CORRADINO

1
Validation of Speeds and Volumes in a Large
Regional Model

• Southeast Regional Planning Model 6
• Florida Department of Transportation
• District IV

2
Consultant Team

• SERPM 6
• Corradino
• AECOM Consult

• With thanks to
• FDOT Districts 4, 6 Turnpike
• MPO staffs of Palm Beach, Broward 7 Miami-Dade
  Counties
• SFRTA

3
SERPM 6

• Palm Beach, Broward and Miami-Dade Counties
• 4,134 TAZs (ZONESA)
• 3 Time periods (AM 630-930, PM 330 630,
  Off-peak)
• 10 Transit modes
• 22 28 hours run time on 3.4 GHZ computer

4
SERPM 6 - Transit Nets
5
SERPM 6 Model Flow

• Setup controls and file names
• External EE by period
• Trip Gen LSGEN by auto ownership and TOD
• Highway Including speed and capacity
  calculators
• Distribution Including equilibration of peak
  period highway speeds
• Transit Nets PT ? Trnbuild and connectors,
  paths, skims and fares

• Mode Choice
• Transit Assignment by access path and TOD
• Highway Trip Tables Combines trip tables for
  assignment
• Highway Assignment by TOD (use most computer
  time)
• Highway Evaluation HEVAL and RMSE by TOD and
  county

6
THE ISSUE

• FTA/FHWA emphasis on correct speeds and travel
  times in models, especially for New Starts
• So, we must
• Match traffic counts and transit ridership in the
  base year.
• Match observed highway travel times
• Match observed transit travel times

7
Old Practice

• Assign uncongested speeds and capacities to links
  from a lookup table indexed by area type (AT),
  facility type (FT) and number of lanes.
• Default lookup table developed by examining
  practice throughout the state.
• Adjust the lookup table to obtain a better match
  of final assigned 24-hour volumes.
• Most adjustments made to speeds.
• Care taken to ensure logical relationships,
  e.g., freeways faster than arterials, rural
  faster than residential.
• Transit speeds derived from highway using INET
  functions based on mode, AT and FT.

8
Well, Whats Wrong With That?

• Lookup table probably better than adjusting
  individual links to get better loads (not dead
  yet what speed does a new link get?).
• While lookup tables are systematic, their
  contents can become arbitrary and skewed.
• Seldom checked to verify they have any
  relationship to reality.

9
New Approach

• Base Capacities on Florida Quality/Level of
  Service Manual, derived from HCM.
• Capacities and speeds determined from roadway
  characteristics, not a subjective determination
  of AT, FT.
• Speeds begin with POSTED SPEED, modified by
  studies of driver behavior.
• If signal controlled, add estimated delays at
  intersections, but not microsimulation (somewhat
  of a slippery slope).
• Some concessions were made to better match counts.

10
Initial Speed Estimation

• Estimates uninterrupted flow initial speed from
  an equation.
• Begin with NCHRP 387
• Modify to make more consistent (remove
  discontinuity at 50 mph).
• Adds delays from traffic control devices
• Later, speeds are changed in the user equilibrium
  assignment and equilibration of travel times.

11
Speed Relationships
FWY12.560.641Postspd0.00340Postspd2 Surface
9.100.657Postspd0.00348Postspd2
12
CAPACITIES

• New with SERPM-6.
• Replaces the lookup tables.
• Allows an OVERIDE coded on the link.
• Based on the FDOT Quality/Level of Service
  Manual.
• Script calculates LOS C and E.
• Calculated as PCEs.

13
Capacity Calculation Variables/Attributes
14
Time of Day

• FSUTMS uses auto speeds from 24-hour assignment
  as for HBW mode choice
• HBO NHB use free-flow speeds
• SERPM6 uses time of day
• Peak period trips use AM congested speeds
• Off-peak period trips use free-flow speeds
• Standard FSUTMS design expected for most models

15
Equilibration

• Peak speeds and time modified from user
  equilibrium assignment.
• Run twice to conserve model running time.
• Speeds found to be stable.
• Produces distribution and mode choice travel
  times for highway and transit that are close to
  the times after the final assignment.
• Times are stable for UB calculations.

16
SERPM 6 Mode Choice
Applied for Peak and Off-peak
17
SERPM 6 Highway Assignment

• By Time-of-Day
• AM Peak
• Off-Peak
• PM Peak
• Composite 24 Hours (add preceding)

18
Volume-Delay CurvesSoutheast Regional Planning
Model VI
19
SERPM 6 Highway Evaluation

• SUMMARIES BY
• SERPM 6 Region by Time-of-Day
• Each County for 24 hours
• Trucks alone each County for 24 hours

20
SERPM 6 - Validation

• Standard FSUTMS
• Plus Speed Validation
• Speed compromises made in favor of Volume/Count
• Good Volume/Count ratios
• Very good RMSE
• Good transit assignment results

21
Highway Sections of Observed Speed Study(from
several sources)
22
2000 Highway Speeds
23
SERPM 6 Speed Validation
24
AM PeriodSpeedValidation
25
PM PeriodSpeedValidation
26
Off-PeakSpeedValidation
27
SERPM 6 Systemwide Comparison
28
SERPM 6 RMSE Validation
29
Highway Screenline Accuracy
30
ScreenlineLocations
31
Volume/Count Ratios of Screenlines
32
Conclusions

• Assigning speeds to network links based on posted
  speeds and traffic control produces good results
  in terms of speeds and travel times.
• The Florida Quality/Level of Service Manual is
  useful for estimating model capacities.
• Speed and capacity calculators appear to work at
  least as well as lookup tables and are less
  arbitrary.
• It is possible to validate speeds and volumes in
  a time-of-day model.

33
Recommendations for Future Study

• Need consistent travel time data.
• Need more travel time data.
• Consider estimating delays at intersections
  dynamically using software intersection or
  junction modeling.
• Better way of estimating the impacts of freeway
  merge/diverge/weaves on model travel times.

34

• Questions ?
• Please use the Microphone.