Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
SUMMARY Chapters 9 and ~ ~ of the 1994 Highway Capacity Manual (HCM) cover signalized intersections and arterial streets, respectively, but no chapter addresses three important related considerations: ~ ~ ~ the times and operational features of signalized (service) interchanges. (21 signal ~ ~ ~ ~ ~ ~ ~ ~ ~ , ~ ~ ~ .. .. . . . . . . .. ~ .. . .. . . . . . .. cooromatlon needs and progression characteristics ot the signalized ramp terminals and other closely-spaced signalizedintersections,and (3) the effects of queue spillback on the output flow of the crossing arterial and exit ramps at service interchanges. The 1994 HCM capacity analysis methodology was judged to be inadequate for accurately evaluating closely-spaced intersections during near-capacity and oversaturated traffic conditions, primarily due to its lack of sensitivity to queue spillback approaching or impeding the subject intersection. Saturation flow estimates would not change in the HCM, as it has no methodology, outside of direct field observation,to deal with spillback. Moreover, the proposed 1997 HCM delay estimation update for intersections and arter~als may over estimate its primary performance measure (delay) by a sizeable amount dunug oversaturation for closely-spacedarter~al signals, even when the downstream output saturation flow is not impeded. In addition, the HCM definition of"effective green" needs to be revisedto: (~) improve delay calculations dunng undersaturated conditions, and (2) cover periods of spilIback and oversaturation. NCHRP 3-47 was conducted to develop and validate an appropriate methodology for determining capacity arid level of service at signalized (service) interchanges, including their ramp terminals and connecting closely-spaced arterial intersections. The research found that mobility at a signalized (service) interchange is dependent on many factors that influence capacity. The importance of the prevailing factors may change depending on the general volume level and degree of existing congestion. Traditional prevailing factors affecting capacity include the interchange geometry, traffic mix, and signal green splits of the serving phases. Dunng periods of oversaturation where the storage links are filled. additionalnon-traditional factors come into play. _ , During oversaturation, the upstream input capacity becomes highly dependent on downstream signal timings and capacity to keep He output link clear. Upstream input capacity cannot exceed the total downstream service capacity available, and may be less than this capacity if demand starvation on short links occurs. In addition to phase capacity, signal offset during oversaturation is the most important factor in the allocation of downstream capacity to upstream phases. Also, this offset is not the offset that provides optimal progression during uncongested flow periods. Signal timing and coordination together with signal spacing (link length) are important variables in determining the ability of the cross arterial to move traffic. Moreover, in the mesosaturation (near capacity~range of traffic conditions,the ultimate determination of the question of "oversaturation?" depends on whether potential flow impediments due to queue spillback can be sufficiently mitigated to the extent that oversaturation can be avoided. V1
Capacity analysis methods of signalized interchanges should be able to determine whether oversaturation wall occur. However, this determination is a complex task in the mesosaturation range near capacity. These conditions not only depend on nominal volumes and phase capacities, but capacity also depends on the interactions of the variables with queue storage, link storage capacity and signal timing. Cr~tena have been provided that indicate the likelihood of queue spilIback affecting nominal output capacity. These critena can address probably SO percent of the operational problems. However, close-calIs can only be solved now using computer simulation models, like NETSIM, or possibly using models like the PDX Mode} developed in this research. Phase capacity is dependent on the prevailing saturation flow rate and effective green time - - r - ~ - r ~ - 7 ~ .. . .. . .. _% . .. . . ~ . ~ . .. . .. ~ ~ . ~ . ~ when motion can occur at the stoplme. Based on thlS research, lt was round that the distance to the downstream queue, the radius of the turn path, and traffic pressure have a significant effect on the saturation flow rate of a traffic movement. Specifically, saturation flow rate decreases when the distance to the downstream queue decreases and is relatively short. This effect is amplified when the signal timing relationship between the two intersections allows queue spillback to occur. Turn radius has a significant effect on the saturation flow rate of a turn-related traffic movement. Saturation flow rates are lower for turn movements with small radii than they are for turn movements w~thiarge radii. The saturation flow rate for the left-turn movements et single-po~nt urban interchanges are more nearly equal to those of through movements because of the large turn radii associated with this interchange type. Traffic pressure, as quantified by traffic volume per cycle per lane, has a significant elect on saturation flow rate. Traffic pressure relates to the presence of aggressive, commuter drivers in the traffic stream. Traffic volume is used as a surrogate measure of the number of these aggressive drivers in the traffic stream. Saturation flow rates of low-volume movements are much lower than those of high-volume movements because the low-volume movements have less traffic pressure. Other factors were examined for their potential effect on saturation flow rate. These factors include: g/C ratio, junction type, downstream signal indication at the start of the upstream phase, and dual versus single left-turn lane. Of these factors, only g/C ratio was found to be correlated with saturation flow rate in a statistically significant manner. Specifically, the saturation flow rate for left-turn movements with low g/C ratios was found to be higher than He rates of similar movements with larger g/C ratios. This effect was also found in the through movements studied; however, it was much smaller in magnitude and not statistically significant. Therefore it was determined that more en , ~. , ~ `~e `_, ~· , ~·, ·, ~r · · , research Is needed to venty the s~gn~cance of this trend and its magnitude before an adjustment factor for g/C effect can be recommended. The definition of effective green time should tee charged to be once time when saturation flow can occur at He stopline for existing conditions. This definition is more robust and can be used in all operating conditions, including periods of oversaturation. Moreover, delay estimates wall be improved using this new definitionofeffective green even curing undersaturated conditions because · e V11
the queue profile wall be estimated more accurately than using the current HCM methodology. The estimation of phase lost time should be improved. Start-up lost time is not a constant value; rather, it is statistically dependent on the prevailing saturation flow rate. Specifically, start-up lost time increases with increasing saturation flow rate. This paradoxical increase is due to the increased acceleration time the discharging queue requires to attain We higher speed associated with a higher saturation flow rate. The term "lost" is a bit of a misnomer in this case, as more lost time occurs as saturation flow gets better. Start-up lost times typically range from 0.61 to 3.] ~ seconds for prevailing saturation flow rates of 1,400 to 2,100 pcphgpI, respectively. The average yellow warning interval used by drivers cleanug the intersection at the end of the phase is termed "green extension, or end use " In the context of chase canacitv. end use is . , . . , . ~ , ~.. red -~~~ - I' ' equivalent to an extension or the elective green period into the yellow. The study of end use indicates that it is a relatively constant value for intersections and interchanges and that it averages about 2.5 seconds for most undersaturated conditions. Thus, this quantity can be subtracted from the signal change interval duration to estimate the lost time at the end of a signal phase. Lane use is almost always uneven (or unbalanced) in intersection lane groups. The degree of this imbalance is expressed in terms of the lane utilization factor. The lane utilization factor vanes depending on the nature of drivers' lane-choice decisions. Lane utilization factors based on travel time minimization tend to be subject to randomness in the lane-choice decision process. The factors stemming from this process range from ~ . ~ to 2.0, depending on the number of lanes in the lane group and its corresponding traffic volume. Lane utilization factors based on driver desire to preposition can vary widely, depending on the volume of traffic that is propositioning in the subject lane group. Neither the signal capacity of various interchange types nor their relative capacity per lane was specifically detennined Within this research. Some examples of this form of analysis are illustrated in a related NCHRP publication. However, examination of parclos as compared to diamonds reveals obviously different traffic volume input patterns that may result in one design being more efficient than another for a given case. The software INTERCHANGE described in Appendix F can readily examine the patterns provided by each interchange type. Parclos versus diamonds also have more right-turn capacity per input lane due to Heir normal signal overlaps, but this feature may tend to overflow downstream closely-spacedlinks more than diamonds. Moreover, single-point diamonds are known to have more arterial r~ght-turn capacity per input lane than on its exit ramps because of unbalanced right-turn signal overlaps using three-phase signal operations. The ideal saturation flow rate recommended for signalized (service) interchanges is 2,000 pcphgpl. In the context of He factors studied for this research, this ideal flow rate applies to through traffic movements that have an infinite distance to the back of downstream queues, operate under non-spilIback conditions, discharge along tangent, level alignments, and have traffic volumes that are relatively high, reflecting those fourld during peak demand periods. . . . vail
The equations providedin Chapters should tee used to estimate the saturation flow rate, start- up lost time, green extension, end lost time, and bale utilization factor. In recognition of the relationship between saturation flow rate and start-up lost time, the equations provided in Chapter 3 should be used to estimate all necessary phase capacity characteristics. The recommenced green extension value is 2.5 seconds for most undersaturated conditions. Other values are possible if the approach speed is outside the range of 64 to 76 km/in or when the volume-to-capacityratio for the analysis period is above 0.~. An equation is provided in Chapter 3 for these situations. The definition of effective green should be changed slightly from that used in the ~ 994 HCM. The new definition should be "effective green is that tune dunug the subject phase when saturation flow at the stopline can occur under prevailing conditions." All "lost" times should be removed from the phase, including: start-up, opposing queue blockage, output blockage due to spilIback, and phase clearance lost times. This definition is very robust and covers all operating conditions, movements, and phases, including protected-plus-permitted. The PDX Mode! should be considered for estimating the output Clear Period and effective green time of the subject phase when oversaturation is likely. High-volume links, which are nominally oversaturated or less than 200 meters long, should be analyzed for queue spilIback blockages using the features provided in the PDX Model. Ramp weaving speeds and crossing capacity can be estimated using the methodology presented in Chapter 3 and Appendix E. Adequate travel distance to the back of the downstream receiving queue must be available for this capacity to be attained. - _ The Highway Capacity Manual should contain a chapter on Interchanges which emphasizes the unique forms and features of interchanges together with the special challenges associated with urban interchange traffic operations in general. Two-leve] signalizedinterchanges operating within a crossing arterial system should be presented together with freeway system integration issues associated with freeway traffic management. Special design and operational issues dealing with continuous one-way frontage roads should be presented. Unsignalized rotary interchanges could be identified as an alternative design concept. Moreover, the selective application of signalized interchanges to upgrade the capacity of a major urban arterial corridor should be noted, as illustrated in Appendix F of this report. A major development of computer software should be funded in support of the Highway Capacity Manual effort. However, this software should not be limited to being just a processor to the HCM Interchange chapter for conducting only "operational analysis of existing conditions." This new software should promote "options analysis" as needed to expediently conduct operational impact analyses for preliminary planning and design activities. The new program INTERCHANGE descnbed in Appendix F illustratesrecommendedanalysis concepts and software features. However, significant funding is still needed to complete a professional-level software package for signalized service interchanges. 1X
