Skip to contentU.S. Department of Transportation/Federal Highway Administration
Asset Management | Bridge Technology | Operations | Pavement

Improving FHWA's Ability to Assess Highway Infrastructure Health

2.0 Phase I Highlights and Recommendations

2.1 PHASE I MILESTONES

In order to achieve the objectives of this project, it has been divided into three phases, shown in Figure 2.1. Phase I focused on defining an approach for assessing infrastructure health, with a focus on pavements and bridges. In Phase II the approach will be finalized and tested via a pilot study on a sample corridor. Phase III is centered on a national meeting to review the project results and discuss the preferred methodology and next steps. This report marks the completion of Phase I (as illustrated by the red arrow in Figure 2.1).

Project Milestones
Figure 2.1 Project Milestones

As part of Phase I, three formal project meetings were held. On October 14, 2010, a Project Kick-off Meeting was held at FHWA Headquarters in Washington D.C. Discussion focused on the project scope and expectations. On December 14, 2010, Project Meeting #2 was held with a focus on preparation for Technical Working Group (TWG) Meeting #1, which occurred on February 4, 2011. Discussions at all three of these meetings shaped the results and direction of Phase I (Sections 2.2 and 2.3 respectively). For more detailed notes on the discussions of each meeting, refer to the meeting minutes.

It is anticipated that the TWG will provide input throughout this project. In many cases, personnel responsible for pavement and bridge management are the best source of information regarding the health and performance of these assets. The TWG members will be asked to identify data that they consider in assessing the overall health of pavements and bridges. The TWG also will provide input regarding the criteria for classifying asset condition as good, fair or poor and information they use to make this determination. The recognition and inclusion of the types of information practitioners use to make assessments of health will be vital to establishing credibility and buy-in for this effort.

FHWA has worked with AASHTO to assemble the TWG to:

  • Provide input on the assessment methodologies;
  • Provide necessary data and support for the pilot; and
  • Meet with the project team to review pilot results.

The TWG includes the following representatives from six State DOTs and AASHTO:

  • Ms. Judith Corley-Lay, North Carolina DOT;
  • Ms. Joneete Kreideweis, Minnesota DOT;
  • Ms. Mara Campbell, Missouri DOT;
  • Mr. David Huft, South Dakota DOT;
  • Mr. Steven Krebes, Wisconsin DOT;
  • Ms. Daniela Bremmer, Washington State DOT; and
  • Mr. Matt Hardy, AASHTO.

Three of the DOTs (Wisconsin, Minnesota, and South Dakota) maintain a portion of the I-90 corridor selected for the pilot study, which is described below.

2.2 KEY FINDINGS FROM PHASE I

Based on the project understanding, literature review, guidance from FHWA, and input from the TWG, several findings regarding asset condition and health have emerged. These findings are summarized below, organized by the two parallel tracks of the project - categorizing asset condition as good/fair/poor, and assessing overall health.

Defining Good/Fair/Poor

The approach being implemented through this project for categorizing asset condition as good/fair/poor requires two separate steps:

  1. Define definitions for good/fair/poor. By design, these definitions relate solely to the condition of a pavement or bridge, and do not consider other factors such as safety, capacity, etc. In addition, they are metric-neutral, meaning that the definitions will remain constant regardless of the metrics selected in step 2.
  2. Define condition metrics and thresholds that can be used to systematically categorize assets based on these definitions. It is anticipated that as new data and modeling capabilities become available, these metrics will evolve.

Based on proposed FHWA definitions and input provided by the TWG, the following definitions have been advanced as part of this project:

  • Good condition - Pavement and bridge infrastructure that is free of significant defects, and has a condition that does not adversely affect its performance. This level of condition typically only requires preventive maintenance activities.
  • Fair condition - Pavement and bridge infrastructure that has isolated surface defects or functional deficiencies on pavements; or minor deterioration of bridge elements. This level of condition typically could be addressed through minor rehabilitation, such as overlays and patching of pavements that do not require full depth structural improvements; and crack sealing, patching of spalls, and corrosion mitigation on bridges.
  • Poor condition - Pavement and bridge infrastructure that is exhibiting advanced deterioration and conditions that impact structural capacity. This level of condition typically requires structural repair, replacement or reconstruction.

These definitions can also be presented in a tabular form, as shown in Table 2.1.

Table 2.1 Defining Good/Fair/Poor
Condition Typical Work Required
Good condition
  • Free of significant defects
  • Condition does not adversely affect its performance
  • Preservation activities
Fair condition
  • Isolated surface defects or functional deficiencies on pavements
  • Minor deterioration on bridge elements
  • Minor rehabilitation
    • Pavement overlays and patching
    • Bridge crack sealing, patching of spalls, and corrosion mitigation
Poor condition
  • Advanced deterioration
  • Conditions impact structural capacity
  • Structural repair, replacement, or reconstruction

These definitions are intended for use by FHWA and State DOTs. They provide a single scale for subsequent measure and threshold discussions. Ultimately, they may need to be simplified for public consumption. In addition, if they are presented to the public, care should be taken to consider potential legal consequences of certain terms.

Ideally, the overall use of these definitions would be considered when finalizing them. For example, if the objective is to hold DOT's accountable for achieving good/fair/poor targets, then there may need to be more technical scrutiny and consensus building than if the main objective is to provide a nationwide reporting mechanism. However, it is unclear at this time how the good/fair/poor definitions may be used by Congress in the next reauthorization, and the focus at this time is on providing a national reporting mechanism.

The definitions and metric thresholds (discussed below) are not meant to vary by functional class. These differences would be addressed during a subsequent target setting process, e.g., where a target could be defined as the percent of a network (or portion of a network) that is in good condition.

Pavement Condition Metrics

Based on the work conducted in Phase I of this study, it is recommended that the following three options for pavement condition metrics be explored during the pilot in Phase II.

Pavement metric option #1 - IRI. There is momentum for IRI to be the initial basis for a national pavement performance measure. For example, IRI thresholds for good/fair/poor are currently being developed as part of NCHRP 20-24(37) G. (At this time, NCHRP 20-24(37) G is still ongoing. Therefore, all references to that work in this document should be considered to be in a draft stage.) These thresholds (which are presented in Table 2.2.) are consistent with the thresholds recommended in a recent FWHA report called, Baseline Performance Data for Federal-Aid Performance Framework, and with the thresholds used in the FHWA's Condition and Performance (C&P) Report.

Table 2.2 Potential IRI Thresholds
Threshold in C&P Report Category Draft Threshold from NCHRP 20-24(37) G Category
≤95 Good < 95 Good
≤170 Acceptable 95 ≤ IRI ≤ 170 Fair
>170 Not Acceptable > 170 Poor
Pavement metric option #2 - New Metric Based on HPMS 2010+ Data.

Although there is interest in IRI as an initial national indicator, there are recognized limitations with using a single indicator to measure pavement condition. IRI is a measure of pavement roughness not of overall condition. Many States DOTs combine ride/roughness data with distress data to represent overall pavement condition. However the details of this approach vary considerably across the country. As part of the pilot study, the project team will investigate using the distress elements required as part of HPMS 2010+ to develop a similar type of combined measure. For example, the new metric could involve a combination of IRI, rutting, faulting, and cracking. Potential issues on using this type of HPMS-based measure include:

  • HPMS 2010+ pavement distress data is reported only for HPMS sample sections.
  • HPMS data are only collected in one direction (e.g., east and south).
  • Some DOTs are investigating the use of transfer functions to convert their existing data into the required HPMS 2010+ format. However, the validity of this approach has not been widely studied.
  • It is anticipated that there is a significant range in how DOTs are collecting HPMS 2010+ condition data. IRI and rutting may be the most consistently collected across agencies.

These issues have been flagged for consideration during the pilot.

Pavement metric option #3 - New Metric Which Incorporates Structural Response. A significant benefit of Options #1 and #2 is that they are possible with existing data, or data that is expected to be available shortly through HPMS 2010+. However, neither option provides a comprehensive view of pavement condition. A more comprehensive pavement metric would also consider the structural response of the layered system. Therefore, as part of this study, the project team will explore opportunities and potential challenges of using state collected falling weight deflectometer data (if available) and deflection data collected with a Rolling Wheel Deflectometer as part of the good/fair/poor assessment in combination with options 1 and 2, presented above.

Bridge Condition Metrics

Three options for bridge condition metrics have been identified for consideration during the pilot.

Bridge metric option #1 - Structural Deficient (SD) Status. SD status is determined by FHWA based on NBI data submitted by State DOTs. A bridge is classified as SD if:

  • The condition of its deck, superstructure, substructure, and/or culvert is rated 4 or less (on a ten-point scale), OR
  • Its structural condition or waterway adequacy is rated 2 or less.3

3FHWA Non-regulatory supplement for 23 CFR 650.409, http://www.fhwa.dot.gov/legsregs/directives/fapg/0650dsup.htm

SD status is being explored as a potential national performance measure through NCHRP 20-24(37) G, and will be evaluated as part of the pilot for this study.

Bridge metric option #2 - NBI Ratings. While SD status is a widely used measure of bridge condition, it is binary (a bridge is either SD or it is not). The ideal metric for bridge condition would be a numeric index that allows for specifying different levels of urgency for addressing a bridge need. For example, the NBI dataset includes ratings for deck, culvert, superstructure, and substructure. These ratings range from 1 to 9. Previous FHWA efforts and a recent comparative analysis study of bridge conditions conducted though NCHRP 20-24(37) E used the following thresholds to categorize bridges as good/fair/poor:

  • Good - Minimum rating ≥ 7;
  • Fair - Minimum rating is 5 or 6; and
  • Poor - Minimum rating < 5. This approach will be explored during the pilot study.

Bridge metric option #3 - New Measure Based on NBI Ratings. Another option for combining the NBI ratings described above is to include additional ratings in the calculation (e.g., inventory load rating and water adequacy rating) and to calculate a weighted average. Each rating would be weighted by its perceived importance to overall bridge condition. This approach is similar to the bridge Health Index metric used by the Pontis bridge management system. During the pilot, the study team will explore options for a new bridge metric that is based on the weighted average of NBI ratings for deck, superstructure, substructure and culvert, inventory load rating, and water adequacy rating. This new metric could be reported on a 1-100 scale, with thresholds established for categorizing bridges as good/fair/poor.

Overall Health

The health assessment is intended to provide a means for FHWA to examine the overall health of specific corridors and respond to requests for information. It will enable FHWA to examine health across States in a consistent manner. State DOT's may also be interested in the health assessment if they would like to know the condition of pavements and bridges in adjacent States, or if they would like to use the data to augment their agency-specific pavement and bridge data.

Several DOTs already compile and report information related to infrastructure health. While the FHWA can learn from these by reviewing existing reports and dashboards, the overall intent of this effort is to develop an approach that can be applied across States. Therefore the approach must be feasible solely with national data sets.

The vision for the health assessment has two components. The first is the ability to develop reports that summarize overall health and identify potential warning signs. The second component is a tool that enables users to review metrics and examine detailed data.

In developing the project approach, the FHWA and the project team narrowed down several options, coming to the following conclusions about the details of the health assessment:

  • Scope - The pilot study will be conducted on an Interstate corridor, but the methodology should be applicable for the entire National Highway System or any other defined roadway network for which the required data is available.
  • Scale - Data should be available by highway segment and flexible for analysis purposes. For example, FHWA would like the option to look at a segment of highway through an urban area, look at a segment of highway across an entire State, look at a segment of highway across multiple States, or examine overall network conditions of several (or all) States.
  • Timing - The methodology will focus on a current snapshot of conditions/health, and where possible anticipate near term issues.
  • Issues to address - The initial effort will focus on system condition, building on the pavement and bridge metrics used to categorize assets as good/fair/poor. It will include other data and metrics where available. For example, asset characteristics and usage may help to provide context and identify red flags. Specific factors identified for consideration include truck weight or truck type, age, output from the FHWA's Pavement Health Track Tool, and the financial demands of maintaining the asset. The health assessment methodology will enable future consideration of additional factors such as operational performance and transportation impacts.

Although potentially useful for DOT's, one concern with the proposed health approach is that they may be put in a position where they have to explain the differences between the results from FHWA's health assessment approach and their own. This issue will be explored further in Phase II of this project.

In addition to finalizing the metrics and data used for the health assessment, a key step in Phase II will be the development of an algorithm for combining these elements. For example, one option is to develop an index which assigns relative weights to each included metric. Another option is to base the assessment on independent thresholds, where any individual threshold could result in a lower overall health.

The display of the health assessment will also be critical to its success. Figures

2.2 and 2.3 represent two sample displays developed for illustrative purposes as part of Phase I.

Sample Health Display, Part 1
Figure 2.2 Sample Health Display, Part 1

Sample Health Display, Part 2
Figure 2.3 Sample Health Display, Part 2

2.3 PILOT PROGRAM

The objective of the pilot program that will be conducted in Phase II of this study is to test the methodologies described above for categorizing assets as good/fair/poor and assessing overall health. As part of the pilot, the study team will compile and compare three data sets for a multi-state, Interstate corridor: 1) data from State DOT databases, 2) HPMS and NBI data submitted to FHWA, and 3) data collected in the field as part of this effort. These data will be analyzed in order to address the following issues:

  • Consistency between the three data sets listed above. Table 2.3 illustrates how these comparisons could be reported.
  • Consistency in data between State DOTs.
  • Implications of the use of automated data collection techniques.
  • The validity of proposed methodologies, including algorithms for calculating the new measures, good/fair/poor thresholds, and the health assessment approach.
Table 2.3 Example of Testing Consistency Between Data Sources and Metrics
State DOT Data Data Submitted to FWHA Data Collected in theField by Project Team
Data Set Good Fair Poor Good Fair Poor Good Fair Poor
Metric Option #1X     X       X  
Metric Option #2  X     X   X    
Metric Option #3    X   X       X

Pilot Corridor Selection

As part of Phase I, potential pilot corridors were evaluated based on consideration of a number of criteria, including the following (more details of the corridor selection process are available in a Phase I project memorandum, dated December 29, 2010):

  1. Do not consider I-95. The FHWA has conducted a previous data compilation pilot on the I-95 corridor.
  2. Adjacency. Ideally, the three States would be adjacent. However, this was not mandatory.
  3. Data compilation expediency. To ensure adherence to the overall project schedule, a major consideration was the extent to which data is readily available. For bridges, the potential good/fair/poor approaches rely on NBI data, which is readily available from all DOTs. Therefore, the data expediency considerations focused on pavement data. The potential good/fair/poor approaches for pavements rely on HPMS 2010+ data. Therefore, the pilot States should have already submitted HPMS 2010+ files to FHWA. In addition, to minimize data compilation requirements, the pilot selection focused on corridors solely owned and operated by DOTs, as opposed to a combination of DOTs and toll authorities.
  4. Overlap with the TWG. Include a representative from each pilot State on the technical working group.

Figure 2.4 provides a geographic view of four potential pilot corridors that were considered.The I-90 corridor through Wisconsin, Minnesota, and South Dakota was selected because HPMS 2010+ data is available in all three States, and because no portion of it is operated by a tolling agency. Figure 2.5 illustrates the pilot corridor in more detail. For more details on the proposed approach to the pilot data collection and data gathering effort, refer to the Phase II work plan in Section 3.0.

Pilot Corridor Candidates
Figure 2.4 Pilot Corridor Candidates

I-90 Pilot Corridor
Figure 2.5 I-90 Pilot Corridor

PDF files can be viewed with the Acrobat® Reader®
Updated: 11/15/2012