Click for next page ( 29

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement

Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 28
28 CHAPTER 8 BASE CASE The PSRC travel model data set was selected for the appli- Freeway HOV lanes are coded as parallel links to the free- cation of the NCHRP 25-21 methodology to case studies. way with HOV/bus-only cross connectors. For each transit The PSRC travel demand model covers four counties of line, the following data are available: USER'S GUIDE the Seattle/Tacoma metropolitan area with a population of about 3 million people. (See the University of Washington Mode, and Cambridge Systematics's "Land Use and Travel Demand Vehicle type, Forecasting Models, Model Documentation," prepared for the Headway (minutes), Puget Sound Regional Council, final report, June 30, 2001, Speed (mph), Length (miles), and The model represents the PSRC region using 852 internal Number of segments. zones, about 19,000 directional road links, and 317 transit lines. The model splits travel demand between three time peri- The projected year 2020 population is 4.3 million people, ods (3-hour AM peak, 3-hour PM peak, and rest of day) and and the projected 2020 employment is 2.3 million jobs. The three modes of travel (drive alone, carpool, and transit). An PSRC model estimated travel demand for 2020 is 12.4 mil- economic forecasting model and a pair of land-use allocation lion daily person trips in nine OD tables by mode and time models (DRAM and EMPAL) are used by PSRC to generate period (summarized in Table 16). the socioeconomic data required by the travel demand model. 8.2 APPLICATION OF THE HCM ASSIGNMENT 8.1 INPUT MODULE TO THE PSRC DATA SET The PSRC model for the year 2020 was selected as the The basic PSRC highway must be modified before the base case for demonstrating the application of the NCHRP HCM Assignment Module can be applied to it. 25-21 methodology. All of the other case studies using the NCHRP 25-21 were run in comparison to this base case for 8.2.1 Step 1: Code Free-Flow Speeds the year 2020. and Capacities Three key inputs are required from the PSRC model for application in the NCHRP 25-21 methodology: the highway Step 1 consists of substituting HCM-based capacities and network (Table 14), the transit network (Table 15), and the free-flow speeds for the planning values in the model. In the base case OD travel demand (Table 16). case of the PSRC model, the capacities and free-flow speeds The highway network contains the following data items are customized for individual links. Each facility type in the for each directional highway link, where ul1, ul2, and ul3 are PSRC model is applied to a wide range of conditions. For user-definable fields: example, ramps are sometimes coded as freeway facility types, arterial street types, or one-way arterial street types. The free- Length (in miles), flow speeds for freeway-type links consequently range from Modes (SOV, HOV, bus, rail, ferry, transit walk access, 20 mph to 70 mph. Similar ranges occur for the other facil- transit auto access), ity types. It is therefore not possible to make a blanket sub- Number of lanes, stitution of capacities and free-flow speeds based upon facil- Volume/delay function, ity type and area type. The substitutions would have to be Capacity per lane (vph) (ul1), made on a link-by-link basis. Because this basis is not practi- Free-flow travel time (minutes) (ul2), and cal for a demonstration of the methodology, the link-specific Facility type (0 = bus/walk link, 1 = freeway, 2 = express- capacities and free-flow speeds will be left unchanged. way, 3 = urban arterial, 4 = urban one way, 5 = centroid The one change made to the current PSRC method was to connector, 6 = rural arterial) (ul3). replace the current link free-flow travel times (ul2) in the AM

OCR for page 28
29 TABLE 14 Base case 2020 highway network Centerline-Miles Lane-Miles No. of Links Capacity-Miles (VMT) Mean Free-Flow Speed (mph) 11,388 17,390 17,711 20,194,252 19.9 TABLE 15 Base case 2020 transit network Network Transit Vehicles Lines Route-Miles 2020 1,286 542 9,716 and PM scenarios (which, in the PSRC model, are computed fd49 = ul2 (1 + .15 (.375 volau/(lanes ul1)) 4) from congested speed output by the daily assignment) with fd59 = ul2 (1 + .15 (.455 - .125) volau/(3 lanes the free-flow speeds from the daily assignment. ul1)) USER'S GUIDE Fd10 is used primarily in the daily assignment for all 8.2.2 Step 2: Replace BPR Equations roads. Although 179 links appear to use fd10 in the AM peak with HCM Equation assignment, the rationale for this use is unclear, so fd10 was replaced with fd59 for these 179 links. Fd10 is not used in the The BPR speed-flow equations used in the PSRC model PM assignment. are replaced with the HCM 2000 speed-flow equation. Fd30 is used for 14 auto-ferry links in both the AM and The existing PSRC volume delay functions (VDFs) for the PM assignments. These VDFs were retained unchanged. daily and off-peak scenarios were not touched. The VDFs Fd40 is used in both the AM and PM assignments for 404 involve 24-hour and 18-hour demand assignments and are nonauto ferry and walk links for the 1990 network. This only moderately capacity constrained (12-hour capacities for function is also used for 1,465 links in the AM assignment the daily assignment and 8-hour capacities for the off-peak and 873 links in the PM assignment for the 2020 network. In assignment). The off-peak assignment currently uses the con- essence, the travel time for the link is fixed at whatever value gested travel times from the daily assignment for its free-flow was originally coded by the PSRC modeler. This VDF was times. This use was unchanged. not changed. The AM and PM peak-hour assignments currently use the Fd47 is used for 10 freeway HOV lane links in the AM and following VDFs, where fd10, fd30, fd40, fd47, fd49, and fd59 PM peak assignments for the 2020 network (not present in are functions and volau is the auto volume: the 1990 network) and was not changed. Fd49 is used for 16 short connector links between the free- fd10 = ul2 (1 + .15 (.08 volau/(lanes ul1)) 4) way HOV lane links and the mixed-flow lane links of the fd30 = ul2 + (((.34 (volau/ul1) / lanes) - 1) .max. 0) freeway for the AM and PM peak 2020 network assignments (60/lanes) (not used in 1990 network). This VDF is also used for some fd40 = ul2 rural arterial links and really short urban arterial links. This fd47 = ul2 (1 + .15 (.125 volau/(lanes ul1)) 4) VDF was not changed. TABLE 16 Base case 2020 person trips Peak Mode Person Trips % Mode AM SOV 1,720,034 79.9% HOV 273,841 12.7% Transit 160,154 7.4% PM SOV 2,766,570 88.9% HOV 345,056 11.1% Transit ?* ?* Off Peak SOV 6,732,642 96.3% HOV 258,595 3.7% Transit ?* ?* Daily SOV 11,219,246 90.6% HOV 877,492 7.1% Transit 287,932 2.3% Total 12,384,670 *The PSRC model does not split transit trips into PM and off peak, but these trips are included in the estimated daily transit trips.

OCR for page 28
30 Fd59 is used for the vast majority of the road links in the (to avoid double counting the signal at the end of one link and AM and PM peak assignments. This VDF will be replaced the beginning of the next link), N is equal to 1. with the HCM speed-flow function. An integer divide is used to obtain the number of signals, since modelers usually terminate a link at a major intersection, which is likely to be signalized. So a 1.5-mile-long link with 8.2.3 Step 3: Generate Additional Network Parameters Required by HCM Equation signals assumed to be spaced an average of 1 mile apart would have one signal at the start, one signal at the end, and no sig- The HCM equation requires several additional parameters nals in between. Thus, the default signal density assumption is not coded in the PSRC network: used as a rough guide for determining whether multiple sig- nals might exist within the stretch of a model link; however, The number of signals on a link (N), if the link length is close to a multiple of the signal density, The zero-flow signal delay (D0), the coded-link length is assumed to be more accurate than the The segment delay between signal (DL), and assumed default signal density. Table 17 shows the signal density (Sd). The table was cre- USER'S GUIDE The calibration parameter (J). ated using local knowledge of typical signal densities on expressways and arterials. Number of Signals The number of signals on a link (N) is computed and stored for each link as follows: Zero-Flow Signal Delay For freeways (ul3 = 1), centroid connectors (ul3 = 5), and The zero-flow delay in hours (D0) is computed and stored rural arterials (ul3 = 6), the number of signals is zero, but for each link. The zero-flow control delay is zero for freeway, because N must be at least 1, N = 1 for these links. centroid, and rural facility types (ul3 = 1, 5, 6). For ul3 = 2, For all other facility types, N is computed as follows: 3, 4, it is computed using the equation in the methodology: N = max[1, INT ( L Sd )] ( ) 2 N C Equation 15 D0 = DF 1 - g C Equation 16 3,600 2 Where: Where: N = the number of signals on the segment, Sd = the signal density for the link (signals/mile), D0 = the zero-flow control delay at the signal (hours); L = the length of the link (miles), N = maximum of 1, or the number of signals on the max = maximum function (outputs the maximum of two segment; values), and 3,600 = conversion from seconds to hours; INT = integer divide function (outputs result truncated to g/C = average effective green time per cycle for signals integer value). on segment; C = average cycle length for all signals on the segment Note that the first signal at the start of a link is excluded (seconds); and from N, so if a link is 1 mile long and signals are spaced 1 mile DF = delay factor, apart, there will be two signals on the link (one at the start = 0.9 for uncoordinated traffic-actuated signals, and one at the end), but because the first signal is excluded = 1.0 for uncoordinated fixed-time signals, TABLE 17 Facility type, free speed, arterial class, and signal density Free Speed Arterial Class Signals/Mile Expressway Urban Arterial Ul3 = 2 Ul3 = 3,4 55+ I 1 2 50 I 1 2 45 I 1 2 40 II 1 2 35 III 3 5 30 IV 6 8 25- IV 8 8

OCR for page 28
31 = 1.2 for coordinated signals with unfavorable The delay per mile (dL) is given in Table 18, which was progression, derived from the assumed signal density and Exhibit 15-3 of = 0.9 for coordinated signals with favorable pro- the HCM 2000. gression, and = 0.6 for coordinated signals with highly favorable progression. The Calibration Parameter (J) A default value of 0.44 is used for the g/C ratio. A default The calibration parameter J is stored for each link. Table signal cycle length of 120 seconds is used. 19 was created from the table provided in the methodology using the facility types and free-flow speeds coded in the PSRC model network. Centroid connectors were given a flat Between-Signal-Segment Delay speed-flow equation taken from freeways (for 75+ mph free- flow speed). The segment delay (DL) is computed and stored as follows: Table 20 shows the final combined set of parameters for USER'S GUIDE the new HCM speed-flow equations for the PSRC model. dL DL = L Equation 17 The selection criteria are used to select the default values 60 used to compute the additional parameters for the HCM equations for each link. The standard BPR parameters (also Where: used by the HCM equations) are already coded in the PSRC L = the length of the segment. model for each link. TABLE 18 Segment delay by facility type and free-flow speed Free-Flow Speed Expressway Urban Arterial ul3 = 2 ul3 = 3, 4 55+ 0 secs 8 secs 50 0 8 45 0 8 40 0 8 35 0 20 30 25 45 25- 60 60 TABLE 19 J Parameters by facility type and free-flow speed Free-Flow Freeway Expressway Urban One-Way Rural Rural Speed Ul3 = 1 Ul3 = 2 Arterial Arterial Arterial Arterial Ul3 = 3 Ul3 = 4 Ul3 = 6 Ul3 = 6 Lanes > 1 Lanes = 1 75+ 29.47E-06 22.1E-06 204E-06 204E-06 2.296E-06 90.43E-06 70 20.03E-06 22.1E-06 204E-06 204E-06 2.296E-06 90.43E-06 65 14.23E-06 22.1E-06 204E-06 204E-06 2.296E-06 90.43E-06 60 8.426E-06 22.1E-06 204E-06 204E-06 2.296E-06 138.5E-06 55 3.306E-06 22.1E-06 204E-06 204E-06 1.821E-06 223.9E-06 50 3.306E-06 22.1E-06 204E-06 204E-06 1.108E-06 389.3E-06 45 3.306E-06 22.1E-06 204E-06 204E-06 2.174E-06 748.4E-06 40 3.306E-06 49.9E-06 200E-06 200E-06 2.174E-06 748.4E-06 35 3.306E-06 802E-06 1780E-06 1780E-06 2.174E-06 748.4E-06 30 3.306E-06 3170E-06 4990E-06 4990E-06 2.174E-06 748.4E-06 25 3.306E-06 3170E-06 4990E-06 4990E-06 2.174E-06 748.4E-06

OCR for page 28
32 TABLE 20 Final parameters for HCM equation VDF 59 Selection Criteria Standard BPR Parameters Additional Parameters for HCM Equations Facility ul3 @fresp Lanes L Ro X Do DL N T J Freeway 1 >70 all len @ul21 volau/(ul1*lanes) 0 0 1 1 2.947E-05 1 65-70 all len @ul21 volau/(ul1*lanes) 0 0 1 1 2.003E-05 1 60-65 all len @ul21 volau/(ul1*lanes) 0 0 1 1 1.423E-05 1 55-60 all len @ul21 volau/(ul1*lanes) 0 0 1 1 8.426E-06 1 <=55 all len @ul21 volau/(ul1*lanes) 0 0 1 1 3.306E-06 Expressway 2 >45 all len @ul21 volau/(ul1*lanes) N*16.93/60 0 max(1,INT(len*1)) 1 2.21E-05 2 40-45 all len @ul21 volau/(ul1*lanes) N*16.93/60 0 max(1,INT(len*1)) 1 4.99E-05 2 35-40 all len @ul21 volau/(ul1*lanes) N*16.93/60 0 max(1,INT(len*1)) 1 8.02E-04 USER'S GUIDE 2 <35 all len @ul21 volau/(ul1*lanes) N*16.93/60 L*25/60 max(1,INT(len*3)) 1 3.17E-03 Urban 3 >45 all len @ul21 volau/(ul1*lanes) N*16.93/60 L*8/60 max(1,INT(len*2)) 1 2.04E-04 Arterial 3 40-45 all len @ul21 volau/(ul1*lanes) N*16.93/60 L*8/60 max(1,INT(len*2)) 1 2.00E-04 3 35-40 all len @ul21 volau/(ul1*lanes) N*16.93/60 L*20/60 max(1,INT(len*2)) 1 1.78E-03 3 <35 all len @ul21 volau/(ul1*lanes) N*16.93/60 L*45/60 max(1,INT(len*5)) 1 4.99E-03 One-Way 4 >45 all len @ul21 volau/(ul1*lanes) N*6.00/60 L*8/60 max(1,INT(len*2)) 1 2.04E-04 Arterial 4 40-45 all len @ul21 volau/(ul1*lanes) N*6.00/60 L*8/60 max(1,INT(len*2)) 1 2.00E-04 4 35-40 all len @ul21 volau/(ul1*lanes) N*6.00/60 L*20/60 max(1,INT(len*2)) 1 1.78E-03 4 <35 all len @ul21 volau/(ul1*lanes) N*6.00/60 L*45/60 max(1,INT(len*5)) 1 4.99E-03 Centroid Connector 5 all all len @ul21 volau/(ul1*lanes) 0 0 1 1 2.947E-05 Rural 6 >60 >1 len @ul21 volau/(ul1*lanes) 0 0 1 1 2.296E-06 Arterial 6 55-60 >1 len @ul21 volau/(ul1*lanes) 0 0 1 1 1.821E-06 6 50-55 >1 len @ul21 volau/(ul1*lanes) 0 0 1 1 1.108E-06 6 1 len @ul21 volau/(ul1*lanes) 0 0 1 1 2.174E-06 6 >65 1 len @ul21 volau/(ul1*lanes) 0 0 1 1 9.043E-05 6 60-65 1 len @ul21 volau/(ul1*lanes) 0 0 1 1 0.0001385 6 55-60 1 len @ul21 volau/(ul1*lanes) 0 0 1 1 0.0002239 6 50-55 1 len @ul21 volau/(ul1*lanes) 0 0 1 1 0.0003893 6 <50 1 len @ul21 volau/(ul1*lanes) 0 0 1 1 0.0007484 ul3, ul21 = user-definable fields. @fresp = at free-flow speed (mph). L = segment length. R0 = segment traversal time at free-flow speed. X = volume/capacity ratio. D0 = zero-flow control delay at the signal. DL = delay per mile. N = maximum of 1, or the number of signals on the segment. T = length of analysis period, in hours. J = calibration parameter. volau = auto volume. INT = integer divide function.