Optimizing Transit Operations in Vancouver, B.C.: TransLinks Rapid Transit Model

1
Optimizing Transit Operations in Vancouver, B.C.
TransLinks Rapid Transit Model

• May 9, 2007

11th National Transportation Planning
Applications Conference, May 7, 2007
Ian Fisher, TransLink Wolfgang Scherr, PTV
2
Overview

• The Context Rapid Transit in the Greater
  Vancouver Area
• Objectives of the Model
• The Model Structure
• The Supply Model
• The Demand Model
• Modelling Future Rapid Transit
• Application in a Fleet Requirement Study
• Methodology
• Volume/Capacity Analysis
• Results Fleet Requirements
• Conclusions and Future Directions
• References

3
Location
4
Greater Vancouver Region

• 21 cities in regional federation (GVRD)
• 2.1 million residents, 3 million by 2031
• 165.1 million transit rides (linked) in 2006
• 78 annual transit rides per resident
• Multicultural, diverse population
• Major gateway for sea, air and rail freight
• No urban freeways/motorways
• Prior to 1999 no regional transportation agency

5
TransLink Mandate Integration of
Intelligent Transportation Systems
Public Transport
Transportation Demand Management
Regional Cycling
Roads Bridges
Vehicle Emissions Testing
6
Multimodal Transit System

• Bus (includes trolleybus)
• B-Line bus rapid transit
• Community Shuttle minibus

7
Multimodal Transit System

• SeaBus ferry
• West Coast Express commuter rail

8
SkyTrain The Backbone

• Automated, driverless 49 km train system
  (Bombardier)
• 210 cars
• 2 Lines in split tail (Expo and Millennium)
• 210,000 passengers (unlinked boardings) per day
• 25 of transit boardings region-wide
• 10,500 passengers/hour peak load
• Peak headway 108 seconds (both lines combined)

9
SkyTrain The Backbone
10
Rail and Ferry Transit (Re)Development
North Vancouver
Coquitlam
City Centre
UBC
Metrotown
New Westminster
Airport
Surrey
Richmond
11
Needs for Model

• SkyTrain at or near capacity in peaks
• Operations are complex and constrained
• Millennium Line introduced a split-tail with
  lower demand
• Tight headways (108 s.)
• Variable train capacities
• Two car types
• 2-, 4-, 6-car trains, possibly 5-car in future
• Financial constraint on fleet size
• Need to review
• Splitting lines to make them independent
• Short-turns
• Variable train lengths
• Future lines will increase intermodal route
  choices

12
Objectives of the Rapid Transit Model

• Operational model of the Rapid Transit Network
• Based on off-the-shelf technology (VISUM)
• In-house use of the model by TransLink
• Types of scenarios
• What If scenarios including variation of routes
  and schedules, vehicle/train capacity,
  decoupling, yard location
• Network extensions (new rapid transit corridors)
• Types of analysis
• Vehicle/train use and fleet requirements
• Cost/benefit analysis
• Ridership/capacity analysis
• Create animations that are comprehensible to lay
  people

13
The Rapid Transit Model Network View
14
24-Hour Time-Dynamic Supply and Demand

• Time-dynamic passenger volumes 2006, colour-coded
  by v/c-ratio

15
Supply Model - Characteristics

• Nodes and links (routable network graph)
• Stops (stations)
• Attached to nodes or to links
• One stop can include several physical stop
  locations (multiple platforms in one station)
• Transit lines
• One line can include many routes
• Several run-time-patterns per route (including
  run, dwell, layover and recovery times)
• Vehicle model
• Vehicle units and combinations (train types)
• Seated and standing capacity per train type
• Trains are assigned to schedule items or blocks
• Schedules
• Exact time of all departures (bus or train runs)
  for each route
• Operating day 500 AM through 300 AM
• Blocks
• Sets of individual schedule runs
• One block represents the work assignment for one
  train for a single workday the total of all
  blocks is generated by the blocking heuristic and
  determines the total number of vehicles and
  trains needed.
• Master network (includes multiple scenarios)

16
Supply Model - Schedules
17
Supply Model 24-Hour Blocking Fleet Assignment
18
Demand Model - Characteristics

• OD matrices for four times of day (AM, Mid, PM,
  Eve)
• Synthetic matrices derived from counts
• Departure time distributions
• Passenger flow model
• Timetable-based assignment
• Multi-path
• Time-dynamic
• Assignment outputs
• Ridership disaggregated for each line, route,
  link, stop
• Time dynamic over 24 hours currently down to
  15-minute intervals

19
Demand Model Calibration (1)
Station
20
Demand Model Calibration (2)
21
Integrated Analysis of Supply and Demand

• Ridership and operations data are integrated in
  the same data platform
• V/C analysis through the entire network
• Interaction between Demand and Supply
• Ridership is elastic to changes in schedules
  affecting transfer time
• Ridership is elastic to changes in route
  alignment affecting travel time and transfers

15-minute link capacities during 24 hours
15-minute link passenger volumes during 24 hours
22
Ridership Forecast 2010 and 2021

• Future ridership shall combine
• Goodness of calibration obtained for 2006
• Forecast by regional travel demand model
• Regional travel demand model forecasts
• Include effects of demographic change and land
  use
• Include mode choice and shifts between automobile
  and transit
• Include demand elasticity on future network
  extensions
• Ridership formula for the Rapid Transit Model

23
Forecast 2010 and 2021 (24-hour Operations)
Unlinked trips
Passenger km
Total Capacity km
Seat km
24
Application SkyTrain Fleet Requirement Study

• Objective
• Maximize capacity on critical network segments
• Constraints
• Fleet (number and types of cars) and network
  topology
• Variables
• Route scheme
• Headway
• Train combinations
• Train assignment
• Methodology
• AM peak only
• Enumeration of alternatives in spreadsheets and
  quick analysis
• Detailed, integrated analysis in VISUM
• Several performance measures

25
Fleet Study Variations of Route Schemes
3) Variations of 1 and 2 with a mix of long and
shorter route patterns
26
Fleet Study Disaggregated V/C Ratios

• Scenario comparison with full time and space
  detail
• Good for
• Bottleneck screening
• Better understanding of bottleneck build-up
• Disadvantage
• Volatility of 15-minute capacity and 15-minute
  v/c ratio
• Un-precise measurement of v/c performance
• Difficult to communicate to decision makers

27
Fleet Study Performance Measure V/C Ratio

• Aggregation over time and space
• Several network sections (each including multiple
  stations)
• Time aggregation one v/c result from 700 900
  AM
• Average 15-minute capacity, 700 900 AM
• Maximum 15-minute link volume, 700 900 AM

28
Fleet Study Other Performance Measures

• Performance indicators 700-900 AM
• Operations Spare Rate, Train km,
• Ridership Transfers
• Volume/Capacity Relationship (focus)
• Aggregated statistics for the entire AM period

29
Fleet Study Summary

• 2006
• The existing operations obtain the best possible
  V/C ratios with the current fleet (150 Mark I
  60 Mark II cars).
• No operations scenario performed better than the
  existing operations.
• One scenario was found that would have equivalent
  performance.(Millennium CO-VC and two Expo
  routes WF-KG, WF-MT)
• 2010/2021
• The planned fleet (150108 cars in 2010 , 248 in
  2021) will relieve the capacity situation for the
  projected ridership. Trains will remain full
  during the peak.
• Peak spreading could add additional relief in the
  future not yet included in the model

30
Current Status

• Calibration and pilot application finished in
  February 2007
• The model is in use by TransLink staff
• Currently the model includes all rail modes in
  the region
• SkyTrain
• West Coast Express
• Canada Line
• Evergreen Line
• Applications
• Evaluation/optimization of operational scenarios
• Vehicle requirements
• Continuous maintenance (new survey data)

31
Future Directions

• Increased network coverage
• B-Lines / Frequent Transit Network
• SeaBus
• Western extension in Broadway corridor
• Fleet facilities analysis
• SkyTrain yards
• Extension of the demand model to the entire
  region
• Streamlined interaction with regional travel
  demand model
• Modelling of all buses
• Fare/revenue model

32
Thank You
For more information, please contact
Wolfgang Scherr PTV 302-654-4384wscherr_at_ptvameric
a.com www.ptvamerica.com
Ian FisherTransLink604-453-3058ian_fisher_at_trans
link.bc.cawww.translink.bc.ca
33
Slide Storage (for QA)
34
Supply Model Stops and Stations

• Stations are modeled in three layers
• Stop (abstract umbrella)
• Stop area (platform)
• Stop point (physical stop location)