Traffic Impact Assessment of Urban Grade-Separated Intersections on the Adjacent Street Network

1. IntroductionEach year, huge amounts of budget are spent for urban transportation projects, in order to improve traffic performance and reduce travel time. In some cases, they may produce adverse impacts on the whole network due to the lack of adequate studies. These adverse effects are in addition to financial and human resource losses and environmental impacts on urban communities. Usually in urban arterial streets, at-grade intersections are changed into grade-separated in order to reduce delays and increase the capacity of the intersection. But experiences have shown that the construction of grade-separated intersections in urban arterial streets may lead to increase travel time and delay in the adjacent network and intersections. Accordingly, deeper and more scientific studies are needed to address this phenomenon and to obtain a reliable criterion for assessing the performance of the network with at-grade or grade-separated intersection.Despite extensive studies on the performance of grade-separated intersections and establishment of necessary rules for assuring the performance and quality of traffic in recent decades, few researches has been done on measuring the traffic impacts of grade-separated intersections [1]. This could be due to the difficulty of field data collection required for this evaluation. In addition, before and after studies of grade-separated intersection are long lasting, meanwhile the traffic parameters may change. In these circumstances, the use of simulation techniques is inevitable. The main objective of this paper is to examine the effect of grade-separation on traffic indices in an urban arterial street network based on a simulation approach. To this end, using micro-simulation software, a hypothetical network is modeled, with and without a grade-separated intersection, and under different traffic and geometric conditions. The simulation results will be used to develop a two-fluid model. The two-fluid model establishes relatively simple relationship between stopping and moving vehicles in an urban street network at aggregate level. The simulation results in terms of traffic indices as traffic delay speed and volume as well as the two-fluid parameters will be compared. 2. Methodology2.1. Theoretical backgroundIn general, traffic flow modeling is possible at micro and macro levels. In disaggregate modeling, micro variables as vehicle’s speed and spacing are used. Simulation is considered amongst disaggregate models, in which based on the specific models, the behavior of vehicles on the network elements regenerates step by step. Aggregate models are developed based on the macro variables in the network, and their relationships. This study was intended to deploy both these models to assess the effects of grade-separated intersection. To this end, the strong and well-known software, CORSIM, was used for traffic flow simulation on a hypothetical network. Based on the simulation results a two-fluid model of network was developed.The two-fluid model gives a relatively simple relationship between stopped and moving vehicles variables [2]. The first assumption in this model is that the speed of moving vehicles is proportional to the ratio of vehicles in motion, and the second assumption the model is that the average ratio of stopping to moving time of vehicles in the network is equal to the average ratio of total vehicles stopped on the network, fs. The travel time, stopped vehicles, and free flow travel time data can be used for logarithmic regression to obtain parameter n, of the network [3].2.2. Simulation DetailsThe simulation model was intended to be similar to actual situation and the results could be generalized to various conditions of urban arterial street network [4]. There fore various geometric and traffic conditions were applied to the model. The model includes the following states: - network, with or without grade-separated intersection- different intersection spacing (300, 400, 500 m), - different network loading (300 500, 700, 900, 1100, 1300, 1500 vehicles per hour),3. Results and discussionThe hypothetical model was established under each of the spacing and loading for both at-grade and grade-separated intersection in order to compare the parameters and changes in these parameters, as well as the effect of each geometric and traffic variable. The hypothetical network which was a grid type street network consisting of 21 nodes. Was connected by two-way, six-lane streets at regular spacing, without right or left turn lines. The network structure and intersections are shown in Fig. 1. All intersections of the grid, with the exception of central intersection (node A), are signalized. But the central intersection was considered in both cases of at-grade and grade-separated, with the possibility of turning movements at single point level.4. ConclusionsThe analysis of simulation results showed that: - Construction of grade-separated intersection at high traffic volumes can decrease the network throughput, up to 30%. This effect increases as the distance between intersections decreases. - Grade-separation only at low volume would reduce the network delay and at short intersection spacing and high intersection volume the delay may increase up to 50%.- Construction of grade-separated intersection merely at low volumes increases the average speed of traffic. With increasing spacing, this effect is reduced.- Percentage of stopped vehicles in the network increases the grade-separation at medium and large volumes. But as spacing increases, this effect is reduced.- Based on the parameter n, the construction of the grade-separated intersection at low to medium volumes results in the improved performance of network; and in other conditions, the impact is negative.5. References[1] Garber N. J., Fontaine M. D., “Guidelines for Preliminary Selection of the Optimum Interchange Type for a Specific Location”, VTRC Report No.99-R15, Charlottesville, Virginia Transportation Research Council, 1999.[2] Amini, B. Shahi, J., Ardekani, S., “An observational Study of the Network – Level Traffic Variables”, Transportation Research, 1998, 32 (4), 271-274.[3] Williams J. C., Mahmassani H., Herman R., “Analysis of Traffic Network Flow Relation and Two-Fluid Model Parameter Sensitivity", Transportation Research Record 1985, 95-106.[4] CORSIM User’s Manual, Version 5.1, 2008.