8 Highway Capacity PDF

Source: HANDBOOK OF TRANSPORTATION ENGINEERING
CHAPTER 8
HIGHWAY CAPACITY
Lily Elefteriadou
Department of Civil Engineering, The Pennsylvania State University, University Park, Pennsylvania
8.1
INTRODUCTION
How much trafc can a facility carry? This is one of the fundamental questions designers and trafc engineers have been asking since highways have been constructed. The term capacity has been used to quantify the trafc-carrying ability of transportation facilities. The value of capacity is used when designing or rehabilitating highway facilities, to obtain design elements such as the required number of lanes. It is also used in evaluating whether an existing facility can handle the trafc demand expected in the future. The denition and value for highway capacity have evolved over time. The Highway Capacity Manual (HCM 2000) is the publication most often used to estimate capacity. The current version of the HCM denes the capacity of a facility as the maximum hourly rate at which persons or vehicles reasonably can be expected to traverse a point or a uniform section of a lane or roadway during a given time period, under prevailing roadway, trafc and control conditions (HCM 2000, 2-2). Specic values for capacity are given for various types of facilities. For example, for freeway facilities capacity values are given as 2,250 passenger cars per hour per lane (pc / hr / ln) for freeways with free-ow speeds of 55 miles per hour (mph), and 2,400 pc / hr / ln when the free-ow speed is 75 mph (ideal geometric and trafc conditions). For a long time, trafc engineers have recognized the inadequacy and impracticality of the capacity denition. First, the expression maximum . . . that can reasonably be expected is not specic enough for obtaining an estimate of capacity from eld data. Secondly, eld data-collection efforts have shown that the maximum ow at a given facility varies from day to day, therefore a single value of capacity does not reect real-world observations. The main objective of this chapter is to provide transportation professionals with an understanding of highway capacity and the factors that affect it, and to provide guidance on obtaining and using eld values of capacity. The next part of this chapter discusses the history and evolution of capacity estimation. The third part discusses the factors that affect capacity. The fourth part presents uninterrupted ow capacity issues, while the fth part presents interrupted ow capacity issues. The last part presents a vision for the future of dening and estimating capacity.
Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright 2004 The McGraw-Hill Companies. All rights reserved. Any use is subject to the Terms of Use as given at the website.
8.1
HIGHWAY CAPACITY 8.2

CHAPTER EIGHT
8.2
8.2.1
CAPACITY DEFINITION AND ESTIMATION METHODS

The Highway Capacity ManualA Historical Perspective
The Highway Capacity Manual is the publication most often used to estimate capacity. The rst edition of the Highway Capacity Manual (1950) dened three levels of roadway capacity: basic capacity, possible capacity and practical capacity. Basic capacity was dened as the maximum number of passenger cars that can pass a point on a lane or roadway during one hour under the most nearly ideal roadway and trafc conditions which can possibly be attained. Possible capacity was the maximum number of vehicles that can pass a given point on a lane or roadway during one hour, under prevailing roadway and trafc conditions. Practical capacity was a lower volume chosen without the trafc density being so great as to cause unreasonable delay, hazard, or restriction to the drivers freedom to maneuver under prevailing roadway and trafc conditions. The second edition of the Highway Capacity Manual (1965) dened a single capacity, similarly to the possible capacity of the HCM 1950. Basic capacity was replaced by capacity under ideal conditions, while practical capacity was replaced by a series of service volumes to represent trafc operations at various levels of service. It is interesting to note that in the HCM 1965, the second chapter is titled Denitions and begins as follows: The confusion that has existed regarding the meaning and shades of meaning of many terms . . . has contributed . . . to the wide differences of opinion regarding the capacity of various highway facilities. . . . In fact, the term which is perhaps the most widely misunderstood and improperly used . . . is the word capacity itself. Thus, the denition of the term capacity allowed for various interpretations by different trafc analysts, and there was a desire to clarify the term. In the HCM 1965, the denition of capacity was revised to read as follows: Capacity is the maximum number of vehicles which has a reasonable expectation of passing over a given section of a lane or a roadway in one direction (or in both directions for a two-lane or three-lane highway) during a given time period under prevailing roadway and trafc conditions. This denition includes the term reasonable expectation, which indicates that there is variability in the numerical value of the maximum number of vehicles. Subsequent editions and updates of the HCM (1985, 1994, and 1997) dene capacity in a similar manner, with the most recent denition (HCM 2000) as stated in the introduction of this chapter. This most recent denition indicates there is an expected variability in the maximum volumes, but it does not specify when, where, and how capacity should be measured, nor does it discuss the expected distribution, mean, and variance of capacity. Capacity values provided in the HCM have increased over time. For example, the HCM 1950 indicated that the capacity of a basic freeway segment lane is 2,000 pc / hr / ln, while the HCM 2000 indicates that capacity may reach 2,400 pc / hr / ln for certain freeway facilities. In addition to the denition of capacity, the HCM has historically provided (beginning with the HCM 1965) relationships between the primary trafc characteristics (speed, ow, and density) which have been the basis of highway capacity analysis procedures, particularly for uninterrupted ow facilities. Figure 8.1 presents a series of speed-ow curves that are provided in the HCM 2000 and illustrate the relationship between speed and ow for basic freeway segments, and for various free-ow speeds (FFS), ranging from 55 to 75 mph. As shown in Figure 8.1, speed remains constant for low ows and begins to decrease as ow reaches 1,3001,750 pc / hr / ln. The capacity for facilities with FFS at or above 70 mph is 2,400 pc / hr / ln and decreases with decreasing FFS. For example, the capacity of a basic freeway segment with FFS 55 mph is expected to be 2,250 pc / hr / ln. Figure 8.2 provides the respective ow-density curves for basic freeway segments. Similarly to Figure 8.2, capacity values are shown to vary for varying free-ow speeds. This gure clearly illustrates the assumption used in the development of these curves that capacity is reached when density is 45 passenger cars per mile per lane (pc / hr / ln).
HIGHWAY CAPACITY
HIGHWAY CAPACITY
8.3
FIGURE 8.1 Speed-ow curves. (Source: HCM 2000, Exhibit 13-2.)
Both gures provide speed-ow-density relationships for undersaturated (i.e., noncongested) ow only. When demand exceeds the capacity of the facility, the facility will become oversaturated, with queues forming upstream of the bottleneck location. The HCM 2000 does not provide speed-ow-density relationships for oversaturated conditions at freeways, because research has not been conclusive on this topic. In summary, the denition of capacity within the HCM has evolved over time. There has been an implicit or, more recently, explicit effort to include the expected variability of maximum volumes in the capacity denition; however, there is no specic information in that document on where, when, and how capacity should be measured at a highway facility.
FIGURE 8.2 Flow-density curves. (Source: HCM 2000, Exhibit 13-3.)

CHAPTER EIGHT
8.2.2
Other Publications
For a long time, researchers have recognized the inadequacy and impracticality of this definition for freeway facilities. Field data collection of capacity estimates has shown that there is wide variability in the numerical values of capacity at a given site. This section summarizes the most recent literature ndings regarding capacity denition and estimation. Persaud and Hurdle (1991) discuss various denitions and measurement issues for capacity, including maximum ow denitions, mean ow denitions, and expected maximum ow denitions. They collected data at a three-lane freeway site over three days. In concluding, they recommend that the mean queue discharge ow is the most appropriate, partly due to the consistency the researchers observed in its day-to-day measurement. Agyemang-Duah and Hall (1991) collected data over 52 days on peak periods to investigate the possibility of a drop in capacity as a queue forms, and to recommend a numerical value for capacity. They plotted prequeue peak ows and queue discharge ows in 15-minute intervals, which showed that the two distributions are similar, with the rst one slightly more skewed toward higher ows. They recommended 2,300 pc / hr / ln as the capacity under stable ow and 2,200 pc / hr / ln for postbreakdown conditions, which corresponded to the mean value of the 15-minute maximum ows observed under the two conditions. The researchers recognized the difculty in dening and measuring capacity, given the variability observed. Wemple et al. (1991) also collected near-capacity data at a freeway site and discuss various aspects of trafc ow characteristics. High ows (above 2,000 vehicles per hour per lane, vphpl) were identied, plotted, and tted to a normal distribution, with a mean of 2315 vehicles per hour (vph) and a standard deviation of 66 vph. Elefteriadou, Roess, and McShane (1995) developed a model for describing the process of breakdown at ramp-freeway junctions. Observation of eld data showed that, at ramp merge junctions, breakdown may occur at ows lower than the maximum observed, or capacity ows. Furthermore, it was observed that, at the same site and for the same ramp and freeway ows, breakdown may or may not occur. The authors developed a probabilistic model for describing the process of breakdown at ramp-freeway junctions, which gives the probability that breakdown will occur at given ramp and freeway ows, and is based on ramp-vehicle cluster occurrence. Similarly to this research, Evans, Elefteriadou, and Natarajan (2001) also developed a model for predicting the probability of breakdown at ramp freeway junctions, which was based on Markov chains, and considered operations on the entire freeway cross-section, rather than the merge inuence area. Minderhoud et al. (1997) discuss and compare empirical capacity estimation methods for uninterrupted ow facilities and recommend the product limit method because of its sound theoretical framework. In this method, noncongested ow data are used to estimate the capacity distribution. The product-limit estimation method is based on the idea that each noncongested ow observation having a higher ow rate than the lowest observed capacity ow rate contributes to the capacity estimate, since this observation gives additional information about the location of the capacity value. The paper does not discuss transitions to congested ow, nor discharge ow measurements. Lorenz and Elefteriadou (2001) conducted an extensive analysis of speed and ow data collected at two freeway-bottleneck locations in Toronto, Canada, to investigate whether the probabilistic models previously developed replicated reality. At each of the two sites, the freeway breakdown process was examined in detail for over 40 breakdown events occurring during the course of nearly 20 days. Examining the time-series speed plots for these two sites, the authors concluded that a speed boundary or threshold at approximately 90 km / hr existed between the noncongested and congested regions. When the freeway operated in a noncongested state, average speeds across all lanes generally remained above the 90 km / hr threshold at all times. Conversely, during congested conditions average speeds rarely exceeded 90 km / hr, and even then they were not maintained for any substantial length of time. This 90 km / hr threshold was observed to exist at both study sites and in all of the daily data samples evaluated as part of that research. Therefore, the 90 km / hr threshold was
HIGHWAY CAPACITY
HIGHWAY CAPACITY
8.5
applied in the denition of breakdown for these sites. Since the trafc stream was observed to recover from small disturbances in most cases, only those disturbances that caused the average speed over all lanes to drop below 90 km / hr for a period of 5 minutes or more (15 consecutive 20-second intervals) were considered a true breakdown. The same criterion was used for recovery periods. The authors recorded the frequency of breakdown events at various demand levels. As expected, the probability of breakdown increases with increasing ow rate. Breakdown, however, may occur at a wide range of demands (i.e., 1,5002,300 vphpl). The authors conrmed that the existing freeway capacity denition does not accurately address the transition from stable to unstable ow, nor the trafc-carrying ability of freeways under various conditions. Freeway capacity may be more adequately described by incorporating a probability of breakdown component in the denition. A suggested denition reads: [T]he rate of ow (expressed in pc / hr / ln and specied for a particular time interval) along a uniform freeway segment corresponding to the expected probability of breakdown deemed acceptable under prevailing trafc and roadway conditions in a specied direction. The value of the probability component should correspond to the maximum breakdown risk deemed acceptable for a particular time period. A target value for the acceptable probability of breakdown (or acceptable breakdown risk) for a freeway might initially be selected by the facilitys design team and later revised by the operating agency or jurisdiction based on actual operating characteristics. With respect to the two-capacity phenomenon, the researchers observed that the magnitude of any ow drop following breakdown may be contingent upon the particular ow rate at which the facility breaks down. Flow rates may remain constant or even increase following breakdown. This may explain the fact that some researchers have observed the two-capacity phenomenon and others have not: it seems to depend on the specic combination of the breakdown ow and the queue discharge ow for the particular observation period. The paper does not discuss maximum prebreakdown ow, however, nor does it directly compare breakdown ows to maximum discharge ows for each observation day. Elefteriadou and Lertworawanich (2002) examined freeway trafc data at two sites over a period of several days, focusing on transitions from noncongested to congested state, and developed suggested denitions for these terms. Three ow parameters were dened and examined at two freeway bottleneck sites: the breakdown ow, the maximum pre-breakdown ow, and the maximum discharge ow. Figure 8.3 illustrates these three values in a time series of ow-speed data at a given site. It was concluded that:
The numerical value of each of these three parameters varies and their range is relatively
large, in the order of several hundred vphpl. The distributions of these parameters follow the normal distribution for both sites and both analysis intervals. The numerical value of breakdown ows is almost always lower than both the maximum prebreakdown ow and the maximum discharge ow. The maximum prebreakdown ow tends to be higher than the maximum discharge ow in one site, but the opposite is observed at the other site. A possible explanation for this difference may be that geometric characteristics and sight distance may result in different operations under high- and low-speed conditions. In summary, several studies have shown that there is variability in the maximum sustained ows observed, in the range of several hundred vphpl. Three different ow parameters have been dened (maximum prebreakdown ow, breakdown ow, and maximum discharge ow), any of which could be used to dene capacity for a highway facility. The maximum values for each of these are random variables, possibly normally distributed. Prebreakdown ow is often higher than the discharge ow. The transition from noncongested to congested ow is

CHAPTER EIGHT
FIGURE 8.3 Illustration of three parameters on time-series plot of ow and speed.
probabilistic and may occur at various ow levels. The remainder of the chapter refers to these three as a group as capacity, with references to a specic one when appropriate.
8.3
FUNDAMENTAL CHARACTERISTICS OF TRAFFIC FLOW

To understand the causes of variability in capacity observations, let us rst review the fundamental characteristics of trafc ow. Figure 8.4 provides a time-space diagram with the trajectories of a platoon of ve vehicles traveling along a freeway lane. The vertical axis shows the spacing (hs, in ft) between each vehicle, while the horizontal axis shows their time headways (ht, in sec). As shown, the spacing is different between each pair of vehicles if measured at different times (time 1 versus time 2). Similarly, the time headway between each pair of vehicles varies as they travel down the freeway (location 1 versus location 2). Flow can be expressed as: Flow 3600 / Average (ht) (8.1)
When Average (ht) is minimized, the ow is maximized (i.e., capacity is reached). Therefore,
HIGHWAY CAPACITY
HIGHWAY CAPACITY
8.7
D ista n c e (ft)
Ve
hi
c le
2 hi c le hi Ve
c le hi c le Ve hi c le 5 Ve 4
hs 1
hs 2
ht 1
FIGURE 8.4 Vehicle platoon trajectories.
Ve
ht 2
T im e (se c)
the distribution and values of ht greatly affect the observed capacity of a facility. In Figure 8.4, the speed of each vehicle can be graphically obtained as the distance traveled divided by the respective time, or: Speed Distance / Time hs / ht ht hs / Speed (8.2)
Throughput is maximized when ht is minimized, or as spacing decreases and speed increases. Therefore, spacing and speeds also have an impact on the maximum throughput observed. In summary, the microscopic characteristics of trafc, i.e., the individual spacing, time headway, and speed of each vehicle in the trafc stream and their variability, result in variability in the eld capacity observations. The remainder of this section discusses factors that affect these microscopic characteristics of trafc, and thus capacity.
8.4
THE THREE COMPONENTS OF THE TRAFFIC SYSTEM AND THEIR EFFECTS ON CAPACITY
The three components of the trafc system are the vehicle, the driver, and the highway environment. The vehicle characteristics, the driver characteristics, and the roadway infrastructure, as well as the manner in which the three components interact, affect the trafc operational quality and capacity of a highway facility. This section describes each of these three components and their specic aspects that affect capacity, along with their characteristics and interactions.
8.4.1
Vehicle Characteristics
As discussed above, highway capacity is a function of the speed, time headway, and spacing, which in turn are affected by the performance and size of the vehicles in the trafc stream.

CHAPTER EIGHT
The variability of these characteristics contributes in the variability in capacity observations. Vehicle characteristics that affect capacity include:
Wt / Hp (Weight-to-horsepower ratio): The Wt / Hp provides a measure of the vehicle load
to the engine power of the vehicle. It affects the maximum speed a vehicle can attain on steep upgrades (crawl speed), as well as its acceleration capabilities, both of which have an impact on capacity. Heavier and less powerful trucks generally operate at lower acceleration rates, particularly at steep upgrades. Slower-moving vehicles are particularly detrimental to capacity and trafc operational quality when there are minimal passing opportunities for other vehicles in the trafc stream. Braking and deceleration capabilities: The deceleration capability of a vehicle decreases with increasing size and weight. Frontal area cross-section: The aerodynamic drag affects the acceleration of the vehicle. Width, length, and trailer-coupling: The width of a vehicle may affect trafc operations at adjacent lanes by forcing other vehicles to slow down when passing. In addition, the width, length, and trailer coupling affects the off-tracking characteristics of a vehicle and the required lane widths, particularly along horizontal curves. The encroachment of heavy vehicles on adjacent lanes affects their usability by other vehicles and thus has an impact on capacity. Vehicle height: The vehicle height, even though not typically included in capacity analysis procedures, may affect the sight distance for following vehicles and thus may affect the resultant spacing and time-headways, and ultimately the capacity of a highway facility.
8.4.2
Driver Characteristics
Individual driver capabilities, personal preferences, and experience also affect highway capacity and contribute to the observed capacity variability. The driver characteristics that affect the capacity of a facility are:
Perception and reaction times: These affect the car-following characteristics within the
trafc stream. For example, these would affect the acceleration and deceleration patterns (and the trajectory) of a vehicle following another vehicle in a platoon. They also affect other driver actions such as lane changing and gap acceptance characteristics. Selection of desired speeds: The maximum speed at which each driver is comfortable driving at a given facility would affect the operation of the entire trafc stream. The effect of slower-moving vehicles in the trafc stream would be detrimental to capacity, particularly when high trafc demands are present. Familiarity with the facility: Commuter trafc is typically more efcient in using a facility than are drivers unfamiliar with the facility, or recreational drivers.
8.4.3
Roadway Design and Environment
The elements included under this category include horizontal and vertical alignment, crosssection, and trafc control devices.
Horizontal alignment and horizontal curves: Vehicles typically decelerate when negotiating
sharp horizontal curves. In modeling speeds for two-lane highways (Fitzpatrick et al. 1999), it has been shown that drivers decelerate at a rate that is proportional to the radius of the curve.
HIGHWAY CAPACITY
HIGHWAY CAPACITY
8.9
Vertical alignment and vertical curves: Steep grades result in lower speeds, particularly
for heavy trucks with low performance characteristics. Crawl speeds can be determined as a function of grade. Steep vertical crest curves would also affect sight distances and may act as local bottlenecks. Cross-section: The number and width of lanes, as well as the shoulder width, have been shown to affect speeds and thus the capacity of a highway facility. Provision of appropriate superelevation increases the speed and thus enhances the efciency of a highway facility. Trafc-control devices: The clarity and appropriateness of trafc-control devices enhance the capacity of highway facilities. Interactions between the three factors are also very important. For example, the effect of a steep upgrade on a heavy vehicles performance is much more detrimental for capacity than generally level terrain. Also, the effect of challenging alignment would be much more detrimental to an unfamiliar driver than to a commuter.
8.5
CAPACITY OF UNINTERRUPTED FLOW FACILITIES

Uninterrupted ow facilities are dened as those where trafc is not interrupted by trafc signals or signs. These include freeway segments, weaving segments, ramp junctions, multilane highways, and two-lane highways. This section provides procedures for obtaining maximum throughput (i.e., capacity) estimates along uninterrupted ow facilities in eld data collection, using microsimulation models. It provides guidance on observing and measuring maximum prebreakdown throughput, breakdown ow, and maximum discharge ow. The last part of the section discusses capacity estimation for uninterrupted ow facilities using the HCM 2000.
8.5.1
Field Data Collection
The four important elements that should be considered when observing breakdown and maximum throughput are site selection and measurement location, denition of breakdown, time interval, and sample size. Regarding site selection, the site should be regularly experiencing congestion and breakdown as a result of high demands and not as a result of a downstream bottleneck. For example, in Figure 8.5, which provides a sketch of a freeway facility with two consecutive
FIGURE 8.5 Freeway facility with two consecutive bottlenecks.

CHAPTER EIGHT
bottlenecks (one merge and one lane drop), the bottleneck at location B will result in queue backup into location A. The downstream bottleneck location (i.e., location B) should be the data-collection point. To illustrate graphically the importance of site and location selection, Figure 8.6 provides the speed-ow relationships and the time-series of speed and ow for locations A and B across the freeway facility shown in Figure 8.5. Figure 8.6 illustrates the maximum throughput at location B is equivalent to its potential capacity, i.e., two-lane segment capacity. At location A the potential capacity is that of a three-lane segment, which, however, cannot be attained due to the downstream bottleneck. As soon as location B reaches capacity and breaks down, the queue created spills back into location A, which also becomes oversaturated (for additional discussion see May 1990). The speed time-series at location B shows the breakdown occurrence, which typically occurs with a relatively steep speed drop. At location A speed drops gradually as a result of the downstream breakdown. The second important element when measuring maximum throughput is the denition and identication of breakdown. In a previous study Lorenz and Elefteriadou (2001) identied and dened breakdown at freeway merge areas as a speed drop below 90 km / hr, with duration of at least 15 minutes. Another study on weaving areas (Lertworawanich and Elefteriadou 2001) showed that the breakdown speed threshold exists at these sites as well, but has a different value (80 km / hr). Given that speed drop and its respective duration can uniquely identify the presence of breakdown, it is recommended that breakdown be quantitatively dened using these two parameters. As illustrated in Figure 8.6, at the bottleneck (location B, part (b)) the speed drop is typically sharp, and the breakdown can be clearly identied. The third element that is important in clearly dening maximum throughput is the selection of an appropriate time interval. Time intervals that are typical in trafc operational analysis range between 5 and 15 minutes. Previous research has demonstrated that maximum throughput increases for smaller intervals, due to the general ow variability. Longer intervals result in averaging the ow over a longer time period, and thus the respective maximum throughput is lower. The last element to be considered is the required sample size. Given the inherent variability of maximum throughput, it is important to observe an adequate number of breakdown
P o ten tial C ap a city (3 la n es )
C ap ac ity (2 la n es )
(a)
(a)
F lo w
(b )
F lo w
(b )
O v e rsa tu ra te d C o n d itio n s
S p e ed
S p e ed
T im e
T im e
FIGURE 8.6 Speed-ow curves and time-series for freeway locations A and B.
B rea k d o w n
C ap ac ity (2 la n es )
L o c atio n A
L o c atio n B
S p e ed
S p e ed
HIGHWAY CAPACITY
HIGHWAY CAPACITY
8.11
events and the respective maximum prebreakdown and maximum discharge ows. Sample size-determination equations should be used to establish the required number of observations for the desired precision in the maximum throughput estimate. Once these four elements are established, ow and speed data can be collected at the selected site(s), and time-series plots prepared (such as the one depicted in Figure 8.3) for each breakdown event. The breakdown can be identied based on the denition selected, and the respective breakdown ow can be obtained from the time series. Next, the maximum prebreakdown ows and the maximum discharge ows can be obtained for each breakdown event.
8.5.2
Additional Considerations for Obtaining Capacity from Microsimulation Models
The elements and procedure outlined above for eld data collection can be followed when maximum throughput information is obtained through a microsimulation model. Additional considerations include simulation model selection and development of simulated demand patterns, which are discussed below. Microsimulation model selection is a very broad topic and is only dealt here very broadly with regard to capacity and breakdown observations. The model selected should have the necessary stochastic elements and the capability to simulate breakdown as a random event. Stochastic elements include the vehicle and driver capabilities described above in section 8.4. Specically, the acceleration and deceleration parameters for each vehicle, including the car-following models, should address and be calibrated for breakdown conditions. Similarly, the lane-changing algorithm of the model selected should consider and be calibrated for breakdown conditions. The second consideration when using simulation modeling for capacity estimation is what demands to use and how to vary them so that breakdown is achieved. The most common technique is to run the model with incrementally higher demands, starting at a sufciently low, below-capacity level. The analyst would need to load the network starting with relatively low demands and increasing them at constant intervals until breakdown is reached. Another, more complicated, technique is to develop random patterns of demand to simulate the demand patterns in the eld. For both techniques the increments employed at each successive demand level would be a function of the desired interval in the capacity observations. Output data can then be collected on breakdown events and maximum throughput, similarly to the eld data collection.
8.5.3
Capacity Estimates in the HCM 2000
The HCM 2000 provides capacity estimates for basic freeway segments, ramp merge segments, weaving segments, multilane highways, and two-lane highways. The HCM provides, for each segment type and set of geometric conditions, a single value of capacity. For example, for freeway facilities, capacity values are given as 2,250 passenger cars per hour per lane (pc / hr / ln) for freeways with free-ow speeds of 55 mph, up to 2,400 pc / hr / ln when the free-ow speed is 75 mph (ideal geometric and trafc conditions). These values represent average conditions at similar sites around the United States, obtained based on general trends of maximum ows observed at various freeway locations. Note that the HCM 2000 capacity denition is more closely aligned with the denition of maximum prebreakdown ow. The HCM 2000 does not dene breakdown ows and maximum discharge ows, nor does it provide estimates for these at various facility types.

CHAPTER EIGHT
8.6
THE CAPACITY OF INTERRUPTED FLOW FACILITIES

Interrupted ow facilities are those where trafc ow experiences regular interruptions due to trafc signs and signals. These include facilities such as signalized and unsignalized intersections, and roundabouts, all of which are discussed in this section.
8.6.1
Signalized Intersections
The capacity of a signalized intersection depends very much on the phasing and timing plan. Trafc ow on a signalized intersection approach is regularly interrupted to serve conicting trafc. Thus, the capacity of the approach is a function of the amount of green given to the respective movements within a given time interval. For example, if the cycle length at a signalized intersection is 90 seconds and the eastbound trafc is given 45 seconds of green, then the total amount of time that the approach is given the right-of-way within an hour is: Total green time No. of cycles per hour Green time
3600 / 90 45 1800 sec
This corresponds to 1800 / 3600 50 percent of the full hour. In addition to this time restriction of right-of-way-availability, the time headways observed at the stop-line of a signalized intersection approach follow a different pattern as the trafc signal changes from green to yellow to red and then back to green. Figure 8.7 illustrates a series of consecutive time headways (also called discharge headways when referring to queued vehicles at signalized intersection approaches) observed as vehicles depart from a single-lane approach. The horizontal axis in the gure represents time (in seconds), while the vertical dashed lines represent events (signal changes and vehicle departures). At the beginning of the green there were 10 vehicles queued at the approach. The volume during this green interval is 16 vehicles. As shown, the rst discharge headway, measured from the beginning of the green to the departure of the rst vehicle in the queue, is also the largest among these rst 10 queued vehicles. Subsequent discharge headways gradually decrease until they reach the saturation headway (sh) level for the approach, which is the minimum time headway observed under conditions of continuous queuingin this example 2 seconds. Next, saturation ow can be dened as the maximum throughput for the signalized intersection approach, if the approach were given the green for a full hour. The saturation ow for a single lane can be calculated as: Saturation flow (vehicles per hour of green) 3600 / Average (sh)
3600 / 2 1800 vehicles per hour of green per lane (vphgpl)
(8.3)
Considering that each approach does not have the green for the full hour, the maximum throughput that can be achieved at a signalized intersection approach depends on the percent of time the approach is given the green. Mathematically: Max. throughput per lane Percent eff. green Saturation flow
g / C 3600 / sh (3600 g) / (sh C)
(8.4)
where g is the duration of effective green for the approach and C is the cycle length for the intersection. The effective green is dened here as the time the approach is effectively used by this movement, or the actual green time minus the lost time experienced due to start-up and acceleration of the rst few vehicles (for discussion of lost time and its precise denition,
HIGHWAY CAPACITY
HIGHWAY CAPACITY
8.13
Eleventh vehicle departure: T=28 sec
First vehicle departure: T=4 sec
Thirteenth vehicle departure: departure T=42 sec
Twelfth vehicle departure: T=34 sec
Second vehicle departure: T=8 sec
Third vehicle departure: T=11 sec
Start of green: T=0
Fourth vehicle departure: T=13 sec Fifth vehicle departure: T=15 sec
Tenth vehicle departure: T=25 sec
Number of vehicles (this cycle): 16 Saturation headway: 2 sec (fourth to tenth vehicle) Saturation flow rate: 1,800 vphgpl For cycle length = 100 sec: Capacity of the cycle: 31.5 vehicles Capacity per hour: 1.134 vph
FIGURE 8.7 Time headways at a single-lane signalized approach.
consult the HCM 2000). In the example of Figure 8.7, the lost time is the extra time incurred by the rst 10 vehicles after subtracting the saturation headways for each of these vehicles. Thus, the effective green time is: Effective green time Actual green time Lost time
68 sec (25 sec 10 vehicles 2 sec) 63 sec
In the example of Figure 8.7, the maximum throughput (i.e., capacity), assuming that the duration of the effective green remains constant through the entire hour and that the cycle length is 100 seconds, is estimated from equation (8.4) as: Capacity (3600 63 sec) / (2 sec 100 sec) 1,134 vph An alternative method to estimate capacity within the hour is to estimate the capacity per cycle and multiply by the number of cycles within the hour. In the example of Figure 8.7, the capacity per cycle is: Capacity per cycle Eff. green / Saturation headway
63 sec / 2 sec 31.5 vehicles / cycle
The total number of cycles in the hour is:
End of green: T=68 sec

CHAPTER EIGHT
Number of cycles 3600 / 100 sec 36 cycles / hr Thus, the capacity within an hour is: Capacity Capacity per cycle Number of cycles
31.5 vehicles per cycle 36 cycles 1,134 vph
8.6.2
Two-Way Stop-Controlled (TWSC) Intersections and Roundabouts
The operation of TWSC intersections and roundabouts (yield-controlled) is different than that of signalized intersections in that drivers approaching a stop or yield sign use their own judgment to proceed through the junction through conicting trafc movements. Each driver of a stop- or yield-controlled approach must evaluate the size of gaps in the conicting trafc streams, and judge whether he / she can safely enter the intersection or roundabout. The capacity of a stop- or yield-controlled movement is a function of the following parameters:
The availability of gaps in the main (noncontrolled) trafc stream. Note that gap is dened
in the HCM 2000 as time headway; elsewhere in the literature, however, gap is dened as the time elapsing between the crossing of the lead vehicles rear bumper and the crossing of the following vehicles front bumper. In other words, the HCM 2000 gap denition (which will be used in this chapter) includes the time corresponding to the crossing of each vehicles length. The availability of gaps is a function of the arrival distribution of the main trafc stream. The gap acceptance characteristics and behavior of the drivers in the minor movements. The same gap may be accepted by some drivers and rejected by others. Also, when a driver has been waiting for a long time, he or she may accept a shorter gap, having rejected longer ones. The parameter most often used in gap acceptance is the critical gap, dened as the minimum time headway between successive major street vehicles, in which a minorstreet vehicle can make a maneuver. The follow-up time of the subject movement queued vehicles. The follow-up time is the time headway between consecutive vehicles using the same gap under conditions of continuous queuing, and it is a function of the perception / reaction time of each driver. The utilization of gaps in the main trafc stream by movements of higher priority, which results in reduced opportunities for lower-priority movements (this is not applicable for roundabouts because there is only one minor movement). The example provided below is a simplied illustration of the capacity estimation process for a stop- or yield-controlled movement. The capacity of the northbound (NB) minor through movement of Figure 8.8 will be estimated based on eld measurements at the intersection. There is only one major trafc stream at the intersection (eastbound, EB) and one minor street movement (NB). The critical gap was measured to be 4 seconds. It is assumed that any gap larger than 4 seconds will be accepted by every driver, while every gap smaller than 4 seconds will be rejected by every driver. The follow-up time was measured to be 3 seconds. Thus, for two vehicles to use a gap, it should be at least: 4 sec (first vehicle) 3 sec (second vehicle) 7 sec Conversely, the following equation can be used to estimate the maximum number of vehicles that can use a gap of size X:
HIGHWAY CAPACITY
HIGHWAY CAPACITY
8.15
Gap Major Traffic Stream (EB) Strea
STOP
m
Minor Traffic Stream (NB)
FIGURE 8.8 Capacity estimation for the NB through movement.
Number of vehicles 1 (Gap Size X Critical Gap) / Follow-up time
(8.5)
Table 8.1 summarizes the eld data and subsequent calculations for the intersection of Figure 8.8, and provides the capacity estimate. Column (1) of Table 8.1 provides the gaps measured in the eld. Column (2) indicates whether the gap is usable, i.e., whether it is larger than the critical gap. Column (3) uses equation (8.5) to provide the number of vehicles
TABLE 8.1 Capacity Estimation for a Stop-Controlled Movement
Gap size (sec) Col (1) 5 2 7 10 4 2 12 17 7 3 2 3 13 16 103 sec
Usable Gap? (Y or N) Col (2) Y N Y Y Y N Y Y Y N N N Y Y 9 usable gaps
Vehicles in NB movement that can use the gap (veh) Col (3) 1 0 2 3 1 0 3 5 2 0 0 0 4 5 26 vehicles
TOTAL

CHAPTER EIGHT
that can use each of the usable gaps. The total number of vehicles that can travel through the NB approach is the sum provided at the bottom of the column (3) (26 vehicles over 103 seconds), or (3600 / 103) 26 909 vph, which is the capacity of the movement. The HCM 2000 methodology for estimating the capacity of TWSC intersections is based on the principles outlined above, using mathematical expressions of gap distributions and probability theory for establishing the use of gaps by higher-priority movements.
8.7
SUMMARY AND CLOSING REMARKS

This chapter provided an overview of the concept of highway capacity, discussed the factors that affect it, and provided guidance on obtaining and using eld values of capacity. The denition of capacity within the HCM has evolved over time. Even in the earliest publications related to highway capacity there is a general recognition that the maximum throughput of a highway facility is random, and thus there has been an effort to include the expected variability of maximum volumes in the capacity denition. Recently, several studies have used eld data and have proven that there is variability in the maximum sustained ows observed, in the range of several hundred vphpl. The cause of this variability lies in the microscopic characteristics of trafc, i.e., the individual spacing, time headway, and speed of each vehicle in the trafc stream and their variability. This chapter also provided the fundamental principles of capacity estimation for uninterrupted and interrupted ow facilities, and provided guidance on obtaining capacity estimates in the eld and using simulation. The discussion provided in this chapter is not exhaustive and is only intended to provide the fundamental principles of capacity estimation for various highway facilities. Detailed methodologies for estimating highway capacity for various facilities are provided in the HCM 2000, as well as elsewhere in the literature.
8.8
REFERENCES
Agyemang-Duah, K., and F. L. Hall. Some Issues Regarding the Numerical Value of Highway Capacity, Highway Capacity and Level of ServiceInternational Symposium on Highway Capacity, Karlsruhe, July 1991, pp. 115. Bureau of Public Roads. 1950. Highway Capacity Manual. U.S. Department of Commerce, Bureau of Public Roads, Washington, DC. Elefteriadou, L., and P. Lertworawanich. 2002. Dening, Measuring and Estimating Freeway Capacity. Submitted to the Transportation Research Board. Elefteriadou, L., R. P. Roess, and W. R. McShane. 1995. The Probabilistic Nature of Breakdown at Freeway-Merge Junctions. Transportation Research Record 1484, 8089. Evans, J., L. Elefteriadou, and G. Natarajan. 2001. Determination of the Probability of Breakdown on a Freeway Based on Zonal Merging Probabilities. Transportation Research B 35:23754. Fitzpatrick, K., L. Elefteriadou, D. Harwood, R. Krammes, N. Irizzari, J. McFadden, K. Parma, and J. Collins. 1999. Speed Prediction for Two-Lane Rural Highways. Final Report, FHWA-99-171, 172, 173 and 174, June, www.tfhrc.gov / safety / ihsdm / pdfarea.htm. Flannery, A., L. Elefteriadou, P. Koza, and J. McFadden. 1998. Safety, Delay and Capacity of SingleLane Roundabouts in the United States. Transportation Research Record 1646, 6370. Highway Research Board. 1965. Highway Capacity Manual. Highway Research Board, Special Report 87, National Academies of Science, National Research Council Publication 1328. Lertworawanich, P., and L. Elefteriadou. 2001. Capacity Estimations for Type B Weaving Areas Using Gap Acceptance. Transportation Research Record 1776, 2434. . 2003. A Methodology for Estimating Capacity at Ramp Weaves Based on Gap Acceptance and Linear Optimization. Transportation Research B 37(5):45983.
HIGHWAY CAPACITY
HIGHWAY CAPACITY
8.17
Lorenz, M., and L. Elefteriadou. 2001. Dening Highway Capacity as a Function of the Breakdown Probability. Transportation Research Record 1776, 4351. May, A. D. 1990. Trafc Flow Fundamentals. Englewood Cliffs: Prentice-Hall. Minderhoud, M. M., H. Botma, and P. H. L. Bovy. Assessment of Roadway Capacity Estimation Methods, Transportation Research Record 1572, National Academy Prem, 1997, pp. 5967. Persaud, B. N., and V. F. Hurdle. Freeway Capacity: Denition and Measurement Issues, Highway Capacity and Level of ServiceInternational Symposium on Highway Capacity, Karlsruhe, July 1991, pp. 289308. Transportation Research Board (TRB). 2000. Highway Capacity Manual. National Research Council, TRB, Washington DC. Wemple, E. A., A. M. Morris, and A. D. May. Freeway Capacity and Flow Relationships, Highway Capacity and Level of ServiceInternational Symposium on Highway Capacity, Karlsruhe, July 1991, pp. 439456.
HIGHWAY CAPACITY

8 Highway Capacity PDF

Caricato da

Informazioni sul documento

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

8 Highway Capacity PDF

Caricato da

Copyright:

Formati disponibili

Source: HANDBOOK OF TRANSPORTATION ENGINEERING

HIGHWAY CAPACITY 8.2

CAPACITY DEFINITION AND ESTIMATION METHODS

FIGURE 8.1 Speed-ow curves. (Source: HCM 2000, Exhibit 13-2.)

FIGURE 8.2 Flow-density curves. (Source: HCM 2000, Exhibit 13-3.)

HIGHWAY CAPACITY 8.4

HIGHWAY CAPACITY 8.6

FIGURE 8.3 Illustration of three parameters on time-series plot of ow and speed.

FUNDAMENTAL CHARACTERISTICS OF TRAFFIC FLOW

HIGHWAY CAPACITY 8.8

Wt / Hp (Weight-to-horsepower ratio): The Wt / Hp provides a measure of the vehicle load

Roadway Design and Environment

CAPACITY OF UNINTERRUPTED FLOW FACILITIES

Field Data Collection

FIGURE 8.5 Freeway facility with two consecutive bottlenecks.

HIGHWAY CAPACITY 8.10

P o ten tial C ap a city (3 la n es )

Additional Considerations for Obtaining Capacity from Microsimulation Models

Capacity Estimates in the HCM 2000

HIGHWAY CAPACITY 8.12

THE CAPACITY OF INTERRUPTED FLOW FACILITIES

Eleventh vehicle departure: T=28 sec

First vehicle departure: T=4 sec

Thirteenth vehicle departure: departure T=42 sec

Twelfth vehicle departure: T=34 sec

Second vehicle departure: T=8 sec

Third vehicle departure: T=11 sec

Start of green: T=0

Tenth vehicle departure: T=25 sec

FIGURE 8.7 Time headways at a single-lane signalized approach.

The total number of cycles in the hour is:

End of green: T=68 sec

HIGHWAY CAPACITY 8.14

Two-Way Stop-Controlled (TWSC) Intersections and Roundabouts

Gap Major Traffic Stream (EB) Strea

Minor Traffic Stream (NB)

FIGURE 8.8 Capacity estimation for the NB through movement.

Number of vehicles 1 (Gap Size X Critical Gap) / Follow-up time

TABLE 8.1 Capacity Estimation for a Stop-Controlled Movement

Gap size (sec) Col (1) 5 2 7 10 4 2 12 17 7 3 2 3 13 16 103 sec

Usable Gap? (Y or N) Col (2) Y N Y Y Y N Y Y Y N N N Y Y 9 usable gaps

HIGHWAY CAPACITY 8.16

SUMMARY AND CLOSING REMARKS

Potrebbero piacerti anche