Transportation Research Record: Journal of the Transportation Research Board

A speed-change lane (SCL) is an auxiliary lane added to the traveled way for the acceleration and deceleration of vehicles entering or leaving a roadway. When the length of an acceleration SCL is adequate, drivers are able to accelerate comfortably from the speed at entrance to a speed appropriate to the road, find a gap in the traffic flow, and merge in a safe and secure manner. The length of an SCL is currently determined in terms of the ramp design speed, the freeway design speed, and the acceleration rate. Embedded in these values are assumptions for the operating speed at the entrance and merging points. This study examined a probabilistic approach instead of such a deterministic approach. The main benefit of a probabilistic approach is that traffic flow characteristics are assumed to be stochastic; therefore, the outcome of a probabilistic methodology is a distribution of drivers’ acceleration distance on the SCL. The reliability-based analysis enables designers to select a specific percentile value of this distribution as a design length that better matches a certain situation and avoids unnecessary extra construction costs. This paper presents analytical and simulation models for the application of the reliability approach, with all parameters based on recently collected field data. Even though the presented model should be superior to the deterministic model adopted in current design guides, additional enhancements are recommended for a full reliability-based, safety-explicit design model.

Geometric design decisions regarding interchange and ramp spacing on freeways have traditionally taken a nominal approach to safety. This approach oversimplifies driver behavior and complex interactions between roadway geometrics, traffic operations, and safety and tends to promote a one-size-fits-all approach to evaluating design alternatives. The objective of this paper was to quantify the relationship between ramp spacing and freeway safety. Data for this study included freeway geometric features, traffic characteristics, and crash counts for 404 freeway segments in California and Washington State. A negative binomial regression modeling approach was used to explore the relationship between ramp spacing and safety. Results indicated that expected crash frequency increased as ramp spacing decreased. The expected proportion of crashes resulting in a fatality or injury appeared to decrease as ramp spacing decreased. The presence of an auxiliary lane was associated with a lower expected frequency of crashes for any given ramp spacing; the safety benefits of providing an auxiliary lane diminished as ramp spacing increased. The model results related to ramp spacing and the presence of an auxiliary lane were transformed into crash modification factors and validated through comparisons with an independent research study on weaving areas in Texas. The safety findings were generally consistent with current Green Book design criteria related to ramp spacing and the presence of an auxiliary lane. The ability to quantify the expected impact on safety of ramp spacing associated with new or modified interchanges will assist transportation agencies in making well-informed assessments of the overall benefits, impacts, and costs of freeway access decisions.

One of the most important tasks in traffic safety is investigating the relationship between motor vehicle crashes and the geometric characteristics of roadways. A large body of previous work provides meaningful results on the impact of geometric design on crash frequency. However, little attention has been paid to the relationship between roadway departure crashes and relevant roadside features such as lateral clearance, side slope condition, and driveway density. The lack of roadside data for use in estimating rigorous statistical models has been a major obstacle to roadside safety research for many years. This study investigated the relationship between single-vehicle roadway departure crashes and roadside features. Two types of models were developed: a negative binomial model of crash frequency and a multinomial logit model of crash severity. The study used field data collected in four districts in Texas. The results showed that shoulder width, lateral clearance, and side slope condition had a significant effect on roadway departure crashes. Crash frequency and severity increased when lateral clearance or shoulder width decreased and when the side slope condition became worse. Driveway density was not found to have a significant influence on crash frequency or severity.

Typical four-phase, four-legged intersections often operate inefficiently and severely restrict vehicle throughput, and thereby cause large delays. These configurations are limited by traditional geometric design. Altering the geometry of the entire intersection can significantly increase the capacity. Two-phase intersections employ an unconventional lane arrangement to maximize vehicular throughput. This arrangement involves displacing left-turn lanes across opposing through traffic before the main intersection is reached. Such an alteration allows left and through vehicles to proceed simultaneously; consequently, the intersection capacity is improved and delays decrease. Numerous studies have validated the operational improvements associated with such a geometric design, but in-depth analysis of this unconventional system can provide a maximized effect. This paper proposes an improvement to the geometric design and develops an optimization model of a two-phase intersection to increase intersection capacity. After main objective functions and a list of constraints were defined, preliminary tests were performed to show the effectiveness of the method in adapting to changes in traffic demand. The results from the optimization were used to run simulation tests that compared the capacity and delay at a two-phase intersection with those of a conventional four-legged geometric design. The results of several trials strongly showed the optimum performance of the proposed unconventional geometric arrangement.

A performance-based approach to the interaction of geometric design, speed, and safety is considered given the availability of two key documents: the Highway Safety Manual and Modeling Operating Speed: Synthesis Report. A historical look at the concept of design speed shows that although the definition of design speed has changed on more than one occasion, the same basic philosophy that related design speed to a safe speed is still reflected in current policy in supplemental guidance related to the selection of design speed. A conservative approach to establishing design criteria, used to address the range of driver, vehicle, and roadway conditions and capabilities that a designer must consider, is demonstrated. Operating speeds are shown to be higher than design speeds for design speeds of approximately 55 mph or less. This outcome may be considered undesirable, but that categorization seems to be based more on subjective judgments of what is desirable than on actual safety findings. Finally, the idea of speed management through the use of roadway geometrics (i.e., geometric designs that influence driver selection of operating speed)—one component of self-enforcing, self-explaining roadway design—is explored. Findings uncover possible challenges to implementing this idea. Five related questions are addressed: (a) What is known about the relationships between road geometry and operating speeds? (b) To what degree does road geometry influence operating speeds? (c) How are safety and security influenced by road geometry? (d) What are the potential impacts on large vehicles? and (e) What is the nature of the speed–safety trade-off?

This paper presents the results of a study to develop improved design guidance for freeway main-line ramp terminals on the basis of modern driver behavior and vehicle performance capabilities. The primary steps included a crash analysis, an observational field study, and a driver behavioral study. This research reached several key conclusions, as follows. Vehicle merging speeds tend to be closer to freeway speeds at tapered acceleration lanes than at parallel acceleration lanes. The recommended minimum lengths for acceleration lanes presented in the 2004 edition of AASHTO's A Policy on Geometric Design of Highways and Streets (Green Book) are conservative and, under certain conditions, could be reduced by 15%. In a situation in which a significant volume of trucks uses an entrance ramp, an acceleration lane length that will better accommodate trucks can be derived by using speed–distance curves developed for a range of weight-to-power ratios. The recommended minimum lengths for deceleration lanes presented in the 2004 Green Book are conservative and do not account for deceleration in the freeway. Providing deceleration lanes longer than the minimum values given in the Green Book may promote casual deceleration by exiting drivers. Several potential changes for consideration in the next edition of the Green Book are recommended on the basis of the findings and conclusions of this research.

This paper evaluates the structural integrity and safety performance of weathered steel beam guardrail (WSBG) in North Carolina. WSBG is a popular alternative to galvanized steel guardrail (GSG) in that it is an aesthetic solution that blends in with the surrounding natural environment. A New Hampshire study, however, found that WSBG deteriorates more quickly. That study prompted an FHWA recommendation discouraging the use of WSBG across the country. This paper describes a comprehensive study of all WSBG installations in North Carolina, where weather conditions and deicing practices are generally much less severe than in New Hampshire. The research results did not show any trends of deteriorating thickness as a function of guardrail age (the oldest installation was almost 30 years old), elevation (the tallest installation was 4,200 ft), and annual average daily traffic (the highest traffic volume was 27,000 vehicles per day). The WSBG at all 25 test sites was shown to be above AASHTO minimum thickness specifications of 0.096 in. at a 95% confidence level. From a safety perspective, the research results suggested a lower percentage of injury collisions at WSBG installations relative to GSG installations at comparison sites, although low crash sample sizes made a more robust comparison difficult. Overall, this research suggests there are no structural concerns about using WSBG in North Carolina. The findings of this study may have broader policy implications for other states.

Pavement edge drop-off can be a serious safety concern when a vehicle leaves the paved roadway surface and encounters a significant difference in vertical elevation between the paved roadway and the adjacent unpaved shoulder. Edge drop-offs are potential safety hazards because scrubbing between the pavement edge and tire can result in loss of control. FHWA developed the Safety Edge on the basis of research results that indicated a sloped pavement edge surface could be more easily traversed by a vehicle leaving its lane and attempting to remount the pavement edge. The Safety Edge is a design feature that creates an approximate 30° fillet along the outside edge of the paved section of a roadway. Although a number of benefits have been attributed to the Safety Edge, agencies in Iowa were slow to adopt it. To accelerate use of the Safety Edge, the research team marketed and monitored it in Iowa during the 2010 construction season. This project provided the opportunity to evaluate the Safety Edge as well as summarize lessons learned from agencies and contractors. Evaluation of the Safety Edge included assessment of the consistency of slope application, identification of common problems during construction, measurement of the density of the Safety Edge, and computation of the approximate additional material needed to construct the Safety Edge. The team also measured drop-off formation along previously constructed Safety Edge sites and compared this with control sites to assess any difference in occurrence.

The varied lateral stiffness between bridge rails and approach guardrail systems may cause vehicle pocketing or wheel snagging to occur near rigid bridge rail ends. To mitigate this potential hazard, an approach guardrail transition (AGT) is used to provide a gradual increase in the lateral stiffness of the barrier between the W-beam guardrail system and the bridge railing. However, these transitions can also cause a propensity for vehicle pocketing or wheel snagging if the change in lateral stiffness occurs too rapidly. Recently, a stiffness transition based on NCHRP Report 350 was developed for use with the Midwest Guardrail System (MGS) and a stiff Thrie beam AGT, and successful testing was performed close to the upstream end of the AGT. The transition was designed with three sizes of steel posts, one of which was nonstandard for state departments of transportation. Thus, a simplified version of the original MGS stiffness transition that used two common sizes of steel posts was developed and was subjected to full-scale crash testing according to Test Level 3 as set forth in the Manual for Assessing Safety Hardware. Subsequently, dynamic post properties obtained from bogie testing and numerical simulations were used to develop an equivalent wood post version of the simplified MGS stiffness transition. Recommendations are made regarding the attachment of the stiffness transitions to FHWA-accepted Thrie beam bridge rail AGTs.

A new fracturing-bolt universal breakaway steel post (UBSP) was developed and evaluated for use as a replacement for the controlled-release terminal (CRT) wood post currently used in the Thrie beam bullnose system. After numerous steel post concepts were investigated, a fracturing-bolt steel post was selected as the most promising design. The fracturing-bolt steel post successfully matched the strength and dynamic behavior of the CRT wood post in three impact orientations. The UBSP was incorporated into the Thrie beam bullnose barrier system and subjected to three full-scale vehicle crash tests according to the Test Level 3 guidelines provided in NCHRP Report 350. Test Designations 3–30, 3–31, and 3–38 were chosen to evaluate the performance of the Thrie beam bullnose system with UBSPs. All three full-scale crash tests demonstrated that the UBSP performed in a satisfactory manner in the bullnose system, as the vehicle was captured and safely brought to a controlled stop. On the basis of the successful completion of the three full-scale crash tests, it is recommended that the UBSP be considered a safe alternative to CRT posts in the original Thrie beam bullnose median barrier system. It is also noted that the performance of the UBSP suggests that it may have additional applications, including in long-span guardrail, end terminals, and guardrail in mow strips or encased in pavement.

Median barriers are commonly used to separate opposing lanes of traffic on divided highways and to separate managed lanes from general purpose lanes. Concrete median barriers (CMBs) are often preferred on urban freeways with narrow medians because of their minimal deflection and low maintenance. However, long, continuous runs of CMBs limit access of emergency and maintenance vehicles to the other side of a roadway or a managed lane. Implementation of crashworthy median barrier gates at these locations can maintain the desired level of median protection for motorists while offering improved cross-median access for emergency and maintenance vehicles. A new median barrier gate was developed and crash tested. The gate spans a 30-ft opening in a CMB and consists of two vertically stacked 12- x 12- x ¼-in. steel tubes connected to steel end brackets with 2¼-in. diameter steel pins. The gate is economical to fabricate and install. It can be manually operated by a single person and is designed to accommodate reversible traffic flow on both sides of the median and be operable in both directions on each end. The median barrier gate satisfies the criteria for impact performance of Test Level 3 of the Manual for Assessing Safety Hardware and is considered suitable for implementation on divided highways at locations where cross-median access is desired.

The National Crash Analysis Center (NCAC) at the George Washington University simulated the crash of a 2,270-kg Chevrolet Silverado pickup truck into a standard 32-in. New Jersey shape concrete barrier under the requirements of Test 3–11 of the Manual for Assessing Safety Hardware (MASH). The new, detailed finite element (FE) model for the Chevrolet Silverado was used as the surrogate for the MASH 2270P test vehicle. An FE model of the New Jersey barrier was drawn from the array of NCAC hardware models. The primary objective of this analysis was to simulate the crash test conducted to evaluate how this commonly used, NCHRP 350–approved device would perform under the more rigorous MASH crashworthiness criteria. A secondary objective was to use newly developed verification and validation (V&V) procedures to compare the results of the detailed simulation with the results of crash tests undertaken as part of another project. The crash simulation was successfully executed with the detailed Silverado FE model and NCAC models of the New Jersey concrete barrier. Traditional comparisons of the simulation results and the data derived from the crash test suggested that the modeling provided viable results. Further comparisons employing the V&V procedures provided a structured assessment across multiple factors reflected in the phenomena importance ranking table. Statistical measures of the accuracy of the test in comparison with simulation results provided a more robust validation than previous approaches. These comparisons further confirmed that the model was able to replicate impacts with a 2270P vehicle, as required by MASH.

Very little is known about the real-world performance of traffic barriers when subjected to impacts by large trucks. This study investigated real-world impacts of large trucks into traffic barriers to determine barrier crash involvement rates, the impact performance of barriers not specifically designed to redirect large trucks, and the real-world performance of barriers specifically designed for large trucks. Data sources included the Fatality Analysis Reporting System (2000 to 2009), the General Estimates System (2000 to 2009), and the Large Truck Crash Causation Study (155 in-depth crashes of large trucks into barriers). Impacts of large trucks into longitudinal barriers constituted 3% of all police-reported impacts into longitudinal barriers and roughly the same proportion of barrier fatalities. A logistic regression model predicting barrier penetration showed that the risk of a large truck penetrating a barrier increased by a factor of 6 for impacts with barriers designed primarily for passenger vehicles. Although barriers specifically designed for impacts by large trucks performed better than barriers not specifically designed for impacts by heavy vehicles, the penetration rate of the former was 17%. This penetration rate is of concern, because barriers used for higher test levels are designed to protect other road users, not the occupants of large trucks. Barriers not specifically designed for impacts by large trucks prevented penetration by a large truck approximately half the time. This finding suggests that adding costlier barriers that meet higher test levels may not always be warranted, especially on roadways with lower truck volumes.

In 2009, the Manual for Assessing Safety Hardware (MASH) was adopted as an updated replacement of NCHRP Report 350. Under MASH, the impact severity of Test Level 4 (TL-4) (Test 4–12) was increased by 56%. Test vehicle mass was increased from 17,640 to 22,050 lb. Impact speed was increased from 50 to 56 mph. Although this is a significant increase in the amount of lateral energy imparted to the barrier, the 54-kip design impact load currently suggested by the AASHTO load and resistance factor design (LRFD) bridge specifications has not been updated to meet MASH TL-4 requirements. Similarly, the AASHTO LRFD specifications require all TL-4 railings to have a minimum height of 32 in. However, because of the increase in test vehicle mass and impact velocity under MASH, the minimum height for TL-4 railings is expected to be greater than 32 in. This height was verified by a recent unsuccessful MASH Test 4–12 of a New Jersey profile rigid concrete barrier 32 in. tall. The research presented in this paper used finite element analysis and subsequent crash testing to determine the minimum rail height for MASH TL-4 impact conditions. The researchers also determined the appropriate lateral impact design load for use with AASHTO LRFD bridge specifications for the design of TL-4 bridge rails.

In 2003, the Minnesota Department of Transportation (DOT) developed a unique method of integrating context-sensitive solutions (CSS) into its design–build program in a process called visual quality management. Although this name implied that the process was only concerned with generating an aesthetically appealing project, the process was not as concerned with visual quality per se as it was with a method that resulted in a project that was acceptable to the public. This method was based on the CSS principles and practices that the Minnesota DOT had adopted as a business model, notably the inclusion of stakeholders; the use of interdisciplinary design teams; the adoption of multimodal solutions; the solving of associated social, economic, and environmental problems; and the creation of a safe, efficient, and effective transportation system. Visual quality management was actually a method for measuring conformance with a project's visual quality goals, which had been generated according to the principles of CSS. This process required the successful bidder to develop a visual quality management plan that outlined how the contractor would define and mitigate visual impacts, the responsibilities and authority of a visual quality manager and a visual quality review committee, and the process that the contractor would use to engage that committee in developing a visual quality manual. The manual established the criteria for developing CSS for 15 design elements that constituted a transportation project's visual environment and to which the contractor had to conform.

Despite a large body of research on operating speeds, there is still much to learn about the factors that affect free-flow speeds, especially with regard to the development of a continuous speed profile for two-lane rural highways. This paper presents a new approach to modeling driver speed profile along two-lane rural roads. Speed data regarding individual drivers traveling on selected two-lane rural roads were sampled during a naturalistic experiment that used an onboard Global Positioning System (GPS). GPS devices facilitate the collecting and processing of continuous speed data. Unlike spot speed measurements, which have typically been collected in similar studies, continuous speed data are useful for studying the actual speed profile with regard to acceleration–deceleration behavior and maximum speed on tangents and curves. New variables are used in the regression equation. These variables refer to the horizontal curvature and the vertical grade as weighted values of the geometric features of the alignment preceding and following the vehicle position. To predict the operating speed, the model takes into account the geometric alignment of the road preceding and following the vehicle, as this information is used by drivers to establish the driving task, which is based on memory expectation, vehicle dynamics, and visual perception. The final result is an algorithm for modeling a continuous speed profile without previous assumption of acceleration–deceleration rate and constant speed on tangents and curves.

The U.S. Environmental Protection Agency is developing a nationwide standard for turbidity in construction site runoff. It is widely expected that conventional measures for erosion and sediment control cannot meet this standard; consequently, innovative practices for managing sediment on construction sites must be developed. One emerging practice is the use of polyacrylamide (PAM) to improve sediment control by promoting flocculation of particles in runoff. Effective use of this practice requires an understanding of how the properties of soil and PAM affect flocculation. The objective of this study was to develop such an understanding. A protocol for creating modified synthetic stormwater runoff from soil samples was developed, and synthetic runoff suspensions were created with soil from six Texas construction sites. Particle size distribution was used to compare the synthetic runoff suspensions with grab samples of stormwater from one site. Flocculation tests were performed on the synthetic runoff suspensions with PAM doses ranging from 0.03 to 10 mg/L. The polymers used included anionic PAMs of 0% to 50% charge density and 0.2 to 14 Mg/mol molecular weight and one copolymer. A neutral PAM and the copolymer were the most effective in reducing the turbidity of all the synthetic runoff suspensions below 200 nephelometric turbidity units at doses of 10 mg/L. Hardness tests indicated interparticle bridging to be the bonding mechanism. The high-molecular-weight (HMW) anionic PAMs were effective on only two of the six synthetic runoff suspensions. As the charge density of these HMW PAMs increased, their effectiveness decreased.

The North Carolina Department of Transportation and the U.S. Geological Survey collaborated on a study to better understand the effects of stormwater runoff from bridges on receiving waters. The following tasks were performed: (a) characterizing the quality and quantity of stormwater runoff from a representative selection of bridges in North Carolina, (b) measuring the quality of stream water upstream of selected bridges to compare constituent concentrations and loads in bridge deck stormwater with those in the stream, (c) determining whether the chemistry of bed sediment upstream and downstream of selected bridges differed substantially according to the presence or absence of a best management practice for bridge runoff, and (d) estimating the rate at which bridge deck runoff mixed with the receiving stream. The investigation measured bridge deck runoff from 15 bridges for 12 to 15 storms, the quality of stream water for base flow and storm conditions at four of the bridge deck sites, and the chemistry of stream bed sediment upstream and downstream of 30 bridges across North Carolina. Runoff and stream samples were analyzed for a wide range of constituents, including dissolved and total recoverable metals and nutrients, major ions, total suspended solids, suspended sediment, oil and grease, petroleum hydrocarbons, and semi-volatile organic compounds. For 64% of comparisons, concentrations in bridge deck runoff were no different or were less than those measured in receiving waters, and the maximum concentrations of constituents in the bridge deck runoff were rapidly reduced to the ambient stream concentrations, generally within 50 ft downstream of the bridge.

Recent improvements and the widening of the I-84 Bridge crossing of the New York Canal in Boise, Idaho, have increased the number of bridge columns from 28 to 60. The resulting structure has two parallel rows of columns that extend across the width of the bridge longitudinally within the canal. After the widening of the bridge and addition of the bridge columns, the canal began experiencing an oscillating wave phenomenon that originated from the bridge columns and caused erosion of upstream and downstream canal banks and bridge abutments. A physical model study was conducted to investigate the wave phenomenon and determine what modifications to the columns or canal would be necessary to prevent the wave oscillations. The physical model was successful in simulating the wave phenomenon, and four different modifications for resolving the wave problem were tested in the model. A unique solution was found that used precast nose cones attached to selected columns. The nose cones have been installed in the prototype bridge crossing, and no wave oscillations have occurred since installation. This paper discusses the study to simulate the wave phenomenon and the four modifications that were evaluated to reduce or prevent wave oscillations.

Detection of utility conflicts as early as possible during the project development process can help to substantially improve the timely relocation of utilities and can allow time to develop alternatives to avoid utility relocation. However, collecting accurate information on the location of underground utilities from utility companies can be challenging. This is one reason that subsurface utility engineering (SUE) has become a critical tool in helping to identify and locate utility installations within the right-of-way. Although the benefits of SUE have been well documented, a disconnect between its full potential and its actual use at the Texas Department of Transportation (DOT) was suspected. A survey of staff in diverse organizational units at the Texas DOT was conducted to learn about current utility investigation practices and to gain a better understanding of how the Texas DOT uses SUE. The team found, among other issues, some unfamiliarity with SUE technology and its best uses, a general need for training and education on the use of SUE, and a recent decline in the use of SUE technology that may be linked to frequent uncertainty about the benefits of SUE.