U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
202-366-4000


Skip to content
Facebook iconYouTube iconTwitter iconFlickr iconLinkedInInstagram

Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations

 
REPORT
This report is an archived publication and may contain dated technical, contact, and link information
Back to Publication List        
Publication Number:  FHWA-HRT-12-035    Date:  November 2012
Publication Number: FHWA-HRT-12-035
Date: November 2012

 

Relating Ride Quality and Structural Adequacy for Pavement Rehabilitation/Design Decisions

CHAPTER 2. LITERATURE SEARCH

The objective of the literature search was to collect, review, and synthesize available information relating ride quality and structural adequacy for pavement rehabilitation and design decisions. Pertinent information was gathered through Web-based searches of State highway agencies, university pavement research centers, TRB, ASCE, industry, and other national and international (e.g., PIARC and AustRoads) organizations. A particularly relevant source of information was the references database prepared under the LTPP program and posted on the program’s Web site (http://www.ltpp.org). In addition to looking at previous attempts to establish such a relationship, the following topics were also considered during the literature search:

The key observations, findings, and conclusions from the literature search are detailed in this chapter.

2.1 OVERVIEW

The purpose of the literature review was to identify, review, and synthesize information for eventual use in the planning and execution of an effort to look for the potential relationship between ride quality and structural adequacy of pavement structures. A total of 62 references were identified from various sources and reviewed. Of those, 16 references were considered of particular value to the project. They have been synthesized and included in the references list of this report. The remaining 47 references were not considered sufficiently relevant and have been included in the bibliography section.

Each of the 16 relevant references was reviewed by identifying the reference type, source, objectives and goals, findings and observations, relevance to the project, and other pertinent information. Table 2 shows the distribution of these references by type and source. As shown, the majority (62 percent) of references were from studies related to the LTPP program. Although the LTPP program has been categorized separately, it could be considered together with the transportation departments since all LTPP test sections are on State highways. Table 2 also shows that the bulk of references were articles (69 percent). Five reports (31 percent) relevant to the current study were reviewed in addition to the articles.

Table 2. Distribution of references by type and source.

Type State Transportation
Department/ University
LTPP Program Total
Report/
guideline
3 2 5
(31 percent)
Article/
presentation
3 8 11
(69 percent)
Total 6
(38 percent)
10
(62 percent)
16
(100 percent)

Table 3 shows the distribution of elements in the references that are relevant to this study on relating ride quality to structural adequacy for rehabilitation and design decisions. The values shown in this table indicate the number of references of the total (16) that contained a topic related to the current study. As expected, many of the references covered more than one topic. The majority of these references relate to factors influencing pavement performance (43 percent). Understanding the factors influencing pavement performance was considered an important first step in the effort to establish a relationship between ride quality and structural adequacy. Only two promising studies were identified in the literature search that directly relate to the research topic at hand. One article on the topic of structural and functional evaluation approaches was reviewed, and six references that looked at relationships between ride and distress were reviewed. Figure 12 shows the distribution of the topics considered essential for evaluating a relationship between ride and structural adequacy.

Table 3. Distribution of references by type and topic.

Type Relationship Between
Ride and Structural Adequacy
Factors Influencing Pavement Performance Structural- and Functional- Based
Approaches for Pavement Evaluations
Relationship Between
Ride and Distress
Total
Report/
guideline
2 1 0 2 5
(31 percent)
Article/
presentation
0 6 1 4 11
(69 percent)
Total 2 (13 percent) 7 (43 percent) 1 (6 percent) 6 (38 percent) 16
(100 percent)

This graph shows the dynamic loads predicted for the leading trailer axle by a truck simulation model. Load is on the y-axis ranging from 5,000 to 12,000 lbf, and distance is on the x-axis ranging from 0 to 425 ft. Predicted dynamic loads for air suspension and leaf suspension are shown in the graph; a separate line is provided for each. There is also a solid horizontal line at 8,500 lbf, which represents the static axle load. The pavement section in question is fairly smooth, with an overall International Roughness Index of 75 inches/mi, but it has wavelengths that are close to the natural frequency of the body bounce motion of the truck, causing high dynamic loads to be applied to the pavement.

Figure 12. Graph. Distribution of references.

 

2.2 SPECIFIC FINDINGS

Two studies relate directly to this study—relating ride quality to structural adequacy for rehabilitation and design considerations. The first was an FHWA study conducted by Von Quintus et al. on LTPP data that looked at the dissipated work obtained from deflection time history data and pavement condition data (distresses, IRI, etc.) to see if a relationship exists between the dissipated work and pavement performance.(4) In this study, dissipated work was defined as the area under the loaded and unloaded portion of the stress-strain curve (hysteresis loop) and was used to define the viscoelastic and inelastic properties of pavement material. Several LTPP sites with varying IRI values, distress magnitudes, and traffic levels were analyzed to determine if a relationship existed between the dissipated work and pavement performance. The dissipated work and hysteresis loop were found to vary extensively by structure and pavement type. Based on the sections evaluated in the study, lower amounts of dissipated work were observed on PCC-surfaced pavements than on AC-surfaced pavements. The findings from the study indicate that the greater the dissipated work, the greater the amount of pavement distress observed in terms of magnitude, severity, and type. The conclusions from this study indicate that dissipated work can be used to determine the performance behavior of pavement structures.

The second study was performed by Zhang et al. and looked at the structural adequacy of pavements in implementing pavement management decisions.(5) The study evaluated various methods to look at the structural condition of the roadway prior to applying thin overlays to correct ride. The study indicated that the Texas Department of Transportation (TxDOT) pavement management information system (PMIS) database shows a yearly decrease of 0.3 points in ride score for the highway network in Texas as a result of the increase in pavement roughness attributed to permanent deformation. The permanent deformation was a result of inadequate pavement strength for existing traffic loads. The study was aimed at determining a structural index using the FWD data to evaluate the structural adequacy prior to planning maintenance.

In the process of determining a suitable methodology for characterizing structural estimators, the TxDOT study looked at the deterioration process of the pavement. As the deterioration process represents the behavior of a non-linear system, it can be characterized by different rates of deterioration in different stages of the pavement service life. The true condition of a pavement at any moment can be described more accurately if the deterioration rate is known. Unfortunately, a mathematical solution to this problem is impossible because there are no models that can precisely represent the true deterioration process of a pavement. Even complicated mathematical models such as sigmoid forms cannot be calibrated accurately enough to represent a true deterioration process.

The TxDOT study points out that if the general mathematical formula describing the transition of a system from one state to another is not available, one can use the finite difference between the states. The study presents a unique way of characterizing the condition deterioration (ride and distress). It also proposes a method to characterize deterioration in terms of differences between condition measures over a unit time period, normalized to the initial condition, to give a more accurate picture of pavement deterioration. Furthermore, the study points out the importance of considering the traffic applied when determining the pavement deterioration. An equal yearly drop in the condition score of a pavement subjected to different ESALs represents different structural conditions of the pavement. The normalized drop in condition divided by the ESALs for the year, called the “unit ESAL deterioration” (UED), provides a more accurate condition of the pavement. Structural failure occurs when the roughness starts progressing at a high rate, and the UED concept helps describe the deterioration rate. In the study’s analysis the UED is calculated as the normalized yearly drop in the PMIS scores (ride, condition, and distress scores) caused by a single ESAL for a consecutive two-year period, as shown in figure 13 to figure 15.

UED open parenthesis Distress Score closed parenthesis equals open parenthesis dDS divided by open parenthesis DS times ESAL subscript y closed parenthesis closed parenthesis times 10 raised to the 6th power.

Figure 13. Equation. UED distress score.

UED open parenthesis Condition Score closed parenthesis equals open parenthesis dCS divided by open parenthesis CS times ESAL subscript y closed parenthesis closed parenthesis times 10 raised to the 6th power.

Figure 14. Equation. UED condition score.

UED open parenthesis Ride Score closed parenthesis equals open parenthesis dRS divided by open parenthesis RS times ESAL subscript y closed parenthesis closed parenthesis times 10 raised to the 6th power.

Figure 15. Equation. UED ride score.

Where:

DS = Distress score in initial year.

RS = Ride score in initial year.

CS = Condition score in initial year.

dDS = Yearly drop in distress score.

dCS = Yearly drop in condition score.

dRS = Yearly drop in ride score.

ESALy = Estimated amount of ESALs in a year.

The results from the TxDOT study indicate that different factors, such as dRS, dCS/CS, and UED(RS), show different levels of sensitivity to the structural estimators, but the outputs from all considered methods indicate some level of sensitivity to the UED values. As deterioration increases, the structural estimators tend to produce smaller values. The factor defined as UED(RS) showed the best trend for all methods (structural estimators considered). The intent of the trend analyses was to visualize whether there was a trend between the deterioration variables (e.g., UED of the PMIS score values) and the pavement structural estimators (e.g., structural number (SN)). However, it was not the intention of the study to quantify the correlation between them through regression analysis or other means. An example of the trend found between SN and UED(RS) is shown in figure 16. The study also puts forth the concept of a structural condition index, a ratio of the effective SN (SNeff) to the required SN (SNreqd), as a measure of structural adequacy.

This graph shows an example of the trend between pavement structural number (SN) and unit equivalent single-axle load (ESAL) deterioration based on the unit ESAL deterioration of ride score (UED(RS)) in the Texas Department of Transportation study based on hundreds of data points. SN is on the y-axis ranging from  4 to 10 (only three of hundreds of points are below 0), and UED(RS) is on the x-axis ranging from 0 to 90. Between 0 and 10 UED(RS), the SN for the data points ranges between 1 and 5, and beyond a UED(RS) of 10, the SN for the data points ranges mostly between 1 and 2.

Figure 16. Graph. Sensitivity of SN to UED of ride score.

In order to investigate a potential relationship between ride quality and structural adequacy, it is important to understand the factors that influence pavement performance. Seven of the reviewed references identify factors influencing pavement performance. The findings from these studies are summarized in table 4.

Table 4. Summary of factors influencing pavement performance.

Pavement Surface Type Performance Indicator Most Influencing Factors Least Influencing Factors Factors Not Influencing Reference No.
Hot mix AC (HMAC) Overlay roughness Overlay thickness, climatic zones Subgrade type 6
HMAC Overlay roughness IRI prior to overlay, overlay thickness, and milling prior to overlay 7
Jointed plain concrete (JPC) Cracking and roughness Base type, pavement drainage, and slab thickness 8
HMAC Roughness Base type, drainage, and climatic conditions Base thickness and subgrade type 9
HMAC Roughness, rutting, fatigue cracking, and transverse cracking 10
HMAC Roughness and rutting Construction 11
JPC Roughness and faulting Dowels 12
Jointed reinforced concrete Roughness Moisture in subgrade, joint spacing, thicker slabs, and concrete modulus 12
Continuously reinforced concrete Roughness 12

— Indicates that the reference did not apply.

A study by Vepa et al. looked at structural- and functional-based approaches for pavement evaluations.(13) In this study, approaches including the 1993 AASHTO procedure based on NDT data, a pavement condition rating procedure, and a survivor curve method were used to evaluate the remaining life of flexible and rigid pavements. The conclusions from the study indicate that structural failure-based remaining life calculations result in conservative estimates compared to functional failure-based methodologies.

Finally, in an effort to identify an indirect relationship between ride and structural adequacy, studies relating ride and distress were reviewed. Six references were identified. The results from four of these studies are summarized in table 5. The other two references included models to relate IRI and distress and are discussed in detail in the following paragraphs.

Table 5. Relationship between pavement performance measures.

Pavement Type Performance Measures Statistically Significant Relationship Strength of Relationship Reference No.
HMAC and HMAC over PCC IRI and distress Yes Weak 14
HMAC and HMAC over PCC IRI and rutting Yes Weak 14
HMAC IRI and Pavement Condition Index Yes Weak 15
HMAC IRI and initial IRI, fatigue, and rutting Yes Strong, but mainly influenced by initial IRI 16
HMAC IRI and distress Yes Strong 17

In the Mechanistic-Empirical Pavement Design Guide (MEPDG), the functional adequacy for both flexible and rigid pavements is quantified by pavement smoothness.(18) The parameter used to define pavement smoothness in the MEPDG is IRI.

In the MEPDG, models are available to empirically predict the IRI at any point in the life of the pavement by adding the increase in IRI of the pavement due to pavement distress and a site factor to a known initial IRI of the pavement. The site factor accounts for the increase in roughness due to the shrink/swell and frost heave characteristics of the subgrade. Distress prediction models are used to predict the distresses that are used as an input to the roughness models.

The MEPDG presents models for predicting the IRI for the following pavement types:

The following is a brief description of the distress types and site factors included in the roughness prediction models for each of the pavement types:

As previously described, in all of the models presented in the MEPDG for predicting roughness, age is included in the site factor, which accounts for the increase in roughness that occurs over time due to the frost heave and shrink/swell characteristics of the subgrade.

The three distress types considered in the roughness prediction model for new HMA pavements and flexible overlay of HMA pavements are fatigue cracking, transverse cracking, and rut depth. The amount of fatigue cracking in the pavement should have a strong influence on the structural strength of the pavement as determined from FWD testing conducted along the wheel path. However, transverse cracks are unlikely to impact the structural strength of the pavement as determined from the FWD testing if the FWD test is carried out away from a crack location. Rut depths may or may not affect the structural strength of the pavement as determined from FWD testing. If the rutting is caused by a weak base or subbase layer, the structural strength of the pavement will be affected. However, rutting caused by the distortion of the HMA layer may not result in a loss of structural capacity of the pavement. Rutting caused by weak subgrade, base, or subbase layers can result in early fatigue cracking of the pavement.

The three distress types considered in the roughness prediction model for JPCP pavements are transverse cracks, joint spalling, and faulting. FWD tests to evaluate the structural capacity of JPCP pavements are conducted at the center of the slab. Transverse cracks present on the slab will affect the structural capacity of a JPCP slab as determined from FWD testing. However, joint spalling and faulting are not related to the structural strength of the concrete pavement. Spalling may be caused by concrete durability issues as well as by low load transfer at joints. Faulting is also related to low load transfer at joints and can be exacerbated by the presence of erodible bases.

The only distress type influencing the roughness of CRCP pavements is punchouts, which are load-related distresses.

In the MEPDG roughness models, some of the distresses that result in an increase in IRI will result in a decrease in the structural strength of the pavement. However, some distresses that contribute to the increase in roughness may not necessarily cause a decrease in the structural strength of the pavement. The MEPDG models account for the increase in roughness due to shrink/swell and frost heave characteristics of the subgrade. Changes in the pavement profile caused by shrink/swell and frost heave of the subgrade may not directly impact the structural strength of a pavement. However, these actions could initiate pavement distress, which would have an impact on roughness.

Lastly, the World Bank has developed software called Highway Development and Management (HDM) System to make comparative cost estimates and economic evaluation of different policy options.(19) The HDM-III model was developed using data collected from a multiyear empirical study carried out in Brazil. The statistical relationships in this model were validated and extended using data from several other deterioration studies carried out in various locations such as Kenya, the Caribbean, India, and Texas. The revised and improved models are included in the latest version, HDM-4. HDM-4 includes a model for predicting roughness of roadways. The roughness prediction model predicts the incremental change in roughness during an analysis year based on contributions to roughness from five sources: (1) change in roughness due to structural deterioration, (2) cracking, (3) rutting, (4) potholes, and (5) environmental effects. Models are provided to predict the incremental change in roughness due to each of the factors. The HDM-4 model has been widely used in many countries with calibration to suit local conditions.

2.3 SUMMARY

Although a few studies were identified during the literature search that indicated a potential relationship between structural response and pavement performance, none of these studies established a direct relationship between ride quality and structural adequacy. As discussed in this chapter, several studies have been successful at relating ride quality to pavement condition or distress. In fact, the roughness models in the MEPDG use distress as an input to predict roughness. Similarly, structural response has been used to predict certain distresses. While it is evident that a multitude of factors influence pavement performance in different ways (i.e., ride condition and structural response), a strong correlation between the three different performance measures has yet to be established such that measurement of any one performance indicator is adequate to identify both the structural and functional condition of the pavement.

 

Federal Highway Administration | 1200 New Jersey Avenue, SE | Washington, DC 20590 | 202-366-4000
Turner-Fairbank Highway Research Center | 6300 Georgetown Pike | McLean, VA | 22101