1. EMME User’s Conference Project Experience of a DYNAMEQ Simulation Model : TRPC – Smart Corridors Project October 4, 2010 Natarajan JANA Janarthanan PhD, PTP Ming-Bang Shyu PhD, PTP Fehr & Peers Jailyn Brown Thurston Regional Planning Council

2. Outline Project Overview Model Development Model Validation and Calibration DYNAMEQ (DTA) Simulation VMT / Emission Calculation Q&A

3. Geography Source: Fehr and Peers (2009); courtesy map Google Area : 727 Sq Miles Population: 245,300 (2009) 373,000 (2030) Olympia is the capitol of Washington State Freeway: 90 Miles Arterials: 220 Miles Collectors: 360 Miles

4. Study Corridors

5. Why DTA Model? TRPC wants a tool –to evaluate ITS and TSP options –to calculate emissions –to create a traffic operations model for its jurisdictions to integrate signal coordination efforts Traditional travel demand models have limitations Micro-simulation models for a larger area is not practical

6. What is Dynamic Traffic Assignment (DTA) Model? Time-dependent methodology Experienced shortest (minimal-cost) path from origin to destination in response to roadway connectivity, capacity, or travel demand changes.

7. Why DYNAMEQ DTA Model? A simulation-based approach capturing system dynamics (many are deterministic) Car following and Lane changing methodology Intersection controls

8. DYNAMEQ Model Development NETWORK –Import network into DYNAMEQ from EMME Model –Run DTA to check convergence and flow problems –Refine network by adding missing intersections on the corridors –Modify centroid connectors for the zones around two corridors to reflect field conditions –Add intersection detail (geometry & turning pockets) –Add signal data / intersection controls –81 signals & 67 stopped controls –Network properties in DYNAMEQ model: – 800 centroids – 2500 regular nodes – 8000 links – 20 transit lines (study corridors only)

9. DYNAMEQ Model Development TRIP TABLES –PM peak hour trip tables brought from Travel Demand Model –30-mimute Pre-peak and post-peak loading applied –The modes are SOV, HOV & Truck

10. DYNAMEQ Model Development Travel Demand Model - Link node basis DTA Model - Lane basis

11. DYNAMEQ Model Development

12. Assign trip tables in DTA model – without any intersection controls and validation / calibration –network check –Flow blockage check –Convergence check

13. DYNAMEQ Model Development EMME’s Static assignment model DYNAMEQ model without intersection controls and validation Link Volume Comparison

14. DYNAMEQ Model Development Run DTA with intersection controls without any validation / calibration

15. DYNAMEQ Model Development DTA with intersection controls DTA without intersection controls

16. DYNAMEQ Model Development General Approaches to Validate / Calibrate the models Static Assignment Model Dynamic Assignment Model Validation Counts Travel times / speeds /queues Network measures (VMT, VHT etc) Network measures (VMT, VHT etc) Traveling paths Calibration Link/node propertiesLink/node/movement properties Turn penalties Driver behavior properties (response time, follow up time, gap acceptance) Intersection control properties Demand adjustment

17. Model Convergence

18. Base Year Model Validation / Calibration – Link Volume Including I-5 R Squared = 0.955, Slope = 1.01

19. Base Year Model Validation / Calibration – Link Volume Excluding I-5 R Squared = 0.894, Slope = 0.97

20. Base Year Model Validation / Calibration – Turn Movement R Squared = 0.900, Slope = 1.00

21. Comparison of Travel Speed PM Peak Hour Weekday Observed Travel Speed (mph) DTA Model Travel Speed (mph)

22. Base Year Model Validation / Calibration – Travel Time Observed Travel Time (sec)Model Travel Time (sec)Travel Time Comparison (Model-Observed) /Observed Martin Way SEWBEBMartin Way SEWBEBMartin Way SEWBEB 500 Ft. E of Marvin Rd NE - I-5 SB Ramps521488 500 Ft. E of Marvin Rd NE - I-5 SB Ramps485447 500 Ft. E of Marvin Rd NE - I-5 SB Ramps-6.9%-8.4% I-5 SB Ramps - Pacific Ave SE397446I-5 SB Ramps - Pacific Ave SE437418I-5 SB Ramps - Pacific Ave SE10.1%-6.3% State Ave NE Pacific Ave SE - Capitol Way S257N/APacific Ave SE - Capitol Way S265N/APacific Ave SE - Capitol Way S3.1%N/A 4th Ave S Capitol Way S - Pacific Ave SEN/A280Capitol Way S - Pacific Ave SEN/A301Capitol Way S - Pacific Ave SEN/A7.5% Capitol Way S/Capitol Blvd SNBSBCapitol Way S/Capitol Blvd SNBSBCapitol Way S/Capitol Blvd SNBSB State Ave NE - Carlyon Ave SE395330State Ave NE - Carlyon Ave SE341370State Ave NE - Carlyon Ave SE-13.7%12.1% Carlyon Ave SE - Linwood Ave SW214192Carlyon Ave SE - Linwood Ave SW196212Carlyon Ave SE - Linwood Ave SW-8.4%10.4% Linwood Ave SW - Tumwater Blvd SW289360 Linwood Ave SW - Tumwater Blvd SW317325 Linwood Ave SW - Tumwater Blvd SW9.7%-9.7% Observed Travel Time Model Output

23. Base Year Model Simulation Source: Movie clip from the DTA model simulation Density

24. Base Year Model Simulation Outflow Source: Movie clip from the DTA model simulation

25. Base Year Model Simulation Queuing

26. Base Year Model Sensitivity Analysis using an Incident Scenario - Tested on I-5 SB in the vicinity with two-lane closure - Separated car and truck demands into two -- external-external trips -- others - Run 10 more iterations with incident lane closure.

27. Source: Snapshot from the DTA model simulation Incident Analysis – Paths Incident location Base Year Model Simulation

28. Source: Snapshot from the DTA model simulation Incident Analysis – Flow change Base Year Model Simulation Incident location

29. Source: Snapshot from the DTA model simulation Incident Analysis – Speed change Base Year Model Simulation Incident location

30. Emission Calculation EMFAC2007 Running Emissions Factors in Grams per Mile for Year 2009 Conditions in North Coast Speed Bin (MPH) Total Organic Gasses (TOG) Sulfur Dioxide (SO2) Diesel Particulate Matter (Diesel PM) Particulate Matter < 2.5 microns (PM 2.5) Particulate Matter < 10 microns (PM10) Oxides of Nitrogen (NOx) Carbon Dioxide (CO2) Carbon Monoxide (CO) 0-52.0740.0130.157550.2120.2313.7761,279.8714.878 5-101.350.010.1091350.1450.1582.87984.06811.86 10-150.8910.0080.0737150.0990.1082.295782.6819.8 15-200.6390.0060.0530150.0720.0782.029646.18.372 20-250.5110.0050.0445050.0590.0641.915561.2977.383 25-300.4250.0050.0380650.0490.0541.838503.4366.673 30-350.370.0050.0336950.0430.0471.793465.4496.177 35-400.3370.0040.031280.040.0431.778443.1485.865 40-450.3230.0040.030820.0390.0421.793434.3155.727 45-500.3260.0040.03220.040.0431.84438.2435.779 50-550.3470.0040.0355350.0430.0471.924455.5926.062 55-600.3880.0050.0405950.0490.0532.051488.5286.664 60-650.4560.0050.0474950.0570.0622.234541.177.742 60-700.5020.0050.056120.0650.0712.41551.1488.576 70-750.5720.0060.06670.0750.0822.663566.51210.013 Note: EMFAC has unusual emissions factors for CO2 for speeds above 65 MPH, modify or use with caution. Source: CTEMFAC 2.6, UC Davis and Caltrans, Sept 21, 2009. Model Run By: Fehr & Peers, 2009

31. Comparison of Speed Output Link Congested Speed (mph) Total Link Volume per hour Speed<=5 - 5<Speed<=10 - 10<Speed<=15 - 15<Speed<=20 1,711 20<Speed<=25 67,544 25<Speed<=30 65,538 30<Speed<=35 94,199 35<Speed - Link Congested Speed (mph) Total Link Volume Total (One Hour) Speed<=5 5,450 5<Speed<=10 20,719 10<Speed<=15 18,250 15<Speed<=20 15,858 20<Speed<=25 17,720 25<Speed<=30 24,168 30<Speed<=35 16,089 35<Speed 85,330 EMME ModelDYNAMEQ Model

32. Comparison of VMT EMME vs. DYNAMEQ

33. Comparison of PM10 Calculation

34. Emissions on Corridor

35. Benefits of DYNAMEQ Model More realistic traffic simulation - Lane based simulation - Traffic congestion / queuing - Intersection delays Region-wide traffic operation model Hot spot identification and problem solving TSP analysis Emission Calculation Congested areas/network analysis

36. Lessons Learned building this Dynameq Model Data needs Network resolution Demand Adjustment Validation/Calibration Emissions Calculations Travel Demand Model Micro Traffic Simulation Model DYNAMEQ Model

37. Do you have any questions on this presentation or related issues? Jana / Ming Fehr & Peers 11410 NE 122nd Way, Suite 320 | Kirkland, WA 98034 425.820.0100 - T | 425.821.1750 – F jana@fehrandpeers.com m.shyu@fehrandpeers.com www.fehrandpeers.com