Dash logoDataONE logo

Identifying Roadway Physical Characteristics that Contribute to Emissions Differences between Hybrid and Conventional Vehicles

Citation

Sullivan, James; Sentoff, Karen (2019), Identifying Roadway Physical Characteristics that Contribute to Emissions Differences between Hybrid and Conventional Vehicles, DataONE Dash, Dataset, https://doi.org/10.15146/R36975

Abstract

 

In this study, a second-by-second (SbS) data set obtained from monitoring vehicle emissions over a series of 75 test runs from 2 test vehicles (a conventional vehicle (CV) and a hybrid-electric vehicle (HEV)) over an 18-month period in 2010-2011 during real-world on-road operations on a specified 32-mile route in Chittenden County, Vermont was used in an innovative new method of analysis to assess emissions differences between the two propulsion systems and attribute these differences to physical roadway/infrastructure characteristics. The K-S test was used to assess the difference between the cumulative distributions of the CV and HEV emissions samples on each link, and the K-S test statistic was regressed against the full set of roadway link characteristics. The regression results allowed the team to identify specific roadway characteristics that contribute to emissions differences between the vehicle types. Overall, the models that included maximum grade and intersection control type performed best, however speed limit and horizontal curvature were also shown to be important. The performance differences identified in this project confirm that engine controls that are responsive to roadway characteristics are necessary.

Methods

This project leveraged prior UTC-funded work that developed methods and unique databases of second-by-second (SbS) on-board vehicle emissions data for estimating emissions differences between conventional vehicles and hybrid-electric vehicles. The team used SbS emissions data obtained from a series of 75 test runs driven by a single driver over 18 months in 2010 and 2011, across all seasons and at a variety of hours in the day. The suites of emissions data collected during the test runs were isolated by the road segment the vehicle was traversing at the time. Spatial processes were used to associate the emissions data with the characteristics of the roadway on which the vehicle was operating at the time, by assigning each emissions data point a Road Segment ID. For each road segment in the project, then, the following data is available:

  • ID                                 TransCAD native feature ID
  • Dir                               Link direction: 0 - two-way; 1, -1 - one-way
  • Length                         Link length (miles)
  • SP2 Link ID                  Link ID corresponding to this study data
  • LINK_NAME                Name of roadway represented by the link
  • Daily_Cap                   Daily roadway capacity (vehicles per day)
  • Hourly_Cap                Hourly roadway capacity (vehicles per hour)
  • SPEED_LIMIT               Speed limit (mph)
  • AADT                           Annualized Average Daily Traffic for 2010 (from traffic counts)
  • AADT/Daily_Cap         AADT / Daily roadway capacity
  • GRADE_min                Minimum grade (%) on the roadway represented by the link
  • GRADE_max               Maximum grade (%) on the roadway represented by the link
  • GRADE_avg                 Average grade (%) on the roadway represented by the link
  • HC Degree                   Horizontal curvature (degrees)
  • Lane Width                 Lane width (ft)
  • Shoulder Width          Right shoulder width (ft)
  • Lane + Shoulder          Width of the travelled way (lane + shoulder)
  • Median Width             Width of the roadway median (null if none)
  • Lanes Each Way          Number of lanes of travel in each direction
  • Control_Type             Intersection- control / turning-movement type at each end of the link (1 thru 9)
  • Speed_Limit_Cat        Speed limit category (low/high) for cross-classification analysis
  • L+S_Cat                       Maximum grade category (negative, flat, low positive, and high positive) for cross-classification analysis

The distributions of specific SbS emissions data were compared using the K-S test, resulting in a D* statistics for each link in the study. These data were regressed against the physical characteristics above.