关 键 词：

水泥 稳定 砂砾 路基 基层 配合 研究

资源描述：

水泥稳定砂砾路基基层配合比研究,水泥,稳定,砂砾,路基,基层,配合,研究

内容简介：

Long-Term Performance of Stone Interlayer PavementHani Titi, P.E.1; Masood Rasoulian, P.E.2; Mark Martinez3; Byron Becnel, P.E.4; and Gary Keel5Abstract: This paper presents an evaluation of long-term performance of an alternative flexible pavement design referred to here asstone interlayer pavement. This pavement design was introduced to reduce/defer reflective cracking experienced with soil-cement bases.The stone interlayer pavement consisted of a crushed limestone base on top of a cement-stabilized base. The performance of the stoneinterlayer pavement was compared to that of the conventional pavement design with a cement-stabilized base. The stone interlayer andconventional pavements were constructed on State Highway LA-97 near Jennings, Louisiana. Both pavements were monitored for 10.2years after construction. During the evaluation period, pavement distress surveys, testing of roughness and permanent deformation, andevaluation of pavement structural capacity using dynamic nondestructive testing were conducted. Additionally, as a part of the LouisianaTransportation Research Center accelerated pavement testing research program, both pavement designs were tested to failure under theAccelerated Loading Facility in Port Allen, Louisiana. The results of this investigation showed a superior performance of the stoneinterlayer pavement over the conventional soil cement pavement as tested in the field as well as under accelerated loading.DOI: 10.1061/ASCE!0733-947X2003!129:2118!CE Database keywords: Flexible pavements; Stones; Pavement design; Performance evaluation.IntroductionSoft and saturated subgrade soils are a common occurrence inLouisiana, especially in the southern part of the state. Such sub-grade soils are not competent to support pavements and theirtraffic loading. To overcome this problem, the Louisiana Depart-ment of Transportation and Development DOTD! has adopted aconventional pavement design method for flexible pavements onstate highways other than the interstate system. The design con-sists of lime treatment of subgrade soil subbase layer!, in-placecement stabilization of soil soil-cement base layer!, and a hotmix asphalt concrete HMAC! surface layer. This pavement de-sign method, with a strong soil-cement base, has the advantage ofproducing pavements that are structurally capable of supportingtraffic loading under weak subgrade support. In addition, the con-struction of pavements with soil-cement base layers is quick andcost effective. While the use of soil-cement base layers effectivelyimproves the structural capacity of flexible pavements, it is themajor cause of reflective cracking, which accelerates pavementdeterioration and decreases pavement life.Researchers at the Louisiana Transportation Research CenterLTRC! have focused their effort on innovative alternative meth-odologies to reduce reflective cracking and improve the long-termperformance of flexible pavements in Louisiana. Among thesemethods is the use of granular materials such as crushed lime-stone! between the soil-cement base and the HMAC surface layer.This pavement type is denoted herein as the stone interlayer pave-ment. This paper describes a research effort conducted to evaluatethe long-term performance of stone interlayer pavement designand to assess the capability of the stone interlayer to reduce re-flective cracking in flexible pavements. Two flexible pavementsections were designed and constructed on State Highway LA-97near Jennings, Louisiana, in 1991. The first section consisted ofconventional pavement design soil-cement base!, and the secondsection consisted of stone interlayer pavement design crushedlimestone over soil-cement base!. The latter design is also re-ferred to as the inverted pavement design. The performance of thetwo pavements was monitored over a period of 10.2 years. Pave-ment distress surveys, evaluation of structural capacity of thepavements, measurement of permanent deformation, and evalua-tion of pavement roughness and serviceability were conducted. Inaddition, the conventional and the alternative stone interlayerpavements were tested under Accelerated Loading Facility ALF!at the Pavement Research Facility PRF! site in Port Allen, Loui-siana.BackgroundSoil-cement has long been used as engineered material in variousapplications including base layers in pavements with weak sub-grade soil. In addition, the use of soil-cement is cost effective inareas lacking aggregate resources. These conditions make south-ern Louisiana a perfect candidate for pavements with soil-cementbases. Indeed, Louisiana has thousands of highway miles withsoil-cement bases, some of which have been in service for morethan 40 years. The use of soil-cement base course layers effec-tively improved the structural capacity of pavements built on1Assistant Professor, Dept. of Civil Engineering and Mechanics, Univ.of Wisconsin-Milwaukee, P.O. Box 784, Milwaukee, WI 53201.2Pavement Systems Specialist, Louisiana Transportation ResearchCenter, 4101 Gourrier Ave., Baton Rouge, LA 70808.3Pavement Research Engineer in Training, Louisiana TransportationResearch Center, 4101 Gourrier Ave., Baton Rouge, LA 70808.4Traffic Operations Engineer, Louisiana Dept. of Transportation andDevelopment, P.O. Box 831, Baton Rouge, LA 70821.5Engineering Technician, Louisiana Transportation Research Center,4101 Gourrier Ave., Baton Rouge, LA 70808.Note. Discussion open until August 1, 2003. Separate discussionsmust be submitted for individual papers. To extend the closing date byone month, a written request must be filed with the ASCE ManagingEditor. The manuscript for this paper was submitted for review and pos-sible publication on December 3, 2001; approved on April 9, 2002. Thispaper is part of the Journal of Transportation Engineering, Vol. 129,No.2,March1,2003.ASCE,ISSN0733-947X/2003/2-118126/$18.00.118 / JOURNAL OF TRANSPORTATION ENGINEERING ASCE / MARCH/APRIL 2003weak subgrade soils and minimized localized failure and perma-nent deformation. However, there is a cracking problem associ-ated with the use of soil-cement bases on flexible pavements.The behavior of soil-cement mixture during hydration varies,depending on many factors including soil type, mineralogicalcomposition, moisture content, density, and cement content.These factors control the development of shrinkage crackingwithin the soil-cement base during hydration. Shrinkage cracking,resulting from tensile stresses during hydration of soil-cementmixtures, usually extends to the pavement surface in the form ofreflective block cracks. Fig. 1a! shows the mechanism in whichshrinkage cracking is reflected to the pavement surface. Sincereflective cracks are extended from the soil-cement base to thesurface, the pavement will be vulnerable to deterioration. Rainfallinfiltration through cracks combined with repeated traffic loadingwill cause washout of underlying materials through a pumpingaction. Loss of pavement base support usually results in rapidpavement deterioration in terms of localized failure zones, settle-ment of pavement surface, cracks, and potholes. These pavementconditions shorten the life of the roadway. Therefore, it is neces-sary to improve the flexible pavement design to reduce/defer thepropagation of soil-cement cracks into the surface layer.Research efforts at LTRC were concentrated on economicalalternative pavement designs to reduce reflective cracking andimprove the long-term performance of flexible pavements inLouisiana. The introduction of an interlayer between the soil-cement base and the HMAC surface layer to absorb tensileshrinkage stresses is among other options used. As shown in Fig.1b!, tensile stresses developed within the soil-cement mass willbe absorbed by the stone particles through their relative move-ment. The magnitude of tensile stresses will be drastically mini-mized at the interface between the stone interlayer and theHMAC surface. This is expected to reduce the amount of reflec-tive cracking.Objective and ScopeThe objective of this research is to determine the effectiveness ofusing the stone interlayer pavement design by comparing thelong-term performance of the stone interlayer pavement designwith that of the conventional pavement design. The scope of thisresearch was to monitor the performance of the conventional andthe stone interlayer pavements constructed on LA-97 by measur-ing pavement distresses and evaluting pavement conditions over aperiod of 10.2 years. The pavement distresses measured werepavement cracking and rutting. The quantity, severity, and pat-terns of cracking were examined. The pavement conditions mea-sured were structural capacity and roughness. The results of thisresearch were compared with the results of full-scale acceleratedloading experiments conducted at the PRF site.MethodologyProject DescriptionA research study was initiated to investigate the effect of acrushed limestone interlayer in effectively reducing/deferringcracks from reflecting to the flexible pavement surface. The planwas to establish two test sections to compare the performance ofthe stone interlayer pavement design and the conventional pave-ment design. The research was conducted on State Route LA-97near Jennings, Louisiana. The LA-97 project extends from Junc-tion LA-100 to Junction LA-1123. Location of the project onhighway LA-97 is shown in Fig. 2. This portion of highwayLA-97 is classified as a rural collector highway with an averagedaily traffic ADT! of 2,000 vehicles in 1990. LA-97 is a lowvolume rural highway with two 7.3 m 24 ft! travel lanes.The project consisted of new construction of 7.56 km 4.7 mi!flexible pavement highway with two 7.3 m 24 ft! travel lanes.The two test sections selected on the project are test section 1,which is the conventional pavement soil-cement base! design,and test section 2, which is the limestone interlayer design. Eachsection is 322 m 1,056 ft! long. Fig. 3a! shows the two testsections and the corresponding station numbers.The original design Fig. 3b!# was to construct a 216 mm 8.5in.! in-place cement stabilized base course layer and 89 mm 3.5in.! HMAC surface layer. A change of plan was initiated to inves-tigate the effect of adding a crushed limestone base interlayer tominimize reflective cracks. Therefore, a 1.61 km 1.0 mi! sectionof the highway was constructed with a 152 mm 6 in.! in-placecement stabilized base course layer, a 102 mm 4 in.! crushedlimestone interlayer, and an 89 mm 3.5! HMAC surface layer.The pavement cross section of limestone interlayer pavement isshown in Fig. 3c!. Both pavement designs were constructed on305 mm 12 in.! lime-treated subgrade soil subbase layer! toprovide stability and support for the pavement structure.Long-Term Pavement Performance MonitoringProgramPavement Distress SurveyThe Pavement Systems Research Group at LTRC conducted pe-riodic pavement distress surveys over a 10.2 year period Table1!. The surveys consisted of visually inspecting the pavementsections and recording the pavement distresses and any other un-usual observations within the pavement structure. Surveyed pave-ment distresses included longitudinal and transverse cracks. InFig. 1. Mechanism of development of reflective cracking in conven-tional pavements with soil-cement base and role of stone interlayer instopping reflective cracking development: a! development of shrink-age cracking within soil-cement base and subsequent propagation toHMAC surface layer: b! introducing limestone interlayer creates me-dium that separates soil-cement base and HMAC surface layerJOURNAL OF TRANSPORTATION ENGINEERING ASCE / MARCH/APRIL 2003 / 119addition, pavement sections were surveyed for raveling, shoving,and potholes. The severity level and patterns for cracks and otherdistresses were determined according to the Distress IdentificationManual for the Long-Term Pavement Performance Project SHRP1993!.The pavement surface cracking surveys were conducted bymapping the longitudinal and transverse cracks on a paper. Thedistance from the beginning to the end of the crack was measuredto determine the length of the crack.Nondestructive Evaluation of Pavement StructureAs presented in Table 1, nondestructive pavement testing NDT!and evaluation of the test sections was conducted six times overthe past 10.2 years. The NDT consisted of measurements of pave-ment deflection due to induced dynamic load using the DynamicDeflection Determination System Dynaflect!. This system Fig.4! is a nondestructive testing device that induces dynamic load onthe pavement surface and measures the corresponding deforma-tion at different locations. The maximum dynamic load, about4.448 kN 1,000 lb!, is induced due to countereccentric rotation oftwo masses at frequency of 8 Hz. The load is transmitted to thepavement through two steel wheels. The deformation is recordedby a set of five geophones installed on a beam and spaced at 305mm 12 in.!, with the first geophone placed between the two steelwheels.Kinchen and Temple 1980! developed a mechanistic approachfor design of HMAC overlays based on deflection analysis. Themethodology was proposed and verified based on comprehensiveevaluation of the structural capacity of Louisiana pavementsusing the Dynaflect system. For more than 20 years, DOTD hasbeen using this methodology for pavement evaluation and design.This method was used in the current study to evaluate the struc-tural capacity of the investigated pavements.A series of nondestructive testing was conducted during thedifferent stages of the pavement construction. This was to evalu-ate the structural capacity of the individual pavement layers. Thefirst set of nondestructive tests was conducted on the 305 mm 12in.! lime-stabilized subbase for both test sections 1 and 2. Thenthe second set of testing was conducted after construction of 216mm 8.5 in.! soil-cement for test section 1 and 152 mm 6 in.! soilcement for test section 2. The third set was conducted only on testsection 2 after the construction of the 102 mm 4 in.! crushedlimestone interlayer. The final set was conducted after construc-tion of the 89 mm 3.5 in.! HMAC surface layer. Tests wereconducted on each section on 30 m 100 ft! intervals.Evaluation of Pavement Roughness and DeformationField tests were conducted to determine the ride quality and per-manent deformation rutting! of the investigated pavements. Aspresented in Table 1, field testing was conducted seven timesduring the last 10.2 years. From 1991 to 1995, the Mays RideMeter MRM! was used to evaluate pavement roughness and theAASHTO A-frame was used to measure pavement rutting. After1995, the high-speed road profiler was used to evaluate pavementroughness and rutting.Adescription of the test equipment is givenbelow.Mays Ride Meter MRMThe MRM is equipment that measures road roughness consistingof a recorder, a photoelectric transmitter, and a special odometer.It is designed to operate in a vehicle with a traveling speed of 80km/h 50 mi/h!. MRM records travel distance from the actualroad profile on continuous feed chart paper. Based on the relativemovement of the car axle with respect to the vehicle body, thephotocell transmitter converts the accepted light rays into electri-cal impulses that are converted into profile on the chart. Specialcharts are developed for interpreting road roughness using MRMtest results. The MRM test results are expressed in terms of thePresent Serviceability Index PSI!.Field testing using the MRM was conducted five times on theentire test section to ensure repeatability of test results. The MRMtest results for the test sections were expressed in terms of PSIand then were converted into the International Roughness IndexIRI! using the correlation established by Sayers et al. 1986!.High-Speed Road ProfilerThe high-speed road profiler is a vehicle equipped with three lasersensors for height measurements, one sensor to measure traveldistance, and two accelerometers to account for vehicle vibra-tions. The accelerometers allow the system to function indepen-dently of the vehicle characteristics and travel speed. The systemis capable of collecting road profile data at a minimum distanceinterval of 76.2 mm 3.0 in!.The LTRC high-speed road profiler was used to collect pave-ment surface data for evaluation of roughness and rutting. Longi-tudinal profiles of the test sections were measured by conducting35 runs on the entire test section to ensure repeatability of thetest results. The profile of each wheel path was obtained, and thepavement roughness was determined in terms of IRI. The averageFig. 2. Location of project on State Highway LA-97 near Jennings,Louisiana120 / JOURNAL OF TRANSPORTATION ENGINEERING ASCE / MARCH/APRIL 2003IRI from both wheel paths was used to evaluate each pavementtype.The mean rut depth was also determined from the laser sen-sors height measurements. Verification of the rut depth was alsomade using the AASHTO A-frame. The mean rut depth of theentire test section the outside wheel path! was used in the analy-sis and evaluation of the test sections.Accelerated Pavement TestingLTRC has established the PRF test site in Port Allen, Louisiana,to bridge the gap between laboratory characterization and fieldperformance of pavements under heavy loading conditions. Thesite houses the Louisiana ALF, the second of its kind in theUnited States. Shown in Fig. 5,ALF is a 30.5 m 100 ft! long, 489Fig. 3. Location and configuration of conventional and stone interlayer pavement test sections: a! pavement test sections and areas of cracksampling; b! typical section A-A! of conventional pavement soil-cement base!; c! typical section B-B! of stone interlayer pavementTable 1.Long-Term Performance Monitoring ScheduleTest datemonth and year!Time after ConstructionPavement Testingand MonitoringMonthYearDistresssurveyNDTaRoughnessand rutting03/199110.08AA08/199160.50A12/1991100.83AA10/1992201.67AA03/1993252.08A09/1993312.58A02/1995484.00AA05/1995514.25A09/1996675.58AA05/1998877.25A07/1998897.42A04/200112210.17AAAaNDT5nondestructive testing.Fig. 4.Picture of Dynamic Deflection Determination SystemDynaflect!JOURNAL OF TRANSPORTATION ENGINEERING ASCE / MARCH/APRIL 2003 / 121kN 55 ton! accelerated loading device used to simulate and ac-celerate truck loads for pavement testing. The ALF device is atransportable, linear, full-scale accelerated loading facility, whichimposes a rolling wheel load on a 1.2312 m (4339 ft) area testpavement. Loading is applied in one direction only, at a constantspeed of 16.7 km/h 10.4 mi/h!. Each loading cycle takes eightseconds and is applied through a standard dual-tire truck capableof applying loads between 43 kN 9,750 lb! and 84 kN 18,950lb!. This indicates that for each pass 1.3819.7 equivalent singleaxle loads ESALs! are applied to the pavement. This allows theALF to traffic a test pavement at up to 8,100 wheel passes11,200160,000 ESALs! per day. The ALF device, by increasingthe magnitude of the load and running the device for 24 hours aday, can compress many years of road wear into just a fewmonths of testing.The PRF site is located on six acres of soft saturated heavyclay with 84% clay and 14% silt. A 1.52-m-thick embankmentwas constructed of silty clay with 23% clay and 70% silt to in-vestigate the performance of flexible pavements under acceleratedloading conditions. Nine full-sized test lanes were constructed in1995 and subsequently tested to failure under accelerated loading.This was the first Louisiana accelerated pavement-testing experi-ment, entitled Construction and Comparison of LouisianasConventional and Alternative Base Courses under AcceleratedLoading Metcalf et al. 1998!. The purpose of this experimentwas to evaluate alternative cement base courses with reduced re-flective cracking but no loss of structural capacity. In order toverify the field performance of the alternative stone interlayerpavement design, two of the nine test lanes were constructed asstone interlayer and conventional soil-cement pavements and thenwere tested to failure under accelerated loading conditions. Thetwo test lanes designated 008 and 009! had the same surfacecourse and base course materials as that of the stone interlayerand conventional soil-cement pavements constructed at highwayLA-97.Results and DiscussionEvaluation of pavement performance is a difficult task due to thecomplex nature of the variables controlling the behavior of pave-ments. Among these factors are the pavement material character-istics, properties of the subgrade soil, pavement construction pro-cess, environmental conditions, and traffic loadings. In thissection, the long-term performance of the investigated pavementswill be analyzed with consideration of these factors.Traffic Loading and Environmental ConditionsThe research project was conducted on State Highway LA-97,which is classified as a rural collector highway with an ADT of2,000 vehicles in 1990. It is important to base the analysis on fielddata in order to make valid comparisons and conclusions. There-fore, historical ADT data were obtained from DOTD files andanalyzed to evaluate the long-term performance of the investi-gated pavement types. As shown in Fig. 6a!, the actual ADT datawas used to predict the future ADT using simple linear regressionanalysis. The AASHTO 1993! design guide indicated that trafficon some minor arterial or collector-type highways tends to in-crease along a straight line.The traffic volume and loading over a period of 30 years wasused to determine the corresponding ESALs. As shown in Fig.6b!, the pavement loading in terms of ESALs was determinedevery year during the period. This data are important to estimatethe remaining life of the pavement and to make valid comparisonswith the accelerated loading tests conducted at the PRF test siteon a similar pavement type.The project is located in the southern part of Louisiana. Farmsand crawfish ponds are located along highway LA-97. Ditchesaround the highway are usually full of water, indicating a satu-rated subgrade soil. In the project area, the average air tempera-ture variations from low to high range between 21.1 and 37.8C30 and 100F). The average pavement surface temperature canbe as high as 62.8C (145F) during summer. The average annualrainfall is about 1,524 mm 60 in.!.Fig. 5. Picture of Accelerated Loading Facility ALF! located atLTRC Pavement Research Facility test site, Port Allen, LouisianaFig. 6. Traffic loading and volume characteristics for State HighwayLA-97: a! prediction of growth of traffic volume from actual averagedaily traffic for highway LA-97; b! growth in traffic loadingexpressed in ESALs for LA-97 over period of 30 years.122 / JOURNAL OF TRANSPORTATION ENGINEERING ASCE / MARCH/APRIL 2003Evaluation of Pavement PerformancePavement CrackingPavement distress surveys conducted over a period of 10.2 yearsshowed no evidence of shoving, raveling, or potholes on bothinvestigated pavements. However, cracking was developed withinboth investigated pavements. Fig. 7 presents the results of crack-ing surveys conducted over the 10.2 year period of the pavementmonitoring program. As shown in Fig. 7a!, it took four years fora total of 4.27 m 14 ft! of low severity level cracking to developwithin the stone interlayer pavement. This is as compared to154.53 m 507 ft! of low and medium severity level crackingdeveloped within the conventional pavement. Inspection of Fig. 7shows that there is a progressive increase in cracking with time inboth pavement types, but with a drastically lower rate in the caseof the stone interlayer pavement. The pattern of cracking withinthe conventional pavement area was 65% transverse and 35%longitudinal. The pattern of cracking within the stone interlayerpavement area was 36% transverse and 64% longitudinal.The latest pavement cracking survey, conducted 10.2 yearsafter construction, showed that the cracking quantity and severitylevels developed within the conventional pavement soil-cement!base are greater than that developed within the stone interlayerpavement. The total cracking length developed within the sur-veyed panels of the conventional pavement is 232.81 m 764 ft!,in which 151.8 m 498 ft! has a low severity level, 64 m 210 ft!has a medium severity level, and 17.1 m 56 ft! has a high sever-ity level. The stone interlayer pavement showed less cracking,with a total length of 118.3 m 388 ft! in which 94.5 m 310 ft!has a low severity level and 23.8 m 78 ft! has a medium severitylevel. No cracking of high severity level was observed within thestone interlayer pavement.The cracking density is used as a criterion to evaluate pave-ment deterioration and failure. The cracking density of 4.92 m/m2(1.5 ft/ft2) is used as a failure criterion for pavements testedunder accelerated conditions. Fig. 7b!, shows the variation ofcracking density versus traffic loading on the investigated pave-ments. The cracking density for the stone interlayer pavementafter 10.2 years of service is 0.177 m/m2(0.053 ft/ft2), which isonly 3.6% of the failure criterion limit. This corresponds to atraffic loading of 296,667 ESALs. The conventional pavementcracking density after 10.2 yeas is 0.348 m/m2(0.106 ft/ft2),which is 7.1% of the failure criterion limit.Acracking map for the conventional and stone interlayer pave-ments is depicted in Fig. 8. The results are from the latest pave-ment distress survey conducted in April 2001. The figure showsthat the stone interlayer base was very effective in reducingcracks from reflecting to the pavement surface. Cracking patternsand characteristics suggest that these are block reflective cracks.There is little evidence of fatigue cracking within the investigatedpavement panels.Pavement distress surveys showed that the stone interlayerbase effectively reduced/deferred the starting time as well as therate of reflective cracking development.Structural CapacityIt is essential to design pavements that are capable of supportingtraffic loading under certain environmental conditions. The struc-tural number SN! is a measure of the pavement capacity to sup-port traffic loading. Design of flexible pavements is usually con-ducted following the AASHTO 1993! design guide procedure. Inthis procedure, input parameters are required to determine thestructural number of the pavement. These input parameters in-clude the estimated traffic, initial and terminal serviceability,overall standard deviation, reliability level, and resilient modulusFig. 7. Results of distress surveys conducted on conventional andstone interlayer pavements 10.2 years after construction: a! totalcracking developed within both pavement types b! cracking densityversus traffic loadingFig. 8. Cracking of conventional and stone interlayer pavementsafter 10.2 years from construction most cracking is low severity, Mdenotes medium severity, S denotes high severity according to SHRP1993! manual#: a! panel #1 section #1! from stations 73136 to74136; b! panel #1 section #2! from stations 30196 to 31196JOURNAL OF TRANSPORTATION ENGINEERING ASCE / MARCH/APRIL 2003 / 123of subgrade soil. The thickness of the different pavement layerscan then be determined by satisfying the following equation:SN5a1D11(i52naiDimi(1)where SN5structural number; ai5structural layer coefficient;Di5layer thickness; and mi5layer drainage coefficient. Theseinput parameters depend on the material characteristics and varydepending on local experience. The structural layer coefficientsused by DOTD for the materials used in both pavements are asfollows: 0.01654/mm 0.42/in.! for the HMAC surface layer,0.00551/mm 0.14/in.! for the soil-cement base course layer,0.00551/mm 0.14/in.! for the crushed limestone base courselayer, and 0.00276/mm 0.07/in.! for the lime-treated subgradesoil. The drainage coefficient was assumed 1.0 for all layers. Eq.1! was used to calculate the design structural number for bothpavement types. The design SN for the conventional pavement is3.86 and the design SN for the stone interlayer pavement is 4.07.The results of the nondestructive testing and the evaluation ofthe pavements since the beginning of the pavement constructionare summarized in Table 2. These values represent the mean foreach pavement type. Fig. 9 presents the results of long-term non-destructive testing on the investigated pavements after construc-tion of the HMAC surface layer. The method developed byKinchen and Temple 1980! was used to determine the structuralnumber and resilient modulus of the subgrade soil from NDTresults. The structural number estimated from nondestructive test-ing SNNDT! ranges from 4.2 to 7.1 for the conventional pave-ment and from 4.1 to 7.2 for the stone interlayer pavement. Theminimum SN determined from NDT evaluation is higher than thedesign SN calculated using Eq. 1! for both pavement types. Therelationship of mean SNNDT! to traffic loading is depicted inFig. 9. On average, the conventional and the stone interlayerpavements showed similar structural capacities.The resilient modulus of the subgrade soil was also evaluatedusing NDT. For the 10.2 year evaluation period, the resilientmodulus varied within the same section ranging from 57.9 to172.4 MPa 8.4 to 25 ksi! for the conventional pavement andfrom 61.4 to 172.4 MPa 8.9 to 25 ksi! for the stone interlayerpavement. These results show the consistent subgrade conditionsand uniformity of lime-treatment construction in both pavementtypes. The variation of the average resilient modulus of subgradeFig. 9. Comparison of structural capacity performance of conven-tional and stone interlayer pavements over 10.2-year periodTable 2.Nondestructive Evaluation of Pavement Layers During and After ConstructionTest sectionTest dateTested pavementlayer surfaceNDT EVALUATIONStructural Number,SNNDT!Resilient Modulus,(Mr)SCIamm31023(in.31023)MeanCOVb%!MeanMPa psi!#COVb%!MeanCOVb%!Conventional pavement,section 102/1991Lime-stabilizedsubbase03/1991Soil-cement base3.31083.88 12,166!152.0 0.08!1803/1991HMAC surface5.36136.74 19,833!130.8 0.03!2308/1991HMAC surface6.74120.66 17,500!110.8 0.03!2003/1993HMAC surface5.97147.37 21,375!101.0 0.04!3502/1995HMAC surface5.41267.17 9,742!181.0 0.04!6009/1996HMAC surface6.213135.02 19,583!140.5 0.02!4604/2001HMAC surface6.2495.00 13,779!90.8 0.03!29Stone interlayer pavementsection 202/1991Lime-stabilizedsubbase0.034275.36 10,930!1716.3 0.64!3203/1991Soil-cement base1.83267.68 9,816!145.1 0.20!3003/1991Limestoneinterlayer3.51669.84 10,130!242.0 0.08!4903/1991HMAC surface4.08120.66 17,500!192.3 0.09!1708/1991HMAC surface6.24124.10 18,000!111.3 0.05!2003/1993HMAC surface5.55144.79 21,000!131.5 0.06!1802/1995HMAC surface4.61075.50 10,950!221.8 0.07!1909/1996HMAC surface6.46128.59 18,650!151.0 0.04!3304/2001HMAC surface5.5793.25 13,525!141.8 0.07!24aSCI5surface curvature index.bCOV5coefficient of variation.124 / JOURNAL OF TRANSPORTATION ENGINEERING ASCE / MARCH/APRIL 2003soils with traffic loading is presented in Fig. 10. The resilientmodulus of the subgrade soil on LA-97 is consistent for bothpavement types and on average is equal to 103.4 MPa 15 ksi!.Pavement Roughness and Permanent DeformationFig. 11a! shows the variation of IRI with time for the conven-tional and stone interlayer pavements. The average IRI for thestone interlayer pavement is low, indicating a smooth pavementsurface and good ride quality. The average IRI 10.2 years afterhighway construction is 1.03 mm/m 65 in./mi!, while the averageIRI for the conventional pavement is 1.25 mm/m 79 in./mi!,which also is considered within the good ride quality range. Thestone interlayer pavement showed a consistently better ride qual-ity as compared to the conventional pavement. The reason for thiscould be related to the amount of reflective cracking developedwithin the conventional pavement.The PSI for the investigated pavement types is shown in Fig.11b!. PSI values for the conventional and stone interlayer pave-ments have decreased slightly since pavement construction. How-ever, the PSI values are higher than 4.0, which put both pavementtypes after 10.2 years of service in the very good category.Measurements of permanent deformation showed that themean rut depth for the conventional pavement is 3.3 mm 0.13in.! and that for the stone interlayer pavement is 3.8 mm 0.15in.!. These results are consistent with typical consolidation undertraffic and indicate that neither type of base design suffered pave-ment rutting.Pavement Performance under Accelerated LoadingThe two test lanes constructed at PRF site to verify the behaviorof the stone interlayer pavement are test lanes S-008 and S-009.The test lanes were tested under accelerated loading using theALF machine. The test lane S-008 is a conventional pavementtype with 216 mm 8.5 in.! of in-placed stabilized soil cementwith a 10% cement content and 89 mm 3.5 in.! of HMAC sur-face layer. The test lane S-009 is a stone interlayer pavement thatconsisted of 89 mm 3.5 in.! of HMAC surface layer and 102 mm4 in.! of crushed stone constructed over 152 mm 6.0 in.! ofin-placed stabilized soil cement with a 10% cement content.The test lanes performance was determined by the amount ofsimulated traffic loading in ESALs received at failure. The pri-mary failure criteria were a rutting of 25.4 mm 1.0 in.! and acracking density of 4.92 m/m2(1.5 ft/ft2) in 50% of the testedarea. The accelerated pavement testing results showed that thestone interlayer pavement failed after 1,294,800 ESAL while theconventional pavement failed after 314,500 ESALs. This indi-cated that the stone interlayer pavement lane received about fourtimes the ESALs of the conventional pavement lane before fail-ure. The stone interlayer pavement also outperformed all eightother lanes tested under the first Louisiana accelerated pavement-testing experiment. The accelerated pavement testing resultsclearly verify the superior performance of the stone interlayerpavement design concept. This is particularly true where there ishigh moisture content in the base, which was the case for thisexperiment.Prediction of Pavement LifeLong-term pavement monitoring as well as accelerated load test-ing of the alternative stone interlayer pavement both demon-strated the outstanding performance of this pavement type. It isessential to perform analysis in order to provide information re-garding how long the good performance will last before pave-ment fails. To answer to this question, analysis was conducted onthe pavements constructed on highway LA-97 and the two pave-ment lanes tested under accelerated conditions at the PRF site.Metcalf et al. 1998! conducted survival analysis on the test lanesS-008 and S-009 to predict the pavement failure mode and thecorresponding traffic loading in terms of ESALs. Results of theanalysis are presented in Fig. 12a!. As shown in the figure, thedominant mode of failure 88%! for the stone interlayer pavementis rutting, while cracking was the characteristic failure mode forthe conventional pavement. The stone interlayer pavement wasFig. 10. Comparison of resilient modulus of subgrade soil for con-ventional and stone interlayer pavements over 10.2 year periodFig. 11. Comparison of ride quality and serviceability of conven-tional and stone interlayer pavements over 10.2-year period: a! com-parison of conventional and stone interlayer pavements roughness;b! evaluation of Pavement Serviceability Index for conventional andstone interlayer pavementsJOURNAL OF TRANSPORTATION ENGINEERING ASCE / MARCH/APRIL 2003 / 125able to support the traffic loading range from 995,000 to1,428,000 ESALs, while the conventional pavement was able tosupport 298,000 ESALs before failure.Analysis of long-term monitoring of the conventional andstone interlayer pavements provided information regarding crack-ing, deformation, structural capacity, and roughness. Based on theanalysis of the collected long-term data, the only mode that mightbring both pavement types to failure at LA-97 is cracking. There-fore, a cracking density criterion of 4.92 m/m2(1.5 ft/ft2) wasused to estimate the pavement life in terms of ESALs. As shownin Fig. 12b!, the regression analysis was used to develop or-thogonal polynomial functions to estimate pavement life just be-fore the failure criterion is satisfied. Based on this analysis, it ispredictedthatthestoneinterlayerpavementwillsupport1,296,371 ESALs and the conventional pavement will support623,911 ESALs. The results of t