Bill Chen

Application of Aimsun Hybrid Model in Transport PlanningBill Chen & Graeme Inglis

Agenda

1. Project background – Wilton Junction Precinct

2. Road network

3. Transport Management and Accessibility Plan

4. Strategic modelling

5. Mesoscopic modelling

6. Hybrid modelling

7. Discussion

8. Conclusion

2

Part of the Potential Housing Opportunities Program

Currently Wilton is a small village (1,890 people and 590

homes)

A new town is planned in Wilton Junction Precinct

Residential (about 12,000 dwellings)

Employment (11,000 jobs)

Retail (at a new town centre)

School and other facilities.

http://www.planning.nsw.gov.au/wiltonjunction

Project background – Wilton Junction Precinct

3

Hume Highway

Interstate highway linking Sydney and Melboune

Main north-south corridor accessing Wilton

Two lanes in each direction and speed of 110 km

B-double route

Picton Road

State road travelling between Picton and Mount

Ousley Road

Main west-east corridor accessing Wilton

one lanes in each direction

freight route to Port Kembla and a commuter traffic

route on weekdays

Other key roads including Remembrance Drive and

Appin Road/Wilton Road

Road network

4

Transport management and accessibility plan

5

TMAP

Support the application for the rezoning of Wilton

Junction Precinct

Travel requirements of the Precinct by road, public

transport, walking and cycling

3-Tier Transport modelling

identify the transport impacts of the Wilton Junction

Precinct

assess the provision and staging of the required

traffic and transport infrastructureone lanes in each

direction

Macro/Meso/Micro

Strategic modelling

6

Sydney Strategic Travel Model (STM)

Maintained by BTS and Sydney GMA

provide origin-destination (OD) demand and

patterns for the existing year

determine the future year trip generation, trip

distribution and mode split for development

Subarea matrices were extracted from STM for

study area to consider impact on wider regional

network

Mesoscopic modelling

7

Select modelling platform

Relatively recent development and has been gaining

popularity in NSW

Intermediate level of detail between Macroscopic

and Microsimulation

Quickly assign traffic to alternative routes at wider

regional area

Test initial intersection design option for the

development area with acceptable level of accuracy

AIMSUN software was selected

Model network development

Cover Bargo to Campbelltown (about 35 km in

distance) and Picton to the Illawarra Escarpment

(about 36 km in distance)

Zone aggregation and disaggreation

Mesoscopic modelling

8

OD demand estimation

Initial cordon subarea matrices extracted from STM

Restructure matrices to fit Meso zone system

Furness matrix estimation

Adjust matrices with static traffic assignment

Fine-tune at dynamic traffic assignment

Traffic assignment

Considering the rural environment of the study area,

stochastic dynamic assignment method

Traffic proportionally assigned to alternative travel

routes with preference over shortest routes

Mesoscopic modelling

9

Calibration and validation

Network coding

Routing

Traffic flows

Travel time

Model stability

Existing road conditions

Heavy congestion on Narellan Road

Long delays for the turning traffic at the Hume

Highway/Picton Road intersection

Heavy but moving flows along Hume Highway and

Picton Road

Hybrid modelling

10

Micro pocket created for Wilton Junction Precint

Wider road network modelled at Meso level

Model network, zone system and database

common to both level

Parameters only applied to Micro level were

calibrated

Detailed analysis of intersection design at Micro

level

Transport infrastructure requirement and staging

plan were determined with Hybrid model results

Discussion

11

Global parameter in AIMSUN Meso

Time reaction is most influencing parameter

Difficult to calibrate it for the existing condition

TSS research indicated lower value leads to aggressive and unrealistic

driver behaviour

SIDRA model was used to cross check Meso model results

Our model review experience showed this parameter was commonly

changed to make model calibration results look right and no appropriate

justification was provided

Intersection Meso SIDRA

Picton Road/Pembroke

ParadeLOS B LOS C

Picton Road/Almond Street LOS B LOSC

Discussion

12

Sensitivity test for different traffic assignment method

Model was reviewed by RMS and found to be fit for purpose

Dynamic User Equilibrium (DUE) was suggested to test to check robustness of

model

Base model was recalibrated and validated using DUE to conduct like with like

comparison for future year scenario

DUE tended to assign more traffic to route with higher hierarchy and speed

Propose infrastructure upgrades would be able to accommodate the increased

demand on key routes

Confidence in our proposed infrastructure requirement was enhanced.

Conclusion

13

First application of Hybrid modelling for large-scale development in NSW

3-tier modelling framework is suitable for capturing development impact in wider regional

context and local envrionment

AIMSUN Mesoscopic modelling is able to quickly assess route choice in large network and

test initial intersection design with acceptable level of detail

Microsimulation pocket can be created to conduct detailed analysis of core study area

Road network, zone system and database are common to both Meso and Micro areas of

network can avoid manual interfacing between different Meso and Micro software packages

Time reaction is one of the most influencing parameters in AIMSUN Meso. Caution should

be taken when calibrating parameter value. Low parameter value tends to underestimate

traffic delays

Sensitivity test for DUE and DTA can be useful to build confidence in model robustness and

proposed upgrade scheme