Travel Time Savings Benefit Analysis of the Continuous Flow Intersection: Is It Worth Implementing? The continuous flow intersection (CFI) has become increasingly popular and credible in recent years. Despite its unusual and unintuitive design, CFI implementation is beneficial to the reduction of traffic congestion. However, replacing every conventional intersection with a CFI design is not feasible. This study investigates whether it is always beneficial and worth the implementation. By Jumrus Pitaksringkarn, P.E. and Shinya Hanaoka INTRODUCTION Traffic congestion has become one of the most prevalent and frustrating problems in growing cities worldwide. A torrent of attempts have been made to alleviate the problem, including a recently proposed groundbreaking intersection design, the continuous flow intersection (CFI). The CFI design is an unconventional intersection that enhances capacity by using special handling of right turns, considering left-hand driving regions under study. Specifically, right-turn traffic crosses oncoming opposing traffic 300 400 feet (100 120 meters) before approaching the main intersection with a two-phase signal-controlled intersection at the crossing point. Although this CFI application may seem unconventional and unintuitive, a series of feasibility studies and analyses has been conducted, and the CFI has gained more credibility and acceptance in the past few years as an advantageous alternative. Before any real conclusion can be made, however, these studies and analyses still need additional components, such as an in-depth analysis of travel time savings and monetary benefits. These two components are considered critical factors in any transportation benefit analysis study. Travel time savings could account for as much as 80 percent of the total benefits in any infrastructure improvement. Therefore, this study was analytically focused on this travel time savings measurement and its monetary benefits estimation, a fundamental factor in the CFI implementation decision. These evaluations help answer questions about how much would be gained from the implementation, how much would be saved at various traffic conditions and volumes, or when benefits would or would not result from a CFI application. WHAT IS A CFI? A CFI is an unconventional intersection improvement. Various naming alternatives have been proposed, such as crossover displaced left turn (XDL) and displaced right turn (DRT), depending on the driving side. 1,2 Despite the different names, they all achieve the same outcome; basically, the CFI removes conflict between right-turning and oncoming traffic at the main intersection by allowing right-turn traffic to enter a right-turn bay and crossing an oncoming throughtraffic lane before approaching the main intersection (for left-hand driving). Traffic accesses the right-turn bay at a midblock signalized intersection on the approach where continuous flow is desired. In Figure 1, a design concept of the CFI is illustrated in a north-south direction for left-hand driving commonly applied in England, Australia and some east-asia countries. Several forms of CFI exist and depend on the number of arms at the intersection and the number and locations of crossovers incorporated into the design. In 1994, a CFI was documented by Goldbatt et al. The study revealed that the mean speed of the CFI was doubled along with an increase in signal operation efficiency of at least 80 percent, as compared to the conventional one. 3 Successive studies also confirmed CFI advantages: It could be a very cost-effective solution for a high traffic volume intersection that needs to be converted to a grade-separated interchange. 4 Another CFI case study in Maryland measured its effectiveness and found that CFIs could reduce intersection delay by more than 60 percent during a peak afternoon traffic condition. 5 ITE Journal on the web / May 2009 69
Although a full evaluation study has not yet been conducted to confirm the benefits of CFIs and their effects on operations and safety, they have gained more and more credibility. They recently were acknowledged by the Federal Highway Administration and have been implemented in a few locations in the United States. 6 PROPOSED WORK To achieve travel time savings benefit estimation, travel times were measured in terms of an average total intersection delay when traffic is passing through each of the conventional intersections and CFI. Figure 2 summarizes the proposed methodology, which is based on a micro-simulation modeling, using the VISSIM micro-simulation software package for both the conventional intersection and the CFI analyses. 7 Figure 1. CFI design concept in a north-south direction (considering left-hand-side driving). Figure 2. Methodology diagram. Micro-Simulation Modeling A scenario of conventional intersection and CFI geometric components was considered; the geometric design and number of lanes of the conventional intersection and CFI under study are shown in Figure 3. All the layouts and design elements were drawn using AutoCAD software, then added as a base model in the VIS- SIM software. Note that urban driving characteristics in Bangkok, Thailand, are adopted in this study, with a traffic composition of 90 percent passenger vehicles and 10 percent heavy vehicles. Based on the actual site observation, this traffic composition does represent the overall speed of motorcycles, Tuk Tuk and other types of vehicles in terms of the volume of passenger car and heavy vehicle units. However, because footbridges are typically implemented in Bangkok, pedestrians are not interfering with the CFI traffic and therefore are excluded from the model. Other crucial simulation modeling parameters include turning movement distributions, traffic volume, desired speed distribution, reduced speed areas, routing decisions and signal control coding. More specifically, a balanced turning movement distribution with 20 percent left-turn, 60 percent through and 20 percent right-turn vehicles was assumed. Traffic volumes were set in a range between 500 and 3,500 vehicles per peak hour per intersection approach, with 70 ITE Journal on the web / May 2009
an increment of 500 vehicles for each scenario. Off-peak-hour volumes were considered at approximately one-fifth of the corresponding peak hourly volumes. In addition, for an unbiased evaluation, optimal signal phasing and timing were determined by Synchro software, with fixed-time signal control operations for both types of intersection. 8 Model Calibration To ensure that the simulation model was as closely mimicking the actual conditions as possible, a model calibration was essential. All simulation parameters to be calibrated fell into two main categories: 1) global parameters that affect the performance of the entire model, such as vehicle characteristics (length, width, desired speed, maximum acceleration/deceleration and minimum headway); and 2) local parameters that only influence specific portions of the roadway, such as approaching intersection speeds and lane changing areas before entering the intersection. In general, the process starts with the calibrations of the global followed by the local parameters. Because the approaching intersection speed is perhaps the most sensitive parameter, the field data collection must be performed. If the approaching speed parameter does not reflect the actual condition, the traffic flow measurements may be inaccurate. In addition, the field location was selected accordingly to adhere with the VISSIM model; the intersection is signalized, four-legged, with three through, one left-turn and two right-turn lanes, as illustrated in Figure 3. Vehicle speed data are collected when the traffic condition is in free-flow or close to free-flow condition. Specifically, the collected speed information includes approaching speed, turning speed and lane crossing speed (for the CFI calibration). The lane changing area is set at a distance from the intersection to imitate a commuter s condition, where lane changes are made relatively early. Further manual adjustment based on experiences and observations were fine-tuned correspondingly. Figure 3. Geometric design of conventional intersection and CFI for left-hand-side driving. Table 1. Comparison of average intersection delay of conventional intersection and CFI at different traffic demands and annual benefits. Measures of Effectiveness VISSIM micro-simulation modeling offers various output options for the measures of effectiveness (MOE), including a number of vehicles, travel time, total average delay, vehicle stopped delay and vehicle queue length. In this study, travel time and average intersection delay were selected as the primary MOEs, with a range of 2,000 to 14,000 total intersection entering vehicles. However, vehicle queue length was also measured during a simulation run to see if the queue interfered with advance crossing maneuvers. In addition, standard deviation (STD) and coefficient of variance (CV) values were also calculated to ensure an accurate delay obtained from the VISSIM microsimulation models. More specifically, the standard deviation measures how widely spread the results are about their mean value; the smaller the standard deviation, the more representative the results. Correspondingly, the coefficient of variance is the ratio of the standard deviation to the mean, as it measures the precision of the data collection or results; the smaller the CV value (closer to zero), the higher the precision and the lower the variance within the data. ITE Journal on the web / May 2009 71
Figure 4. Average intersection delay of conventional intersection and CFI during peak hour (left) and off-peak hour (right) at various traffic demands. RESULTS AND ITS MONETARY BENEFITS Based on the VISSIM output results, average intersection delays for both CFI and conventional intersections of the traffic condition are shown in Table 1 during peak and off-peak hours along with their STDs and CV values that evidently are very low. In Figure 4, average total intersection delay plots are presented for conventional intersection and CFI at various peak and offpeak hour demands. Note that the delays are plotted for the actual intersection entering volumes. Because this conventional intersection reaches its capacity at around 8,000 vehicles per hour, the intersection delays of the demands over this amount will be limited to the intersection-entering capacity, as shown in Figure 4 (left) and Table 1 (Actual Entering columns). Overall travel time savings of the CFI compared with the conventional intersection during the peak-hour condition range from 35 to 75 percent and from 33 to 52 percent during the off-peak hour condition. This finding considers a range of traffic demands entering the intersection during the peak and off-peak hours of 2,000 to 14,000 vehicles per hour and 440 to 3,080 vehicles per hour, respectively. Generally, the low average vehicle queue length measured during simulation runs demonstrates that it does not interfere with the CFI s advance crossing at peak-hour intersection-entering traffic demands. However, at very high demands (12,000 14,000 vehicles), congestion may occur, causing some interference. Figure 5. Annual benefits estimation results from CFI s travel time savings compared with the conventional intersection. From this finding, the design of advance crossing must take traffic demands and the vehicle queue length into account to ensure uninterrupted operation. After travel time savings are obtained, they can be converted to annual monetary benefits (AMB) in Thai Baht per year according to Equation 1, assuming 243 weekdays and 122 holidays/weekend days per year, a vehicle occupancy rate of one person per vehicle, an exchange rate of US$1 = 34 Thai Baht and average values of time for Thai people in Bangkok of 75 Baht per hour on a weekday and 18.75 Baht per hour on a weekend. 9 AMB = (MB wkd 243) + (Mb wke 122) (1) where: MB wkd = monetary benefits of travel time savings during one typical weekday (Baht per day) MB wke = monetary benefits of travel time savings during one typical weekend day (Baht per day) According to the outputs from VISSIM micro-simulation modeling software, av- 72 ITE Journal on the web / May 2009
erage intersection delays of each scenario and some other essential parameters were used to calculate the annual benefits. Annual benefits from the CFI s travel time savings compared with the conventional intersection are illustrated in Figure 5. The benefits range from 0.7 million to 86 million Baht, depending on different entering intersection traffic demands. It is apparent that the annual benefits increase exponentially with the intersection entering volumes; higher benefit gains are achieved at higher traffic demands. IS A CFI WORTH IMPLEMENTING? The overall travel time savings of the CFI compared with the conventional intersection range from 35 to 75 percent during the peak hour and from 33 to 52 percent during the off-peak hour. A range of traffic demand entering the intersection from 2,000 to 14,000 vehicles per hour and 440 to 3,080 vehicles per hour during the peak hour and off-peak hours, respectively, was considered in this simulation study. However, this information does not really indicate whether it is worth implementing the CFI. Therefore, an annual monetary benefit was estimated from the total travel time savings from the CFI application. The annual benefits from the travel time savings from the CFI as compared with the conventional intersection ranged from 0.7 million to 86 million Baht at the same traffic demands. From the results, the CFI application contributes a significant travel-time savings when traffic at the conventional intersection is reaching its capacity. Still, is the CFI worth implementing? Based on these findings, CFI implementation is always beneficial, but the monetary benefits could be very low. Specifically, it appears not worth implementing the CFI for low entering traffic demands up to 6,000 vehicles per hour. On the other hand, the monetary benefits appear to be fairly high and worth implementing at higher entering traffic demands, between 8,000 and 14,000 vehicles per hour, considering the implementation cost of two to three times the conventional cost. 10 CONCLUSIONS This research study re-emphasizes the benefits of the CFI over the conventional intersection by measuring the travel time savings and monetary benefits through VISSIM simulation modeling. The CFI can be a very cost-effective solution for traffic congestion reduction at major intersections for certain traffic conditions. Specifically, it is worth implementing the CFI under high traffic demand, but not at lower demand. Nonetheless, the VIS- SIM micro-simulation results shows that the CFI itself does have limitations in accommodating the intersection entering traffic as well (14,000 vehicles per hour in this study), where the CFI is eventually congested causing possible interference of vehicle queue with the CFI s advance crossing, and other more effective solutions must be applied. n References 1. Jaganathan, R. and J.G. Bared. Design and Operation Performance of Crossover Displaced Left Turn Intersections. Transportation Research Record, No. 1881 (2004): 1 10. 2. Simmonite, B.F. and M.J. Chick. Development of the Displaced Right-Turn Intersection. Transportation Research Record, No. 1881 (2004): 11 18 3. Goldbatt, R., F. Meir and J. Friedman. Continuous Flow Intersections. ITE Journal, Vol. 64, No. 7 (July 1994): 35 42. 4. Ibid. 5. Pitaksringkarn, J. Measures of Effectiveness for Continuous Flow Intersection: A Maryland Intersection Case Study. ITE Compendium of Technical Papers, No. 383 (2005). 6. Bared, J.G. Signalized Intersections: Informational Guide. Washington, DC, USA: Federal Highway Administration, 2005. 7. PTV Planung Transport Verkehr AG. VIS- SIM Version 3.6 Manual Guideline. Karlshure, Germany, 2001. 8. Synchro User Reference Guild Version 6. Trafficware, Berkeley, CA, USA, 2000. 9. Study Group on Road Investment Evaluation. Guidelines for the Evaluation of Road Investment Projects. Japan Research Institute, Tokyo, Japan, 2000. 10. Goldbatt, Meir and Friedman, 1994. JUMRUS PITAKSRINGKARN, P.E., is a principal engineer with Systra MVA (Thailand) Ltd. and is completing his doctorate of engineering in transportation engineering at the Asian Institute of Technology, Thailand. He received a master of science in civil engineering from the University of Massachusetts, Lowell. He has extensive experience in traffic engineering and transportation planning, including micro-simulation modeling. He is a member of ITE. SHINYA HANAOKA is an associate professor in the Department of International Development Engineering at the Tokyo Institute of Technology. He received a doctorate in information sciences, a master of information sciences and a bachelor of engineering from Tohoku University, Sendai, Japan. His research interests include sustainable transport, air transport and transport logistics. ITE Journal on the web / May 2009 73