te

ey

h i g h l i g h t s

odulus of a Class F y ash by the means of cyclic triaxial testing.ied by air was practical to use, it was difcult to maintain exact target pressure.aving higher cement or lime contents exhibited lower strains.ndingween a

udy is to investigate the effect of stabilization on resilient modulus of a Class F

waste production, with billions of tons of waste being generated

the last few decades, however, the increasing amount of waste pro-duction and the diminishing availability of land ll sites have re-strained the dumping of waste products. Therefore, the industriesproducing waste materials now face an increasing nancial burden(cost of disposal) and community concern regarding the potentialrisk to the environment.

electricity. Before combustion, the coal is pulverized in coal millst air. During theme physicalls througform botto

which is a coarser material. The majority (about 8085%) isout of the furnace by means of the ue gases, where it hardea ne grained material known as y ash.

As the demand for energy increases, so does the amount of yash that have the potential of being utilized in various applications.Construction of road base, subbase and subgrade provides anopportunity to use high volumes of this material. There is a grow-ing demand for road materials in every country, yet the supply ofthe classical base and subbase materials such as crushed rockand gravel is diminishing as government regulations restrict

Corresponding author. Tel.: +90 535 6119529.

Construction and Building Materials 54 (2014) 1016

Contents lists availab

B

evE-mail address: lav@itu.edu.tr (A. Hilmi Lav).each year around the world. As a result of this, considerable atten-tion has been given to the utilization of waste materials and otherby-products from different production processes. In the recent pastthese materials were simply stockpiled and dumped as waste. In

and injected into the furnace by compressed hocombustion process coal minerals undergo sochemical changes. About 15% of the burned coal fgrids into the furnace oor where it sinters to0950-0618/$ - see front matter 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.conbuildmat.2013.12.029al andh openm ashcarriedns intoFly ashResilient modulusPavement analysesMechanistic-empirical designCyclic triaxial testingIndirect tensile testing

y ash. 2013 Elsevier Ltd. All rights reserved.

1. Introduction

One of the most signicant detriments for the modern world is

Increasing demand for electricity has rendered coal red powerstations indispensable for many countries. In these stations, thecoal is used to provide the steam that drives turbines to generateKeywords:Highway

modulus, the aim of this st The effect of stabilization on resilient m Although the conning pressure suppl Under the same stress paths samples h The rate of axial strains under correspo There was a nonlinear relationship bet

a r t i c l e i n f o

Article history:Received 27 July 2013Received in revised form 5 December 2013Accepted 16 December 2013Available online 9 January 2014stress paths decreased with increasing cement or lime content.xial strain and deviator stress.

a b s t r a c t

The disposal of industrial by-products constitutes a mounting problem of global dimensions. Amongthese, the disposal of y ash is very problematic, for the increasing demand for electricity from coal burn-ing power stations has been resulting in growing amounts of stockpiled y ash, inevitably causing envi-ronmental problems. The use of this y ash in road pavements provides an opportunity to use highvolumes of this material, however, the ash needs to be stabilized to improve its performance when uti-lized in upper pavement layers. Because the principal input in mechanistic-empirical analyses is resilientEffects of stabilization on resilient characmaterial

M. Aysen Lav, A. Hilmi Lav Istanbul Technical University, Faculty of Civil Engineering, 34469 Maslak, Istanbul, Turk

Construction and

journal homepage: www.elsristics of y ash as pavement

le at ScienceDirect

uilding Materials

ier .com/locate /conbui ldmat

materials and reduction of greenhouse gas emissions). Hence, y ashhas the potential of replacing classical road building materials

Stiffness is the most important fundamental property of a pave-ment materials and it is expressed by resilient modulus, which

ranked the method as the fourth preferred procedure for testingthe resilient characteristics of stabilized materials. Since the test

n ancan be obtained by advanced laboratory tests explained below.

3. Properties of y ash and stabilizing agents used in the study

The y ash used in this study is a Class F [1] material with lowcalcium content (1.61%) and very low loss on ignition (LOI) of 1.28,which is related to the grinding method used to pulverize the coalin power station. Grain size distribution of the material wasdetermined with a laser operated particle size analyzer as thehydrometer test is unsuitable due to its pozzolanic reactivity. AHoriba LA-500 analyzer was utilized by a professional laboratoryand the results were presented. The y ash was found to besmoothly graded, being the maximum and the minimum particlesizes were 174 and 0.15 lm. A total of 83% of the material is nerwhen circumstances permit.

2. Structure of highways

In general, highways consist of two parts: the pavements andfoundation. The pavements are engineering structures that sepa-rate the trafc loading (the tires of vehicles) from the foundation(which is usually natural soil, also called subgrade) of roads. Sub-grades are generally weak materials and may not be able to bearthe stresses induced by wheel loads. Therefore, pavements areneeded to decrease the stress to a tolerable level in the subgrade.As an engineering structure, pavements should not fail during theirdesign life. Unlike most civil engineering structures, pavement fail-ure does not happen unexpectedly like the collapse of a bridge orbuilding, but gradually deteriorates over time. That means thatthe emergence of cracks and the accumulation of plastic deforma-tion up to an unallowable level are considered as design failures.This pavement failure eventually decreases or totally eliminatesthe function of the structure. Pavement design aims to preventthe failure of the pavement over its design life which is usually2030 years.

Regarding the design of y ash pavements, mechanistic-empir-ical methods are preferred as empirical design methods are re-stricted to the range of classical pavement materials. In themechanistic-empirical method, the stress induced by wheel load-ing is calculated in order to identify the mechanical responses inthe structure (pavement analysis), usually by means of purposebuilt computer programs. The mechanical responses are basicallydeections, stresses, strains and displacements. In order to performa pavement analysis, the fundamental properties of the y ash(along with other pavement materials used) should be measured.quarry operations. As a result, natural materials are becoming evermore scarce and expensive. The consumption of high quality natu-ral quarry materials in pavement building continues, however, tobe the standard practice. If our goal, however, is to support an envi-ronmentally sustainable world, it makes more sense to use thesevaluable natural materials in structural concretes (high rise build-ing, bridge, pre-stressed concrete structures, etc.), where highstrength and good quality material are essential. Furthermore, uti-lizing materials that are already produced results in less energyand emission in total highway construction, resulting in a GreenHighway (A generic for a highway that is produced with minimumor even no harm to the environment in terms of protection of natural

M. Aysen Lav, A. Hilmi Lav / Constructiothan 45 lm; therefore, it could be considered as a ne graded yash. The stabilizing agents, namely, cement and lime, were alsotested to determine their particle size distribution. The maximumis relatively easy to conduct and sample preparation is easy as well,it was therefore decided to conduct an indirect tensile test to ob-serve the variation of resilient modulus over time and the inuenceof stabilizing agent content. The Universal Materials Testing Appa-ratus (UMATTA) manufactured by Industrial Process Control Lim-ited [7] was used. Samples having a diameter of 105 mm and alength of between 51.5 and 70.5 mm were compacted at optimummoisture content and maximum dry density. The lime or cementcontents in the samples were 2%, 4%, 8% and 10%. The tests werecarried out on samples cured for 7 days, 28 days, 180 days and300 days. The testing procedure conforms to the ASTM D 4123-82 [8].

Initially, different loading forces up to 700 N (the maximumapplicable force was 738 N) were applied to examine the variationin results. Discrepancies due to different load levels were insignif-icant and it was decided to apply 300 N of loading force throughoutthe entire test. Testing of samples at a constant tensile strain (con-stant strain test) was found to be inappropriate for stabilized yash, as some samples broke during the trial of constant strain. Inthe test, ve consecutive loading forces having a pulse period of3000 ms and a rise time of 30 ms were applied in the form of a tri-angular wave shape. Average modulus was calculated from theand the minimum particle sizes were 116 and 0.13 lm for cement,and 152 and 0.2 lm for lime.

As mentioned previously, y ash should be stabilized to im-prove its performance when utilized in upper layers of pavementssince stress and strains due to vehicle loadings are signicant atthe top layers of pavements. In addition, Class F ashes exhibit lowerstiffness than Class C ashes due to their low calcium content. How-ever, an uncommon application of 100% y ash base with no addi-tive or aggregate was attempted in 1988 at Fulshear, Texas, USA.After placement and compaction, the average compressivestrength increased only 255 kPa, with between 7 days and 28 daysof curing. Also, in four months time it was seen that the asphaltlayer at the top became corrugated in some sections. The trialwas considered to be a failure [2] and, henceforth, that kind of test-ing was not again attempted.

In our study, cement and lime were used separately to stabilizethe y ash. These were mixed as percentages by total weight (e.g.10% lime or cement stabilized y ash means that 10 g stabilizingagent was mixed with 90 g of y ash). All of the stabilized samplesthat required curing before testing were wrapped in plastic bagsand cured in a room with controlled humidity and temperature(23 C and 50% humidity) until the test date.

4. Repeated load indirect tensile test of stabilized y ash

The repeated load indirect tensile test is based on Frochts equa-tion [3] for the stress distribution due to a point load. The pointload on a ctitious disk is extended to line loading to create a cyl-inder. In practice, the load is distributed over a loading strip. Thestrip reduces the vertical compressive stresses and prevents failureunder the line load. The load results in a uniform tensile stress(along the vertical diametric plane) and a corresponding horizontalstrain that is used to calculate the properties of the material tested.The test is carried out by repeated pulse loading the sample todetermine resilient characteristics. According to Kennedy andHudson [4,5], indirect tensile test can be used to evaluate all pave-ment materials including stabilized materials. The only exceptionis unbounded granular materials. Moreover, AUSTROADS [6] has

d Building Materials 54 (2014) 1016 11recoverable diametric strain. Each sample was then rotated 90and the same procedure was repeated. The mean value of the aver-age of the moduli at 0 and the average of the moduli at 90

positions was calculated as the typical modulus. The repeated loadindirect tensile test results are presented in Fig. 1 for cement stabi-lized samples and in Fig. 2 for lime stabilized samples.

The effects of stabilization on lime stabilized samples weresmall during the rst 7 days, whereas cement stabilized samplesclearly showed the distinctive effect of the stabilizing agentamount in terms of modulus improvement. After 28 days, the ef-fect of the amount of stabilizing agent in y ash on the moduli ofsamples became apparent. The higher content gave higher modu-lus except for the case of 2% lime which had a higher modulus va-lue than 4% lime at 180 days. This possibly happened due to highpozzolanic reactivity of the y ash used in that particular sample(this is a signicant point to be considered y ash stabilization).The modulus of the lime and cement samples in correspondentcontents become closer as the curing time increased. Conse-quently, the response of pavement structure to trafc loading interms of tensile strength development became similar over timein both cement or lime stabilized layers. Therefore, the fatiguebehavior of lime and cement stabilized y ash will be determinant

cylindrical sample. The system is not able to represent all condi-tions (triaxial) because two of the principal stresses are equal (con-ning stress or cell pressure). There are a few true triaxialmachines available with rigid boundaries using a cube sample[11], but the common triaxial system gives good results and is suit-able for testing of pavement materials. Although the method isranked number one by AUSTROADS [6] for the testing of granularpavement materials, it is equally effective on stabilized pavementmaterials, including waste materials [12,13].

In pavement layers under trafc loading, each stress has a con-stant value from overburden, and the passing wheel load causes anadditional transient value. Transient stress may be simulated bydeviator stress (vertical), while the overburden is simulated byconning pressure (horizontal). Consequently, the repeated loadtriaxial test corresponds to actual stress conditions, however, forthe lower section of the stabilized y ash layers, tensile stresses

Fig. 2. Repeated load indirect tensile modulus test results for lime stabilized y ashsamples.

Fig. 3. Effect of cement content and curing on indirect tensile strain for single pulseof 300 N.

12 M. Aysen Lav, A. Hilmi Lav / Construction anin the performance of each type of pavement.It is also signicant to observe the effect of hydration over time

(increased modulus and decreased tensile strain over single pulse),which is shown in Figs. 3 and 4.

5. Stress strain relationship and triaxial testing of pavementmaterials

In general, a point of a material in a pavement structure is sub-jected to nine stress components in Cartesian coordinates. If thispoint is located somewhere beneath the center of the wheel load,the point is subjected to three principal stresses alone, becausethere are no shear components of stress under the center line ofthe wheel load.

The repeated load triaxial testing method has been extensivelyused to characterize the stressstrain behavior of pavement mate-rials. In other words, it measures the mechanical response of thematerials under simulated, repeated trafc loadings. In the USA,Seed and McNeill [9] carried out experiments on unbound materi-als in a repeated load triaxial machine. In the UK, the instrumentwas used by Grainger and Lister [10] at the Transportation Re-search Laboratory to characterize the performance of soils and un-bound pavement materials. The test applies a repeated verticalstress and either a repeated or constant conning stress to aFig. 1. Repeated load indirect tensile modulus test results for cement stabilized yash samples.d Building Materials 54 (2014) 1016may be developed in the horizontal direction (due to exure) andthis situation cannot be reproduced in the repeated load triaxialtest. According to Brown [14], this case may not diminish the

n anM. Aysen Lav, A. Hilmi Lav / Constructioimportance of the triaxial testing to characterize the behavior ofsuch materials (i.e. materials subject to tensile stresses underwheel load).

5.1. Cyclic triaxial testing of stabilized y ash

The other experimental part of this research was achieved by astress path triaxial test system which is capable of characterizingthe stress dependent (resilient) characteristics of pavement mate-rials under different stress paths by simulating the in situ condi-tions in laboratory. This system was manufactured by IndustrialProcess Controls Limited [7] and was originally congured to applyconning pressure through a liquid medium (water or silicon oil),but for this study compressed air was used as to provide conningpressure.

The test is carried out by applying a repeated vertical stress(deviator stress) and constant conning stress to cylindrical sam-ples having a diameter of 100 mm and a height of 200 mm as rec-ommended by FHWA [15]. The stress was transmitted to thesample by haversine shape cyclic load. Initially samples are

Fig. 4. Effect of lime content and curing on indirect tensile strain for single pulse of300 N.

Fig. 5. Nonlinear resilient modulus of 2% cement stabilized y ash samples.conditioned, applying between 500 and 2000 cycles. Cement orlime content was 2%, 4%, 8% and 10% by weight. During the testthe samples were subjected to ve constant conning pressuresin the following order: 250 kPa, 200 kPa, 150 kPa, 100 kPa and50 kPa. For each conning pressure, cyclic deviator stress, namely,50 kPa, 100 kPa, 150 kPa and 200 kPa was applied consecutively (atotal of 20 different stress paths were applied). These stress levelswere well below failure lines and thus the possibility of the devel-opment of plastic (permanent) strains was greatly reduced. Vari-ous researchers suggest applying different pulse cycles beforemeasuring the resilient strain (recoverable strain). Some recom-mended that the measurement should be made between 10 and100 cycles [1618]. According to AASHTO [19] and TRB [20] elasticstrain should be measured after applying 200 cycles. In this studythe test continued until six consecutive identical strains were ob-tained. Otherwise 200 vertical pulses were applied for each stresspath as recommended by AASHTO and TRB.

Fig. 6. Nonlinear resilient modulus of 2% lime stabilized y ash samples.

d Building Materials 54 (2014) 1016 13Another parameter to be considered is loading frequency. Fre-quencies between 0.01 Hz and 25 Hz have been tested on granularmaterials by various researchers [18,21,22]. However, in this studysingle frequency (repetition period) and loading duration (pulsewidth) of 1 Hz and 100 ms applied all samples, respectively. Theremaining time in each cycle was left intentionally to allow thesamples to recover (rest period). Recoverable axial strains weremeasured for each stress path. The predetermined stress paths(particularly conning stress) could not be always achieved pre-cisely due to the difculty of keeping air pressure at the desiredlevels.

In the most simplistic terms, resilient modulus (MR) is the elas-tic modulus utilized in mechanistic-empirical pavement analysesand design. It is based on the recoverable strain under repeatedload as the loading and unloading cycle exhibits a hysteresis loop.Therefore, it is dened as the ratio of the amplitude of the repeatedaxial stress to the amplitude of the resultant recoverable axialstrain [23] as emphasized by researchers [24].

The in situ resilient modulus of pavement materials greatly de-pends on the level of stress applied by trafc loading. Therefore, amathematical model (constitutive equation) that is a function ofstress and resulting strain state should be developed. There areseveral constitution equations from simple to complex that maybe employed to estimate resilient modulus of pavement materials.In this study, the resilient behavior of stabilized y ash is identiedby the following well known, nonlinear model:

Fig. 7. Nonlinear resilient modulus of 4% cement stabilized y ash samples.

Fig. 8. Nonlinear resilient modulus of 4% lime stabilized y ash samples.

Fig. 9. Nonlinear resilient modulus of 8% cement stabilized y ash samples.

Fig. 10. Nonlinear resilient modulus of 8% lime stabilized y ash samples.

Fig. 11. Nonlinear resilient modulus of 10% cement stabilized y ash samples.

Fig. 12. Nonlinear resilient modulus of 10% lime stabilized y ash samples.

14 M. Aysen Lav, A. Hilmi Lav / Construction and Building Materials 54 (2014) 1016

%)

7,28,180,300 Resilient modulus

n ankh model was the rst to be introduced in order to establishnonlinear relationship from cyclic triaxial test results with con-stant conning pressure [25,26]. There are several other modelsdeveloped to estimate resilient modulus, however, the kh modelis quite reliable and widely used [27,28]. The general presentationof the model is as follows:

MR k1frk2

where MR is the Resilient modulus, f(r) is the Stress parameterwhich may be chosen as follows k1 and k2 = regression constantsrepresenting nonlinear elastic properties of material tested.

The so called kh model was widely used to describe the non-linear behavior of unbound granular pavement materials[27,29,30]. A version of this relationship was also used for unstabi-lized y ash by Lee and Fishman [31] and they obtained a trendsimilar to that obtained this study. Moreover, Fossberg [32] re-ported a relationship between resilient modulus and rst stressinvariant for cement stabilized soil.

In this study the following stress parameter was chosen forcalculations

MR k1hk2

where MR is the Resilient modulus, h is the sum of the principalstresses (sum of the rst stress invariant/bulk stress; r1 + r2 + r3),k1 and k2 are the constants depend on the cement or lime content.

Although it is generally used for granular materials, the khmodel closely ts the triaxial results of stabilized y ash as pre-sented in Figs. 512. According to Raad and Figueroa [33], the mod-el is valid even for h values that approach failure for granularmaterials, but the usual practice is to apply stresses well belowfailure to avoid excessive permanent strains with which this studycompiles this rule.

6. Summary of testing program

In order to keep the readers follow testing program easier, allthe laboratory testing program is summarized in Table 1.

7. Environmental concerns

While the aim of current studies has been to investigate thesuitability of y ash as pavement material where high amounts

Table 1Laboratory testing program.

Test method Complying standard Cement or lime content (

Grain size distribution Plain (y ash only)Repeated load indirect tensile ASTM D 4123 2,4,8,10 (C)

2,4,8,10 (L)Repeated load triaxial 2,4,8,10 (C)

2,4,8,10 (L)

M. Aysen Lav, A. Hilmi Lav / Constructioof the material may be used, this utilization also raises some con-cerns that the toxic elements in y ash may leach out to the envi-ronment. Although this concern is beyond the scope of this paper, adetailed leaching study of this material has been carried out andpublished. This study found that the material is not hazardous interms of leaching characteristics [34].

8. Discussion and conclusions

In this study, repeated load indirect tensile and repeated loadtriaxial (cyclic triaxial) tests were carried out to investigate the ef-fects of stabilization on resilient behavior of y ash as pavementmaterial with the following outcomes:The stress strain behavior of stabilized y ash was non-linear,namely, stress dependent. The relationship may be explainedby a resilient strain model. A well-known relationship was usedto relate resilient modulus to stress levels.The repeated load indirect tensile test was easy to conduct. Theonly problem encountered was that the samples were fragileunder tensile stresses and they easily split into two pieces.Therefore, the applied force needed to be adjusted carefully,especially for lime-stabilized samples. Once the test was carriedout properly, the results were consistent with other test resultsobtained under compressive stresses.Since the repeated load triaxial apparatus capable of applyingcompressive stresses only, the results are valid in pavementregions where all the principal stresses are in compression (ten-sile characteristics of the material should be investigated bycyclic exural test).Although the conning pressure supplied by air was practical touse, it was difcult to maintain exact target pressure. There wasan approximate 5% difference between intended and appliedconning pressure.Under the same stress paths, samples having higher cement orlime contents exhibited lower strains.The rate of axial strains under corresponding stress pathsdecreased with increasing cement or lime content. This behav-ior showed similarity to the strain changes under indirect ten-sile stresses. However, unlike indirect tensile results thevariation of strains between cement and lime stabilization weresmall. This is probably due to loading type and connement ofthe samples.There was a nonlinear relationship between axial strain anddeviator stress (the stress due to the vertical loading pulse)and that the nonlinear component of the stress strain relationof the samples increased as the conning stress decreased.

In conclusion, stabilized y ash is a stress dependent materialand cyclic triaxial results showed that an apparent relationship be-tween resilient modulus and the rst stress invariant (the sum ofthe three principal stresses). Therefore, the resilient modulus ofthe material increases with the increasing magnitude of the rststress invariant. It should be noted that throughout the testing pro-gram, after removing the conning stresses, the strains under con-ning pressures were recovered. This proved that the sampleswere not subjected to signicant permanent strains, hence in

90 Resilient modulus in terms of stressstrain relationshipCure duration (days) Parameters obtained

Fineness, max. and min. grain size

d Building Materials 54 (2014) 1016 15in situ applications it is recommended that this material shouldbe designed in thicker layers than conventional pavement layersto keep strains lower. Otherwise, it may end up with prematurecracking over shorter times, which signicantly reduces pavementlife.

References

[1] ASTM C 618. Fly ash and raw or calcined natural pozzolan for use as a mineraladmixture in portland cement concrete. Am Soc Test Mater1996;04(02):3046.

[2] Wong C, Michael KH. Experimental y ash base farm to market road 1093fulshear. Texas: State Department of Highways and Public Transportation,Report no. DTH-17; 1989.

[3] Frocht MM. Photoelasticity. 2nd ed. New York: John Wiley & Sons Inc.; 1964.[4] Kennedy TW, Hudson WR. Application of the indirect tensile test to stabilized

materials. Highway Res Rec 1968;235:3648.[5] Kennedy TW. Characterization of Asphalt pavement materials using the

indirect tensile test. Assoc Asphalt Pavement Technol 1977;46:13249.[6] AUSTROADS Pavement Design A. Guide to the structural design of road

pavements. AUSTROADS: Publication; 1992: No. AP-17/92.[7] De Vos KB, UMATTA. Universal materials testing apparatus for asphalt and

unbound specimens. Reference and Operating Manual. IPC Limited; 1992.[8] ASTM D 4123. Standard test method for indirect tension test for resilient

modulus of bituminous mixture. Am Soc Test Mater 1996;04(03):4235.[9] Seed HB, McNeill RL. Soil deformation under repeated stress applications. In:

Proceedings of conference on soils for engineering purposes. Mexico; 1957.[10] Grainger GD, Lister NW. A laboratory apparatus for studying the behaviour of

soils under repeated loading. Geotechnique 1962;12:314.[11] Lanier J, Zitouni Z. Development of a data base using the grenoble true triaxial

apparatus. In: Proceeding of the international workshop on constitutiveequations for granular non-cohesive soils. Cleveland; 1988. p. 4758.

[12] Ullidtz P. Pavement analysis. Elsevier Science Publisher B.V.; 1987.[13] Leite FC, Motta RS, Vasconcelos KL, Bernucc L. Laboratory evaluation of

recycled construction and demolition waste for pavements. Constr Build Mater2011;25:29729.

[14] Brown SF. Improved framework for predicting permanent deformation inasphalt layers. Trans Res Rec 1975;537:1829.

[15] FHWA. Predictive Design Procedures. VESYS Users manual. Report No. FHWA-RD-77-154. US Federal Highway Administration; 1978.

[16] Hicks GR, Monismith CL. Factors inuencing the resilient response of granularmaterials. Highway Res Rec No. 45 1971:1531.

[17] Allen JT, Thompson MR. Resilient response of granular materials subjected totime dependant lateral stresses. Trans Res Rec No. 510 1974:113.

[18] Brown SF, Selig ET. The design of pavement and rail truck foundations, cyclicloading of soils: from theory to design, In: OReilly, Brown (Editors). Blackie;1991. p. 249305.

[19] AASHTO Guide for design of pavement structures. American Association ofState Highway and Transportation Ofcials. Washington (DC); 1993.

[20] TRB Test Procedures for characterising dynamic stress strain properties ofpavement materials, special report 162. Transportation Research Board.Washington (DC); 1975.

[21] Kalcheft JV, Hicks GR. A Test procedure for determining the resilient responseproperties of granular materials. J Test Eval 1973;1:4729.

[22] Brown SF. Repeated load testing of a granular material. J Geotech Eng Div,ASCE 1974:82541.

[23] Mallick RB, El-Korchi T. Pavement engineering. 1st ed. Boca Raton: CRC; 2009.[24] Bilodeau JP, Dore G. Relating resilient behaviour of compacted unbound base

granular materials to matrix and interlock characteristics. Constr Build Mater2012;37:2208.

[25] Brown SF, Pell PS. An experimental investigation of the stresses, strains anddeections in layered pavement structure subjected to dynamic loads. In: 2ndInternational conference on structural design of asphalt pavements. Ann Arbor(USA); 1967. p. 487504.

[26] Hicks RG, Monismith CL. Prediction of the resilient response of pavementscontaining granular layers using non-linear elastic theory. In: Proceedings ofthe 3rd international conference on asphalt pavements; 1972; 1: 41029.

[27] Huang YH. Pavement analyses and design. 2nd ed. New Jersey: Pearsons-Prentice-Hall; 2004.

[28] Kim D, Kim JR. Resilient behaviour of compacted subgrade soils underrepeated triaxial test. Constr Build Mater 2007;21:14709.

[29] Brown SF, Pappin JW. Analysis of pavements with granular bases. Trans ResRec 1981;810:1722.

[30] Brown SF, Pappin JW. Modelling of granular materials in pavements. Trans ResRec 1985;1022:4551.

[31] Lee SW, Fishman KL. Waste products as highway materials in exiblepavement system. J Trans Eng, ASCE 1993;119:43349.

[32] Fossberg PE. Load-deformation characteristics of three-layer pavementscontaining cement-stabilized base. PhD Dissertation. University of Californiaat Berkeley; 1970.

[33] Raad L, Figueroa JL. Load response of transportation support systems. TransEng J, ASCE 1980;106:11128.

[34] Lav AH, Kenny PJ. Leaching characteristics of eraring (NSW) y ash as roadbase material. Road&Trans Res, ARRB 1997;6:5263.

16 M. Aysen Lav, A. Hilmi Lav / Construction and Building Materials 54 (2014) 1016

Effects of stabilization on resilient characteristics of fly ash as pavement material1 Introduction2 Structure of highways3 Properties of fly ash and stabilizing agents used in the study4 Repeated load indirect tensile test of stabilized fly ash5 Stress strain relationship and triaxial testing of pavement materials5.1 Cyclic triaxial testing of stabilized fly ash

6 Summary of testing program7 Environmental concerns8 Discussion and conclusionsReferences