Incorporating Truck Analysis into the Highway Capacity Manual (2014)

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.

72 S e c t i o n 7 This section presents example applications of the recommended truck LOS model and framework in several case studies. These case studies illustrate the application of three of the truck LOS models considered in the prior section: • Model 2—The Streamlined Utility Model, • Model 3—The Logistic Model with Truck Friendliness Index, and • Model 4—A Reliability Plus Friendliness Model. In all cases, it is assumed that the analyst has estimated truck speeds and reliability using one or more appropriate methodologies. The following sections describe recommended methodologies for making these estimates. 7.1 Study Site 1—California Class I Interstate Freeway The application of three of the truck LOS models is illustrated for a 20-mile-long section of a major interregional Interstate freeway in California. The high volume of trucks and the fact that this freeway is a critical inter-regional link between the San Francisco Bay Area and the San Joaquin Valley makes this a Class I facility for truck LOS purposes. The study section is split into two seg- ments (one mountainous, the other level). The selected study period is the 7–9 a.m. peak period. Exhibit 38 shows the facility-specific data required by the truck LOS models. 7.1.1 Case Study 1.1—Computation of Existing Truck LOS In this case study, the truck LOS is computed using three candidate LOS models for two segments of the Site 1 freeway: one in mountainous terrain and the other on level terrain. See Exhibit 39. Model 2—Streamlined Utility Model For Segment 1, the free-flow utility is computed using Equation 24, a length of 8.9 miles, a probability of on-time arrival of 90%, a toll of zero dollars, and a TTI of 1.00. The resulting free- flow utility for Segment 1 is –8.49. The worst case utility is computed using the same equation with a policy speed of 10 mph and probability of on-time arrival of 10% for the facility. The worst case utility is –9.01. The actual utility for Segment 1 is computed using the same equation and length, but with a 10% probability of on-time arrival and a 36.9 mph actual speed. The actual utility is –8.75. The difference between the actual and the ideal utilities is 0.26, which is about 50% of the difference between best utility and the worst utility. The actual utility is thus LOS F according to the scale given in Exhibit 29. The computations for Segment 2 proceed similarly. Truck Level-of-Service Case Studies

Truck Level-of-Service Case Studies 73 Model 3—Logistic Model with Friendliness Index Model 3 uses the actual probability of on-time arrival, TTI, and toll. The “ideal” free-flow reli- able condition is incorporated in the variables it uses (POTA-1, and 1-TTI). It does not require the worst-case conditions to estimate LOS. Model 3 requires a truck friendliness assessment, which is not required by Model 2. In this case, the facility is an Interstate freeway designed to modern standards; thus, the TFI is set at 1.00. Model 4—Reliability and Friendliness Logistic Model Model 4 does not require the toll or TTI information required by Models 2 and 3. Like Model 3, Model 4 also requires a truck friendliness assessment. The TFI is identical to that for Model 3. Comparison of LOS Model Results for Case Study 1.1 The truck LOS results for the facility using each of the three LOS models are compared in Exhibit 40. All three models agree that Segment 1 (the long-grade section) is operating at LOS F for trucks. The models disagree regarding the truck LOS for Segment 2: Model 2 says it is C, Model 3 says it is F, and Model 4 says it is E (but not too far from the LOS E/F threshold). Data Item Segment 1 Segment 2 Length 11.1 miles 8.9 miles Terrain Mountainous Level Major Grades (% Grade, Length) 3% up, 4.2 miles 3% down, 2.3 miles None Max/Min (Best/Worst) Speeds 65/10 mph 65/10 mph Actual Speed (A.M. Peak Period) 36.9 mph 42.3 mph Best/Worst Probability of On-Time Arrival 90%/10% 90%/10% Probability of On-Time Arrival 10% 60% Note: Best/worst speeds and probability of on-time arrival are set by agency policy for the facility. Exhibit 38. Data for Case Studies 1.1 and 1.2. Exhibit 39. Study section for Case Studies 1.1 and 1.2.

74 incorporating truck Analysis into the Highway capacity Manual Note that Models 3 and 4 find a greater difference in performance between Segments 1 and 2 than is found by Model 2 as evidenced by the wider range in indices (percent ideal) output by Models 3 and 4 than by Model 2. 7.1.2 Case Study 1.2—Sensitivity Tests Segment 1 from Case Study 1.1 was selected for sensitivity testing. The effects of chang- ing the state in which the facility is located are shown in Exhibit 41. This exhibit also shows the effect on the letter grade truck LOS of the facility class (Class I or Class III). Model 2 would rate the segment at LOS D if it were a Class III facility located in Alaska. Otherwise, the models all agree that the segment would rate LOS F regardless of class or location in United States. Arbitrarily shortening or lengthening Segment 1 from Case Study 1.1 had no effect on the results produced by Models 2, 3, and 4. As shown in Exhibit 42, reliability has a significant effect on the computed truck LOS. This test was performed on both segments from Case Study 1.1. Model 2 would rate both segments at LOS A if reliability were improved to 90% probabil- ity of on-time arrival. Model 4 would rate both at LOS B with improved reliability. Model 3 would rate both segments at LOS C. All three models show similarly large sensitivities to reli- ability. Reliability has a slightly greater effect on the TLOS indices in Models 3 and 4 than for Model 2. Segment 1 Segment 2 Model 2 – Streamlined U lity Model LOS F C Percent of ideal 50% 78% Model 3 – Logis c LOS Model F F Percent of ideal 5% 44% Model 4 – Reliability and Friendliness LOS F E Percent of ideal 10% 58% Exhibit 40. Comparison of LOS model results for example application facility. Facility Class Class I Class III State Model 2 Model 3 Model 4 Model 2 Model 3 Model 4 Continental U.S. F (50%) F ( 5%) F (10%) F (50%) F ( 5%) F (10%) Alaska E (56%) F (16%) F (29%) D (56%) F (16%) F (29%) Hawaii F (15%) F ( 0%) F ( 0%) F (15%) F ( 0%) F ( 0%) Note: The test segment is an 11.1-mile segment from Case Study 1.1. All variables held constant except region. Exhibit 41. Effect of region and facility class on Models 2, 3, and 4 results.

Truck Level-of-Service Case Studies 75 7.2 Study Site 2—Virginia Class I Interregional Freeway The second study site for case studies is a 29.4-mile-long section of the Interstate freeway in Virginia (see Exhibit 43). This is a high-truck-volume, critical, interregional facility that is rated by the agency as a Class I facility for truck LOS analysis purposes. The selected study period is the weekday 6–10 a.m. peak period. 7.2.1 Case Study 2.1—Predict the Effects of High-Occupancy-Vehicle Lane on Truck LOS This case study involves predicting the effects on northbound a.m. peak-period truck LOS of a project adding a third northbound HOV lane to the Interstate freeway. The number of Segment 1 Segment 2 Original/Improved On-Time Arrival 10%/90% 60%/90% Model 2 – Streamlined U
lity Model LOS Original/Improved LOS (%TLOS) F (50%)/A (92%) C (78%)/A (94%) Model 3– Logisc LOS Model Original/Improved LOS (%TLOS) F (5%)/C (74%) F (44%)/C (78%) Model 4 – Reliability and Friendliness LOS Original/Improved LOS (%TLOS) F (10%)/B (86%) E (58%)/B (86%) Note: The test segments are from Case Study 1.1; all variables held constant except probability of on-me arrival. Exhibit 42. Effect of reliability on Models 2, 3, and 4 results. Exhibit 43. Study Site 2—Class I facility in Virginia.

76 incorporating truck Analysis into the Highway capacity Manual HOV lanes will be expanded from two to three, and those HOV lanes will be converted to high- occupancy-toll (HOT) operation. While the mainline mixed flow lanes are not undergoing any capacity expansion or improve- ments, with the addition of a third HOV lane and an expansion of service to include both HOV and HOT customers, it is expected that the reduction in vehicles using the mixed flow lanes during the morning peak period will benefit freight traffic in the area. A combination of travel demand forecasting and traffic operations models is used to estimate existing and future mixed flow lane mean speeds for the four segments. A methodology such as that developed by the SHRP2-L08 project (Kittelson and Vandehey, 2012) is used to estimate existing and future weekday a.m. peak-period travel time distributions for the segments. The input data are shown in Exhibit 44. The truck LOS results using each of the LOS models are compared in Exhibit 45 for existing conditions and future conditions. The models all agree on the trend of improvement for truck LOS caused by adding the HOV lane. Model 2 shows the most extreme improvement of the models for Segment 1, going from LOS E to A. Model 3 is slightly more conservative than the other models, often rating the segments one letter grade poorer than Models 2 and 4. 7.2.2 Case Study 2.2—Effects of Tolling on Truck LOS One option being considered for better managing traffic congestion on the I-95 freeway is congestion pricing for the full facility. The question is how much is reliability worth to the ship- pers and carriers using the freeway? Only Models 2 and 3 (which are sensitive to price) can be applied to this case study. Since this case study is to determine the value to shippers and carriers of improved reliability on the freeway, the reliability is improved to an 85% probability of on-time arrival and then the Segment Characteristics Segment 1 Segment 2 Segment 3 Segment 4 Length (miles) 8.3 7.0 11.9 2.2 Policy Best/Worst Speed (mph) 65/10 65/10 65/10 65/10 Existing Speed (mph) 53 48 46 42 Future Speed (mph) 58 55 53 51 Policy Best/Worst Probability On-Time 90%/10% 90%/10% 90%/10% 90%/10% Existing Probability On-Time 50% 38% 35% 33% Future Probability On-Time 83% 65% 55% 48% Exhibit 44. Case Study 2.1 input data. Segment 1 Segment 2 Segment 3 Segment 4 Model 2 – Streamlined Ulity Exist/Future LOS (%TLOS) C(76%)/A(95%) D(69%)/B(85%) D(66%)/C(79%) D(64%)/C(75%) Model 3– Logisc LOS Model Exist/Future LOS (%TLOS) F(40%)/C(79%) F(24%)/E(59%) F(21%)/F(46%) F(17%)/F(36%) Model 4 – Reliability/Friend Exist/Future LOS (%TLOS) F(45%)/B(81%) F(31%)D(63%) F(28%)/E(51%) F(26%)/F(43%) Exhibit 45. Comparison of LOS models for adding HOV lane to Class I freeway.

truck Level-of-Service case Studies 77 new utility and LOS compared with the no-toll condition. The toll is then added to the cost of the shipments on each segment until the original utility is obtained. Exhibit 46 summarizes the results by segment. Both models value an 85% probability of on-time arrival (reliability) at $14.75 for the 29.4-mile facility—approximately 50 cents per mile. This agreement between the two models is to be hoped for since both models incorporate the same cost parameters, although they employ different func- tional forms. 7.3 Study Site 3—Urban/Rural Highway The third study site is a 16.6-mile-long section of highway that extends from a small town to a nearby airport and then continues on to serve an agricultural area. The highway is a mix of urban arterial, multilane highway, and two-lane rural highway (see Exhibit 47). It starts out as a four-lane signalized urban street within a small town, then transitions to a four-lane divided highway between the town and its airport. Beyond the airport, the multilane highway becomes a two-lane rural highway. The selected study period is the weekday 4–6 p.m. peak period. 7.3.1 Case Study 3.1—Assess LOS for Urban/Rural Highway with Growth The highway is a locally important connector between the local population center, its airport, and agricultural areas beyond the airport. The portion between the small town and the airport is rated Class II by the agency. The portion beyond the airport is rated Class III. Segment Original Reliability, No Toll Good Reliability, No Toll Good Reliability, Toll POTA Toll LOS POTA Toll LOS POTA Toll LOS Model 2 1 83% $0 A(95%) 85% $0 A(96%) 85% $0.25 A(95%) 2 65% $0 B(85%) 85% $0 A(95%) 85% $3.50 B(85%) 3 55% $0 C(79%) 85% $0 A(95%) 85% $9.00 C(79%) 4 48% $0 C(75%) 85% $0 A(94%) 85% $2.00 C(75%) Total $14.75 Model 3 1 83% $0 C(79%) 85% $0 B(81%) 85% 0.50 C(79%) 2 65% $0 E(59%) 85% $0 C(80%) 85% $3.50 E(59%) 3 55% $0 F(46%) 85% $0 C(79%) 85% $8.75 F(46%) 4 48% $0 F(36%) 85% $0 C(78%) 85% $2.00 F(37%) Total $14.75 Note: Model 4, since it lacks a toll component, cannot be used for such a test. Exhibit 46. Results for Case Study 2.2.

78 Incorporating Truck Analysis into the Highway Capacity Manual The state DOT is anticipating a shift from primarily agricultural crops on this highway to a mix of electronics and agricultural goods with higher intensities of traffic associated with light industry. As a result, this highway is being evaluated by the state to determine its ability to con- tinue to provide acceptable freight LOS as the area continues to develop. Exhibit 48 shows existing conditions, and Exhibit 49 shows future conditions. Exhibit 50 shows the existing TTI distribution for the highway segments. The TFI is set by the agency at 0.75 on the two-lane highway section to reflect an at-grade railroad crossing and load limits on a couple of bridges on the highway. For urban streets, a larger tolerance of 3.33 TTI for on-time arrival is set because the HCM definition of free-flow speed on urban streets excludes all signal delay (causing a wide disparity between the posted midblock speed limit and the actual achievable through-speed on the street even in low-flow conditions). For multilane and two-lane highways, the original 1.33 TTI tolerance for on-time arrival is retained. Exhibit 47. Case Study 3 site. Truck LOS Input Data Segment 1 Segment 2 Segment 3 Segment 4 Truck Facility Type Class II Class II Class III Class III HCM Facility Type Urban Street Multilane Hwy Multilane Hwy 2-Lane Hwy Limits First Street to City Limits City Limits to Airport Airport to 2-Lane Multilane to End of Highway Length 0.6 2.3 3.8 9.9 Free-Flow Speed (mph) 35 55 55 45 Worst Speed (mph) 10 10 10 10 Actual Speed (mph) 12 42 43 37 85% TTI 3.27 1.39 1.35 1.32 Best POTA 90% 90% 90% 90% Worst POTA 10% 10% 10% 10% Actual POTA 85% 70% 90% 95% Truck Friendliness Index 1.00 1.00 1.00 0.75 Notes: POTA = probability of on-time arrival; TTI = travel time index. Exhibit 48. Case Study 3.1 data—existing conditions.

Truck Level-of-Service Case Studies 79 Exhibit 51 shows how the existing and future truck LOS results vary by segment by truck LOS model. The models are in general agreement as to trends in truck LOS. All show truck LOS worsening in the future: • For Segment 1—the urban street segment with a high probability of on-time arrival (85%) but a low mean speed of 12 mph in comparison with the midblock free-flow speed of 35 mph—is rated LOS B for existing conditions by Model 4. Models 2 and 3, however, rate this segment at LOS E/F for existing conditions due to the low mean speed on the segment (which Model 4 does not include). For future conditions on this street segment, all three models agree at rating Segment 1 at LOS F. This is primarily due to the degradation in reliability for this seg- ment in the future. Truck LOS Input Data Segment 1 Segment 2 Segment 3 Segment 4 Truck Facility Type Class II Class II Class III Class III HCM Facility Type Urban Street Multilane Hwy Multilane Hwy 2-Lane Hwy Limits First Street to City Limits City Limits to Airport Airport to 2-Lane Multilane to End of Highway Length (miles) 0.6 2.3 3.8 9.9 Free-Flow Speed (mph) 35 55 55 45 Worst Speed (mph) 10 10 10 10 Actual Speed (mph) 10.8 40.7 41.7 35.4 85% TTI 3.68 1.44 1.40 1.42 Best POTA 90% 90% 90% 90% Worst POTA 10% 10% 10% 10% Prob. On-Time Arrival 60% 50% 70% 75% Truck Friendliness Index 1.00 1.00 1.00 0.75 Notes: POTA = probability of on-time arrival; TTI = travel time index. Exhibit 49. Case Study 3.1 data—future conditions. Exhibit 50. Case Study 3.1 TTI distributions—existing conditions.

80 incorporating truck Analysis into the Highway capacity Manual • Model 3 appears to be the most conservative among the three models in rating the truck LOS for all 4 segments under both existing and future conditions. • Model 2 rates Segment 4 (the two-lane highway segment) as LOS A under existing conditions, primarily because of the excellent probability of on-time arrival (95%). The other two models rate this segment at LOS C/D, primarily because of the TFI for this segment, which is not taken into consideration in Model 2. Exhibit 51. Comparison of LOS model results for Case Study 3.1. Segment 1 Segment 2 Segment 3 Segment 4 Model 2 – Streamlined Ulity Exist/Future LOS (%TLOS) D(56%)/F(35%) A(85%)/C(74%) A(96%)/A(85%) A(99%)/A(87%) Model 3– Logisc LOS Model Exist/Future LOS (%TLOS) F(12%)/F(1%) D(61%)/F(35%) A(81%)/C(61%) D(56%)/F(30%) Model 4 – Reliability/Friend Exist/Future LOS (%TLOS) B(83%)/F(43%) C(69%)/E(45%) A(86%)/C(69%) C(63%)/F(39%)