Key Findings From LTPP Analysis 2000-2003

STRUCTURAL FEATURES

Producing an effective pavement design is a complex process. The obvious decisions–asphalt or concrete, and how thick–are critical, but there are other equally important decisions regarding other structural features that can assist with a pavement design. The following are key findings from several LTPP analyses on the effects of structural features on pavement performance. These findings are grouped into two areas: rigid and flexible pavements.

1. Rigid Pavements
   • Slab Thickness
   • Slab Widening
   • Base
   • Subgrade
   • Subdrainage
   • Joint Faulting
   • Load Transfer Efficiency
   • Joint Spacing
   • Joint Type
   • CRCP
   • Variability of PCC Pavement Design Parameters
   •  
2. Flexible Pavements
   • Layer Thickness
   • Base
   • Subgrade
   • Subdrainage
   • Variability of AC Pavement Design Parameters
   •  
3. Guide to Pavement Layer Thickness Data

Rigid Pavements

1. Slab Thickness
   Report No. FHWA-RD-01-167
   In the SPS-2 experiment:
   •  
   • Thinner slabs (203 mm (8 inches)) develop more transverse cracking than do thicker slabs (279 mm (11 inches)). These data support a similar finding from earlier studies that slab thickness has a strong effect on transverse cracking.
   •  
   • Thinner slabs (203 mm (8 inches)) have more longitudinal cracks than do thicker slabs (279 mm (11 inches)). Sections with a thinner slab and widened slab show the highest level of longitudinal cracking.
   •  
   • Thinner slabs (203 mm (8 inches)) with lower (3.8 megapascals (MPa) (550 poundforce per square inch (psi)) 14-day strengths were found to be smoother than thicker slabs (279 mm (11 inches)) with higher (6.2 MPa (900 psi)) 14-day strengths.
   •  
2. Slab Widening
   Report Nos. FHWA-RD-00-076 and FHWA-RD-01-167
   For jointed concrete pavements (JCP) with nondoweled joints, the PCC slabs originally built wider by 0.6 m (2 ft) (i.e., 4.26 m (14 ft) in total) as compared to conventional-width (3.66 m (12 ft)) slabs appear smoother and show reduced joint faulting. The reduction in joint faulting is achieved by moving wheel loads further away from the corner of the slab, reducing the frequency of traffic encroachment to the slab corner or pavement edge. This confirms a similar finding from a previous study. pavement performance. These findings are grouped into two areas: rigid and flexible pavements.
3.  
4. Base Report No. FHWA-RD-01-167
   •  
   • In the SPS-2 experiment, the sections with permeable asphalt-treated base developed the fewest transverse and longitudinal cracks, whereas the sections with lean concrete base developed the most transverse and longitudinal cracks over the first 10 years of pavement life.
   •  
   • This confirms a similar finding from previous studies. In the SPS-2 experiment, those sections built over a permeable asphalt-treated base with edge drains or aggregate base were found to be smoother than those over lean concrete base.
   •  
5. Subgrade
   Report No. FHWA-RD-01-167
   The SPS-2 JCP without dowels built over fine-grained subgrades are rougher after construction than those built over coarse-grained subgrades. Stiffer foundation appears to be a key factor.
6.  
7. Subdrainage
   Report No. NCHRP Project 1-34B
   •  
   • Based on 7 years (on average) of SPS-2 data, for JCP with dowels, both permeable bases and edge drains were not found to reduce joint faulting significantly.
   •  
   Report No. NCHRP Project 1-34C
   In the SPS-2 experiment:
   •  
   • For IRI, transverse cracking, and longitudinal cracking with all other factors matched:
     •  
     • Pavement sections with undrained dense-graded aggregate bases tend to perform more poorly than those with drained permeable asphalt-treated bases.
     •  
     • Pavement sections with undrained lean concrete bases tend to perform more poorly than those with drained permeable asphalt-treated bases.
     •  
   • The faulting levels were too low for analysis.
   •  
8. Joint Faulting
   Report No. FHWA-RD-00-076
   •  
   • Based on the faulting data collected from 307 test sections in the SPS-2, SPS-4, SPS-6, SPS-8, GPS-3, GPS-4, and GPS-9 experiments, the presence of dowels was found to be the most effective of the design features examined in controlling joint faulting. When doweled joints are used, the effects of design features such as subdrainage, tied-concrete shoulders, and joint spacing are not as significant. The analysis supports a similar finding from a previous study that JCP with dowels had 50 percent less joint faulting than those without dowels.
   •  
   • Widened lanes showed significant reduction in joint faulting for both doweled and nondoweled joints.
   •  
   • Effective subdrainage designs reduce faulting for all types of pavements and designs, especially for nondoweled sections.
   •  
   • For nondoweled JPCP, the following design features were found to significantly reduce faulting:
     •  
     • Use of widened lanes.
     • Effective subdrainage system.
     • Stabilized base/subbase.
     • Shorter joint spacing.
     •  
   • The analysis results confirm similar findings from previous studies reviewed in the Key Findings from LTPP Analysis 1990 ­1999 brochure.
   •  
   • The use of load transfer devices (circular steel dowels) has the greatest effect on the amount of joint faulting for all concrete pavements. Sections with dowels having larger diameters (e.g., 38 mm (1.5 inches)) exhibited much less faulting. The use of properly sized dowel bars reduces joint faulting of JPCP by more than a factor of two.
   •  
   • The nondoweled JPCP sections located in the colder and wetter climates exhibit the worst faulting among all sections examined. This also confirms a similar finding from previous studies.
   •  
   • In the SPS-2 experiment:
     • Doweled joint faulting occurred mostly in the dryfreeze climates, followed by the dry no-freeze climates, and the wet-freeze climates. Sections in the wet no­freeze climates have the least faulting.
     •  
     • Sections with an unbound aggregate base have the highest doweled joint faulting level. Sections with a lean concrete base and permeable asphalt-treated base have the lowest doweled joint faulting.
     •  
     • Widened (4.27 m (14 ft)) slab sections have less faulting than conventional-width (3.66 m (12 ft)) slabs.
     •  
   Report No. FHWA-RD-01-167
   
   • The data support similiar findings from other studies
     
     
9. Load Transfer Efficiency
   •  
   • Report No. FHWA-RD-02-088
     • Poor correlation was found between LTE and design parameters such as PCC thickness, PCC strength, steel content, joint spacing, and joint orientation.
     •  
   • Report No. FHWA-RD-96-198
     
     • The presence of properly sized dowels at the joint will eliminate corner breaks and transverse cracking near the joint as well as minimize joint faulting.
     •  
   • Report No. NCHRP 20-50(5)
     
     • JCP with nondoweled joints:
       

       The variability of the average LTE of a given section measured over time is inversely correlated to the average LTE. As the average section LTE increases, the variability decreases.

       The average joint spacing, base type, and outside shoulder type (PCC or AC) show no effect on the variability of the average LTE measured over time.

       The average LTE of pavements with subsurface drainage systems has higher variability measured over time than that of pavements without subsurface drainage systems.

       The average LTE of pavements with a granular subgrade has higher variability than pavements with a silty clay subgrade.

       The amount of annual precipitation, the number of annual freeze-thaw cycles, and the average mean annual temperature do not have any effect on the variability of the average LTE measurements over time. However, the variability of the average LTE seems to decrease as the annual freezing index increases.

       There is no direct relationship between pavement age and the variability of LTE measurements over time.

     • JCP with doweled joints:
       

       The variability of the average LTE measured over time is inversely correlated to the average LTE. As the average LTE increases, the variability decreases.

       The average joint spacing and base type do not show any effect on the variability of the average LTE measured over time.

       The average LTE of pavements with a concrete shoulder has higher variability than that of pavements with an asphalt shoulder.

       The average LTE of pavements with subsurface drainage systems has lower variability than that of pavements without subsurface drainage systems.

       The amount of annual precipitation, the number of annual freeze-thaw cycles, and the average mean annual temperature do not show any effect on the variability of the average LTE measured over time. However, the variability of the average LTE seems to decrease as the annual freezing index increases.

       There is no direct relationship between pavement age and the variability of LTE measurements over time.

     • CRCP:

       No apparent relationship was found between the variability of the mean transverse crack LTE measured over time and the mean crack LTE. Ranges of both crack LTE and its variability are very small. This is expected because the transverse cracks are strongly reinforced and changes in crack width and LTE are minimal.

       The average crack spacing of the CRCP slabs, base type, slab stiffness, and outside shoulder type does not have any effect on the variability of the average LTE measured over time.

       The average crack LTE of CRCP with subsurface drainage systems has lower variability than those of pavements without subsurface drainage systems.

       The amount of annual precipitation does not show any effect on the variability of the average crack LTE. However, the variability of the average LTE seems to decrease as the annual freezing index increases.

       No direct relationship between pavement age and the variability of crack LTE was observed. This is expected because an increase in variation would indicate a deterioration of transverse cracks in CRCP, which would lead to rapid failure.
       

10. Joint Spacing
    Report No. FHWA-RD-00-076
    
    •  
    • For JRCP sections in good condition, the average joint spacing is about 13 m (43 ft). For JRCP sections in poor/normal condition, the spacing is approximately 18 m (59 ft). The LTPP data shows that the joint spacing of sections in good condition is significantly shorter. This confirms a similar finding from previous studies reviewed in the Key Findings from LTPP Analysis 1990­1999 brochure.

      Report No. FHWA­RD­96­198

    •  
    • Thermal gradients, moisture gradients, and built-in construction curling are important considerations in overall slab support. These considerations should be used along with traffic loadings and slab thickness in the selection of appropriate joint spacing for JCP with nondoweled joints.
    •  
11. Joint Type
    
    •  
    • Large-diameter dowel bars reduce joint faulting more than do small-diameter dowel bars. For example, pavements with 38-mm (1.5-inch) diameter dowels have very little faulting regardless of other design features.
    •  
    • Results show that doweled joints do not need to be skewed to control faulting and provide reinforcement. Skewed joints are primarily introduced to minimize the impact slab curling and joint faulting have on vehicles. This provides reinforcement to the same finding from a previous study.
    •  
12. CRCP Report No. NCHRP 20-50(8/13)
    CRCP has the potential to provide long-term, smooth, and low maintenance service life, as evidenced by many well-performing sections in the GPS-5 experiment (existing CRCP).
    •  
    • Steel content was concluded to be the most important factor in CRCP performance. This is consistent with previous findings.
    •  
    • The CRCP sections in the GPS­5 experiment have shown little change in roughness over the monitored period, and many of the pavements are very old.
    •  
    • The parameters with higher parameter values found to contribute to higher levels of roughness in CRCP include:
    •  
    • For both wet-freeze and wet no-freeze weather zones: the PCC elastic modulus and the ratio between the PCC elastic modulus and tensile strength.
    •  
    • For the wet-freeze weather zone: the content of fines in the subgrade, the moisture content in the subgrade, and the number of wet days per year.
    •  
    • For the wet no-freeze weather zone: the number of days per year greater than 32 °C (90 °F).
    •  
13. The parameters with higher parameter values found to contribute to lower levels of roughness in CRCP pavements include:
    •  
    • For both wet-freeze and wet no-freeze weather zones: the PCC water-cement ratio.
      
    •  
    • For the wet-freeze weather zone: none of the factors studied was found to contribute to lower levels of roughness.
    •  
    • For the wet no-freeze weather zone: the number of wet days per year.
    •  
14. Variablility of PCC Pavement Design Parameters
    Report No. NCHRP 20-50(5)
    To help pavement engineers incorporate the variability of PCC pavement parameters into design, NCHRP Project 20-50(5), Variations in Pavement Design Inputs, presents variability recommendations for four key laboratory concrete strengths and four backcalculated PCC pavement design parameters, respectively.
    •  
    • Variability of concrete strengths (laboratory)

      Compressive strength data variability: For the GPS sections, the variability of the 7-day and 28-day strengths was found to be similar. For the SPS test sections, it was found that the variability was also independent of test age (7 days, 21 days, and 1 year) and specimen type (cylinders versus cores).

      Flexural strength data variability: For both the GPS and SPS sections, the variability of flexural strength appears to be independent of age at time of testing.

      Split-tensile strength data variability: Only SPS-2 data were available for analysis. The variability of split-tensile strength data was found to be independent of age at testing.

      Modulus of elasticity data variability: Only SPS-2 data were available for analysis. The variability of concrete modulus of elasticity was found to be independent of age at time of testing.

      The recommended magnitudes of acceptable variability for the above four laboratory concrete strength measurements are available in the report for pavement design consideration.

    •  
    • Variability of backcalcultated moduli data for PCC pavements
      •  
      • The variability associated with four backcalculated parameters for PCC pavements was found to be relatively low and consistent from one testing time to another over a period of several years. The four backcalculated parameters investigated were: Modulus of subgrade reaction.
        •  
        • Modulus of subgrade reaction.
        • Concrete modulus of elasticity-dense liquid foundation.
        • Subgrade modulus of elasticity-elastic solid foundation.
        • Concrete modulus of elasticity-elastic solid foundation.
        •  
      • The recommended magnitudes of variability for the above four backcalculated PCC pavement design parameters are available in the report for design consideration.

Flexible Pavements

1. Layer Thickness
   Report No. FHWA-RD-01-166
   •  
   • The SPS-1 test sections with thick (178-mm (7-inch)) AC surface layers appear to be smoother and develop less fatigue cracking than those sections with thin (102-mm (4-inch)) surface layers. This confirms a similar finding from earlier studies.
   •  
   • In the SPS-1 experiment, AC surface thickness and the age of the project appear to influence the amount of fatigue cracking that occurs. The test sections that are younger and have thicker AC surface layers have the least fatigue cracking.
   •  
   Report No. NCHRP 20-50(5)
   For pavement designers to predict that a given layer thickness will be constructed at least as thick as was assumed in design, several layer thickness adjustment recommendations are available in NCHRP Project 20-50(5), Variations in Pavement Design Inputs, for the following layer materials falling within qualified thickness ranges:
   •  
   • Unbound granular base layers from 100 to 360 mm (4 to 14 inches).
   •  
   • Treated (with either portland, asphalt, or lime cement in small quantities) base layers from 100 to 150 mm (4 to 6 inches) and 180 to 250 mm (7 to 10 inches).
   •  
   • Asphalt-bound layers (surface AC and asphalt-treated base) from 100 to 180 mm (4 to 7 inches) and 200 to 300 mm (8 to 12 inches).
   •  
   • AC overlays from 50 to 130 mm (2 to 5 inches) over existing AC.
   •  
   • AC overlays around 100 mm (4 inches) over existing PCC.
   •  
2. Base
   Report No. FHWA-RD-01-166
   In the SPS-1 experiment:
   •  
   • Hot-mix asphalt (HMA) pavements with unbound aggregate base layers show greater rut depths than those sections with asphalt-treated base layers. This suggests that a portion of the rutting measured at the surface is a result of permanent deformations in the unbound aggregate base layer, which is consistent with a previous finding from analysis of the GPS test sections.
   •  
   • The HMA pavements with unbound aggregate layers have slightly more fatigue cracking and higher IRI values than those sections with asphalt-treated base layers.
   •  
   • The test sections with coarse-grained soils, asphalt-treated base layers, permeable base layers, thicker bases, and thicker HMA layers were found to be smoother.
   •  
   • The test sections with permeable asphalt-treated base layers exhibit more fatigue cracking than those without permeable base layers.
   •  
   Report No. NCHRP 20-50(8/13)
   GPS-2 (AC pavements on stabilized base):
   •  
   • For the asphalt-treated bases in the GPS-2 projects, a higher rate of increase of IRI is noted for sections with high AC void ratios (percentage of air voids per unit AC volume).
   •  
   • For the cement-treated bases in the GPS-2 projects, a higher rate of increase of IRI is observed for sections in warmer climates.
   •  
3. Subgrade
   Report No. FHWA-RD-01-166
   In the SPS-1 experiment:
   •  
   • HMA pavements built over coarse-grained subgrade soils are smoother than pavements built over fine-grained subgrade soils. This is consistent with the finding in the SPS-2 JPCP: A stiffer foundation contributes to smoother pavements.
   •  
   • HMA pavements built over coarse-grained subgrade soils and in a no-freeze climate are smoother and stay smoother over a longer period of time than do those built over fine-grained subgrade soils in a freeze climate. HMA pavements built over fine-grained subgrades and in a wet-freeze climate are substantially rougher than those built in other climates.
   •  
   • HMA pavements built over fine-grained subgrade soils have more fatigue cracking than those projects built over coarse-grained subgrade soils.
   •  
   • Subgrade soil type and, to a lesser degree, age are important to the amount of transverse cracking measured at each site. More transverse cracking has occurred on the HMA pavements built on fine-grained soils than on pavements built on coarse-grained soils.
   •  
4. Subdrainage
   Report No. NCHRP Project 1-34B
   •  
   • Based on 7 years (on average) of SPS-1 data, those HMA sections built on permeable bases without edge drains were found to perform better than those with edge drains.
   •  
   Report No. NCHRP Project 1-34C
   In the SPS-1 experiment:
   •  
   • In terms of IRI and cracking with all other design features matched, from poor to good performance: undrained dense-graded aggregate bases, drained permeable asphalt-treated bases, and undrained dense-graded asphalt-treated bases.
   •  
   • In terms of rutting, the results for the above three subdrainage designs are inconclusive, thus far.
     
     
   •  
5. Variability of AC Pavement Design Parameters
   Report No. NCHRP 20-50(5)
   To help pavement engineers incorporate the variability of AC pavement parameters into design, NCHRP Project 20-50(5), Variations in Pavement Design Inputs, presents variability recommendations for backcalculated surface AC layer and for subgrade modulus of elasticity.
   •  
   • The variability of the backcalculated AC modulus is independent of the falling weight deflectometer (FWD) drop height.
   •  
   • The data analysis indicates that the variability of the backcalculated subgrade modulus is not related to the environmental zone in which the pavements are located.
   •  
   • AC layers, as well as granular base layers, have less variability in backcalculated moduli than other underlying layers.
   •  
   • The variability associated with AC layer backcalculated modulus is primarily dependent on the season in which the FWD measurements are taken. Layers underlying the AC layer in AC-surfaced pavements display no discernable patterns in relation to the season.
   •  

Guide To Pavement Layer Thickness Data

Report Nos. FHWA-RD-03-040 and FHWA-RD-03-041 A user's guide has been developed to provide guidance for selecting layer material and thickness data from the LTPP database. The LTPP database contains extensive information for pavement layer material type and thickness (as-designed versus as-constructed) for both rigid and flexible pavements. Such information is very important for many types of analyses including backcalculation of layer moduli, mechanistic analysis of pavement structures, and performance modeling. Layer thickness variability and the comparisons of thickness design versus constructed values for various pavement layer types are also available in the user's guide.

Previous | Table of Contents | Next