Evaluation of Incident Management Impacts on an Integrated Motorway and Arterial Networks Using Traffic Simulation

Mr William Gondwe (2008). Evaluation of Incident Management Impacts on an Integrated Motorway and Arterial Networks Using Traffic Simulation MPhil Thesis, School of Engineering, The University of Queensland.

Attached Files (Some files may be inaccessible until you login with your UQ eSpace credentials)
Name Description MIMEType Size Downloads
s3081039_MPhil_abstract.pdf Abstract application/pdf 24.17KB 0
s3081039_MPhil_totalthesis.pdf Total Thesis application/pdf 3.71MB 0
Author Mr William Gondwe
Thesis Title Evaluation of Incident Management Impacts on an Integrated Motorway and Arterial Networks Using Traffic Simulation
School, Centre or Institute School of Engineering
Institution The University of Queensland
Publication date 2008-11-29
Thesis type MPhil Thesis
Supervisor Dr Hussein Dia
Prof Phil Charles
Total pages 120
Subjects 290000 Engineering and Technology
Formatted abstract
Background: The task of quantifying the impacts of Incidents is an important
and necessary undertaking in order to justify the expenditure of public funds on
ITS and other road projects. For example, roadway incidents impose a
substantial cost to society when delays, congestion, secondary accidents, and
environmental emissions are taken into consideration. Incident impacts can be
substantially reduced through implementation of incident management
programs. In recent times, there has been a growing interest among
researchers and practitioners in developing traffic management plans for
integrated freeway and arterial road network. However, comprehensive
research tools for evaluating the benefits of these systems have not been fully
developed. For example, the impacts of incident management programs will
depend on the extent of existing congestion on the road network, and will also
vary according to the severity of the incident, its duration, and the time of day
when it occurs. The use of computer modelling, and microscopic traffic
simulation in particular, offers a cost-effective approach in which input
conditions can be varied (e.g. to reflect incidents during peak and non-peak
conditions) and the impacts of incidents on network performance can be
Aims and Objectives: This research aims to quantify the traffic and emissions
impacts of freeway incidents that occur on the Pacific Motorway (also called the
M1 Motorway between Brisbane and the Gold Coast), and evaluating the
benefits of incident management strategies. The primary objective of this
research was to demonstrate the feasibility of using AIMSUN NG microscopic
traffic simulation models to evaluate the impacts of incidents using field data
and other related information from the Nerang Traffic Management Centre
Methodology and Model Development: Unlike other previous similar studies
that have depended on limited data and field tests, this research involved the
development of a large scale simulation model covering an approximately 122
square Kilometres, including 43 kilometres of the Motorway and about 85
kilometres the arterial roads on the surrounding network. The model was
extensively calibrated and validated for the AM and PM Peak periods with
approximately 90,000 vehicles modelled inside the network during a 2-hourperiod. The calibration process involved a detailed examination of the global and
local parameters to ensure that the selected parameters produced results for
modelled traffic counts and travel times which were in close agreement with
field data. The validation process involved testing the calibrated model using a
traffic data set which was not used in model calibration. The M1 and arterial
model was validated using Traffic flow in different sections and link travel times,
in addition to average speeds.
Calibration and Validation Results: The calibration results for the M1 showed
errors of around 14 and 6 percent during the AM and PM peak periods,
respectively. For the arterials, the calibration errors were around 10 and 13
percent for the AM and PM peak periods, respectively. The calibrated M1 model
was validated using average speeds collected from loop detector sites on the
motorway. The validation results showed overall validation errors of around 8.5
percent, which is an excellent result given the large scale of the model. For the
arterial roads, three performance measures were used for validation. These
included mid-block traffic counts from 2006 and speeds and travel times from
2004. The results showed overall travel time errors between 0.6 and 18
percent; speed errors between 8 and 20 percent; and mid-block traffic count
errors between 12 and 24 percent. All of these errors are reasonable and
acceptable given the large scale of the model coverage and also given the
random nature of traffic. The results of the calibration and validation showed
that the model provided an accurate representation of field conditions and that
its precision is consistent with findings reported for similar large scale models.
These results provided a good degree of confidence in the model’s ability to
replicate field conditions and its suitability for use as a valid tool for modelling
traffic management and ITS applications on the M1 and surrounding arterial
network on the Gold Coast.
Evaluation of Incident Impacts: This research quantified the traffic and
emissions impacts of simulated freeway incidents which occurred on the Pacific
Motorway. A large number of incidents were simulated with durations of 1.0 to
1.5 hours where incidents were modelled to block either one or two lanes. A
total of 54 incidents were simulated for the morning AM Peak and 66 incidents
for the PM peak (total of 120 incident cases). A number of localised and
network-wide key performance indicators including average speeds, traveltimes, delays, number of stops together with fuel consumption, operating costs
and emissions were collected.
Local impacts: At the local level, the results for a major incident that blocked
two lanes for a duration of 1.5 hours showed that traffic flows in the section
immediately before the location of the incident reduced from 3,931 to 1,783
vehicles per hour due to the incident. This meant that the incident resulted in a
reduction of around 55 percent in the volume of traffic that passed through the
section. The results also showed that average section speed dropped from 95 to
60 kilometres per hour (a reduction of around 37 percent in average speeds)
and that the average section travel time increased from 19 to 145 seconds (an
increase of around 660 percent due to the incident). The average delay per
vehicle increased from 1 to 127 seconds and the average stopped time
increased from zero to 111 seconds per vehicle.
Network Impacts: At the network-wide level, the results showed that the 54
AM peak period incident cases resulted in average increases of 2.2 percent in
travel time; 5.7 percent in delays; 10 percent in densities; 11.5 percent in stop
times; and 11.1 percent in number of stops. The results for the PM model also
showed a similar trend where the 66 incident cases resulted in around 2.5
percent increase in travel time; 4.5 percent increase in delays; 8.3 percent
increase in densities along the freeway; 3.4 percent increase in stop times; and
25 percent increase in number of stops.
Fuel Consumption, Operating Costs and Emissions Impacts: The fuel
consumption, operating costs and emissions were collected using a powerengine
four-mode elemental model. The results showed that on average, the 54
incident cases which were modelled for the duration of the 2-hour AM peak
period resulted in 1.5 percent increase in CO emissions and fuel consumption
and 5.0 percent increase in operating costs. Similar results were obtained for
the PM peak period where the CO emissions and fuel consumption increased by
around 1.9 percent and the operating costs increased by 2.7 percent. The AM
incidents resulted in an average increase in operating costs of around $21,000;
and the PM incidents resulted in an average increase of around $13,000. These
results, when extrapolated to the real incident data set which was obtained from
the Nerang TMC, translate into a cost of around $2M for 117 incidents which
occurred between January 2004 and September 2006Evaluation of Effectiveness and Efficiency of Selected Incident
Management Strategies: This research examined the effectiveness and
efficiency of selected traffic management strategies in reducing the negative
impacts of incident-induced congestion. The research considered the following
incident management strategies: ramp metering; VMS information
dissemination and route diversions; and variable speed limit systems.
Ramp Metering: The ramp metering results showed that metering under
current conditions does not produce any substantial benefits on the mainline.
This was probably due to the fact that existing ramp volumes were not too
heavy. However, it was shown that ramp metering starts to provide substantial
benefits when traffic demands increase. For example, when the demand was
increased by 25 percent, the delays on the mainline were reduced by 10.5
percent; number of stops decrease by 23 percent and travel times are reduced
by 3 percent as a result of implementing ramp metering.
VMS Information and Route Diversion: This study also evaluated the
benefits of incident response in terms of provision of VMS information on the M1
and implementation of dynamic signal plans on diversion routes. An incident
was simulated on an arterial road (Smith Street) and VMS information was
provided on the Motorway advising motorists of incident conditions. The results
revealed that traffic adjustments due to the diversions and dynamic signal plans
resulted in equilibrium conditions on both the normal route (Smith Street) and
alternative diversion route (Nerang-Southport Road) when the diversion rate did
not exceed 30 percent. When the percentage of drivers diverting to Nerang-
Southport Road increased above 30 percent, the traffic volumes on the diversion
route increased and started to exceed the traffic volume on Smith Street. The
results also showed that the benefits were only realised when the two incident
management responses (VMS route diversion and implementation of incident
cycle plan 160 seconds) are implemented at the same time. The best benefits
were realised for diversion rates of 30% and resulted in the reduction of delays
by 8.8 percent (from 159 to 145 seconds per vehicle); increase in speeds by 4.5
percent (from 44 to 46 kph); decrease in number of stops by 22 percent (from 9
to 7); and decrease in travel time by 3.3 percent (from 451 to 436 seconds).
Furthermore, the maximum queue length reduced to around 70 metres which is
a substantial reduction that can be solely attributed to the VMS diversions and
incident signal plans.Variable Speed Limits: This study also conducted a preliminary investigation
of variable speed limits as a means to reduce the negative impacts of incidents.
The M1 microscopic traffic simulation model was used to test an 8 kilometre
section of the Motorway with the aim to ascertain whether VSL had a positive
impact on the safety and efficiency of the motorway. This was explored through
examination of flow homogenisation (reducing the variation of the speeds
between vehicles, both within a lane and adjacent lanes) and reducing
decelerations at the back of queued vehicles. The results showed that VSL has
the potential to provide an 11% improvement in efficiency and indicated that
VSL can be used as a measure to increase the efficiency of congested sections
on motorways. VSL was also found to provide safety and efficiency benefits by
homogenising the flow in higher speed regimes. The number of stops per vehicle
on the motorway reduced by 64% following the speed limit being reduced from
110 kph to 70 kph as a result of the incident. Clearly, these are preliminary
results using a basic implementations of a VSL system.
Overall, this thesis was successful in meeting the stated aims and objectives and
demonstrated the feasibility of using traffic simulation models to evaluate the
impacts of incidents and the benefits that can be obtained from selected incident
management strategies. The thesis also highlights the limitations of the
approach and emphasises that the impacts and benefits reported in this work
are network-specific and will vary according to the network geometry, traffic
demands and extent of existing congestion on the main and alternative routes.
Additional Notes Not necessary

Citation counts: Google Scholar Search Google Scholar
Created: Mon, 08 Dec 2008, 16:05:59 EST by Mr William Gondwe on behalf of Library - Information Access Service