TRANSVERSE CRACKING PROGRESSION IN ASPHALT SHOULDERS ADJACENT TO CONCRETE PAVEMENTS

Transcription

1 TRANSVERSE CRACKING PROGRESSION IN ASPHALT SHOULDERS ADJACENT TO CONCRETE PAVEMENTS Sam Owusu-Ababio and Robert Schmitt University of Wisconsin-Platteville, U.S.A. Abstract The shoulder forms an integral part of the highway pavement system. It promotes safe traffic operations and provides lateral support for the adjacent mainline pavement when properly designed and maintained. A literature survey conducted on shoulders suggests that, the design procedures currently in use by State Highway Agencies (SHAs) have developed gradually through experience rather than from a rational pavement design approach. Consequently, most shoulders may be considerably under-designed and have resulted in unsatisfactory performance. This paper presents an examination of transverse cracking propagation on in-service asphalt shoulders adjacent to dowel-jointed plain concrete pavements in Wisconsin. On the basis of the examination, it is concluded that a minimum asphalt surface thickness of 100 mm (4 in.) combined with sealing of the longitudinal joint between the concrete pavement and the shoulder, will minimize the extent of transverse cracking. Using a crushed aggregate base course as opposed to an open-graded base course can minimize the severity of transverse cracking. In addition, using a wider asphalt shoulder and sealing of the longitudinal joint can minimize the combination of transverse cracking severity and extent. 1. Introduction and Background Maintenance staff in both the Districts and Central Office of the Wisconsin Department of Transportation (WisDOT) discussed the poor performance of the current asphalt shoulder design and standard being constructed on concrete pavement projects in Wisconsin. It was indicated that excessive transverse cracking forces maintenance crews to address these shoulders early in their life and expose the crew to continual, unwanted high-volume traffic roadways, such as the interstate system. Hence, in 2001, WisDOT initiated a study to develop a set of guidelines for consideration in paved shoulder practice in Wisconsin. As part of the study, a survey was conducted on paved shoulders adjacent to 465 centerline-km (289 centerline-miles) of concrete pavements in Wisconsin. The pavements included: continuously reinforced concrete, dowelled and non-dowelled jointed plain concrete (JPC). The objective of this paper is to present an analysis of transverse cracking progression in asphalt shoulders adjacent to the doweljointed plain concrete pavements.

2 2. Field Survey of In-Service Paved Shoulders Field survey of in-service paved shoulders involved the identification and selection of shoulders to be surveyed as well as the distress types to be measured and how to quantify them. With the aid of a Geographic Information System (GIS), maps showing the general locations of all rural concrete pavement projects completed over the last 30 years in Wisconsin were identified. These maps were then used in tracking all as-built plans for the identified projects in the various WisDOT district offices having jurisdiction over the identified projects. A total of 82 projects were identified for the dowelled JPC. A statistical summary of the shoulder characteristics is shown in Table 1. Table 1. Statistical Characteristics of Asphalt Shoulders Adjacent to Dowelled JPC Shoulder Property Range Mean Std. deviation Surface age (years) Surface thickness mm.(in.) (3-4) 3.24 (82.30) 0.43 (10.92) Asphalt Width m (ft.) (6-9) 1.86 (6.20) 0.20 (0.67) Soil Support Value, SSV Adjacent Mainline PCC thickness, mm (in.) (9-12) (10.25) 22.9 (0.9) 2.1. Transverse Cracking Measurement The majority of distress types that occur on mainline pavements are similar to those occurring on shoulders. For asphalt shoulders, the following distresses were considered: cracking (alligator, block, longitudinal, transverse), patching, outside edge raveling, longitudinal joint deterioration, settlement, and heave. Transverse cracking was the main form of cracking that appeared on the asphalt shoulders that were observed in the field. The measurement descriptions are as provided in Table 2. Figures 1, 2, and 3 show various severity levels for the transverse cracking measured in the field on asphalt shoulders. In Figure 1, a severity level 1 transverse crack developed as a continuation of the PCC joint from the mainline. The measured crack width was less than 13 mm (½-inch) with slight spalling at the edges. In Figure 2, a severity level 2 transverse crack exhibiting some spalling at the edges is shown, while Figure 3 shows a severity level 3 transverse crack with some dislodgement of pavement material and band cracking appearing close to the longitudinal joint. Figure 1: Asphalt Shoulder Transverse Cracking Severity Level 1

3 Table 2: Measurement Description of Transverse Cracking on Asphalt Shoulders SEVERITY 0 = None; 1 = less than 13-mm (1/2-inch) in width 2 = greater than 13-mm (1/2-inch) in width 3 = band cracking. A transverse crack is banded if the pavement area affected is within 0.305m (one foot) of the crack. EXTENT The extent of Transverse Cracking is determined from the average number of transverse cracks per station in the survey segment. A transverse crack length should be at least 25% of the shoulder width to be counted. 0 = None; 1 = 1 to 5 cracks per 30-m (station) 2 = 6 to 10 cracks per 30-m (station) 3 = greater than 10 cracks per 30-m (station) Figure 2: Asphalt Shoulder Transverse Cracking Severity Level 2 Figure 3: Asphalt shoulder Transverse Cracking Severity Level 3 3. Paved Shoulder Performance Various researchers 1,2,3 have attributed the problems associated with asphalt concrete shoulders to be related mostly to structural performance. Insufficient asphalt concrete shoulder thickness or heavy traffic loading tends to create excessive stresses in the shoulder, especially when the foundation materials have softened under the mainlineshoulder area due to water infiltration and/or freeze-thaw cycles. This increased deformation will produce accelerated fatigue cracking in both the outside edge of the mainline pavement and in the paved shoulder. Poor drainage conditions can also

4 accelerate the development of this distress. Field observations have shown that shoulder distress location is primarily within a 0.61-m (2-foot) distance of the longitudinal mainline pavement-shoulder joint. Barksdale and Hicks 3 reported the occurrence of transverse cracking at locations where the transverse pavement joints intersect the asphalt shoulder. The transverse cracking developed in this manner, appears to be related to the expansion and contraction of the PCC pavement. The difference in material properties and the effects of weather on those properties create inconsistent movements and stresses resulting in transverse cracking Performance Indices Traditionally, SHAs combine measured distresses to yield a single index for describing pavement performance. WisDOT uses the Pavement Distress Index (PDI), which is a mathematical expression for pavement condition rating based on observable surface distresses. It reflects the composite effects of various distress types on the condition of a pavement segment. 4 Distresses are assigned a value, based upon type of distress, severity, and extent. Distress values are summed up for a pavement, with a value of 0 reflecting no distresses, rising to a maximum value of 100. There are, however, major concerns associated with the use of combined indices such as the PDI, Pavement Serviceability Index (PSI), Pavement Condition Index (PCI) to indicate performance. These problems have been outlined by Paterson 5 and include: a) Different types of maintenance are appropriate for different levels of each distress type. b) The relative seriousness of different defects varies with the pavement type, environment, the rate of deterioration and the maintenance program in place. c) Each distress type evolves at different rates in different pavement types and under different traffic and environmental conditions. The problems outlined by Paterson 5 suggest that modeling the performance of shoulders using a combined index, such as the PDI, requires determining the average amount of distress effects from the many different combinations of distresses encountered on the shoulder. This method has the potential to yield results that have wide variances that, in turn, may suppress the very effects of interest. The modeling approach adopted in this study, therefore, is a shift from the combined index (PDI) approach to a more versatile approach in which major distress modes are individually modeled to better analyze and explain the relationship between distress progression and their influential factors Transverse Cracking Modeling The analysis of transverse cracking consisted of examining the factors believed to influence three characteristics of transverse cracking, namely the extent, severity, and the combination of severity and extent. A Shoulder Distress Index Factor (SDIF) denoted the combination of severity and extent. The SDIF is analogous to the PDI factors established for transverse cracking by WisDOT. The SDIF values for transverse cracking are as shown in Table 3 and are determined as a function of the extent and severity. Descriptions for the extent and severity have already been presented in Table 2. The lower the SDIF value, the poorer the shoulder surface in terms of transverse cracking.

5 Table 3. Shoulder Distress Index Factor for Transverse Cracking Transverse Cracking Extent Transverse Cracking Severity (No. per Station) Level 1 Level 2 Level The modeling process consisted of two phases: a preliminary phase and a model-building phase. The former phase used analysis of variance (ANOVA), scatter plots, and correlations to identify key input variables believed to have an effect on the extent, severity, and SDIF values for transverse cracking. The extent provided information on the frequency of occurrence while the severity indicated the seriousness of the distress. From a designer s point of view, the influential factors for the extent and severity can provide a basis for design modifications. The combination of the severity and extent was also needed for determining the type and level of maintenance work to be performed, and consequently, aid in the life cycle cost analysis associated with specific maintenance alternatives. A summary of analysis of variance conducted on the main factors considered having influence on the three transverse cracking characteristics are summarized in Table 4 and Figures 4 through 8. Table 4 suggests that transverse cracking was highly influenced by the shoulder surface age and quality of subgrade, which was denoted by the soil support value (SSV). The extent reduced with a thicker surface (Figure 4) as well as sealing of the longitudinal joint between the shoulder and the adjacent concrete (Figure 5). The severity was highly influenced by the type of base material used under the shoulder. Crushed aggregate base courses minimized the severity compared to opengraded bases (Figure 6). Using a wider shoulder and sealing the longitudinal joint can also reduce the combination of severity and extent (Figures 7 and 8). Table 4. ANOVA Summary of Transverse Cracking Progression Factors. Independent Variables Dependent Variable Shoulder Base Gradation Shoulder Width Shoulder Surface Thickness SSV Outside Lane PCC Thickness Age Longitudinal.Joint Seal Extent n/s n/s XX XXX X XXX XX Seveity XXX XX n/s XXX n/s XXX n/s SDIF X XX n/s XXX n/s XXX XX XXX = Highly Significant, p-value 0.01 XX = Moderately Significant, 0.01 < p-value < 0.05 X = Marginally Significant, 0.05 p-value 0.1 n/s = Not Significant, p-value 0.1 Total number of observations = 150

6 1.05 Transverse Cracking Extent Shoulder Surface Thickness (in.) Figure 4: Transverse Cracking Extent Variation with Shoulder Thickness Transverse Cracking Extent Longitudinal Joint Seal (1=Sealant present, 0= No sealant) Figure 5: Transverse Cracking Extent Variation with Longitudinal Joint Treatment Transverse Cracking Severity Base Material Type (1=Crushed Aggregate Base, 0= Open-graded base) Figure 6: Transverse Cracking Severity Variation with Base Type

7 Transverse Cracking Index Factor AC Shoulder Width (ft) Figure 7: Transverse Cracking Index Variation with Asphalt Shoulder Width Transverse Cracking Index Factor Longitudinal Joint Seal Presence (0 =No, 1 =Yes) Figure 8: Transverse Cracking Index Variation with Longitudinal Joint Treatment The latter phase of the model building consisted of developing quantitative relationships among inputs and resulting performance output, as measured by the SDIF. The overall statistical model developed for transverse cracking is presented in Equation 1. The progression of transverse cracking was dictated by a combination of the age of the shoulder surface and its width. Wider widths minimized the combination of severity and extent while the progression increased with age. SDIF TC = W *Age (1) Model R 2 =50.3%, degree of freedom =150, F-ratio =73.4, and P-value = Where, SDIF TC = transverse cracking extent and severity level combination Age = asphalt shoulder surface age in years W = asphalt shoulder width in meters.

8 3.3. Engineering Implications of Transverse Cracking Model The transverse cracking model developed for the asphalt shoulders adjacent to dowelled JPC pavements could be used to determine the timing and, consequently, the type of maintenance or rehabilitation intervention based on critical levels of severity and extent combination as determined by agency policy. Table 5 shows the expected age in years to reach specific combinations of severity and extent for specified shoulder widths. The severity and extent levels have been previously described in Table 2. Table 5: Expected Age to Reach Specified Combination of Severity and Extent AC Extent and Severity Level Combinations* Width(m) S1E1 S1E2 S1E3 S2E1 S2E2 S2E3 S3E1 S3E2 S3E *SiEi =Severity level extent and I level i. 4. Summary and Conclusions This paper examined the progression of transverse cracking for in-service asphalt shoulders adjacent to dowelled JPC pavements in Wisconsin. The analysis suggests that the main factors affecting the progression of transverse cracking extent and severity combination are the width and the surface age of the shoulder. Using a wider shoulder can minimize the combination of extent and severity. Sealing the longitudinal joint between the shoulder and the adjacent JPC pavement and using a thicker surfacing can minimize the extent alone. Using a crushed aggregate base course, as opposed to an open-graded base course, can minimize the severity of transverse cracking. 5. References 1. Benekohal, Rahim F., Kathleen T. Hall and Harlan W. Miller. Effect of Lane Widening on Lateral Distribution of Truck Wheels, In Transportation Research Record 1286, pgs , Hicks R.G., Richard D. Barksdale and Donald K. Emery. Design Practices for Paved Shoulders, In Transportation Research Record 594, pgs , Barksdale, R.D., and R. G. Hicks. Improved Pavement-Shoulder Joint Design, NCHRP Report 202, Wisconsin Department of Transportation; Pavement Surface Distress Survey Manual, Division of Highways-Central Office Materials, Pavement Management Section, Madison, WI., Paterson, W.D.O. Road Deterioration and Maintenance Effects: Models for Planning and Management, John Hopkins University Press, Baltimore, Maryland, 1987.