U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations
|This report is an archived publication and may contain dated technical, contact, and link information|
|Publication Number: FHWA-HRT-12-035 Date: November 2012|
Publication Number: FHWA-HRT-12-035
Date: November 2012
State highway agencies spend billions of dollars each year on transportation infrastructure assets to meet legislative, agency, and public expectations. Pavements are a major component of those transportation assets, and pavement rehabilitation—preserving pavements to extend their service life and, more importantly, to improve motorists' safety and satisfaction—is one of the most critical, costly, and complex elements. This is especially true at a time when a large percentage of pavement networks are reaching the end of their serviceable life and pavement rehabilitation has become even more daunting given the funding constraints faced by highway agencies.
In recognition of the importance of pavement rehabilitation, Federal, State, and local policies, mandates, procedures, and initiatives are in place to help ensure pavement rehabilitation is done in a systematic, well-thought-out manner. All State transportation departments, for example, must prepare a Statewide Transportation Improvement Program (STIP), which is a multimodal, multiyear financial document listing all projects expected to be funded with Federal participation and updated periodically. Only those projects for which construction and operating funds can reasonably be expected to be available are included in STIP, and, without STIP inclusion, a project is ineligible for Federal funding.
Although the policies, procedures, mandates, and initiatives vary from one State transportation department to another, the generic approach to pavement rehabilitation decisions is similar. Decisions on pavement rehabilitation projects to be carried out in any given construction year are done centrally at the top of the department organization, where pavement rehabilitation projects compete for funding with the needs of other transportation assets, including new pavement construction. In the case of federally funded rehabilitation projects, these decisions are made as part of STIP. In addition, decisions are typically made based on work performed over the course of many years (e.g., pavement condition surveys on an annual basis or every 2 years), and those efforts enable the decisions. Moreover, these decisions generally rely on a pavement management system (PMS), which provides network-level condition scores for each pavement segment in the system and, on the basis of the score, establishes an initial action to be performed, ranging from no action to major rehabilitation or reconstruction.
After the selection of projects for pavement rehabilitation in a given construction year, the actual design of the rehabilitation is typically, but not always, turned over to districts or regions. The designs generally begin with a review of available historical data, collection of new data, and the generation of alternative rehabilitation strategies (often referred to as "pavement type selection"). The approach and level of sophistication in this design step varies from one State transportation department to another, but the principals are generally the same.
Once the various rehabilitation strategies have been defined, the final step typically entails the evaluation of those strategies and the selection of the optimal strategy using life-cycle cost analyses and other considerations. Again, the approach and level of sophistication in this evaluation step vary among State transportation departments, but the principles are generally the same.
The discussion so far has focused on the generic approach to rehabilitation decisions by State highway agencies. PMSs are at the center of those decisions. The term "PMS" was introduced in the mid-1960s to define a set of rational procedures that provide optimum pavement strategies based on predicted pavement performance, incorporating feedback regarding the various attributes, criteria, and constraints involved. Pavement management is, in essence, a coordinated systematic process for carrying out all activities related to providing pavements; it is a support tool that enables pavement engineers and managers to make better, more cost-effective decisions concerning pavement assets and their preservation. Key pavement management outcomes include the following:
Formalization of pavement decisionmaking.
Entire process to provide quality pavements.
Strong emphasis on economics.
Involvement of all associated groups—planning, design, construction, maintenance, materials, and field groups.
Use of advanced tools and analysis techniques.
The first generation of PMSs was largely driven by pavement ride quality and distress as a direct result of the American Association of State Highway Officials (AASHO) Road Test, which introduced the concept of the Present Serviceability Rating (PSR) and the Present Serviceability Index (PSI). PSR represents the subjective rating of pavement serviceability on an arbitrary scale of 0 to 5. PSI is a more objective measure established through statistical analyses that relates the PSR to various pavement physical measurements—roughness, rutting, cracking, and patching. Without question, PSI is largely driven by pavement roughness.
With advances in technology, PMS began to move away from the present serviceability concept to the use of distress (cracking, rutting, etc.) and longitudinal roughness (typically in the form of the International Roughness Index (IRI)) as key pavement performance indicators in the decisionmaking process. Although both are important indicators that merit emphasis within the PMS process, they are not the only indicators. Structural adequacy, for example, is another important pavement performance indicator that is critical to making rational pavement rehabilitation decisions. Indeed, many State highway agencies are incorporating deflection testing as part of their routine PMS activities in recognition of the need to know about the structural adequacy of their pavements. This key pavement performance parameter and others, such as surface friction and noise, are being introduced in newer PMS generations.
In 1996, the U.S. Federal Highway Administration (FHWA) conducted a survey of how highway users judge roadways. The survey identified that the most important issue for highway users is roadway condition. Furthermore, studies of the AASHO Road Test showed that subjective evaluation of roadway condition by highway users is primarily judged by pavement roughness. Clearly, ride quality is a key performance indicator to highway users. Moreover, research studies have shown that pavements that are built smooth generally have a longer service life. A study analyzing roughness trends in pavements under study by the Long-Term Pavement Performance (LTPP) program showed that with all other factors being equal, pavements follow a generally parallel trend of roughness development over time. Hence, those pavements built with better ride quality take longer to reach unacceptable levels of ride quality, and, subsequently, longer time will be required before rehabilitation.
Figure 1 shows the progression of roughness for a randomly selected set of asphalt pavements with an asphalt overlay from the LTPP database.(1) Each line in the figure represents a separate test section. As shown, test sections in the lower portion of the graph at an age of 0 years are also generally on the lower end of the graph toward the end of the timeline. The sections shown in the graph include sections from all over the United States and represent different environmental zones, subgrades, and rates of traffic. Accordingly, the progression of roughness would not be expected to be the same for all sections.
Figure 1. Graph. Progression of roughness on asphalt sections with asphalt overlay.
Prior to the development of inertial profilers, roughness data were collected using response-type roughness measuring systems, commonly referred to as roughometers (e.g., Portland Cement Association Road Meter, Mays meter, etc.). The CHLOE profilometer was used to collect roughness data in the AASHO Road Test. State highway agencies started using inertial profilers to collect network-level roughness in the 1980s. The most common measure of roughness for many years was slope variance, in large part because of the various correlations that were developed relating roughness to the American Association of State Highway and Transportation Officials (AASHTO) PSI. At present, the state of the practice in roughness data collection involves the use of inertial profilers with laser sensors, typically mounted on the front bumper at transverse locations corresponding to the wheel-paths. The roughness data shown in figure 1 were collected using this device. In addition, the index value most commonly used to represent pavement roughness is IRI.
As with roughness, nondestructive deflection testing for the structural evaluation of pavements has been used for more than half a century. Prior to the development of devices that applied a load onto the pavement surface and measured deflections at several locations, simpler devices, such as the Benkelman Beam, were used to measure the response of the pavement to a load in terms of a maximum deflection. With time and advances in technology, the pavement community started to measure multiple deflections at various radial distances from the center of the applied load (i.e., deflection basins). Since the 1980s, the most commonly used device has been the falling weight deflectometer (FWD). A number of equipment manufacturers produce or market FWDs, but all of them rely on impact loads to produce a response in the pavement similar to that produced by actual traffic loadings, which is then measured by multiple deflection sensors located at varying distances from the center of the load.
Although FWDs represent significant progress from the Benkelman Beam in terms of both quantity and quality of data gathered, FWDs are not without shortcomings. For one, they require stop-and-go rather than continuous operation. Lane closures are required, causing traffic disruptions. The amount of testing is significantly less than with continuous testing, which affects operational costs. To overcome these shortcomings, several organizations in the United States and Europe have developed devices that can continuously measure pavement deflections, such as the FHWA-funded rolling wheel deflectometer. However, FWDs presently represent the state of the practice as far as the structural capacity evaluation of pavements is concerned. Hence, further discussions in this report regarding deflections and structural adequacy are limited to FWDs.
As equipment technology has evolved, so have methods for analyzing deflection data, which allow for multiple complex algorithms to be used in real-time analysis of the deflection data. Many of the analysis techniques developed in the past, such as Boussinesq's one-layer and Burmeister's two-layer solutions, are still used today.(2,3) More complex and rational techniques have also been developed, such as finite-element solutions and dynamic analysis techniques, but because of their complexity, those procedures are most often used in research. By far, the method most often used for the analysis of deflection data is layered elastic theory, which generally works well but may have problems under certain conditions (e.g., composite pavements, thin layers, etc.). To make matters more complicated, most analysis techniques that use layered elastic solutions are heavily dependent on the user or the inputs provided. Results vary depending on the specific software used even if the inputs are all the same, and the results produced by the software require a significant amount of familiarity with the analysis technique used as well as engineering judgment.
Another factor making assessment of structural adequacy challenging is that pavement material properties change, sometimes drastically, with changes in surface and subsurface moisture and temperature conditions. Figure 2 shows changes in deflection for seven sensors at various radial distances over time at an LTPP Seasonal Monitoring Program (SMP) test section. The impact of moisture and temperature changes over time is clearly shown in this figure. Do ride quality or other performance indicators (e.g., distress and friction) change in a similar fashion?
Figure 2. Graph. Changes in deflection over time at LTPP SMP test section.
In summary, ride quality and structural adequacy are key performance indicators, but the relationship between the two is a topic of frequent and continuing discussion in the pavement community. It can be argued that ride quality is not an indicator of structural strength because a pavement with low structural capacity can exist with adequate ride quality. However, if the pavement is not properly designed (for anticipated traffic, ambient conditions, etc.), distresses will likely develop quickly, which could result in an increase in roughness. It is possible that this increase in roughness could be related to structural capacity.
Alternatively, there is the case of a distressed pavement that has a high IRI value (e.g., 150 inches/mi). If a thin overly (e.g., 1.5 inches thick) is applied properly on that pavement, it will likely reduce the IRI (e.g., to 75 inches/mi). The overlay will certainly increase the structural capacity of the pavement, but only by a little. Moreover, the overlay could fail rapidly. Hence, it can be argued that there is no relationship between ride quality and structural adequacy, but there may be a relationship between deterioration of ride quality and structural adequacy of the pavement.
Table 1 shows the pavement structure of the LTPP Specific Pavement Study (SPS)-1 (new hot mix asphalt (HMA) pavement) test sections, and figure 3 and figure 4 show the roughness progression at some of the LTPP SPS-1 projects. All sections within a project appear to have close initial IRI values, but do sections that show a high rate of IRI deterioration have low structural capacity?
|Test Section Number||HMA Thickness (inches)||Layer 2||Layer 3|
— Indicates that there is no layer 3.
DGAB = Dense-graded aggregate base.
ATB = Asphalt-treated base.
PATB = Permeable asphalt-treated base.
Figure 3. Graph. Changes in IRI over time for LTPP SPS-1 test sections in Iowa.
Figure 4. Graph. Changes in IRI over time for LTPP SPS-1 test sections in Arkansas.
Moreover, by looking at plots of performance over time, such as those shown in figure 5 through figure 11 for the LTPP SPS-1 test sections in Virginia and similar plots for both IRI and FWD deflections, it is possible to determine if there is a general relationship between changes in IRI and structural adequacy. The following figures are shown according to base material type—dense-graded aggregate base (DGAB), asphalt-treated base (ATB), and permeable asphalt-treated base (PATB).
Figure 5. Graph. Changes in IRI over time for LTPP SPS-1 test sections in Virginia (DGAB).
Figure 6. Graph. Changes in IRI over time for LTPP SPS-1 test sections in Virginia (ATB).
Figure 7. Graph. Changes in IRI over time for LTPP SPS-1 test sections in Virginia (PATB).
Figure 8. Graph. Changes in IRI over time for LTPP SPS-1 test sections in Virginia (ATB/PATB).
In the case of portland cement concrete (PCC) pavements, the IRI value can increase because of faulting. If FWD testing is performed at the center of the slab, the results could show that the pavement has sufficient structural capacity. Moreover, PCC pavements can be permanently curled, which can cause high IRI. In PCC pavements, center-slab FWD testing will also likely show that the pavement has suffcient structural capacity. PCC slabs can also have a convex shape, where the center of the slab is at a higher elevation compared to the joints. Because of the gap in the center of the slab, FWD testing may indicate that the pavement has a low structural capacity, but the IRI may be acceptable.
References to ride quality have focused almost entirely on IRI, but IRI is not intended to capture pavement profile frequencies known to induce dynamic loading from heavy trucks, which is more likely to be associated with pavement deterioration. It is thought that dynamic truck loadings cause accelerated pavement damage. The dynamic loads applied by heavy vehicles on pavements fall into two distinct frequency ranges: 1.5 to 4 Hz (sprung mass bounce, pitch, and roll vibration modes) and 8 to 15 Hz (unsprung mass bounce and roll). The sprung mass bounce motion is commonly referred to as "body bounce," and the unsprung mass bounce motion is commonly referred to as "wheel hop."
At 65 mi/h, body bounce at 2 and 2.5 Hz corresponds to spatial frequencies of 48 and 38 ft/cycle, respectively. Similarly, at 65 mi/h, axle hop at 10 and 12 Hz corresponds to spatial frequencies of 9.6 and 8 ft/cycle, respectively. On new pavements that do not have distress, dynamic loads will be mainly influenced by body bounce. As a pavement deteriorates, dynamic loads due to axle hop will occur. There is a relationship between frequency, spatial wavelength, and speed.
As previously noted, a truck traveling at 65 mi/h that has a body bounce frequency of 2.5 Hz will have a spatial frequency of 38 ft/cycle. If there is significant spectral content in the roadway near this wavelength, resonance motion can occur in the truck, which can result in high dynamic loads being applied to the pavement.
Figure 9 shows a 25-ft base length continuous IRI plot of a section of roadway. Any point on this plot shows the IRI of a 25-ft-long section that is centered at that location. For example, the IRI shown at 50 ft is the average IRI from 37.5 to 62.5 ft. The overall IRI of this roadway is 73 inches/mi.
Figure 9. Graph. Continuous IRI plot of road profile.
Figure 10 shows the power spectral density (PSD) plot of the profile data. A PSD function is a statistical representation of the importance of the various wavelengths contained in the profile. The PSD plot shows that there is significant spectral content close to a wavelength of about 30 ft. This value is close to the spatial frequency of the body bounce motion of the truck and can result in high dynamic loads being applied on the pavement.
Figure 10. Graph. PSD plot of road profile.
Figure 11 shows the dynamic loads that were predicted for the leading trailer axle by a truck simulation model. The dynamic loads predicted for a truck with air suspension and for a truck with leaf suspension are shown. This pavement section is a fairly smooth pavement with an overall IRI of 75 inches/mi. However, it had wavelengths that were close to the natural frequency of the body bounce motion of the truck. This caused high dynamic loads to be applied on the pavement. There could be other pavement sections that have an IRI of 75 inches/mi that do not have significant wavelength content close to the natural frequency of the body bounce motion of this section. On such a section, the magnitude of the dynamic loads applied on the pavement is expected to be lower. Therefore, the IRI level by itself is not a predictor of dynamic loads imparted on the pavement.
Figure 11. Graph. Dynamic loads applied to road profile.
As previously noted, if a new pavement has dominant wavelengths that will cause resonance body bounce motion that can cause high dynamic loads, pavement distress may result. Such distresses will influence the axle hop motion of the truck, resulting in accelerated distress due to both body bounce and axle hop.
IRI is influenced by wavelengths ranging from 3 to about 100 ft. So, IRI will include the wavelengths that influence dynamic truck loads. The IRI has maximum sensitivity to sinusoids with wavelengths of 7.9 and 50.5 ft. The wavelength of 7.9 ft is close to spatial wavelengths that have a significant impact on axle hop. The wavelength of 50.5 ft is somewhat close to spatial wavelengths that influence body bounce.
From the information presented thus far, it could be argued that the relationship between ride quality and structural adequacy is tenuous at best. However, the discussion has revolved almost entirely around IRI as the surrogate to ride quality, which is not intended to specifically capture pavement profile features known to induce dynamic loading from heavy trucks, which is more likely associated with pavement deterioration. In addition, no clear specific definition of a structural adequacy measure has been provided. Accordingly, a more rational look at ride quality and structural adequacy is needed.
The investigation of the relationship in question must begin with a clear understanding of pavement performance and the factors that affect it. More specifically, the separate and combined effects of the following four factors define the performance of pavements:
Pavement type—HMA, new PCC, HMA overlay over existing HMA, HMA overlay over existing PCC, PCC overlay over existing PCC, PCC overlay over existing HMA, and others (paver blocks, white topping overlay, etc.).
Pavement layers—Thicknesses, material types, material properties, drainage, shoulders, joints and steel reinforcement in PCC pavements, quality of construction and related issues, ambient conditions at time of construction, and others.
Subgrade soil—Material types, material properties, stabilization, embankment, cut/fill, depth to bedrock, drainage, and others.
Traffic—Traffic volumes (design versus actual), traffic loads/load spectra (design versus actual), traffic growth (design versus actual), seasonal trends, load restrictions, and others.
Environmental conditions—Air and surface temperatures, precipitation, wind, solar radiation, subsurface moisture, subsurface temperature, construction ambient conditions, unusual or catastrophic events, freeze-thaw cycles, freeze days, and others.
Without question, the best source of data to explore the ride quality-structural adequacy relationship is the LTPP program, which was established to provide the data necessary to explain how pavements perform and why they perform as they do. The LTPP database contains the most complete and comprehensive set of pavement performance data and associated factors, including the following:
Pavement performance information, including roughness/elevation and deflection data.
Pavement structure and subgrade soil information obtained through one or more of the following methods: test pits and coring/boring, ground penetrating radar, dynamic cone penetrometer, drainage surveys (video), field materials sampling and testing activities, laboratory materials testing, specialized testing, and other destructive and nondestructive testing (NDT) techniques.
Traffic information obtained through one or more of the following methods: automatic vehicle classifier counts, weigh-in-motion (WIM) measurements, average daily traffic, and estimated equivalent single-axle load (ESAL) estimates.
Environmental information obtained from one or more weather stations (e.g., National Climatic Data Center, Canadian Climatic Center, and LTPP-installed stations) or through the use of surface or subsurface instrumentation at the SMP test sections.
In particular, the use of data associated with the SPS-1 (new asphalt concrete (AC) pavements), SPS-2 (new PCC pavements), SPS-5 (rehabilitation of existing AC pavements), SPS-6 (rehabilitation of existing PCC pavements), and SPS-8 (study of environmental factors in the absence of heavy loads) project test sections are considered relevant. Unlike most General Pavement Study (GPS) test sections, the SPS data capture pavement performance and the factors that affect it over the entire performance life cycle.
The objective of this project was to identify and verify the relationship, if any, between ride quality and structural support or between ride deterioration and structural adequacy using LTPP and other pavement performance data sources. This was done in an effort to improve the evaluation and use of pavement condition data in pavement rehabilitation and design decisions.
More specifically, this project was intended to develop and document a mechanism to include both ride quality and structural adequacy values within the context of current network-level PMS practices for highway agency implementation. The results of the project are intended for use by pavement management engineers to ensure smooth pavements that are also structurally adequate.
To accomplish the project objective, the following three tasks were initially performed under phase I, "Identification and Demonstration of the Ride-Structure Relationship," which was intended to establish the foundation for actual development of the ride quality-structural adequacy relationship under phase II, "Guidance for Implementing the Ride-Structure Relationship into PMS":
Literature search—Available information relating ride quality and structural adequacy for pavement rehabilitation and design decisions was gathered, reviewed, and synthesized. Pertinent information was gathered through Web-based searches of State highway agencies, university pavement research centers, the Transportation Research Board (TRB), the American Society of Civil Engineers (ASCE), industry, and other national and international (e.g., World Road Association (PIARC) and AustRoads) organizations. The information gathered under this task was synthesized and presented by relevant topics to serve as the foundation for the work to be carried out under tasks 2 and 3. As part of this effort, consideration was given to how the ride quality-structural adequacy relationship and decisionmaking processes contribute to improved resource allocation through the highway agency's network-level PMS.
Data review and assessment—This effort focused on the review and assessment of relevant data from the LTPP SPS-1, SPS-2, and SPS-5 project test sections because of data completeness and because the sections capture pavement performance and the factors that affect it over entire performance life cycles. General ride quality-structural adequacy trend analyses were performed on those data to see if promising relationships could be identified. In the pursuit of relationships, both changes in ride quality or structural capacity over time and changes in ride quality or structural capacity within the test sections were considered.
Phase I report—Under this task, a draft phase I report was prepared to document the findings from the research performed under tasks 1 and 2. The report was to include the proposed work plan for the phase II effort. However, because a promising ride quality-structural adequacy relationship could not be identified, the project team recommended not proceeding with the phase II effort.
This report documents the phase I results and findings, as detailed under the task 3 summary.
The information presented in this report is organized into the following sections:
Chapter 1. Introduction—Includes project background information, the project goal and objective, and the report organization.
Chapter 2. Literature Search—Provides the results of the literature search, including the sources of information and a summary of information contained in the more relevant references.
Chapter 3. Data Review and Assessment—Details the results and findings from the review and assessment of relevant data from the LTPP SPS and GPS test sections, including the selection of test sections and the pursuit of promising ride quality-structural adequacy relationships. The primary criteria used in the selection of test sections and subsequent analyses were changes in ride quality or structural capacity over time and changes in ride quality or structural capacity within the test sections.
Chapter 4. Other Data Analysis Considerations—Presents additional data analyses performed as part of the study to validate the findings and conclusions presented in chapter 3.
Chapter 5. Summary and Conclusions—Highlights the major observations, findings, and conclusions from the phase I effort.
Four appendices contain plots generated during the data review and assessment effort as well as the data analysis validations, which are not included in the main text of the report.
The references list contains relevant references identified from the literature search or used in the preparation of this report.
The bibliography contains references identified from the literature search that were not considered of sufficient relevance to include in the project.