Performance evaluation of porous asphalt with granulated synthetic lightweight aggregate

, Chia-Ming Wu , Jia-Chong DuTel.: +886 2 27376573; fax: +886 2 27376606.Construction and Building Materials 22 (2008) 902910Constructionand BuildingMATERIALS/locate/conbuildmatPerformance evaluation of porous asphalt with granulatedsynthetic lightweight aggregateDer-Hsien Shen a,1 a,* b,2a Department of Construction Engineering, National Taiwan University of Science and Technology, No. 43, Sec. 4, Keelung Road,Taipei City 10672, Taiwan, ROCb Department of Civil Engineering, Tung-Nan Institute of Technology, No. 152, Sec. 3, Peishen Road, ShenKeng, Taipei County, 22202, Taiwan, ROCReceived 7 October 2006; received in revised form 16 December 2006; accepted 19 December 2006Available online 2 February 2007AbstractThis paper presents the results of a laboratory study evaluating the mixture characteristics and durability of porous asphalt made withgranulated synthetic lightweight aggregate (GSLA). Porous asphalt specimens incorporating 0%, 5%, 10%, 15% and 20% GLSA by vol-ume as a coarse aggregate (retained on a No. 4 sieve) replacement were prepared. Test results indicating that GSLA exhibits well particleshape, lower specic gravity, higher soundness and higher water absorption than conventional crushed stone (CS) does. The mix designresults showed that GSLA mixtures have high asphalt absorption due to porous nature of the GLSA as evidenced in scanning electronmicroscope (SEM) image. Compared with conventional CS mixtures in terms of skid resistance, permeability, particle loss resistance,moisture susceptibility and rutting resistance, the GLSA mixtures perform very favorably. The results show that the mixture of 15%GSLA replacement was determined to be the most suitable mixture.2006 Elsevier Ltd. All rights reserved.Keywords: Synthetic lightweight aggregate; Porous asphalt; Performance evaluation; Mixture characteristics; Durability1. IntroductionTaiwan has a unique geographical environment andtopographical characteristics which is characterized by itssteep mountains and rapid rivers. Reservoirs which regu-late water and generate electric power become very impor-tant resources of water and energy supply in Taiwan.However, a large quantity of soil sediment has accumu-lated in reservoir due to ood scouring, erosion andtyphoon induced debris ow. According to the report ofstatistical data from the Water Resources Agency, Ministryof Economic Aairs 1, around 20% of the 46 main reser-voirs have soil deposits problems and the total volume ofsediment is approximately 4.7 hundred million cubicmeters. Furthermore, there is average sediment of 0.15hundred million cubic meters increased in each year. Reser-voir sediment is not suitable for land lling disposalbecause of the adverse eect on the environment. Thus, uti-lization of synthetic lightweight aggregate (SLA) manufac-tured from reservoir sediment may form an attractivealternative to land lling disposal in Taiwan and reducethe consumption of natural aggregate.SLA usually is produced from expanded clay, shale, slateor by-products from rotary kiln process and must meet therequirements of ASTM C330. In general, the particle shapeof SLA, which is nearly rounded and without any fracturedfaces, is not suitable for use in hot mix asphalt (HMA).Angular shaped particles which are preferred in HMA exhi-* Corresponding author. Tel.: +886 2 27370456; fax: +886 2 27376606. bit greater interlocking and internal friction, and result inE-mail addresses: .tw (D.-H. Shen),D9305401.tw (C.-M. Wu), .tw (J.-C.Du).1greater mechanical stability than do rounded particles 2.Recently, a new manufacturing process is developed to pro-duce granulated synthetic lightweight aggregate (GSLA)2 Tel.: +886 2 86625921x119; fax: +886 2 26629583. and enhance the application of SLA in asphalt mixtures.0950-0618/$ - see front matter 2006 Elsevier Ltd. All rights reserved.doi:10.1016/j.conbuildmat.2006.12.008D.-H. Shen et al. / Construction and Building Materials 22 (2008) 902910 903GSLA not only has well particle shape and fractured faces,but also benet from skid resistance. Several US highwayagencies have reported that the GSLA used in chip sealshas positive results 3,4. Study also showed that GSLAhas excellent potential for being used as 15% by weight ofaggregate replacement of mineral aggregates in HMAappearing to be an optimum amount for use 5.Porous asphalt, which was developed in Europe and hasa porous structure, can reduce trac noise, drain waterfrom the pavement surface and reduce thermal conductiv-ity. However, there is very few research focused on thetopic of utilizing the GSLA in porous asphalt. Therefore,the objectives of this study were to evaluate the mixturecharacteristics and the durability of porous asphalt whichused GSLA as coarse aggregate replacement for mixtures.2. Materials and test methods2.1. Aggregates and asphalt binderThe coarse GSLA used in this study was produced fromsintering ne sediment excavated from reservoirs. A rigidcoating of glazing shell on aggregates during burning processwas formed to reach lower asphalt absorption and higherstrength. After cooling process, the product was made intoGSLA by the granulator as illustrated in Fig. 1. Rivercrushed stone (CS) and crusher nes used as mineral llerwere obtained from a local aggregate supplier. A styrenebutadienestyrene (SBS) modied polymer asphalt binderwas obtained from a commercial asphalt manufacturer.2.2. Test programThe aggregate blending procedure of HMA presumesthat all aggregates have the same specic gravity. However,Fig. 1. The photograph of GSLA product.when specic gravities of individual aggregates dier signif-icantly (by 0.20 or more) they should be adjusted for spe-cic gravity variation before blending to obtain correctvolumetric proportions in the mixture 6. The GSLA usedin this study has a lower specic gravity (1.515) than that ofnatural CS (2.675). The blending of GSLA as a replace-ment of coarse CS aggregates was done on the volumebasis to maintain the volumetric properties in target grada-tion. Porous asphalt specimens incorporating 0%, 5%, 10%,15% and 20% GLSA by volume as a coarse aggregate(retained on a No. 4 sieve) replacement were prepared.The test ow chart in this study consisted of three tasksas illustrated in Fig. 2. Task I was to investigate the phys-ical properties of aggregates (GSLA and CS) and asphaltbinder. Task II was to perform the gradation design basedon a theoretical particle packing method and the mixdesign of porous asphalt. Task III was to evaluate the mix-ture characteristics and durability of porous asphalt withand without GSLA. The fractured section of the GSLAmixture was further investigated to identify the microstruc-ture of GSLA and asphalt binder.2.3. Test methods2.3.1. Skid resistanceSkid resistance was evaluated based on tests performedin accordance with ASTM E303 procedure. The Britishpendulum skid resistance device was used to measure theBritish Pendulum Number (BPN) value of the specimen.2.3.2. Water permeabilityPermeability test was conducted using a constant-headtype equipment which developed by the Japan Road Asso-ciation (JRA) 7. The coecient of permeability measuredwas corrected to a standardized value at 20 C for the vis-cosity of the water. A minimum permeability coecient of10 2 cm/s is usually specied in order to ensure good per-formance of draining water from the road surface.2.3.3. Particle loss resistanceThe Cantabro test was used to evaluate the resistance toparticle loss of the mixtures 8. This test was carried out inLos Angeles abrasion machine without steel ball. A com-pacted specimen, unaged or aged, was tumbled inside thesteel drum for 300 revolutions at a speed of 3033 rpm.The specimen is weighed before and after the test, andthe percentage loss by weight of original specimen is calcu-lated as the Cantabro abrasion. Aged specimens were con-ditioned by placing in an oven at 60 C for 168 h (sevendays), and cooled to 25 C and stored for 4 h prior to test-ing. The specications of European require that the per-centage loss by weight must be less than 25% for unagedspecimens and 30% for aged specimens.2.3.4. Moisture susceptibilityMoisture produces a loss of adhesion between theasphalt binder and the aggregate surface, and accelerates904 D.-H. Shen et al. / Construction and Building Materials 22 (2008) 902910Fig. 2. Flow chart of the test program.the development of other distresses such as pothole, crack-ing and raveling. The moisture susceptibility test was car-ried out in accordance with AASHTO T283 procedure.The tensile strength ratio (TSR) was determined by theaverage indirect tensile strength (ITS) of conditioned sub-set divided by the average ITS of control subset. A mini-mum TSR of 70% is recommended for this test method.2.3.5. Rutting resistanceRutting resistance test was performed using the wheeltracking device 9. Slab specimens, 300 300 50 mm insize, were fabricated by steel roller machine, and testedunder 1.12 and 1.65 MPa repeated wheel loading at a testtemperature of 60 1 C. The repeated loading was per-formed by a solid rubber tire (200 mm diameter and50 mm width) and set at 42 passes per minute. The dynamicstability (DS) was dened by number of wheel pass per onemillimeter deformation between 45 and 60 min during thetest period. The JRA suggested that the calculated DS ofporous asphalt specimen should be more than or equal to3000 passes/mm for heavy trac 10.3. Mix design of porous asphaltThe concept of packing grading design based on the ideato ensure coarse aggregates stone on stone contact (inter-locking) was adopted in this study to provide stable aggre-gate skeleton structure. Previous research focused on theD.-H. Shen et al. / Construction and Building Materials 22 (2008) 902910 905Fig. 3. Packing grading design for porous asphalt.particle packing of porous asphalt also found that a tightlypacked aggregate skeleton structure in porous asphaltwould exhibit higher durability and rutting resistance thanordinary porous asphalt 11,12. The algorithm of theaggregate packing method that based on the concept ofparticle packing and lling process was separated intotwo phases, (1) the main skeleton structure and (2) the l-ler, as illustrated in Fig. 3a.The main skeleton structure phase consisted of coarseaggregates in dierent sieve size to achieve the minimumvoids in coarse aggregate (VCAmin). The procedure of aggre-gate packing was performed by considering each twoselected sieve components at a time. The compositions withdierent proportions were compacted by vibrating tablemethod in accordance with ASTM D4253. When the mini-mum voids and the optimum proportion of the two compo-nents were obtained, then the composition was regarded as anew component to be considered with the next component,until the VCAmin was achieved. The packing results of coarseaggregate are illustrated in Fig. 3bd.906 D.-H. Shen et al. / Construction and Building Materials 22 (2008) 902910After the VCAmin was obtained, the ller phase wasadded into main skeleton structure phase to achieve theTable 2Test results of polymer modied asphalt binderdesign air voids. In other words, the air voids content inporous asphalt was governed by amount of ller materials.The lling materials included ne aggregates and mineralller which were added incrementally into the previouslydetermined gradation, and asphalt binder until a designair void was satised. The design air voids correspondingon optimum adding amount of lling materials can beobtained from the plotted graph, which corresponds tothe amount of ller material added and air voids. ResultsPropertiesSpecic gravity (25 C)Penetration (1/100 cm)Viscosity (60 C, poise)Viscosity (135 C, cSt)Flash point ( C)Solubility (%)Loss on heating (%)Recovery on heating (%)Penetration of residue (4 C, 1/100 cm)Specicationa35 min8000 min3000 max232 min99 min70 min10 minTesting1.032388653286533599.70.047212of the lling process are illustrated in Fig. 3e. a Reference CNS 14184 polymer modied asphalt specication (TypeThe optimum asphalt contents in porous asphalt weredetermined according to the standard procedure proposedby JRA which is based on the draindown test and abrasiontest 10. Draindown test was conducted on uncompactedporous asphalt mixture at 175 C according to theAASHTO T305. In practice, the optimum asphalt contentselected is on the upper limit. Afterward all Marshall spec-III).Table 3Porous asphalt gradation used in this studyimens were compacted by 50 blows on each side with the Coarse aggregate Fine aggregatestandard Marshall hammer.4. Test results and discussion4.1. Physical properties of aggregates and asphalt binderPhysical properties of GSLA and CS aggregate areshown in Table 1. Test results indicating that GSLA exhib-Sieve size(mm)25.019.05Percent passing(%)10084.6Sieve size(mm)2.30.150.075Percent passing(%)its lower specic gravity, higher soundness and higherwater absorption than conventional CS does. The GSLAmanufactured exhibits higher shape factor (angularity)and higher rounded index and lower percentage of atand elongated particles than CS does. Results showed thatphysical properties of GSLA met the requirements of theASTM D692 and AASHTO M283 specications.Table 1The properties of GSLA and CS aggregatePolymer modied asphalt binders are widely used in theproduction of porous asphalt mixtures for increasing adhe-sion between binder and aggregate in order to reducedraindown in the mixture. Test results of the SBS polymermodied asphalt binder used in this study met all specica-tion requirements as shown in Table 2.Properties Specication GSLA CSCoarse FineBulk specic gravityWater absorption (%)L.A. abrasion (%)Sodium soundness (5 cycles) (%)40 max12 min1.5155.927.069.082.6722.5272.3Percent fractured faces (%)One or moreTwo or more90 min75 min1009810096Flat and elongated particlesa (%)1:31:510 max15 max4.570.628.260.75Crush compressive strengthb (MPa) 15.2 Rounded indexShape factor (%)0.650.620.510.55abThe at and elongated particles in coarse aggregate are determined in accordance with ASTM D4791.Reference CNS 14779 test methods for SLA.D.-H. Shen et al. / Construction and Building Materials 22 (2008) 902910Table 4Mixture properties of porous asphalt mixture at optimum asphalt contents907Mixture properties Specicationa Mixture seriesbOpt. asphalt contents (%)Stability (kN)Flow (0.01 cm)Air voids (%)Unit weight (kg/m3)VCAminc (%)VCAmix (%)Draindown (%)Asphalt absorption (%)3.5 min204020 1%0.3 maxPA-05.854.5235.120.8207236.327.70.680.12PA-55.945.1236.420.6199635.827.70.470.13PA-106.355.2135.220.2193336.527.40.870.41PA-156.575.3033.719.818630.54PA-206.885.5934.819.2181135.927.10.980.97abTANEEBs porous asphalt specication.Mixture series denoted by PA-X, where X refers to the replacement percentage of GSLA by volume.Fig. 4. Skid resistance test results for the mixture series. Fig. 6. Abrasion potential test results for the mixture series.Fig. 5. Water permeability test results for the mixture series. Fig. 7. Moisture susceptibility test results for the mixture series.908 D.-H. Shen et al. / Construction and Building Materials 22 (2008) 902910Table 5Summary of ANOVA tests for mixture characteristic and durability(a = 0.05)Source of variation SS d.f MS F FcriticalMixture characteristics:BPN (Dry condition)Between 11.87Within 2.754102.9670.27510.803 3.478Total 14.62BPN (Wet condition)14Between 27.22Within 2.474106.8060.24727.591 3.478Total 29.69Coecient of permeability14Fig. 8. Rutting resistance test results for the mixture series.4.2. Mix designThe tightly packed aggregate gradation based on theconcept of particle packing and lling process is shownin Table 3. Results of mix design for the mixture seriesare summarized in Table 4. To verify the existence ofthe coarse aggregates stone on stone contact within aporous asphalt mixture, an essential criterion of VCAof the compacted mixture (VCAmix) should be checkedif it were equal to or less than the VCAmin. Resultsshowed that the VCAmix for the mixture series were farless than the VCAmin. The interlocking mechanism ofBetween 0Within 0Total 0Cantabro abrasion loss (unaged)Between 168.93Within 7.05Total 175.98Cantabro abrasion loss (aged)Between 117.57Within 14.84Total 132.41Durability:TSRBetween 147.04Within14.39Total161.43410144101441014410140.001042.2330.70529.3941.48436.7591.43917.81959.94219.80725.5373.4783.4783.4783.478the coarse aggregate in porous asphalt was veried 8. RD (1.12 MPa wheel load)Optimum asphalt contents of GSLA mixture were higherthan that of conventional CS mixture due to highBetween 7.47Within 0.044101.8670.004504.694 3.478absorption rate of GSLA. Marshall Stability of GSLAmixtures also were higher than that of CS mixture forTotal 7.51RD (1.65 MPa wheel load)14the same reason. Between 13.89Within 0.02 410 3.4710.002 2019.749 3.4784.3. Mixture characteristics Total 13.90DS (1.12 MPa wheel load)14Test results of the skid resistance were illustrated inFig. 4. Results showed that higher percentage of GSLAresulted in better skid resistance even at wet condition.The values of the skid resistance are increased due to theangularity and rough surface texture of GSLA.Between 30004312Within 29701.33Total 30034014DS (1.65 MPa wheel load)Between 24798318Within 10591.334101441075010782970.13361995801059.1332525.5025853.4463.4783.478Test results of the water permeability were illustrated inFig. 5. A minimum permeability coecient of 10 2 cm/s isgenerally specied to ensure good surface drainage. Resultsshowed that the whole mixture series had excellent perme-ability except a little drop for a mixture consisting of 20%Total 248089094.4. Mixture durability14GSLA. The permeability coecient drops due to possibleaggregate breakdown during compaction resulted in theloss of air voids.Fig. 6 illus