐 𝑙𝑛
𝑘
𝑢𝑠 = 𝑐 𝑙𝑛 𝑘𝑗 − 𝑐 𝑙𝑛 𝑘 Equation 22
Comparing Eq. 22 and the estimated regression function Eq. 18, us in the Greenberg
expression is represented by y in the estimated regression function, c ln kj is represented by a, c
is represented by b, and ln k is represented by x. Table 4 shows values for xi, xiyi, and x2i (Note
that these values are computed to a higher degree of accuracy since they involve logarithmic
values.)
Table 4 Computations
therefore,
∑ 𝑦𝑖 = 404.8
∑ 𝑥𝑖 = 56.72
𝑥̅ = 4.05 𝑦̅ = 28.91
∑ 𝑥𝑖 𝑦𝑖 = 1547.03
∑ 𝑥𝑖2 = 233.04
𝑎 = 28.91 − 4.05𝑏
56.72(404.8)
1547.03 −
𝑏= 14 = −28.67
(56.72)2
233.04 −
14
then,
Since a = 145.02 and b = - 28.67, the speed for maximum flow is c=28.67 mi/h.
Finally, since
𝑐𝑘 ln 𝑘𝑗 = 145.02
145.02
ln 𝑘𝑗 = = 5.06
28.67
𝑘𝑗 = 157.32 𝑣𝑒ℎ/𝑚𝑖
𝑘𝑗 157.32
𝑢𝑠 = 𝑐 𝑙𝑛 = 28.67 ln
𝑘 𝑘
Obtaining ko, the density for maximum flow from Eq. 17, then use Eq. 1 to determine the value
of the maximum flow.
𝑘𝑗
𝑙𝑛 =1
𝑘0
ln 𝑘𝑗 = 1 + ln 𝑘0
ln 157.32 = 1 + ln 𝑘0
𝑘0 = 57.87
𝑣𝑒ℎ
𝑞𝑚𝑎𝑥 = 𝑢𝑘 = 28.67(57.87) = 1660
ℎ𝑟
Using Eq. 21 to determine R2
2
∑𝑛𝑖=1(𝑌𝑖 − 𝑦̅)2
𝑅 = 𝑛 = 0.92
∑𝑖=1(𝑦𝑖 − 𝑦̅)2
The R2 based on the Greenberg expression is 0.92 which indicates that the Greenshields
expression is a better fit for the data in Table 2. Figure 5 shows plots of speed versus density for
the two estimated regression functions obtained and also for the actual data points. Figure 6
shows similar plots for the volume against speed.
Solution
𝑉𝑠 = 57.5 − 0.46𝑘
𝑘
Recall EQ. 2 𝑢 = 𝑢𝑓 (1 − 𝑘 )
𝑗
𝑢 = 𝑢𝑓 − 𝑘𝑘 𝑢𝑓
𝑗
𝑢
𝑢 = 𝑢𝑓 − 𝑘𝑓 𝑘
𝑗
𝑢𝑓 𝑢 𝑓
𝑉𝑠 = = 𝑢; 𝑢𝑓 = 57.5 𝑚𝑝ℎ = 0.46, then 𝑘𝑗 = 0.46 = 125 𝑣𝑝𝑚
𝑘𝑗
𝑞 = 𝑉𝑠 𝑘 (EQ.1)
𝑞 = (57.5 − 0.46𝑘)𝑘
𝑞 = 57.5𝑘 − 0.46𝑘 2
57.5−𝑉𝑠
𝑉𝑠 = 57.5 − 0.46𝑘; isolating k 𝑘= = 125 − 2.17𝑉𝑠
0.46
Substituting k to EQ1.
𝑞 = 57.5𝑘 − 0.46𝑘 2
𝑑𝑞
= 57.5 − 0.46(2)𝑘 = 0
𝑑𝑘
57.5𝑘 = 0.46(2)𝑘
𝑘𝑐𝑎𝑝 = 62.5~63𝑣𝑝𝑚
𝑞 = 57.5𝑘 − 0.46𝑘 2
𝑞 = 57.5(63) − 0.46(63)2
𝑞 = 1796.76
References
Garber, N. J., & Hoel, L. A. (2019). Traffic and highway engineering. Cengage Learning
Homburger, W. S., Hall, J. W., Reilly, W. R., & Sullivan, E. C. (2007). Fundamentals of traffic engineering.