bring back the tram again! steve miller transport for london
TRANSCRIPT
Bring Back the Tram Again!
Steve Miller
Transport for London
Why “Bring Back the Tram Again!”?
• To distinguish from “Bring Back the Tram!” - my presentation to the 20th “International” Emme/2 User Group Meeting (Oct. 2006 – Seattle);
• To update the Emme/2 community on a project’s evolution;
• Not a cry for a nostalgic panorama of pre-war UK trams!
Transport for LondonTransport for London (TfL) is the integrated body responsible
for London's transport system.
Its role is:• to implement the Mayor's Transport Strategy for London• To manage the transport services across the capital for which the
Mayor has responsibility.
TfL manages:– London's buses, the Underground, the Docklands Light
Railway (DLR), Croydon Tramlink and London River Services
– Victoria Coach Station and London's Transport Museum – The red route network, Congestion Charge, and London's
4,600 traffic lights – A range of alternative transport initiatives including walking
and cycling
Public Transport Issues for London
Limited investment in new rail infrastructure
Limited potential to increase rail capacity
“Sweat the Assets”; i.e.:
• Spread the Peak (demand and supply)
• Increased crowding
Need to demonstrate Value for Money – Cost/Benefit Analysis
The Tram Scheme:Old Tram or New Tram?
Why a Tram?
• Relieve tube crowding
• Stimulate regeneration
• Improve accessibility
• Increased reliability
• Better connection between mainline stations
• Environmentally friendly mode of transport
• Cost efficient
• Connect target areas south of river
• Improve overall system efficiency of highway based public transport
A Tram is Efficient
Trams can move passengers far more efficiently than other vehicles
1 tram 2 - 4 buses 220 cars
Possible Route
New 16.5 km tramway
Core alignment:
Euston to Waterloo
30 trams per hour
Branches
North: to Kings Cross and
Camden Town
South: to Peckham and Brixton
15 trams per hour each branch
Alternative routes
Alternative routes
Alternative routes
Initial Assessments
Started in 1997: Intermediate Modes Studies
Then further Studies:– Tram v high-quality bus services– Tram was feasible and beneficial– Ruled out bus option due to capacity and attractiveness
Refined Progressively to demonstrate:
• Sound Business Case (BCR: 2.64:1)
• Road Traffic Impacts can be managed
Models available to TfL
• LTS (London Transportation Studies) Model
• Railplan
• SALT (SATURN Assignment of London Traffic)
• SALT-C (Congestion Charging)
• Micro-Simulation and Junction Models
LTS (London Transportation Studies) Model
• Strategic 4 Stage Model, based on legacy software and TRIPS/Cube
• Covers London and South East England
• Network representation is not that detailed
• 1016 zones
• Run for 2001, 2006, and then every 5 years to 2031
The Railplan Model
• Runs on EMME/2 usually under UNIX
• Public Transport Assignment model of London – Underground– Rail– Tram– Bus– Docklands Light Rail (DLR)
The Railplan Model
• Uses Size 16 licence
• 1,500-3,000+ Zones
• 50,000+ Links
• Distribution-Mode Split model outside Railplan – usually reliant on LTS
Railplan Features
Developed since 1988
Initially for the AM Peak Period only
All Public Transport sub-modes
Congestion in Route Choice
Quantifies social benefit of reduced crowding
Forecasts station flow patterns to aid station planning
Standard Railplan Model: Demand
Derived from LTS Model
Year Trips (3 Hrs)
2001 1,836,318
2011 2,195,962
Growth 2% p.a.
Standard Railplan: Model Dimensions
Network (2001 Base)
No. Zones 1,571
No. Nodes 14,648
No. Links 49,596
No. Services 1,836
Network Usage (2001 Base)
Mode Services Serv. Kms Pass. Kms
N Rail 949 41,470 9,195,488
Underground 125 13,778 2,826,332
Light Rail 6 615 67,120
Tram 4 471 24,575
Bus 752 77,433 998,301
SALT (SATURN Assignment of London Traffic)
• Covers the complete (Greater London) area in detail
• Slow to run
• No longer directly supported or used – difficult to find resources to develop and maintain a detailed network model over such a large area
• Demand derived from LTS
SALT-C (Congestion Charging)
• Designed to model the “Boundary Route” around CC Scheme
• Based on a cordoned SALT model for run-time efficiency
• Matrix Estimation, prior matrix from LTS via SALT
• Demand externally split between “YACS” and “NACS” on a sector – sector basis
• 615 zones
Micro-Simulation and Junction Models
• TRANSYT is the normal basis for such models
• Micro-simulation (VISSIM) increasingly adopted for the more complex areas
• Individual models are developed and validated to represent specific junction groups
New Models For the Tram
A New family of models: CRISTAL
(Cross River Study of Trams Across London)
• CRISTAL-H (Highway – SATURN)
• CRISTAL-P (Public Transport – EMME/2)
• CRISTAL-J (Junctions – TRANSYT and LINSIG)
• CRISTAL-M (Micro-Simulation – VISSIM)
• CRISTAL-D (Demand or Mode Shift)
CRISTAL Models: Overview
RailplanPT
SATURNStrategic Highway
TRANSYTJunction
VISSIMMicrosimulation
Mode Share / Dem
Revised Signal Timings
Revised Signal Timings
Flow Changes
Flow Changes
Demand across the Thames
Westminster, Waterloo and Blackfriars Bridges: Summary of Travel Demand (AM Peak Hour)
0
5000
10000
15000
20000
25000
30000
35000
2006 Base 2016 Reference 2016 With CRT
Scenario
Ve
hic
le/P
as
se
ng
er
Flo
w (
AM
Pe
ak
Ho
ur)
CRT
Bus
Taxi
HGV
Car/LGV
Current State of Models : CRISTAL_H
Based on SALT-C
615 zones expanded to 697 zones
New treatment of taxi user class
AM Peak hour model validated
PM Peak Hour model validated
IP model not yet developed
AM Peak Hour Demand (PCUs/hr) YACS
(Accept C.Charge)
NACS (Divert around
Zone)
Total
LIGHT 77,000 254,000 331,000
HEAVY 36,000 12,000 48,000
TAXIS 20,000 20,000
Total 133,000 266,000 399,000
Highway traffic flows (AM Peak Hour)
Current State of Models : CRISTAL-P
Based on Railplan
• 1571 zones expanded to 1645 zones
• AM Peak model validated
• PM Peak and IP models not yet developed
Public Transport Demand TransferredForecast LU Crowding Relief (AM Peak Period)
4-8% reduction in demand on most crowded services in central London – Northern, Victoria and Piccadilly
over 400,000 travellers on Tube in AM Peak will benefit
Slight increase in demand along a stretch of the Central line, Bank to Holborn
Current State of Models : CRISTAL-J
• 14 Individual local models developed and validated for AM and PM Peak hours
• Forecasts produced on operational aspects of traffic flows with or without Tram
• Mainly using TRANSYT and LINSIG for individual signal groups
Current State of Models : CRISTAL-M
• VISSIM Model was developed from a series of local models used previously
• Adequate to confirm general principles of scheme operation – interaction between tram and general traffic not readily handled in static models
• Route corridor being confirmed at this stage
• Initial model will be expanded and developed for the chosen route corridor
Current State of Models : CRISTAL-M VISSIM photo - BAYLIS ROAD / WATERLOO ROAD- Work in progress
Current State of Models : CRISTAL-D
• Previous Demand model was developed using EMME/2
• This has been refreshed for CRISTAL
• Now complies with DfT Advice (“WebTAG”)
Trip Frequency
Mode Choice (m)
Trip Distribution (d)
Assignment
Assignment Costs
Composite Costs (d)
Composite Costs (d,m)
HIGH
LOW
sensi
tivity
of re
sponse
s
Trip Frequency
Mode Choice (m)
Trip Distribution (d)
Assignment
Assignment Costs
Composite Costs (d)
Composite Costs (d,m) Trip Frequency
Mode Choice (m)
Trip Distribution (d)
Assignment
Assignment Costs
Composite Costs (d)
Composite Costs (d,m)
HIGH
LOW
sensi
tivity
of re
sponse
s
Convergence
CRISTAL-D Convergence Criteria:
• WebTAG % relative GAP < 0.1
• Benefits as % of network costs >= 10 * GAP
• Change in CRT demand < 0.5%
• Change in PT demand < 0.5%
• Change in highway demand < 0.5%
Future Year Modelling – Results2026 IBC RUN, SC359 – Iteration 3Post-Demand Model Results Compared to the Fixed Matrix Assignment
2026 IBC RUN, SC359 – Iteration 3
Post-Demand Model Results Compared to the Fixed Matrix Assignment
Around 800 Highway trips have transferred to PT and ‘slow modes’.
Overall increases in demand (181 trips) and benefits (959 pass-hrs) indicate that improvement in PT times is sufficient to compensate for increased journey times on highway
Fixed Demand Variable Demand Difference
DM SC515 DS SC359 (DS-DM)
PT Demand, pass 2,421,060 2,422,043 983CRT Demand, pass 63,836 66,925 3,089PT Benefits, pass-hrs -6,707 -6,500 207PT %GAP - 0.0002 -
Highway CAR Demand, pass 1,093,904 1,093,101 -803Highway Benefits, pass-hrs 2,349 1,184 -1,165Highway %GAP - 0.0147 -
Total Demand, pass 3,514,964 3,515,144 181Total Benefits, pass-hrs -4,357 -5,316 -959
Statistics, AM peak period
Model Operation
– Operation Platform – DOS / Windows– Batch File to operate– 3 to 5 iterations to converge
CRISTAL-HSATURN
Assignment models
CRISTAL-PEMME2
Assignment models
CRISTAL-DDemand model
Costs &DM
Demands
DS Demands
c on
verg
en
ce
Model Operation: Run Times
Highway Model 9.5 hours
PT Model <0.5 hours
Demand Model<0.25 hours
Per iteration 10 hours
So, a run still takes approx. 30 hours!
Scheme Costs and Benefits
Costs and Benefits (£m PV 2004 prices)
Tram Costs (600)
Revenue 250
User Benefits 1500- Time saving on trips 75%
- London Buses/Underground relief 25%
Economic Benefits 500
Non-User Benefits (200)
Benefit : Cost Ratio 2.8:1
Conclusions
• There seems to be a commercial case for a new Tram in central London
• Congestion Charging has offered some highway capacity for the core section
• The tools and data are available to develop robust forecasts of demand, traffic impacts and benefits
• EMME is an important part of the toolkit
• However, despite the commercial case, there remains issues of political risk and availability of funds
A Vision of the Tram
www.tfl.gov.uk