Traffic Flow Theory & Simulation: S.P. Hoogendoorn

Download as pdf or txt
Download as pdf or txt
You are on page 1of 41
At a glance
Powered by AI
The key takeaways are about shockwave theory in traffic flow and how congestion occurs from both a microscopic and macroscopic perspective.

Shockwaves are boundaries between different traffic conditions (free flow and congested flow) that form when demand exceeds capacity, such as at bottlenecks. They can form due to stationary or temporary bottlenecks as well as moving bottlenecks.

Congestion occurs microscopically as a queuing process when the service time of vehicles exceeds the time between arrivals, leading to an accumulation of waiting times. Macroscopically, congestion occurs when demand exceeds the capacity of a road section or bottleneck, leading to excess demand being stored in queues.

Traffic Flow Theory

& Simulation
S.P. Hoogendoorn

Lecture 4
Shockwave theory
Shockwave theory I:
Introduction
Applications of the Fundamental Diagram

February 14, 2010

Vermelding onderdeel organisatie


Intro to shockwave analysis

Introduce application of fundamental diagram to shockwave


analysis with aim to understand importance of field location

Shockwave analysis:
Vehicles are conserved
Traffic acts according to the fundamental diagram (q = Q(k))
Predicts how inhomogeneous conditions change over time

FOSIM demonstration
Example 3 -> 2 lane drop and emerging shockwaves
(roadworks, incident, etc.)

February 14, 2010 2


FOSIM example

Extremely short introduction to FOSIM


Build a simple network (8 km road with roadworks at
x = 5 km to 6 km)
Implement traffic demand
Assume 10% trucks
Suppose that upstream traffic flow > capacity of bottleneck
What will happen?

February 14, 2010 3


Questions

Why does congestion occur


Macroscopically?
Microscopically?

Photo by My Europe Photo by Juicyrai on Flickr / CC BY NC

February 14, 2010 4


Microscopic description

Congestion at a bottleneck
Simplest is to compare the system to a (sort of) queuing system
Drivers arrive at a certain rate (demand) at specific time intervals
The n servers needs a minimum amount of time to process the
drivers (each lane is a server)
Service time T is a driver-specific (random) variable depending on
weather conditions, road and ambient conditions, etc.
(= minimum time headway of a driver)
Note that service time is directly related to car-following behavior
When another driver arrives when the server is still busy, he / she
has to wait a certain amount of time
Waiting time accumulates -> queuing occurs

February 14, 2010 5


Macroscopic description

Compare traffic flow to a fluidic (or better: granular) flow though


a narrow bottleneck (hour-glass, funnel)

If traffic demand at a certain location is larger


than the supply (capacity) congestion will occur
Capacity is determined by number of lanes,
weather conditions, driver behavior, etc.
Excess demand is stored on the motorway
to be served in the next time period

Photo by My Europe
Remainder: focus on macroscopic description

February 14, 2010 6


Questions

Why does congestion occur


Macroscopically?
Microscopically?

Where does congestion first occur?


Which traffic conditions (traffic phases) are encountered?
Where are these conditions encountered?

February 14, 2010 7


Free flow

Capacity

Congestion

space

Free flow

time
February 14, 2010 8
Definition of a shockwave

Consider over-saturated bottleneck


Traffic conditions change over space and time

Boundaries between traffic regions are referred to as shockwaves


Shockwave can be very mild (e.g. platoon of high-speed vehicles
catching up to a platoon of slower driver vehicles)
Very significant shockwave (e.g. free flowing vehicles approaching queue
of stopped vehicles)
February 14, 2010 9
Definition of a shockwave2

Shockwave is thus a boundary in the space-time domain that


demarks a discontinuity in flow-density conditions
Example: growing / dissolving queue at a bottleneck

February 14, 2010 10


Fundamental diagrams

What can we say about the FD at the different locations?


Some standard numbers (for Dutch motorways)
Capacity point ! n * 2200 pce/h/lane
Critical density ! n * 25 pce/km/lane
Jam-density ! n * 2200 pce/km/lane
Critical speed ! 85 km/h
Free speed: to be determined from speed limit
pce = person car equivalent

Exact numbers depend on specific characteristics of considered


location (traffic composition, road conditions, etc.)

February 14, 2010 11


Emerging traffic states

space
4 Free-flow
End of
bottleneck Stationary shockwave
3 Capacity between capacity
Bottleneck and congested conditions
location
2 What is the flow
Congestion inside congestion?

And upstream of the


1 congestion?

time

February 14, 2010 12


Emerging traffic states

space
4 Free-flow
End of
Q(k)
bottleneck
1
3 Capacity
4
Bottleneck
location 3 2
2
Congestion

1 k

Remainder:
focus on the
dynamics
time of this
shock!

February 14, 2010 13


Shockwave equations

Assume that flow-density relation Q(k) is known for all location x


on the road (i.e. different inside and outside bottleneck!)
Consider queue due to downstream bottleneck
Consider conditions in the queue
Flow q2 = Cb-n (capacity downstream bottleneck)
Speed, density follow from q2, i.e. u2 = U(q2) and k2 = q2/u2

Farther upstream of queue, we have conditions (k1,u1,q1)


Speed u1 > u2 " upstream vehicles will catch up with vehicles in
queue

February 14, 2010 14


Shockwave equations2
Bottleneck
upstream cond.
(q1,k1,u1) x
Vehicles
downstream cond. in queue
(q2,k2,u2)
1

Shockwave
Bottleneck
2
capacity
u2
Vehicles
u2 upstream
u1

u1

February 14, 2010 15


Shockwave equations3

February 14, 2010 16


Shockwave equations4

Relative speed traffic flow region 1 with respect to S: u1 #12


Thus flow out of region 1 into shock S equals (explanation)

Relative speed traffic flow region 2 with respect to S: u2 #12


Thus flow into region 2 out of the shock must be

Conservation of vehicles over the shock (shock does not destroy


or generate vehicles)

February 14, 2010 17


Shockwave equations5

Shockwave speed #12 thus becomes

The speed of the shock equals the ratio of


jump of the flow over the shock S and
jump in the density over the shock S

February 14, 2010 18


Shockwave equations6
Bottleneck
x
Vehicles
Jump in flow
in queue

1
Shockwave
2
u2
Vehicles
u2 upstream
u1 Jump in density

u1

February 14, 2010 19


Shockwave equations7

Remarks:
If k2 > k1 sign shockwave speed negative if q1 > q2 (backward
forming shockwave)
If k2 > k1 sign shockwave speed positive if q1 < q2 (forward
recovery shockwave)
If k2 > k1 sign shockwave speed zero if q1 = q2 (backward
stationary)

Classification of shockwaves

February 14, 2010 20


Final remarks

Shockwave theory is applicable when


Q(k) is known for all location x
Initial conditions are known
Boundary conditions (at x1 AND x2) are known

Shockwaves occur when


Spatial / temporal discontinuities in speed-flow curve
(expressed Q(k,x)), e.g. recurrent bottleneck
Spatial discontinuities in initial conditions
Temporal discontinuities in boundary conditions at x1 (or x2)

February 14, 2010 21


Applications of shockwave theory
Temporary over-saturation
Traffic lights

February 14, 2010

22

Vermelding onderdeel organisatie


Shockwaves at a bottleneck

Temporary over-saturation of a bottleneck


Traffic demand (upstream)

q2

q1
t1 t2

February 14, 2010 23


Application of shockwave analysis

Three simple steps to applying shockwave theory:


1. Determine the Q(k) curve for all locations x
2. Determine the following external conditions:
initial states (t = t0)
boundary states (inflow, outflow restrictions, moving
bottleneck).
present in the x-t plane and the q-k plane
3. Determine the boundaries between the states (=shockwaves)
and determine their dynamics
4. Check for any ommisions you may have made (are regions with
different states separated by a shockwave?)

February 14, 2010 24


Shockwaves at bottleneck2

3 3 capacity of
bottleneck
1
1

t1 t2 t

February 14, 2010 25


Exercise temporary blockade
Which shocks emerge
Fundamental diagram Duration of disturbance?
without capacity drop Draw a couple trajectories
What if q1 = 2200?

February 14, 2010 26


Studies of the fundamental diagram

Need for complete diagram or only a part of it? Will it in general


be possible to determine a complete diagram at a cross-section?
Is the road section homogeneous? Yes: observations at a single
cross-section. No: road characteristics are variable over the
section and a method such as MO might be suitable
Mind the period of analysis:
too short (1 minute): random
fluctuations much influence;
too long (1 hour): stationarity cannot
be guaranteed (mix different regimes)
Estimate parameters of the model chosen
using (non-linear) regression analysis

February 14, 2010 27


Studies of the fundamental diagram2

Many models will fit


your data
Hints for choosing
models?

Simplest model
possible
(parsimony)
Interpretation of
parameters
Theoretical
considerations

February 14, 2010 28


Studies of the fundamental diagram3

Demonstration
FOSIM
Fundamental
diagram
determined from
real-life data, by
assuming
stationary
periods
Dependent on
measurement
location
Flow per lane

February 14, 2010 29


Studies of the fundamental diagram4

Estimation of free flow capacity using fundamental diagram


Approach 1:
Fit a model q(k) to available data
Consider point dq/dk = 0
Generally not applicable to motorway traffic because
dq/dk = 0 does not hold at capacity
Approach 2:
Assume fixed value for the critical density kc
Estimate only free-flow branch of the diagram
More for comparative analysis

February 14, 2010 30


Studies of the fundamental diagram5

Application example: effect of roadway lighting on capacity


Two and three lane motorway
Before after study
Difficulties due to different conditions (not only ambient conditions
change)
See e.g. site SB daylight before after
Effect lighting on capacity approx 2.5% (2 lane) or 1.6% (3 lane)

February 14, 2010 31


Studies of the fundamental diagram6

Effect on rain on capacity / fundamental diagram

February 14, 2010 32


Studies of the fundamental diagram7

Effect on rain on capacity / fundamental diagram

February 14, 2010 33


Studies of the fundamental diagram8

Estimating queue discharge rate


Only in case of observations of oversaturated bottleneck
Three measurement locations (ideally)
Upstream of bottle-neck (does overloading occur?)
Downstream of bottle-neck (is traffic flow free?)
At the bottle-neck (intensities are capacity measurements if
traffic state upstream is congested and the state downstream
is free)
Flow at all three points equal (stationary conditions) and at
capacity; use downstream point

February 14, 2010 34


Summary of lecture

Fundamental diagram for a lane and a cross-section


Shockwave equations
Application of shockwave analysis
shockwave at bottleneck
Establishing a fundamental diagram from field observations

February 14, 2010 35


Shockwaves signalized intersections

B
B A D

B
A C

A
A

D C

February 14, 2010 36


Shockwave classification

6 types of shockwaves

Which situations do
they represent?
Examples?

Rear stationary

February 14, 2010 37


Shockwave classification2

1. Frontal stationary: head of a queue in case of stationary /


temporary bottleneck
2. Forward forming: moving bottleneck (slow vehicle moving in
direction of the flow given limited passing opportunities)
3. Backward recovery: dissolving
queue in case of stationary or
temporary bottleneck
(demand l.t. supply); forming
or dissolving queue for
moving bottleneck

February 14, 2010 38


Shockwave classification3

1. Forward recovery: removal of temporary bottleneck (e.g.


clearance of incident, opening of bridge, signalized intersection)
2. Backward forming: forming queue in case of stationary,
temporary, or moving bottleneck* (demand g.t. supply);
3. Rear stationary: tail of queue
in case recurrent congestion
when demand is approximately
equal to the supply

February 14, 2010 39


Flow into schockwave

Consider a shockwave moving with speed


Flow into the shockwave = flow observed by moving observer
travelling with speed of shockwave
Number of vehicles
x observed on S =
+ Veh. passing x0 during T
X - Vechicles on X at t1

x0

T
t1 t
February 14, 2010 40

You might also like