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.
The study teamâs assessments of SHRP 2 product functionality and decision maker perceptions includes input from all of these efforts and is organized into the following sections: ⢠Technical Functionality of Products ⢠Decision Maker Perceptions ⢠Impacts on Decision Making 9.1 Technical Feasibility of Products The study team considered the relative ease of using the SHRP 2 products, their usefulness in the analysis process, and consistency among the tools. The technical evaluation includes input from study team members (including SCAG), Caltrans, and other stakeholders. How well does the work completed to date help Southern California agencies better understand the causes of baseline reliability? Prior to the SHRP 2 product pilot testing, California already had travel time reliability monitoring capabilities through PeMS. This system includes several years of highway performance data collected by sensors throughout the state and calculates reliability performance measures such as travel time reliability. In addition, several Southern California facilities are covered by CSMPs, which have reliability performance measures included as part of the analysis. The L02 guide provided a number of new analysis techniques and ways of looking at reliability factors. The cumulative distribution functions (CDFs) and the semi-variance tables help to show the relative contribution of reliability factors and visualize their impact on facility reliability. The study team found that the order in which factors are assigned affects the results of the analysis. For the best results, the analysis should include categories for multiple factors (e.g., incidents and weather issues occurring simultaneously) and specifically consider factors such as work zones and lane closures. These new techniques could be incorporated into future CSMPs. Do the tools provide reasonable results for a variety of improvement strategies focused on operations? The bulk of the testing documented in Chapters 5 through 7 is devoted to calibrating the three reliability analysis tools (from L07, L08, and C11) and using these tools to test improvement strategies. The study team found that the tools could be calibrated to baseline conditions once appropriate calibration levers (e.g., capacity, hourly demand, and adjustment factors) were identified. None of the tools had built-in capabilities to model the types of operational projects most likely to be tested in California, including ramp metering, ramp improvements, auxiliary lanes, and freeway connectors. These operational strategies are a key component of the Caltrans Mobility Pyramid shown in Figure 9.1. A critical update to the SHRP 2 reliability tools will be the ability to handle these types of improvements. Caltrans staff expressed interest in some of the design strategies modeled in the L07 tool, such as incident screens and drivable shoulders. However, other strategies such as snow fences 190
and wildlife crash reduction measure do not make sense for Southern California (at least within the core urban area covered by the pilot testing). The tool also omits newer active management strategies, such as dynamic or harmonized speed limits. Figure 9.1. Caltrans Mobility Pyramid. Since the tools were unable to model key operational strategies used in Southern California, the study team had to rely on microsimulation modeling to estimate the changes in capacity needed to test the strategies in the SHRP 2 tools. With these inputs, the tools estimated reliability impacts lower than expected. These impacts appeared to be highly correlated with mobility (or recurring delay) results. Which SHRP 2 tools were easier to use? As shown in Figure 9.2, the study team found that there was a clear order in terms of ease of use among the SHRP 2 reliability tools. The C11 and L07 tools were much easier to use than the FREEVAL-RL tool. The study team had intended to split the analysis among SCAG and other team members, so that agency and consulting staff used all of the tools. The team quickly discovered that calibrating the FREEVAL-RL tool was as complicated and time-consuming as 191
calibrating a microsimulation model, so the calibration was assigned to modelers. In addition, the FREEVAL-RL tool took a long time to complete its probabilistic scenario runs. Figure 9.2. SHRP 2 reliability tools by ease of use. In contrast, the C11 and L07 tools were simple to use and generated results quickly. Caltrans planners and engineers noted that they prefer simple tools. Tools with steep learning curves or that take a long time to run are not practical options. SCAG staff expressed a similar preference. What tools provided more reasonable results and how did these compare to baseline and microsimulation results? The study team was able to calibrate all three tools to baseline conditions on both facilities. The C11 tool could be calibrated by adjusting the capacity and changing the demand- by-hour distribution to match the facility distribution of volume by hour. This was a fairly easy process, once the study team found the delay-by-hour table hidden among the model parameters. The L07 tool could be calibrated using the same levers, but the hourly demand table was easily accessible, while the capacity was hidden in the parameters. These limitations and others are discussed in detail in Chapters 5 through 7 of the report. Despite similarities in calibrating the tools, the C11 tool produced more reasonable results when modeling the scenarios. In comparison, the L07 tool produced very small reliability benefits. The FREEVAL-RL tool was much harder to calibrate to baseline conditions, due to an inability to handle limited-access high-occupancy vehicle lanes and significant congestion along the facility that exceeded the maximum extent and duration supported in FREEVAL-RL. Despite these limitations, the study team was able to calibrate FREEVAL-RL to a shorter test section. Both the baseline calibration and the scenario results seem reasonable for this test section. How did technical members of the supporting agencies react to the work, and did it make sense to them? SCAG and Caltrans thought that the reliability factor analysis provided useful information for understanding the causes of reliability issues along the facilities. SCAG technical staff found the C11 and L07 tools fairly easy to use. 192
Technical staff members were positive about the potential for improved travel time reliability analysis. However, they recognized the need to improve the analysis tools so that they are ready for agency implementation. These changes are described in detail in Chapters 5 through 7 and summarized in Chapter 10. Was the analysis too complicated to duplicate internally at the supporting agencies? SCAG staff participated in testing the L07, C11, and FREEVAL-RL tools. SCAG was able to use the L07 and C11 tools internally with very little support. The FREEVAL-RL tool was more complicated and would be difficult for agency planning staff to use internally (at least for Southern California facilities). The study team conducted the reliability factor analysis in a Microsoft Access database using PeMS data. SCAG staff was able to download PeMS data, but the L02 analysis guidelines would be difficult to follow for simple planning analysis. What problems did the study team have using the different tools? Chapters 5 through 7 document the specific problems encountered while using the tools. More generally, the calibration process took longer than anticipated because the study team needed to learn the tools and find appropriate levers for calibrating the tools to baseline conditions. Now that these levers have been identified, baseline calibration would be much faster for other facilities. The study team had difficulty modeling the reliability impacts of projects without obvious capacity improvements or built-in methods for analyzing. Examples include ramp metering and auxiliary lane projects. The study team had to estimate the potential capacity improvement using the results of the CSMP microsimulation analyses. If the microsimulation models were not available, the study team would have been unable to model these projects. The tools need improvements to support scenario analysis and provide better data handling (import and save scenario data). In some cases, the study team had to re-enter data because the tools would not allow analysis periods to be modified, facility segments to be adjusted, or new scenarios to be built upon previous ones. The FREEVAL-RL tool was difficult to run for the I-5 facility. The study team encountered a number of problems with calibrating the model. Many of these problems were related to limitations in the underlying HCM 2010 methodology. For example, congestion extended over a longer section of I-5 and for a longer time period than FREEVAL-RL was able to handle. The facility also included limited-access high-occupancy vehicle lanes and an on-ramp with three lanes. What changes would the study team recommend for the tools and why? The beginning sections of Chapters 5 through 7 describe the limitations of the reliability tools and suggested improvements. In addition, the study team provided the SHRP 2 program with a list of quick fixes that are high priority and could be implemented quickly. Can the tools be used without modification for benefit-cost analysis? None of the tools produce results that can be used directly in benefit-cost analysis, but each could be made ready with minor adjustments: 193
