By: John Collura, Ph.D., PI, and Andrew Berthaunme, EIT
Much of the Dwight D. Eisenhower National System of Interstate and Defense Highways are more than 30 years old.
As the National Highway System (NHS) continues to age and reach the end of its serviceable life, the focus of roadwork has shifted from new construction to rehabilitation and maintenance of existing roads.
The challenge faced by transportation officials and contractors is to reduce the negative impacts of work zones on driver mobility. Motorists throughout the U.S. have cited work zones as second only to poor overall traffic flow as being the major cause of traveler dissatisfaction.
It is essential to recognize the impacts that proposed reconstruction or rehabilitation work can have on traffic well before construction begins. This allows for appropriate cost-effective mitigation strategies to be developed and implemented prior to delays occurring. Work-zone mobility assessments are necessary to understand the type, severity and extent of impacts associated with different project alternatives. By aggressively anticipating and mitigating congestion caused by work-zone activity, positive impacts of relieving such congestion can be realized.
Despite the increasing frequency of work zones, the effects of a project are not usually considered until the design phase. Moreover, user costs are rarely considered during the planning and development phases of many projects.
Being that agency and user costs are significantly affected by the timing and configuration of a work zone, it has become highly desirable to optimize work-zone scheduling so as to minimize total cost. It is in the interest of transportation engineers to be able to present reliable information regarding impacts that may occur with the implementation of a work-zone strategy. One of the major tools used to realize these impacts is computer simulation.
A traffic backup plan
During the last 20 to 30 years, a large number of sophisticated traffic-simulation models have been developed. Simulation is a powerful tool that can be used in the analysis and assessment of transportation facilities. Simulation models have the capability to incorporate a number of analytical techniques into their framework for simulating complex components, providing users with a greater knowledge and understanding of the system being analyzed. The low-cost, low-risk environment allows users to test a number of assumptions and alternatives and analyze the effects immediately.
States have used computer simulation to predict traffic conditions in work zones as part of the decision-making process on large, highly visible projects. Simulation is not routinely used, however, in either the project planning or design phases of many of the nation’s roadway reconstruction or rehabilitation activities. Simulation models can aid transportation officials and agencies in the prediction of queue lengths, delay times and travel speeds. The Federal Highway Administration (FHWA), however, revealed that many simulation packages are not user-friendly and are not readily adaptable to local traffic conditions experienced during construction activities.
According to the FHWA, many agencies are making an effort to use more advanced tools such as simulation for work-zone analysis. Different tools may be appropriate for different situations, with decisions being based on the size and scope of the project. Work-zone-specific simulation models include QUEWZ and QuickZone. QUEWZ analyzes traffic conditions on freeway segments with and without lane closures, providing estimates of additional road-user costs and of queuing as a result of work-zone lane closures. QuickZone compares traffic impacts for work-zone mitigation strategies, estimating the costs, time delays and potential backups associated with these impacts.
Almost real
The simulation models were evaluated in terms of their ease of use, data requirements and ability to simulate and assess work-zone strategies, shedding light on their relative reliability and accuracy as well as their user-friendliness. The case studies include work-zone projects along I-91 in Greenfield, Mass.; I-91 in Windsor, Conn.; I-95 in West Greenwich, R.I.; I-95 in Bangor, Maine; and I-93 in Manchester, N.H. Where possible, the evaluation portion of this research includes a comparison of the results simulated by QUEWZ and QuickZone. For example, field estimates on queue length were made by state DOT project engineers while the work zone was in place.
Table 1 summarizes a comparative review of the queue lengths estimated with QuickZone and QUEWZ simulation packages and queue lengths observed in the field along the interstate projects. It can be seen from these results that QUEWZ and QuickZone produced queue-length estimates close to the queue-length estimates observed in the field. It also should be noted that the QUEWZ estimates of percent error in this research are comparable with those produced by research done at the University of Iowa. The percent error in this research for QUEWZ queue lengths was 0-6.25% and percent error in the University of Iowa study QUEWZ volume estimates were 1-19.2%. Figures 1 and 2 present the results graphically.
It is interesting to note that along I-91 in Greenfield, Mass., QuickZone estimated a maximum queue of 3.85 miles to occur on a Sunday due to the recreational ski traffic returning home. The queue begins to generate around 11 a.m., reaching its maximum at approximately 4 p.m. The queue was estimated to be totally dissipated by 7 p.m. Comparing these estimates to real-world data provided by past research, QuickZone provides a fairly accurate estimate of the actual queue length. The research reported that, “On most Sundays, the queue would be 4 to 6 miles with propagation beginning at about 11:30 a.m. The queues would dissipate between 4 and 6 p.m., depending on demand for that afternoon.”
It also was reported by the local media and the Massachusetts State Police that queues of approximately 12 miles had formed in the early stages of the project. The estimation provided by QuickZone does not confirm this portion of the reported real-world data. It is believed that this may be due to other factors such as different work-zone staging strategies, driver unfamiliarity, work-zone intensity and poor mitigation strategies.
The QuickZone analysis of I-95 in West Greenwich, R.I., suggests that no queue should have been experienced. These results confirm the observations made on a site visit by the research project team as well as information gathered from a construction worker during the same visit. On Tuesday, June 19, 2007, researchers visited the work-zone site from 2:30 p.m. to 4:30 p.m. During this time period, no queue formation was observed nor did there appear to be any sign of a queue developing. The increase in travel demand was noticeable during this time, but traffic continued to flow steadily through the work zone at an estimated 65 mph.
A RIDOT official stated that a queue would not form, as two lanes of travel are available and maintained through the work zone. Additionally, the area is in a rural setting in which travel demands are not very high. It should be noted that a RIDOT official did reveal that the only time a queue forms for this particular work site is when a crash occurs or when the workers must shut down one or more travel lanes for construction activity. The worker stated that in the occurrence of a traffic incident or lane closure, traffic may back up as far as I-295, approximately 10 miles from the work zone. In an analysis of alternative lane-closure conditions, QuickZone estimated that a 24-hour lane closure would produce a maximum queue of 12.73 miles on a Friday. Additionally, the maximum estimated queue length for a 1-hour lane closure was 5.47 miles on a Friday. The simulation results appeared to be consistent with the RIDOT official’s estimate.
The QuickZone analysis of I-91 in Windsor, Conn., (one lane closed) suggests no queue on Monday and a maximum queue of 0.3 miles on Friday. The queue begins to build at around 9 p.m, and is estimated to dissipate by early morning of the following day. Comparing these estimates to field observations provided by the resident engineer working for the Connecticut Department of Transportation, Quick Zone provided a fairly accurate estimate of the actual queue length. QuickZone also yielded a queue length in Bangor, Maine, that was similar to that estimated based on field observations.
Table 2 presents a comparison of the general characteristics, parameters and constraints of three software packages and also provides a summary of the time required to assemble, input and analyze the data for the QuickZone and QUEWZ simulation models. It is hoped that the information presented will lend insight to the general functional purposes and user-friendliness of each package. It should be noted that time requirements may vary from project to project due to the availability of the necessary data. The times also will vary relative to the user’s familiarity with a given package and fundamental traffic-flow concepts.
From Table 2, it can be seen that QUEWZ takes significantly less time to assemble and model data than QuickZone. QuickZone has more purposes and capabilities and QUEWZ has fewer.
Low risk at a low cost
This article has focused on the application and evaluation of QUEWZ and QuickZone to simulate and assess work-zone strategies implemented in New England. An overview of these simulation models has provided a means for potential users to gain a broad perspective of the requirements and capabilities of each model. The research has illustrated the use of both the data input and output procedures for QuickZone and QUEWZ.
Where possible, the simulated results have been compared directly to observed field data collected in this study and by others, allowing for a judgment to be made as to the reliability and accuracy of the estimation ability of these models. Additionally, the use of these models to conduct this research has shed light on a number of other factors of interest to potential users, including software/hardware requirements, user friendliness, convenience and flexibility.
QuickZone can be obtained from McTrans at the University of Florida. The model runs as a Microsoft Excel macro and can be accessed directly from the computer’s desktop. QuickZone requires a minimum of Microsoft Windows 95 with Microsoft Excel 97 or newer. Along with being a generally accurate simulation model, QuickZone also appears to be rather user-friendly. Although initial data entry may be a time-consuming process, alternative work-zone strategies can be analyzed with relative ease. This allows the user to compare several viable options and select the most optimal. The required base input also is relatively easy to obtain. More detailed input such as seasonal traffic demands and pre-construction travel behaviors may be more difficult to gather, but may provide the user with more accurate results. The results produced by QuickZone do provide the user with meaningful information, from queue length to time delay to user costs. The benefit of QuickZone is that these results are provided in both tabular and graphical form, allowing users to have multiple means of interpretation. Future research involving QuickZone could include:
- Applying and evaluating QuickZone to various roadway classifications (i.e., higher-volume interstates, rural or urban arterials, two- or three-lane interstates, local roads, etc.);
- Analyzing the effect of work-zone intensity as adjusted within the HCM capacity reduction function;
- Analyzing the effect of full road closures with the use of detour routes;
- Analyzing the effects of altering pre-construction travel behaviors and work-zone mitigation strategies; and
- Developing a way to account for speed differentials upon approach, passage and exit of the work zone and analyzing the associated effects related to speed.
This research has shown that some simulation models provide a low-risk, low-cost environment in which to test and analyze a variety of work-zone alternatives. For example, QUEWZ and QuickZone were able to provide reasonable order of magnitude queue-length estimates on interstate highways comparable to observations made in the field. In addition, such estimates required little data including hourly volume and roadway geometry information.
Care must be taken, however, in using simulation results to make concrete decisions. It is strongly recommended that users of these simulation models have a fundamental understanding of highway capacity analyses and traffic-flow fundamentals. Users must trust their intuition and use their knowledge when results appear to be out of the ordinary. Simulation does, however, give the transportation world a better understanding of the impacts of highway work-zone strategies.
About The Author: Collura is professor of Civil and Environmental Engineering at the University of Massachusetts-Amherst and is the director of the UMass Transportation Center. Berthaunme is a research assistant at the UMass Transportation Center.