Embed Size (px)
Citation preview
Construction and Building Materials 48 (2013) 700–707
Contents lists available at ScienceDirect
Construction and Building Materials
journal homepage: www.elsevier .com/locate /conbui ldmat
Moisture sensitivity of asphalt mixtures under different load frequenciesand temperatures
0950-0618/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.conbuildmat.2013.07.059
⇑ Corresponding author. Tel.: +98 21 6640 0243; fax: +98 21 6641 4013.E-mail addresses: [email protected] (M.H. Dehnad), [email protected]
(A. Khodaii), [email protected] (F. Moghadas Nejad).
M.H. Dehnad, A. Khodaii ⇑, F. Moghadas NejadDepartment of Civil & Environmental Engineering, Amirkabir University of Technology, Tehran, Iran
h i g h l i g h t s
�Moisture sensitivity of HMA under various environmental and traffic conditions were evaluated.� Dynamic creep test was performed to compare permanent deformation of dry and wet samples.� At 40 �C with decreasing frequency, percent of increase in permanent strain was more in wet samples.� At 40 �C moisture has more detrimental effect on the HMA compared to 5 �C.� Ratio of creep to resilient modulus can be a good indicator for evaluating moisture susceptibility.
a r t i c l e i n f o
Article history:Received 26 February 2013Received in revised form 1 July 2013Accepted 21 July 2013Available online 24 August 2013
Keywords:Asphalt mixturesMoisture sensitivityPermanent deformationDynamic creep test
a b s t r a c t
Effects of different parameters including aggregate type, asphalt mix design and construction, traffic andenvironmental conditions on moisture sensitivity of asphalt mixtures have been investigated by a largenumber of researchers. But, only a few researchers have reported combined effects of these parameters.The aim of the current study is to evaluate moisture sensitivity of asphalt mixtures under various envi-ronmental and traffic conditions and to assess their combined effect on permanent deformation. Dynamiccreep test was carried out on the saturated and dry samples of asphalt mixtures made with dense gradedaggregate to compare their permanent deformation behaviors in the presence and absence of moisture.The results of examining different temperature and loading combinations indicated that, at 40 �C withdecreased frequency, the rate of increase in permanent deformation was more in saturated comparedwith the dry samples. Moreover, at this temperature, moisture had more detrimental effect comparedwith the temperature of 5 �C. Also, at 5 �C, the effect of moisture damage on the samples increased withthe increase in frequency. According to the test results, the ratio of creep modulus to resilient moduluscan be an appropriate indicator for evaluating moisture susceptibility at high temperatures.
� 2013 Elsevier Ltd. All rights reserved.
1. Introduction
Water is always expected to be present in an asphalt pavement.Several sources can lead to the presence of water in the pavementand cause premature as well as severe pavement failures [1]. Mois-ture damage can be defined as the loss of strength and durability inasphalt mixtures due to the effects of moisture [2]. When seriousenvironmental conditions act together with poor and/or unsuitablematerials and traffic, premature failure is expected [3].
Moisture damage of asphalt concrete may be resulted due totwo basic failure mechanisms of adhesion and cohesion, whichare essential to be distinguished from each other while discussingmoisture damage. When the asphalt film is separated from theaggregate surface completely, adhesion failure becomes apparent
and bare aggregate is visible when pavement is broken apart.Cohesion failure manifests itself as softening of the asphalt binderin an emulsification process as water penetrates into the binder.Cohesion failure may not create bare aggregate; but asphalt mix-ture will have low strength. A combination of these two mentionedfailure mechanisms may result in moisture damage; however, rel-ative effects of each mechanism on the failure is difficult to be dis-tinguished [4].
Factors that can affect moisture sensitivity of a mixture can beclassified in three main categories. The first category is materialproperties including physical and chemical properties of asphaltand aggregate. The second one is mixture properties including as-phalt content, film thickness and permeability of the mixture(interconnectivity of air voids). The third one is external factorsincluding construction, traffic and environmental factors [1].
Not only the moisture damage is an independent damage, butalso it can be a prelude to other damages. Rutting is one of thecommon types of distresses that can be related to moisture [5].
Fig. 1. Aggregate grading of asphalt mixtures.
M.H. Dehnad et al. / Construction and Building Materials 48 (2013) 700–707 701
Rutting or permanent deformation is accumulation of small defor-mations caused by densification and/or repeated shear deforma-tions under the applied wheel loads. Permanent deformation canbe attributed to plastic properties of asphalt mixture under re-peated loading but it cannot be related to elastic properties of as-phalt mixture [6]. A number of different methods can be employedin the laboratory to evaluate resistance of asphalt mixtures to per-manent deformation, which include axial compression creep, shearand wheel-tracking tests [6]. Effects of evaluation method on theprediction of moisture damage of hot mix asphalt were investi-gated by Abo-Qudais [7]. His results demonstrated that creep testwas the best method to monitor influence of the used asphaltand aggregate grading on hot mix asphalt (HMA) moisture damage.Moreover, the creep test has been identified as the most appropri-ate method for evaluating effect of additives on reduction of mois-ture damage. A complete review of the state of the art of moisturesensitivity of bituminous mixtures and methods of its evaluation isgiven by Mehrara and Khodaii [8].
Permanent deformation tests on asphalt mixture were per-formed using repeated loading by Cheng et al. under the presenceand absence of moisture [9]. Their findings indicated that the spec-imens tested after moisture conditioning accumulated more dam-age than those tested before moisture conditioning.
The effect of moisture on strength and permanent deformationof foamed asphalt mixes incorporating reclaimed asphalt pave-ment (RAP) materials was assessed by He et al. [10]. Indirect ten-sile and dynamic creep tests were carried out under dry andsoaked conditions to find moisture effects on indirect tensilestrength and susceptible–resistance to permanent deformation ofthese mixes. They selected repeated loading axial creep test toevaluate susceptibility of mixes to permanent deformation becausethe plastic deformation caused by repeated loading can be derivedfrom the dynamic creep curve. The result of these tests indicatedthat grade of bitumen and ageing of RAP material had considerableeffect on moisture susceptibility in permanent deformation.
Another study carried out by Abo-Qudais et al. investigated ef-fects of chemical and physical properties of aggregate on the creepand moisture damage behavior of hot-mix asphalt [11]. In theirstudy, the percent of increase in creep strain of mixtures due tomoisture conditioning was utilized to assess moisture sensitivity.Moreover, Xiao et al. conducted a laboratory investigation of rut-ting resistance in WMA mixtures containing moist aggregates.Their experimental design included two aggregate moisture con-tents and three aggregate sources. Test results indicated that theaggregate source significantly affects the rutting resistance regard-less of the WMA additive. Also, they showed the influence of mois-ture on rut depth can be neglected and it even results in a betterrut resistance in some cases [12]. Xiao et al. at another research[13] evaluated the moisture susceptibility of mixtures using indi-rect tensile strength (ITS). Their finding indicated that the dryand wet ITS values of HMA mixtures are higher while the WMAmixtures show lower ITS values regardless of the aggregate andstorage duration.
Permanent deformation of unmodified and SBS modified as-phalt mixtures were assessed by Khodaii and Mehrara using dy-namic creep test. The findings indicated that coarse gradedasphalt mixtures had more resistance to permanent deformationthan dense graded mixtures [14]. In another study, Mehrara andKhodaii made an attempt to assess moisture sensitivity and itsinteraction with permanent deformation through performing a dy-namic creep test on coarse graded and dense graded asphalt mix-tures [15]. According to their results, the coarse graded mixtureshad lower moisture sensitivity and could better resist permanentdeformations than the dense graded mix. They also concluded thatlow stress levels could not appropriately show the behavior of as-phalt specimens.
2. Statement and objectives
Conventional moisture susceptibility tests are not able to simu-late actual process of moisture damage formation in the field. Thisis partly because testing sample goes through conditioning andloading processes in two separate steps throughout the test. But,effects of environmental conditions and traffic loading cannot beseparated according to the extensive field investigations on propa-gation of moisture damage [16]. Selecting the dynamic creep testover other available tests in the present study was done to considerthe combined effect of loading and moisture simultaneously. Also,since moisture affects mechanical properties of asphalt mixturesuch as resilient modulus and deformations, the dynamic creeptest was considered as one of the more appropriate tests availablefor evaluating susceptibility of asphalt moisture.
The purpose of this study was to investigate effect of moistureon permanent deformation and assess moisture susceptibility ofasphalt mixtures using the outcomes of dynamic creep test underdifferent environmental and traffic conditions. The following stepswere followed in this study:
1. Comparison of permanent deformation potential of conditionedand unconditioned mixtures at different temperatures underconstant applied load frequency.
2. Comparison of permanent deformation potential of conditionedand unconditioned mixtures under different applied load fre-quencies at a fixed temperature.
3. Simultaneous evaluation of effects of moisture and loading onpermanent deformation.
4. Derivation of resilient and creep modulus parameters in differ-ent combinations of temperature and applied load frequencies.
3. Materials
Aggregate grading has an effect on the magnitude of moisture damage. Compar-ing a dense and coarse graded aggregate, Khodaii and Mehrara reported densegraded mixtures as more susceptible to moisture damage and permanent deforma-tions [14]. The load transfer in coarse graded mixtures depends more on the contactbetween stones due to the presence of a large number of coarse aggregates in themixture; consequently, behavior of asphalt mixture is less dependent on the masticproperties. But, properties of mastic in dense graded mixtures have a significantrole in determining behavior of asphalt mixtures and also have an influence onits deformation, particularly at high temperature when viscosity of bitumen de-creases. Thus the samples made with dense graded mixture are more susceptibleto permanent deformation.
A dense graded aggregate that is used for topeka and binder layers according toASTM D3515 [17] was selected for performing the test. Fig. 1 presents the aggregategrading used in this study. The two dotted curves in the figure represent the upperand lower limits of the permitted grading for pavement surface layer based onASTM D3515 [17]. The used aggregates were siliceous crushed with 85% brokenin two faces and the asphalt used in the study had 60/70 penetration grade compa-rable with PG 64-22. Tables 1 and 2 list mechanical and physical properties of the
Table 1Mechanical properties of the tested siliceous aggregate.
Test Standard Values (%) MS – 2 specifications (%)
LA abrasion loss ASTM C131 25 <30Crushed in one face ASTM D5821 87 –Fractured particles in two face and more ASTM D5821 93 90<Coating of aggregate AASHTO T182 95 956Flakiness BS-812 10 <25Sand equivalent ASTM D2419 85 50<Sodium sulfate soundness ASTM C88 0.4 <8
Table 2Physical properties of the tested siliceous aggregate.
Fraction Standard Specific gravity (g/cm3) Absorption
Apparent Bulk
Retained on 2.36 mm (No. 8) AASHTO T85 2.502 2.325 1.6Passed from 2.36 mm and retained on 0.075 mm AASHTO T84 2.498 2.316 1.6Passed from 0.075 mm (No. 200) AASHTO T100 2.425 2.312
Table 3Physical properties of the tested asphalt.
Test Standard AC 60/70
Ductility at 25 �C (cm) D 113 100Penetration at 25 �C, 100 g (0.1 mm) D 5 60Softening point (�C) D 36 49.4Flash Point (�C) D 92 320Specific gravity at 25 �C D 70 1.02
702 M.H. Dehnad et al. / Construction and Building Materials 48 (2013) 700–707
siliceous aggregate source employed in this study. Using Marshall Test, the opti-mum amount of asphalt for dense graded mixtures was estimated as 5.4%. Physicalproperties of the asphalt are also presented in Table 3.
4. Test procedures
4.1. Preparing the samples
The testing samples were constructed according to ASTMD1559 [18]. To obtain almost the same bulk specific gravity for dif-ferent types of samples, trial and errors were carried out, throughwhich the number of blows in Marshall Compactor was selected. Inthis way, decrease in the effect of variation in volumetric parame-ters on the results was ensured; consequently, reasonable compar-ison could be made between different mixtures. Air void of thesamples in this research was selected as 7 ± 1% to simulate asphal-tic pavements in their early ages after the construction.
Saturation of the samples was carried out according to proce-dure suggested by ASTM D4867 [19]. The wet samples were keptinside a container filled with water to maintain their saturatedstate during the long hours of testing and to prevent drainage. Itis believed that water could frequently enter and exit from porespaces of the sample during each application of dynamic loading.
4.2. Dynamic creep test
Dynamic creep test was performed using UTM25 and applyingrepeated axial stress pulse to asphalt specimens while measuringvertical deformation with Linear Variable Displacement Transduc-ers (LVDTs). In servo hydraulic UTM25 machine, the stress/load ap-plied to the specimen was feedback controlled, providing thecapability for the operator to select a loading wave shape (haver-sine, triangular or square pulse), a pulse width duration, a rest per-iod, a deviator stress/load during each loading pulse and a contactstress/load to ensure that the vertical loading shaft did not lift offthe test specimen during the rest period. Loading mechanism ofUTM machine was placed within an environmental chamber tocontrol ambient temperature of the testing samples.
An integrated software was responsible to control input dataincluding dimensions of sample (height and diameter), preload/stress, deviator stress, frequency of stress application and contactstress as mentioned earlier. A series of pilot tests were carriedout before the main test to determine the appropriate pulse wave.In this series of tests, square and haversine pulse waves were usedwith frequencies of 1 and 0.5 Hz. According to the results obtained
from the pilot tests, application of haversine pulse waves at highfrequencies in the creep test did not provide a condition for waterto be drained during load application. On the other hand, in thecase of pulse wave application, enough constant maximum pres-sure during loading time was provided for water to squirt out ofor get into air voids of the sample. Thus, a square pulse wavewas selected in this study. Fig. 2 presents a sample of pulse waveand measured axial displacements of the specimen for one pulse.The diagram indicates that deformation of the specimen increasesby applying stress during the pulse width and climbing up to itspeak as loading was finalized. Resilient deformation which con-tained a significant portion of deformation disappeared duringthe rest period and the permanent deformation was actually theremaining deformation [20].
Resilient modulus and creep modulus are the most importantoutputs of dynamic creep test. Based on the definitions suggestedby a number of studies [14,21] and UTM25 software referencemanual [20], resilient modulus and creep modulus were derivedusing Eqs. (1)–(3), respectively:
Mr ¼rd
erð1Þ
Mc ¼rd
eðtÞ ð2Þ
eðtÞ ¼ ðee þ ep þ emeðtÞ þ evpðtÞÞ ð3Þ
In the above equations rr is deviator stress (kPa); er is resilientdeformation at a certain number of load applications (ls); eðtÞ is to-tal deformation (including elastic, visco-elastic, plastic and visco-plastic deformations) up to a certain number of load applications(ls); ee is elastic strain, recoverable and time-independent (ls);ep is plastic strain, irrecoverable and time-independent (ls); emeðtÞ
Fig. 2. A pulse wave and the measured displacement.
M.H. Dehnad et al. / Construction and Building Materials 48 (2013) 700–707 703
is viscoelastic strain, recoverable and time-dependent (ls); andempðtÞ is viscoplastic strain, irrecoverable and time-dependent (ls).
Resilient modulus, as derived from Eq. (1), indicates the sampleresistance to resilient deformations and creep modulus indicatesits resistance to permanent deformation. By the application of loadto the samples in each cycle, three sets of diagrams consisting ofpermanent deformation, resilient modulus and creep modulus ver-sus load cycles were drawn by UTM software.
4.3. Input data for testing
To assess the samples’ behavior under different applied loadfrequencies and ambient temperatures stress controlled tests werecarried out, with three different frequencies of 0.5, 1 and 10 Hz andat three different temperature levels of 5 �C, 20 �C and 40 �C. Thefindings of Kim and Coree [22] indicated that, under the stress lev-els of above 200 kPa, the failure of samples was more probable dueto their materials’ properties than moisture effect. On the otherhand, the low stress levels of 100 kPa might not be powerful en-ough to induce permanent deformations in the samples, especiallyin the presence of water pressure [15]. Comparing the creep curvewith Zhou’s model indicated that the samples did not obey themodel and were inclinedto reach an elastic state at 100 kPa stresslevel. Therefore, 100 kPa could not be considered an ample stresslevel to assess moisture sensitivity in compressive creep test[15]. According to these investigations, the most appropriate loadto be applied on the samples was selected as 200 kPa. Moreover,a 5-min preloading process was set into the testing sequenceswhich included static stress with a magnitude equivalent to 10%of dynamic stress. The preloading applied before the dynamiccreep ensured that the samples’ surface and loading plate werein full contact with each other and that the loose parts in the sam-ples’ surface had their displacements. Hence, the deformation mea-sured during the creep tests was related to the samples’ resilientand creep modulus. Samples were kept inside the constant temper-ature chamber for two hours to ensure that they are at the desiredtemperature prior to testing.
Two thermal sensors were installed to monitor chamber andsample temperature simultaneously; one was inside a dummysample and another in the environment around the testing samplein the chamber. This action provided the possibility of recordingthe exact temperature in the sample and the surrounding environ-ment. Testing only started after the environment and samplereached the desired testing temperature. To increase reliability ofthe results, each test was performed with three replicates andthe average results were shown in the diagrams.
5. Results and discussion
Deformation behavior of dry and saturated samples in differentcombinations of temperature and frequency as well as the poten-tial of moisture damage occurrence is investigated in this section.Average physical properties of the tested samples in different com-binations of dry and saturated samples are presented in Table 4.
5.1. Temperature of 40 �C and frequencies of 0.5, 1 and 5 Hz
As is shown in Fig. 3, the permanent strain in the saturated sam-ple is smaller than that in dry sample in the first 3200 cycles andthen supersedes it so that, at the 4000th cycle, the permanentdeformation in the saturated sample is about 20% more than thatin the dry sample.
Higher resistance of the saturated samples in initial cycles isattributed to the fact that the adhesion between asphalt and aggre-gate as well as the cohesion in asphalt structure is not threatenedwith a serious breakdown in the early stages of loading. On theother hand, the Presence of water in the pores of asphalt samplecreates pore water pressure that absorbs a part of applied loadand consequently reduces the influence of dynamic loading. But,with continuing dynamic load application and repeated entranceand exit of water to and from the asphalt mixture pores, watergradually penetrates into the surface between aggregate and as-phalt and causes moisture damage. This obviously speeds up the
Table 4Volumetric properties of the samples.
Moisture condition Test temperature (�C) Frequency (Hz) Bulk specific gravity Maximum specific gravity Air void (%) Percentage of saturation (%)
Dry 40 0.5 2.3 2.48 7.17 –Saturated 40 0.5 2.3 2.48 7.12 75.6Dry 40 1 2.3 2.48 7.15 –Saturated 40 1 2.3 2.48 7.15 79.8Dry 40 5 2.31 2.48 6.69 –Saturated 40 5 2.31 2.48 6.68 79.4Dry 20 0.5 2.31 2.48 7.04 –Saturated 20 0.5 2.3 2.48 7.4 84Dry 20 1 2.31 2.48 6.95 –Saturated 20 1 2.31 2.48 6.89 62Dry 20 5 2.32 2.48 6.61 –Saturated 20 5 2.3 2.48 7.1 78.8Dry 5 0.5 2.31 2.48 7.04 –Saturated 5 0.5 2.3 2.48 7.11 63.1Dry 5 1 2.31 2.48 6.79 –Saturated 5 1 2.31 2.48 6.66 85Dry 5 5 2.3 2.48 7.19 –Saturated 5 5 2.3 2.48 7.37 53
704 M.H. Dehnad et al. / Construction and Building Materials 48 (2013) 700–707
damage and hence increased permanent deformation in saturatedsamples is observed after the initial stages of loading.
Furthermore, as it can be seen in Fig. 3 the second phase in thedry sample tested at 40 �C and 0.5 Hz frequency starts at 700 andfinishes at 4000 where the third phase starts, so it has 3300 cycleduration. The saturated sample tested at the same condition startsits second phase of the creep curve at 700 and finishes at 2800hence has 2100 cycle duration and also enters its third phase1200 cycles earlier than the dry sample.
A more accurate estimation of beginning of each phase of thecreep behavior can be found in the diagram of strain rate versusthe loading cycles. As can be seen in Fig. 4, the comparison be-tween behavior of the dry and saturated samples demonstratedthat increase in the rate of deformation in the saturated sampleis more than the one in the dry sample in third phase. For instance,in the 4400th cycle, the rate of permanent deformation in the sat-urated sample is 4.5 times that of the dry sample. This indicates afaster deformation rate in the saturated sample which is due to thepresence of moisture and increased loss of adhesion between as-phalt and aggregate.
A similar behavior in the creep curve of dry and saturated sam-ple is observed when tested using 1 Hz frequency (Fig. 5). The rateof increase in strain of the samples tested at frequencies of 0.5 and1 Hz clearly indicates that the strain in the dry and saturatedsamples at frequency of 0.5 Hz is more than the strain in the dryand saturated samples at 1 Hz. This difference was expected dueto the fact that, at high frequencies, the actual loading time
0
5000
10000
15000
20000
25000
30000
35000
40000
45000
50000
0 1000 2000 3000 4000 5000 6000
Per
man
ent
Stra
in (
µs)
Cycle
Dry Wet
Fig. 3. Permanent strain versus the number of loading cycles at 40 �C and frequencyof 0.5 Hz in the dry and saturated states.
application decreases and there is less opportunity for the occur-rence of permanent deformation in the sample.
From the comparison of the amount and rate of strain at 0.5 and1 Hz, it can be concluded that the ratio of strain rate in the satu-rated to dry samples in the 4000th cycle and at 0.5 Hz is equal to1.23 while the same ratio for the 1 Hz frequency is 1.07. On theother hand, the ratio of the strain rate in the saturated to dry sam-ples in the same cycle is equal to 3.5 and 2.2 at 0.5 Hz and 1 Hzrespectively. Therefore, it can be concluded that, at 0.5 Hz, mois-ture had a more detrimental effect.
Increasing the frequency from 1 to 5 Hz and applying 10,000 cy-cles of loading at a constant temperature as shown in Fig. 5 indi-cates that unlike the two previous test results, permanentdeformation in the saturated sample is more than the one in thedry sample from the beginning and also the saturated samplesdemonstrates less resistance to permanent deformation. In thiscase, with the increase in loading speed, moisture influences theadhesion between asphalt and aggregate significantly and de-creases its resistance and hence an increase rate of permanentdeformation in the saturated sample is detected.
5.2. Temperature of 20 �C and frequency of 0.5, 1 and 5 Hz
The permanent deformation of dry and saturated samplestested at 20 �C and 0.5 Hz frequency is shown in Fig. 6. It is obser-vable that the dry sample nearly reaches a stable state at the end of
0
5
10
15
20
25
30
35
0 1000 2000 3000 4000 5000 6000
Stra
in S
lope
Cycle
Dry Wet
Fig. 4. Strain slope versus the number of loading cycles at 40 �C and frequency of0.5 Hz in the dry and saturated states.
0
5000
10000
15000
20000
25000
30000
35000
40000
45000
0 2000 4000 6000 8000 10000 12000
Per
man
ent
Stra
in (
µs)
CycleDry, F=1 Hz Wet, F=1 Hz Dry, F=5 Hz Wet, F=5 Hz
Fig. 5. Permanent strain versus the number of loading cycles at 40 �C andfrequencies of 1 and 5 Hz in the dry and saturated states.
0
2000
4000
6000
8000
10000
12000
14000
0 2000 4000 6000 8000 10000 12000
Per
man
ent
Stra
in (
µs)
Cycle
Dry, F=1 Hz Wet, F=1 Hz Dry, F=5 Hz Wet, F=5 Hz
Fig. 7. Permanent strain versus the number of loading cycles at 20 �C andfrequencies of 1 and 5 Hz in the dry and saturated states.
M.H. Dehnad et al. / Construction and Building Materials 48 (2013) 700–707 705
the 10000th cycles and the slope approached to zero; however, thisbehavior is not mirrored in the saturated samples and the rate ofstrain is higher. In this case, through exerting hydraulic pressureresulting from dynamic application of load to the saturated sample,the adhesion between asphalt and aggregate is disturbed, whichcause more damage to the saturated samples.
Comparison of the strain behavior at two temperatures of 40and 20 �C and frequency of 0.5 Hz (as demonstrated in Figs. 3and 6) show that strain values and their rate increases in the satu-rated samples with temperature increase. This indicates that, at40 �C, effect of moisture damage is more serious than that at20 �C. Another difference that could be seen in strain behavior intwo mentioned temperature conditions is that, at 40 �C, strain inthe dry sample is more than the strain in the saturated one inthe beginning but this state gradually reverses due to the presenceof moisture.
Fig. 7 demonstrates that, at both frequencies of 1 and 5 Hz,deformation in the saturated sample is more than the one in thedry sample. It can be noticed that, at temperature of 20 �C and fre-quency of 1 Hz, the sample did not enter its third phase up to the10000th cycle of the loading, which is obviously due to reducedpermanent deformation in lower temperature.
5.3. Temperature of 5 �C and frequency of 0.5, 1 and 5 Hz
Fig. 8 exhibits creep curve of the dry and saturated samples at5 �C and different loading frequencies. It can be noted that, withthe increase of frequency, the difference between strain in the sat-urated and dry samples grew up and the saturated samples exhib-ited less resistance to the applied load compared with the drysamples. As naturally expected, with the increase in frequency,strain in the samples decreased.
At this temperature, as the frequency increases, effect of mois-ture damage on the samples also increases. This is completely
0
2000
4000
6000
8000
10000
12000
14000
0 2000 4000 6000 8000 10000 12000
Per
man
ent
Stra
in (
µs)
Cycle
Dry Wet
Fig. 6. Permanent strain versus the number of loading cycles at 20 �C and frequencyof 0.5 Hz in the dry and saturated states.
opposite the result obtained at 40 �C. It is evident that, at 40 �Cas the frequency increases effect of moisture damage on theincrease of strain in the saturated samples decreases. This is be-cause, at high temperatures when the viscosity of asphalt is low,moisture penetration into the mix is easier and cause significantdecrease in adhesion. Also, longer loading period provides longeropportunity for the moisture to enter and exit the pores and reachthe aggregate surface; at high frequencies, less opportunity isavailable for moisture penetration and therefore smaller decreasein adhesion is observed. On the other hand, at low temperaturedue to high viscosity of asphalt binder, the moisture and water in-side the pores have a minor effect on the decrease of adhesion be-tween asphalt binder and aggregate.
5.4. Variations of creep and resilient modulus
At higher temperatures, the asphalt binder became less viscous.This lower viscosity produces less stiff pavement that could be sus-ceptible to creep attributed to traffic loads. Indeed, creep resistanceis a property of the mix, desirable at high temperatures; Hencecreep susceptibility of the mixtures was studied at high tempera-ture (40 �C).
Figs. 9 and 10 represent variations in resilient modulus (Mr) ofthe dry and saturated samples at 40 �C and frequencies of 0.5and 5 Hz respectively. As is apparent at both frequencies, resilientmodulus of the dry samples was more than that of the saturatedsamples. A falling trend in resilient modulus curves could be seenat both frequencies but the descending rate at frequency of 5 Hz isless than the rate for 0.5 Hz and these curves reach a steady stateafter about 4000 cycles.
Resilient modulus of the samples at 0.5 Hz frequency decreaseswith the increase in loading cycles while, at frequency of 5 Hz, withincrease in the number of loading cycles, resilient modulus almostreaches a fixed status and keeps its elastic properties.
0
2000
4000
6000
8000
10000
12000
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Per
man
ent
Stra
in (
µs)
Cycle
Dry, F=0.5 Hz Wet, F=0.5 Hz Dry, F=1 Hz
Wet, F=1 Hz Dry, F=5 Hz Wet, F=5 Hz
Fig. 8. Permanent strain versus the number of loading cycles at 5 �C andfrequencies of 0.5, 1 and 5 Hz in the dry and saturated states.
0
50
100
150
200
250
300
0 1000 2000 3000 4000 5000 6000
Res
ilien
t m
odul
us (
MP
a)
Cycle
dry wet
Fig. 9. Resilient modulus changes for the dry and saturated samples tested atfrequency of 0.5 Hz at 40 �C.
0
50
100
150
200
250
300
350
400
0 2000 4000 6000 8000 10000 12000
Res
ilien
t m
odul
us (
MP
a)
Cycle
dry wet
Fig. 10. Resilient modulus changes for the dry and saturated samples tested atfrequency of 5 Hz at 40 �C.
0
10
20
30
40
50
60
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Cre
ep m
odul
us (
MP
a)
Cycle
Dry, F=0.5 Hz Wet, F=0.5 Hz Dry, F=5 Hz Wet, F=5 Hz
Fig. 11. Creep modulus changes for the dry and saturated samples tested atfrequency of 0.5 and 5 Hz at 40 �C.
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0.2
0 2000 4000 6000 8000 10000
Mc/
Mr
Cycle
Dry, F=0.5 Hz Wet, F=0.5 Hz Dry, F=5 Hz Wet, F=5 Hz
Fig. 12. Changes of the ratio of creep to resilient modulus for the dry and saturatedsamples tested at frequencies of 0.5 and 5 Hz.
706 M.H. Dehnad et al. / Construction and Building Materials 48 (2013) 700–707
Although the general trend of resilient modulus of the sampleswas descending, a slight ascent in initial cycles can be observed.Previous studies have indicated that the deformation due to com-paction of asphalt mixture has considerable influence on resilientmodulus [14]; therefore, the increase in the initial stage of loadingcould well be due to increased compaction of the sample. However,this increase is more at frequency of 0.5 Hz and also it is morenoticeable in the dry samples compared with the saturatedsamples.
Considering Fig. 11, it can be observed that creep modulus had afalling trend. At frequency of 5 Hz, the rate of falling is low and italmost reached a smooth slope. With frequency decrease to0.5 Hz, variations in creep modulus (Mc) decreases with a high rateuntil the sample reaches a complete failure.
Gokhale et al. [23] showed that the permanent deformation ofasphalt mixtures usually consisted of two parts: the compactiveand shear flow (plastic). In dynamic creep test performed byUTM, it is not possible to distinguishing permanent deformationcaused by compaction from the permanent deformation caused
by shear deformations. However, it is believed that, by continuingthe loading process, the rate of compactive deformation decreasesand the rate of plastic deformation increase. The results presentedin Figs. 9 and 10 show increased resilient modulus of the samplesin the first few repetitions of load, as an indication of the samplesbecoming more compacted; on the contrary, UTM considers thisdeformation as a permanent deformation and indicates a reductionin creep modulus.
According to the findings of Gokhale et al. [23] and Khodaii andMehrara [14], the ratio of creep modulus to resilient modulus is anindex of the ratio of plastic deformation to compactive deforma-tion in the asphalt mixture samples. The higher rate of decreasein this ratio demonstrates that samples have more vulnerabilitypotential to permanent deformation. To investigate the effect ofmoisture on permanent deformation of the samples, the abovementioned ratio was used in this study. Fig. 12 indicates ratio ofcreep modulus to resilient modulus in the dry and saturated sam-ples at two frequencies of 0.5 and 5 Hz. As can be noticed in Fig. 12,at frequency of 0.5 Hz, the falling rate of this ratio in the saturatedsamples was faster than the one in the dry samples. But at fre-quency of 5 Hz, these rates were nearly the same. Process of com-paction occurred earlier than growth of plastic deformation whenthe ratio of plastic deformation to compactive deformation isdecreased.
At frequency of 5 Hz, the samples reached a fixed rate ratiowhich indicates that a minor variation in Mc and Mr occurred andthey had a similar rate. At frequency of 0.5 Hz, the dry and satu-rated samples had separate falling rates, which implies that thefalling rate in Mc is higher than the falling rate in Mr Since Mc/Mr
ratio of the samples clearly decreases at frequency of 0.5 Hz andalso Mr decreases in them, it can be deducted that the ratio ofcompactive to shear deformation decreases due to increase incompactive deformation. This finding was similar to the one ob-tained by Khodaii and Mehrara [14] and Gokhale et al. [23].
It can also be understood from Fig. 12 that, when Mc/Mr isdecreasing slowly (like approaching a constant value), compactionwas still happening or the growth of shear deformation was negli-gible. When the Mc/Mr ratio falls considerably (not in early cyclesof dynamic load application), the samples are going through theshear deformation. It can therefore be concluded that in Fig. 12,the samples experience shear deformation at frequency of 0.5 Hz.
6. Conclusions
Using dynamic creep test on the saturated and dry asphalt mix-tures under controlled environmental conditions and also applyingdifferent frequencies, the present study attempted to determinemoisture sensitivity of asphalt concrete under different conditions.
The most important results obtained from this study could besummarized as follows:
M.H. Dehnad et al. / Construction and Building Materials 48 (2013) 700–707 707
� At high temperature (40 �C) the comparison between behaviorof the dry and saturated samples demonstrated that increasein the rate of deformation in the saturated sample was morethan the one in the dry sample in third phase. Moreover at thistemperature, with increased frequency, effect of moisture dam-age on the permanent strain of the samples decreases.� At low temperature (5 �C), with the increase of frequency, the
difference between permanent strain in the saturated and drysamples grew up and the saturated samples exhibited less resis-tance to the applied load compared with the dry samples. Alsoas frequency increased, effect of moisture damage on the sam-ples enhanced.� Although moisture damage appeared earlier at 20 �C and it had
a main role in the occurrence of permanent strain in the sample,its effect was more detrimental at 40 �C.� The ratio of creep modulus to resilient modulus could be used
as an index of the ratio of plastic deformation to compactivedeformation in asphalt mixtures.� Dynamic creep test could be used as a proper method to evalu-
ate moisture sensitivity of asphalt mixtures and effect of mois-ture on permanent deformation could be determined using thistest.
References
[1] Santucci L. Moisture sensitivity of asphalt pavements. Tech Topics 2002.[2] Little DN, Jones I. Chemical and mechanical processes of moisture damage in
hot-mix asphalt pavements. In: moisture sensitivity of asphalt pavements – anational seminar; 2003.
[3] Brown ER, Kandhal PS, Zhang J. Performance testing for hot mix asphalt.National Center for Asphalt Technology Report 2001(01-05).
[4] Maupin Jr G. Assessment of stripped asphalt pavement; 1989.[5] Kandhal PS. Moisture susceptibility of HMA mixes: identification of problem
and recommended solutions. National Asphalt Pavement Association; 1992.
[6] He G, Wong W. Laboratory study on permanent deformation of foamed asphaltmix incorporating reclaimed asphalt pavement materials. Constr Build Mater2007;21(8):1809–19.
[7] Abo-Qudais S. The effects of damage evaluation techniques on the prediction ofenvironmental damage in asphalt mixtures. Build Environ 2007;42(1):288–96.
[8] Mehrara A, Khodaii A. A review of state of the art on stripping phenomenon inasphalt concrete. Constr Build Mater 2013;38:423–42.
[9] Cheng DX, Little DN, Lytton RL, Holste JC. Moisture damage evaluation ofasphalt mixtures by considering both moisture diffusion and repeated-loadconditions. J Transport Res Rec 2003;1832(-1):42–9.
[10] He GP, Wong WG. Effects of moisture on strength and permanent deformationof foamed asphalt mix incorporating RAP materials. Constr Build Mater2008;22(1):30–40.
[11] Abo-Qudais S, Al-Shweily H. Effect of aggregate properties on asphalt mixturesstripping and creep behavior. Constr Build Mater 2007;21(9):1886–98.
[12] Xiao F, Amirkhanian SN, Putman BJ. Evaluation of rutting resistance in warm-mix asphalts containing moist aggregate. J Transport Res Rec 2010;2180(1):75–84.
[13] Xiao F, Zhao W, Gandhi T, Amirkhanian SN. Laboratory investigation ofmoisture susceptibility of long-term saturated warm mix asphalt mixtures. IntJ Pavement Eng 2012;13(5):401–14.
[14] Khodaii A, Mehrara A. Evaluation of permanent deformation of unmodifiedand SBS modified asphalt mixtures using dynamic creep test. Constr BuildMater 2009;23(7):2586–92.
[15] Mehrara A, Khodaii A. Evaluation of Asphalt mixtures’ moisture sensitivity bydynamic creep test. ASCE J Mater Civil Eng 2010;23(2):212–9.
[16] Kandhal P, Rickards I. Premature failure of asphalt overlays from stripping:case histories. Asphalt Paving Technol 2001;70:301–51.
[17] ASTM D 3515-01. Standard specification for hot-mixed, hot-laid bituminouspaving mixtures: annual book of ASTM standard; 2004.
[18] ASTM D 1559. Method for resistance of plastic flow of bituminous mixturesusing marshall apparatus: annual book of ASTM standard; 2003.
[19] ASTM D 4867/D4867M. Standard test method for effect of moisture on asphaltconcrete paving mixtures: annual book of ASTM standard; 1996.
[20] UTM software manual; 2006.[21] Witzcak MW. Simple performance test for superpave mix design: transportation
research board; 2002.[22] Kim S, Coree BJ. Evaluation of hot mix asphalt moisture sensitivity using the
Nottingham asphalt test equipment; 2005.[23] Gokhale S, Choubane B, Byron T, Tia M. Rut initiation mechanisms in asphalt
mixtures as generated under accelerated pavement testing. J Transport Res Rec2005;1940(-1):136–45.
