Baltimore Red Line: Challenges of a Large Model Area

1
Baltimore Red Line Challenges of a Large Model
Area

• Presented by Mahmood Shehata, P.E.
• McCormick Taylor, Inc.
• Philadelphia, PA
• Presented to PTV Users Group Meeting
• May 16, 2008
• Philadelphia, PA

2
Presentation Topics

• Introduction to Baltimore Red Line
• Work Sharing and Modeling Standards
• Downtown Base Model Development
• Model Calibration Techniques and Obstacles
• Coding of Alternatives Including TFlow Fuzzy
• Lessons Learned

3
Introduction to Baltimore Red Line
4
Introduction to Red Line

• 12-mile East-West Corridor
• Western Suburbs through Downtown Baltimore to
  Eastern End of Baltimore City
• Potential BRT or LRT Corridor
• Alternatives Analysis for DEIS and New Starts

5
Red Line Corridor
6
Existing Transit Services

• Local, Express, and Commuter Bus
• Metro Subway NW Suburbs to Downtown Baltimore
• Central Light Rail Line Northern Suburbs
  through Downtown to Southern Suburbs and Airport
• MARC Commuter Rail From NE Suburbs to SW
  Suburbs and Washington, DC
• Existing E-W service limited, including only one
  high-speed bus line

7
Comprehensive System

• Red Line Part of Regional Plan
• Expand Existing System

8
Three Distinct Areas

• Proposed Line Serves 3 Distinct Areas
• West Baltimore and Western Suburbs
• Downtown Baltimore
• East End (Fells Point / Canton)
• 22 Proposed Stations

9
West Baltimore and Western Suburbs
Security Boulevard MD 122
Eastern Terminus of I-70 Proposed Station
Location
US 40 in West Baltimore
10
The Trench
US 40 Between West Baltimore Commuter Rail and
Downtown
11
Downtown / Inner Harbor
Baltimore St - Location of Most Alignments
Pratt St Source of Congestion in VISSIM Models
12
Fells Point / Canton
Wider Boston Street to south of City Grid in
Canton Area
Narrow Street Grid in Fells Point
13
Work Sharing and Modeling Standards
14
Work Sharing and Modeling Standards

• Synchro initially utilized to evaluate corridor
• MTA and project team agreed upon VISSIM to
  analyze entire corridor, including preemption and
  TSP
• Five Separate Consultants to complete VISSIM
  models
• Single consultant chosen for oversight

15
Work Sharing and Modeling Standards

• Each consultant assigned a portion of study area
  to model
• Model Areas chosen using logical break points
• 1-2 Park and Ride (Baltimore County and West
  Baltimore)
• 2-3 W.Baltimore MARC Station
• 3-4 US 40 / MLK Blvd.
• 4-5 President St / Central Avenue

16
Model Areas 1-2
Modeling Team 1 Area
Modeling Team 2 Area
17
Model Areas 2-3
Modeling Team 2 Area
Modeling Team 3 Area
18
Model Areas 3-4
Modeling Team 3 Area
Modeling Team 4 Area
19
Model Areas 4-5
Modeling Team 4 Area
Modeling Team 5 Area
20
Work Sharing and Modeling Standards

• Standards Developed for Modeling Consistency
• Speed profiles, turning speeds, links and
  connectors, etc.
• Vehicle profiles uniform including articulated
  buses and LRT vehicles
• CD Distributed with base .inp file
• Provided identical vehicle profiles for all
  models.
• Aerials with .bgr - georeferenced models

21
Work Sharing and Modeling Standards

• Spreadsheet Developed and Distributed to Ensure
  Model Consistency

22
Work Sharing and Modeling Standards

• Standards were developed for most components of
  VISSIM models

23
Downtown Model Development
24
Downtown Model Development

• Approximately 95 signalized intersections
• N-S Arterials on east and west end
• Alternating one-way E-W Streets

25
Downtown Model Development

• Synchro base model previously developed
• Small blocks, Unique Lane and Turn Restrictions
• VISSIM model developed from scratch
• Divided into 3 sections to setup model
• MLK to Howard (600-650)
• Howard to Calvert (650-700)
• South to Central (700-750)
• Read Additionally used

26
Unique Characteristics of Downtown Model

• On-Street Parking / Loading Zones in some
  locations
• Off-Street Parking as source/sink
• Initial Model did not include Conway Street
• At times, I-95 and I-395 backs into Conway St

27
Unique Characteristics of Downtown Model
Initial Model Area
Conway St.
Intersection of Howard and Conway Constricting
Point
28
Unique Characteristics of Downtown Model

• Routing decisions through 2,3, even 4
  intersections (in existing conditions model)

29
Unique Characteristics of Downtown Model

• Model accounts for high pedestrians, including
  Leading Pedestrian Intervals

30
Public Transportation System

• Bus routing and headways obtained with
  established schedules
• Bus stops field verified
• Central Light Rail Line included
• Dwell Times
• 15-25 sec. per bus
• 30-50 sec. for Central Light Rail

31
Central Light Rail Line
32
Result Comprehensive Downtown Baltimore Model
33
Model Calibration Techniques
34
Model Calibration Techniques

• Calibration Standards
• Within 10 of counted link volumes
• Within 20 of Travel Times

35
Model Calibration Techniques

• Travel Times calculated in field to mimic bus
  movements
• Travel Times in model extracted to reflect field
  collection
• For all vehicles, VISSIM expected to be higher
• Buses comparable, less dwell time

36
Model Parameters Adjusted

• Driver Behavior Parameters
• Safety Distance Adjustment
• Lane Changing Behavior
• Saturation Flow Rate Additive and
  Multiplicative Distances
• Some Links altered more than others
• Creation of more than one parameter set per link
  type (Urban1, Urban2)
• Account for more friction (e.g. Pratt St)
• Minimum Lane Changing Distance (Connectors)

37
Results

• All 5 modeling teams met all calibration
  parameters set!
• Higher confidence in model essential for
  alternatives analysis

38
Alternatives Analysis
39
Alternative Model Coding

• Design Year of 2030
• All initial volume growth based on regional MPO
  travel demand model
• Western Area models (Groups 1-3) utilized growth
  rate from MPO model
• Downtown and East End account for parallel links
  and alternate routing by using TFlow Fuzzy Logic

40
Use of TFlow Fuzzy

• Existing VISSIM network routing and MPO growth
  used to develop base VISUM network
• TFlow Fuzzy accounted for diversion alternate
  one-way streets
• Aided in redistribution of traffic in both
  Downtown and Fells Point

41
Alternatives Considered

• Shared Use BRT
• Exclusive BRT
• Exclusive LRT
• Exclusive BRT with Downtown Bus Tunnel
• Exclusive LRT with Downtown Tunnel

42
Exclusive BRT Trench Area
43
Exclusive BRT Trench Area
44
Exclusive BRT Baltimore / Lombard Transit
Couplet
45
Exclusive BRT Baltimore / Lombard Transit
Couplet
46
Modeling Approach and Consistency

• BRT/LRT Vehicle Characteristics consistent
• Headways Based on Operating Plan for EIS
• Dwell Times 15-30 Sec. for BRT 30-50 sec. for
  LRT
• Vehicle Change and Clearance Intervals based on
  operating speeds and vehicle characteristics

47
Modeling Approach and Consistency

• Removal of existing travel lanes for exclusive
  transit lanes is anticipated to cause trip
  diversion
• CBD, Fells Point both are based on city grid
• Use of T-Flow Fuzzy Logic in VISUM to estimate
  diversions

48
Lessons Learned
49
Lessons Learned

• Establish a consistent approach at start of model
  process
• Project controls / standards allow for easier and
  more accurate QA/QC review
• Field verification a must for all elements
• Understand how calibration data was collected
• Regular communication between model teams

50
Lessons Learned

• VISUM / VISSIM integration utilizing TFlow Fuzzy
  ideally should be included early in model
  development
• Take care in optimizing signal timings to
  minimize transit delay, vs. roadway corridor
  delay
• Traffic signal timing plans should reflect how
  surface transit travels through the intersection
• Communicate level of modeling effort early in the
  project

51
Lessons Learned

• Expect the unexpected!!!

52

• Thank you!
• Mahmood Shehata, P.E.
• Project Engineer
• McCormick Taylor, Inc.
• 2001 Market St, 10th Floor
• Philadelphia, PA 19103
• [phone]
• mshehata_at_mtmail.biz