The $2.46 billion Oregon Transportation Investment Act of 2003 includes $1.3 billion to repair hundreds of aging state-owned bridges over the next several years. Keeping traffic and freight moving during this time of unprecedented construction is one of the Oregon Department of Transportation’s (ODOT) top priorities.
There are two key aspects to mobility management from a statewide perspective: physical restrictions (such as weight, width, height and length) and delays. Managing physical restrictions in a state with many topographical barriers is challenging and could be the focus of an article on its own. This article focuses on managing delays.
Successful mobility management requires three key components:
- Establishing a practical, measurable and meaningful benchmark;
- Estimating operational issues created by upcoming projects; and
- Measuring actual impacts of projects in the field.
Delay your hands on me
Traffic impacts resulting from work zones are most easily measured by delay. Excessive delays cause economic impacts to freight, user costs to all drivers and potential for economic losses due to reduced tourism.
To manage delay requires a definition different from what typically is used in traffic engineering. For purposes of mobility management, delay is defined as the additional travel time required to travel from one point to another as a result of construction activities. With this definition, recurring delay caused by existing capacity or geometric deficiencies is excluded. Similarly, incident-specific delay is also excluded, although traffic control should be designed to minimize these incidents.
The statewide roadway network was examined and key mobility corridors designated. The corridors were broken into segments ranging from 30 to 70 miles in length, with each segment assigned a delay or travel-time threshold. The threshold governs the number of impacts that can be sustained on a segment or corridor at any given time. Any potential project or group of projects that may cause these thresholds to be exceeded results in closer examination of whether changes in staging, schedule or scope can be made to rectify the situation.
Two threshold types are used to manage mobility along the key corridors. One of the thresholds is based on delay, the other on travel time. Delay threshold is a constant value throughout the day. Current travel times, accounting for congestion, incidents, weather, etc., fluctuate throughout the day. The maximum allowable travel time equals the sum of current travel times plus the delay threshold and also fluctuates.
Delay thresholds are used on most of the freeway system and some long-distance rural corridors. Travel-time thresholds are used in the Portland metro area as well as the coastal corridors, which include U.S. 101 and all of the roadways traversing the Coast Range to connect I-5 and U.S. 101. Travel-time thresholds are more appropriate for areas where diversion is likely (e.g., the Coast Range) and alternative routes are available (e.g., Portland). Travel-time thresholds are used in the Portland area because of the number of available alternative routes.
Though it appears that using travel-time thresholds in place of delay thresholds may allow for lengthy construction-related delays during nighttime hours, projects taking place then still have to adhere to ODOT’s work-zone traffic analysis methodologies, which dictate when lanes should or should not be closed. The methodology bases the allowable work windows on traffic volume levels that do not induce significant delays.
Both travel-time and delay thresholds are an attempt to quantify the maximum delay the traveling public would tolerate as a result of work-zone activities. ODOT personnel and key stakeholders established travel-time thresholds based on maximum tolerable delays before triggering economic impacts to freight and tourism. Delay thresholds were established using the following methodology:
1. Using segment lengths and speed limits for off-peak travel conditions, off-peak travel times were estimated for each segment;
2. Travel times for peak travel conditions were then estimated by increasing off-peak travel times. Off-peak travel times were increased by 45% for urban areas, 30% for semiurban areas and 10% for all other areas; and
3. Delay thresholds were then calculated to be 10% of the peak travel times.
Staging a delay
When managing mobility along segments and corridors, the delay caused by a single project is not as important as the sum of delays on the segment or corridor.
ODOT’s established work-zone traffic analysis methodology was designed to predict the hours during which lanes or shoulders can be safely closed and the approximate traffic queue length and resulting delays that would develop during such closures.
The new methodology began with ODOT’s established work-zone traffic analysis practices and existing data sources for traffic characteristics, which were augmented by software and regression analysis principles as explained in the following paragraphs.
The resulting methodology is used to predict lane closure windows, expected queuing and approximate delays during any hour of any month of the statewide construction program and has been organized into a spreadsheet compilation designed to automate analysis calculations. The methodology makes it possible to report nearly instantaneous lane closure windows and work-zone delays.
The newly developed Work Zone Traffic Analysis methodology incorporates the delay and travel-time thresholds. To manage these thresholds, delays must be estimated during the planning and design stage and measured during construction. The construction delay for projects that have significant estimated delays would be monitored 24 hours a day, seven days a week to be sure that the thresholds are not exceeded. To put these policies into practice, a methodology was created to estimate the delay at any time of day for any month and any year throughout construction. Furthermore, delays would need to be summed for each segment of roadway.
Construction staging strategies were simplified into three general categories: bidirectional, slowdowns and lane drops. In addition, Oregon Bridge Delivery Partners (OBDP), ODOT’s consultant on the bridge delivery program, added a modifier, referred to as a crossover.
In a bidirectional work zone, a two-lane road (carrying a single lane in each direction) is reduced to a single lane on which each direction of traffic alternately uses the remaining lane as controlled by flaggers or temporary signals.
A slowdown is a work zone that does not involve reducing the number of available traffic lanes, but whose impact on traffic flow primarily results from the reduction in free-flow speed caused by construction. A shoulder closure or lane narrowing would be an example of a slowdown.
A lane drop is a work zone in which one or more lanes are closed, thereby reducing the number of available traffic lanes in one direction.
Crossovers refer to a traffic shift into or across a median away from the construction area. Traffic may be required to further reduce its speed during the shift.
Traffic microsimulation was performed using the Federal Highway Administration Traffic Software Integrated System software (also known as CORSIM, short for corridor simulation). Each project scenario was modeled twice, once with no restrictions on traffic flow and once with construction restrictions in place. In this manner, the additional travel time resulting from construction activities could be estimated. Each model run is simplistic, taking into account construction restrictions only, without consideration of project-specific characteristics such as access points, ramps or signals that also may impact traffic flow.
Model runs for the preconstruction scenario utilized industry-accepted lane capacities and a free-flow speed equal to the posted speed limit plus 5 mph. Model runs for construction scenarios utilized a free-flow speed equal to the free-flow speed of the preconstruction scenario less 10 mph. Work-zone lane capacities for the construction scenarios were set equal to 1,400 PCEs per hour for work zones with a lane drop, consistent with ODOT practices.
For construction scenarios that maintain the existing number of lanes, a 5% reduction in the preconstruction lane capacity was utilized, based on field data and a literature review conducted by OBDP.
Within CORSIM, OBDP modeled simple work-zone scenarios for over 34,000 combinations of roadway types, traffic volumes, truck percentages, terrain and staging strategies. The additional travel time between two points could be determined, yielding the travel delay for the work zone. This methodology also avoids the need to calibrate each of the 34,000-plus models.
A rubbernecking factor was used to restrict the capacity of the work zone by a given percentage. A sensitivity analysis was performed to determine the percent reduction in the base roadway capacity that yields a capacity of 1,400 PCEs per hour per lane. This figure was historically used by ODOT as construction zone capacity. It was found that a 24% reduction in capacity in conjunction with the reduction in free-flow speed provided this level of capacity in the work zone. In future work, OBDP intends to increase this percent reduction in capacity to enable the modeling of high-intensity work and the increased gawking effects of such work within close proximity or high visibility from the traveled lanes.
The results of each individual analysis were grouped by model characteristics to allow for the development of volume vs. delay graphs for sets of model runs. For example, one set contains all of the freeway runs with two lanes in each direction in level terrain with a lane drop, no crossover and truck percentages between 10 and 15% with hourly traffic volumes between zero and 3,500 vehicles per hour.
The data within these groupings were exponentially regressed. The plot of the regression results forms a best-fit exponential curve through the microsimulation results. The regression results are compiled in lookup tables that allow a delay estimate to be easily provided for any combination of staging type, traffic volume, truck percentage and terrain type.
In 2006, ODOT published a revised Work Zone Traffic Analysis Manual that incorporated new elements of the methodology. Work-zone traffic analysis training courses were offered to train public- and private-sector analysts in methodology and the use of the spreadsheets to complete this analysis. These courses continue to be offered quarterly with plans for a user group also in development.
While initial field data has indicated that the delay tables are fairly accurate, data collection has continued and will be used to adjust the regressed volume to delay curves to improve the overall delay model’s accuracy. Future versions will include modifiers for the intensity of work performed in close proximity to the roadway. (Intense work tends to further decrease capacity by increasing the degree of gawking that occurs.) Future versions also will include a methodology for estimating delays for rolling slowdowns and stop-and-hold conditions.
A project is already under way to convert the Excel spreadsheets to a web-based SQL application that would allow analysis to be performed using a web browser and Internet connection. Migration to SQL also would allow for integration with ODOT’s existing GIS data, which would provide the analyst vast amounts of information on which to base his or her decisions. It is anticipated that this new application will be completed in August 2008.
Taking measurements
A program is under way to deploy automated vehicle identification (AVI) technology on multiple corridors in the state before the 2008 construction season. This technology tracks commercial vehicles equipped with truck transponders already in use for Oregon’s Green Light program. The Green Light program uses similar AVI technology to prescreen trucks at Oregon Green Light weigh stations, allowing trucks to pass if the station determines they do not need to stop. The deployment of AVI systems throughout the state will provide truck travel times along key segments of roadway that may be up to 60 miles in length.
Previous studies have indicated that truck travel times and the travel times of the entire vehicle stream can be correlated.
Using AVI and side-fire radar technology, it was determined that the average travel time reported by the AVI system was only 0.80% higher than the average travel time recorded for all vehicles by the side-fire radar. In this manner, we expect to get live data by which we can measure predicted impacts against actual impacts. These results will help us calibrate our delay models as well as monitor actual performance.
Managing the operational side of statewide mobility involves tracking delays associated with construction activities.
In Oregon, this has led to creation of new concepts and methodologies as well as finding new applications for existing technology, all focused on keeping traffic moving efficiently and safely.
Thresholds were created to establish a tolerable limit for construction impacts. These thresholds will change the way projects are scheduled, planned and constructed.
With myriad location and roadway characteristics, staging and scheduling strategies and varying traffic patterns, it was necessary to develop a new methodology for calculating delays. The new application combines the existing ODOT methodology with the newly created delay process into a responsive and flexible tool that will aid in the planning and design process for construction projects for years to come.
The corridor-level monitoring will be a true large-scale test of leveraging existing technology and equipment to serve a new purpose.
RSS: Roads & Bridges Articles
