An Asset-Management Framework for the Interstate Highway System (2009)

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.

66 This section describes the pilot conducted of the Interstate Asset Management Framework. The pilot was conducted pri- marily to test application of the framework using existing data and tools. Section B.1 details the approach of the pilot. Section B.2 describes the data obtained. Section B.3 details the analysis performed, and Section B.4 summarizes conclu- sions from the pilot. B.1 Approach Participants in the pilot effort were selected by the research panel, and included representatives from the DOTs of Califor- nia, South Carolina, and Wisconsin. Each participant identified an IHS corridor from their DOT’s IHS network to include in the analysis. The corridors included in the pilot were as follows: • California—Interstate 80; • South Carolina—Interstate 95 (approximately 100 miles of this highway, extending south from the North Carolina border); and • Wisconsin—Interstate 94. The research team requested the same basic data and per- formed, to the extent feasible, the same analyses on the data from each pilot participant. The research team requested the following types of data from each participant, at a minimum: • Pavement/Highway Inventory—Inventory and condition data for each pavement section, including the agency’s HPMS file and additional pavement condition measures not included in the HPMS; • Bridge—Pontis database or NBI file; • Other Structures—Listing of other structures, including structure type, annual maintenance cost, replacement cost, estimated remaining service life, and any available listing of deficiencies for the asset; • Roadside Assets (signs, striping, traffic operations equip- ment, guardrails, right-of-way, shoulders, and any of the other structures listed above for which the agency lacks detailed asset data)—Inventory of roadside assets includ- ing asset type, extent (e.g., counts of the asset over the cor- ridor), and annual maintenance cost for the asset; • Facilities (rest areas, weigh stations, toll booths and any other buildings or other facilities associated with the pilot corridor)—Inventory of facilities including facility type, location, and annual maintenance cost; and • Other—Information on the status, scope and cost of planned, programmed or in-progress projects for the pilot corridor, counts of crashes and fatalities by year for the past five years on the pilot corridor. Agencies were requested to provide whatever data they had already collected and were readily available from the set of data items described above, as well as additional relevant data (e.g., a description of risk scenarios or travel demand data that could be used to facilitate risk analysis). Data were requested in December 2007 and received between January and February 2008. Once data were obtained, the re- search team performed a series of analyses to characterize exist- ing conditions of each corridor, and predicted future conditions under different funding and risk scenarios. These analyses are described in Section B.3 and are considered typical of that re- quired to develop an Interstate Asset Management Plan. B.2 Summary of Data Obtained All of the pilot participants were extremely helpful in facil- itating collection of the data for the pilot, despite the fact that this frequently required coordination with a number of dif- ferent groups within their respective agencies. Table B.1 sum- marizes the data obtained from the pilot participants. The following paragraphs further detail the data obtained for each asset category. • Pavement/Highway Inventory—All of the pilot partici- pants had HPMS data and additional pavement condition A P P E N D I X B Pilot Program

data. One participant (South Carolina) had an in-house capability to simulate future conditions using its Pavement Management System (PMS) and provided PMS results for the pilot corridor. • Bridges—All of the participants had NBI data and addi- tional element-level data. Two participants (South Carolina and California) had developed deterioration and costs models for use in predicting future bridge conditions. One participant (California) provided additional data on a vari- ety of risk-related bridge needs, including seismic vulnera- bilities, scour mitigation needs, and safety improvement (guardrail) needs. • Other Structures—Limited data were available on other structures. All of participants store some amount of struc- ture data for nonbridge structures (e.g., tunnels and some culverts) in their bridge management system (BMS). Except for the data available through their BMS, none of the partic- ipants had available additional detailed structure inventory and condition data. Roadside Assets. Limited data were available for roadside assets. One participant (Wisconsin) had inventories avail- able for signs and pavement markings. Another participant (California) had data on roadside vegetation needs, as well as system-level data on roadside assets to support a maintenance budgeting process, though this was not localized to the pilot corridor. In some cases, participants indicated that individual districts likely had additional inventory data (e.g., in the form of spreadsheet inventories), but this information was not read- ily available and attainable within the timeframe requested. Facilities. Facilities specific to the IHS for the pilot corri- dors included rest areas along each corridor. Two of the par- ticipants (South Carolina and Wisconsin) provided details on their rest areas, including maintenance and renewal costs. Other Data. All of the participants had project information for programmed projects on the pilot corridors. All of the par- ticipants provided data on crashes and fatalities on the pilot corridors. All of the participants had additional data beyond the minimum data requirements described above. For example, California recently updated its process for developing its Strategic Highway Operations and Protection Plan (SHOPP) and had data and results available from that effort. South Carolina had detailed data on highway interchanges and interstate highway sections, including potential interchange improvements and performance measures related to highway sections and interchanges in its Interactive Interstate Man- agement System (IIMS). Figure B.1 shows an example screen from the IIMS showing an interchange diagram and predicted performance measures for an I-95 interchange. Wisconsin provided information on its passenger and freight demand models, as well as additional freight analyses. Regarding risk, two of the participants had considered risk in operational planning in some manner (California through identifying “emergency lifeline routes” and South Carolina through developing hurricane evacuation plans). Also, Cal- ifornia had significant additional data available regarding structure-related risk, as described above. All of the partici- pants either had (California, Wisconsin) or were developing 67 Table B.1. Summary of pilot data obtained. Category Data Obtained Pavement/ Highway Inventory All participants had HPMS data All participants had data on existing/historic conditions One participant had PMS runs available Bridges All participants had NBI data All participants had additional element-level data Two participants had cost models needed to run Pontis One participant had supplemental information on structure-related risk (e.g., seismic, guardrails, scour, etc…) Other Structures Limited data available Roadside Assets Signs/pavement markings for one-third of the participants Facilities Two participants had basic inventory and cost data for rest areas Other All participants had crash data All participants had project data All participants have (or will soon have) travel demand model data, but none were in a position to define risk scenarios to analyze using travel demand data One participant had some level of risk-related data – details on structure vulnerabilities of different types, with scores

(South Carolina, through participating in the I-95 Corridor Coalition travel demand model development effort) travel demand models that could be used to support risk analysis. However, none of the participants had identified risk sce- narios that could be used in conjunction with their travel de- mand models for quantifying potential consequences of different risks. B.3 Analyses Performed Using the data described in the previous section, a set of analyses was performed for each of the pilot corridors. The goal of the analysis was to predict future performance of the pilot corridors for different budget scenarios in terms of the core measures recommended for a typical Interstate Asset Management Plan. Also, to the extent feasible, analyses were performed to demonstrate application of the risk assessment approach recommended previously. The paragraphs below describe the analyses performed and specific analytical tools used for the pilot. Pavement/Highway Inventory. This category of data can be used to predict pavement conditions, as well as a range of other measures, including mobility and safety-related mea- sures. The research team used the FHWA Highway Economic Requirements System—State Version (HERS-ST)—to predict mobility and safety-related measures. HERS-ST is available at no cost and requires HPMS data as input. All of the pilot par- ticipants had this data available, and it was thus feasible to run HERS-ST for each of the pilot corridors. HERS-ST predicts a limited number of measures of pave- ment condition, including Present Serviceability Rating (PSR), and International Roughness Index (IRI). However, HERS-ST’s modeling of pavements is rudimentary compared to that fea- sible in most agency pavement management systems (PMS). Thus, where feasible, supplemental analysis results were used as an alternative to HERS-ST for predicting pavement condi- tion. In the case of South Carolina, agency PMS results were available for the pilot corridor. For Wisconsin, research team members working with Wisconsin DOT on a separate effort used the remaining service life approach described in Interim Report 1 to analyze pavement conditions for the pilot corri- dor. For California, supplemental analyses were available at a system level, but not for the pilot corridor. Thus, the research team relied upon HERS-ST for predicted pavement conditions. Bridges. For South Carolina, a set of Pontis simulations was run for the pilot corridor using South Carolina Pontis models. For California DOT (Caltrans), the bridge needs analy- sis approach defined for Caltrans’ SHOPP update process was 68 Figure B.1. Example South Carolina IIMS screen.

used. This entailed using Pontis to generate functional im- provement needs, predict deterioration over time, and predict the benefits of certain types of bridge needs. For Wisconsin, state-specific models were not available and element data, though available, were not stored in the form of a Pontis data- base. Thus, NBIAS was used to predict bridge conditions for the Wisconsin corridor. Other Structures and Roadside Assets. As noted previ- ously, limited data were available for other assets besides pave- ment and bridges. Where such data existed, they were either in terms of remaining service lives of assets (e.g., of facilities) or average condition/level of service of some population of assets. The research team developed and tested spreadsheet models for applying a remaining service life approach to assets for which remaining life and rehabilitation/replacement costs were available, and for applying a maintenance level-of-service model to assets for which levels of service were defined. Figure B.2 illustrates the results of applying a simple remain- ing service life model. In this example, an asset (e.g., a facility) is deemed to be “functioning as intended” if its age is less than the service life for that asset, and is deemed to be not function- ing as intended if it has no remaining service life. Given an annual budget, it is a straightforward calculation to predict the age distribution over time for a population of assets, aging the assets each year of the analysis period and simulating renewal of an asset that has reached the end of its service life, to the ex- tent this is feasible given the available budget. Performing this calculation requires remaining service life for each asset, re- newal cost, and service life for a newly renewed asset. The figure shows the percent of a group of assets functioning as intended. In this example a $2M annual budget results in 63 percent of the assets functioning as intended at the end of 20 years (com- pared to a starting value of 75 percent), a $1M annual budget results in having 31 percent functioning as intended at the end of the period, and a $0.5M annual budget results in approxi- mately 16 percent functioning as intended. Where other agency systems and data were available, such as the SCDOT IIMS and data associated with the Caltrans SHOPP update process, the research team reviewed this in- formation, but did not integrate these data sources into the analysis process except as otherwise noted. Risk. The research team demonstrated the risk assessment approach for examining structure-related risk using data pro- vided by Caltrans. Structure risks considered systematically by Caltrans include risks of seismic events, bridge scour, and guardrail failure. Previously Cambridge Systematics worked with Caltrans to develop a utility measure that included measures of vulnerability and consequence for each of these risks, as well as for other bridge investment needs. For exam- ple, for seismic risk, Caltrans has developed a list of proposed seismic retrofit projects, and calculated seismic vulnerability scores for each bridge. These scores provide an indication of the degree of risk, and data on the functionality of the bridge (e.g., traffic, detour distance around the bridge) provides an 69 Figure B.2. Example calculation of percent functioning as intended versus annual budget using service life data.

indication of the consequences in the event of a seismic event. These measures were combined to predict the utility of address- ing each identified seismic need, and to develop a relationship between utility achieved and predicted spending. AssetManager NT. AssetManager is a set of analytical tools initially developed through NCHRP Project 20-57, and now licensed through AASHTO. AssetManager NT is a network-level tool that combines analysis results generated through other systems, displaying predicted conditions for a range of different performance measures and budget assump- tions. The system is well-suited for applications such as sup- porting development of an Interstate Asset Management Plan, where results from different systems are combined with the goal of providing an integrated view of predicted conditions. The research team imported into AssetManager NT data on pavement conditions, bridge conditions, mobility, safety, and risk. In the system one can view predicted conditions over time for a pilot corridor for a specified distribution of funds. Figure B.3 shows an example screen from the system. In this example results are shown for six different perfor- mance measures (pavement condition, bridge health index, number of Structurally Deficient bridges, delay, user costs, crash rate) for four different budget allocations. Each pane in the figure represents results for a different performance measure, and each series graphed shows results for a different budget allocation. METIS. The major limitation of AssetManager is that the system shows results for different user-specified budget alloca- tions, but does not provide support for optimizing a budget be- tween different objectives. Performing a cross-asset allocation is a difficult proposition. Typically, one would solve such a multi-objective optimization problem by first defining a utility measure that encompasses all of the objectives one attempts to achieve by investing in IHS assets (as was performed for help- ing characterizing benefits of mitigating structure-related risks described above). Previously Cambridge Systematics devel- oped a tool for solving multi-objective optimization problems using an evolutionary process, in which a user views different candidate solutions (which correspond to different potential objective functions), and gradually converges upon an optimal solution through manual evaluation of potential solutions. The research team used this tool, the Multi-Objective Evolu- tionary Tool for Interactive Solutions (METIS), to pilot the process of optimizing resources allocated to IHS assets. Figure B.4 shows an example screen from METIS. In this screen, alternative resource allocation results are shown for four different candidate solutions with different weights on four dif- ferent objectives: maximizing pavement condition (measured 70 Figure B.3. Example AssetManager screen.

71 using a pavement quality index), minimizing delay, minimizing the number of structurally deficient bridges, and mitigating seis- mic risk. As is the case for AssetManager, METIS uses analysis re- sults generated from other systems to support its projections. The end user reviews each solution, selecting one solution to rule out at each analysis step, and the system uses this informa- tion to narrow in on an optimal allocation of resources between assets for best meeting the competing objectives. B.4 Conclusions The research team developed the following conclusions on the basis of the pilot analysis: 1. IHS owners typically have data readily available concern- ing highway inventory and traffic, pavement conditions, bridge conditions, crashes, rest areas and planned projects. IHS owners have limited data available on other structures besides bridges, as well as on roadside assets. Further, IHS owners typically lack at least some of the data required to support the risk analysis approach recommended as part of the Interstate Asset Management Framework. 2. It is feasible to predict basic measures of pavement condi- tion, bridge condition, mobility, and safety for IHS assets using readily available data and tools. Key analytical tools for such an exercise include an agency’s PMS and BMS (typically the AASHTO Pontis BMS), FHWA’s HERS-ST, and AASHTO’s AssetManager. Where an agency does not have a management system implemented to predict future IHS conditions, one can use HERS-ST as an alternative for predicting pavement conditions, and FHWA’s NBIAS for predicting bridge needs. 3. It is feasible to predict measures related to other structure- related risk using available tools, and supplemental analy- ses, provided an agency has compiled data on these risks. One of the three pilot participants had the data available to support such an exercise. Chapter 4 describes available tools for assessing structure-related risks. 4. The three pilot participants had sufficiently detailed data and tools to support analysis of pavement and bridge con- ditions. They had data on some other assets besides pave- ments and bridges, but generally lacked analytical tools or models for predicting conditions of these assets over time. Chapter 4 describes data and tools needed for modeling needs of these assets. 5. It is feasible to use a utility maximization approach to ap- proximate the optimal allocation of resources between different IHS assets to achieve a set of objectives. However, supplemental analysis outside of existing tools available from AASHTO and FHWA is needed to implement such an approach. Figure B.4. Example METIS screen.