Construction and Building Materials 155 (2017) 531–541

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Analysis on the fatigue properties of vertical vibration compacted

lime–fly ash-stabilized macadam
Jinshun Xue ⇑, Yingjun Jiang
Key Laboratory for Special Area Highway Engineering of Ministry of Education, Chang’an University, Xi’an 710064, Shaanxi, China

h i g h l i g h t s

Fatigue performances of the LFASM compacted by Vibration Testing Method are evaluated.

Fatigue equations of LFASM are established.

Suggested values of tensile strength structure coefficient of the LFASM are proposed.

a r t i c l e i n f o a b s t r a c t

Article history: Although lime–fly ash-stabilized macadam (LFASM) is a common material for the road base course of
Received 7 March 2017 semi-rigid pavements, the fatigue performance of compacted LFASM has been rarely studied through
Received in revised form 23 July 2017 the vibration testing method (VTM). In this work, LFASM was designed using the VTM instead of the
Accepted 10 August 2017
quasi-static testing method. The vibration compaction test method was used to determine optimum
Available online 23 September 2017
water content and maximum dry density, and cast specimens were analyzed through the vibration pres-
sure producing specimen method. The fatigue performances of the LFASM specimens were evaluated, and
Keywords:
the fatigue equations of LFASM were then established. The suggested values for the coefficients of the
Road engineering
Lime fly-ash stabilized macadam
tensile strength structure of LFASM were proposed on the basis of the derived fatigue equations.
Vibration Testing Method Results show that the splitting strength of LFASM increased with increasing lime and fly ash contents.
Splitting strength However, the rates of the increase in the splitting strength of LFASM gradually declined. The splitting
Fatigue equation strength of LFASM with skeleton-dense (SD) gradation was higher than that of LFASM with gradation
Tensile strength structure coefficient of specification (SG). The fatigue life of LFASM conformed to the two-parameter Weibull distribution,
and the established fatigue equations accurately reflected the fatigue life of CRPMs. LFASM with SD gra-
dation showed potentially better fatigue resistance than LFASM with SG. As the contents of lime and fly
ash increased, the effect of gradation on the fatigue performance of LFASM became negligible. The failure
probability based on the fatigue equations was 50%, the values of the regression coefficient a of the fati-
gue equations ranged from 1.198 to 1.278, and the values of coefficient b ranged from 0.051 to 0.059. The
formula of the tensile strength structure coefficient Ks was proposed. The coefficients for tensile strength
structure were recommended after considering various factors. The value of Ks obtained in this work can
be used to effectively guide the design of pavement structures. The pavement designed with VTM offers
many advantages in terms of crack resistance and durability.
Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction to pozzolanic reactivity [5]. Lime–fly ash-stabilized macadam

Fly ash, a waste material of coal-fired power stations, is widely
used in civil engineering [1–4]. Conventional construction materi-
als are mainly replaced with fly ash because of economic consider-
ations and waste utilization. Additionally, the strength of materials
that contain fly ash may continue to increase over long periods due
⇑ Corresponding author.
E-mail address: jinshunx@chd.edu.cn (J. Xue).
(LFASM) is a common semi-rigid base material that is extensively
applied in highway base courses because of its good pavement per-
formance [6–8].
The physical and mechanical behaviors of LFASM, including
maximum dry density, optimal water content, compressive
strength, splitting strength, resilient modulus, shrinkage perfor-
mance, and fatigue performance, have been extensively studied
[9–13]. The fatigue cracking of LFASM in response to seasonal
changes in ambient temperature and repeated vehicle load is a

http://dx.doi.org/10.1016/j.conbuildmat.2017.08.057
0950-0618/Ó 2017 Elsevier Ltd. All rights reserved.
532 J. Xue, Y. Jiang / Construction and Building Materials 155 (2017) 531–541

Nomenclature

LFASM lime fly-ash stabilized macadam MTS material testing system

LFA lime fly-ash Ri splitting strength
VTM vibration testing method rm tensile stress
QSTM quasi-static testing method rS limit splitting strength
SG specification gradation KS tensile strength structure coefficient
SD skeleton dense gradation Ne cumulative equivalent axle loads
VVTE vertical vibration test equipment Ac coefficient of highway classification
VCTM vibration compaction testing method
VPSM vibration pressure producing specimen method

common issue. Thus, the fatigue performance of LFASM should be
studied for material evaluation and pavement design. Mallela et al.
optimized the design method of LFASM layers on the basis of fati-
gue properties [14]. Panda analyzed the fatigue fracture standards
of LFASM by conducting laboratory tests and then applied the
obtained results to the reconstruction of pavement design [15].
Judycki established fatigue formulas in accordance with specific
criteria and compared the fatigue curves of flexible and semi-
rigid pavement [16]. Lav et al. utilized accelerated full-scale road
test data for the fatigue performance of cement-stabilized fly ash
and then performed a mechanistic empirical design procedure to
obtain the appropriate layer thicknesses for pavement with differ-
ent amounts of cement content to improve service life [17]. The
effects of structure type and flexural tensile strength on the fatigue
life of LFASM have been experimentally studied with the aim of
improving the service life of pavement structure [18]. Li studied
the fatigue damage forecast model of LFASM [19]. Gao studied
formance of LFASM based on the VTM should be evaluated. In the
current work, LFASM with different gradations and contents of
lime–fly ash (LFA) were prepared and compacted in the laboratory
using the VTM. Then, the fatigue performances of the LFASM spec-
imens were evaluated, and the fatigue equations of LFASM were
established. The suggested values for the coefficients of the tensile
strength structure of LFASM were proposed on the basis of the
derived fatigue equations.
2. Materials and experiment program
2.1. Raw materials
2.1.1. Lime and fly ash
Lime and fly ash was obtained from Tongchuan city, Shaanxi province, China.
The chemical and physical properties of these materials and the requirements of
the Technical Guidelines for the Construction of Highway Road Bases (JTG/TF20-
2015) are shown in Tables 1–3.

the fatigue performance and fatigue fracture mechanism of LFASM

on the basis of linear fatigue–damage accumulation theory and 2.1.2. Aggregate
The macadam was composed of four types of aggregates with different size
fracture mechanics [20].
ranges. The aggregates were graded into four types of coarse and fine aggregates
In practice, inorganic-binder-stabilized materials are com- as follows: 37.5–19 mm, 19–9.5 mm, 9.5–4.75 mm, and less than 4.75 mm. The
pacted through the vibration testing method (VTM) and quasi- technical performances of the aggregates were tested in accordance with The Test
static testing method (QSTM) [21–24]. The QSTM is commonly Methods of Aggregate for Highway Engineering (JTG E42-2005). The test results are
shown in Table 4.
used to compact LFASM because it is easy to employ and it does
not require the use of expensive equipment. However, the correla-
2.1.3. Mixtures
tion between laboratory specimens prepared using the QSTM with
LFASM is multi-phase composite material that is composed of lime, fly ash, and
in-field specimens is less than 40% because the static pressure that aggregates. Thus, LFASM is composed of two dispersed systems: the LFA (lime–fly
acts on aggregates during QSPM prevents aggregate motion. As sta- ash) mortar microdispersed system and the aggregate dispersed system. The LFA
tic pressure increases, the friction force between the aggregates specimens exhibit the optimal mechanical and shrinkage ratios when the mass ratio
increases. The aggregates are thus crushed, which consequently of lime to fly ash is 2:5; thus, this ratio of lime to fly ash was selected in this
research.
affects the mechanical property of the specimens. The VTM was
As shown in Table 5, two aggregate gradations were selected in the test: the
proposed to meet the rapid growth of traffic volumes, axle loads, gradation of specification (SG) [30] and skeleton-dense gradation (SD) [31]. The
and field compaction conditions [25,26]. Aggregate rearrangement mass ratio of the aggregates with particle sizes of 19–37.5 mm, 9.5–19 mm, 4.75–
is favored given the high-frequency liquefaction that acts on the 9.5 mm, and less than 4.75 mm was 45:30:18:7 with SD and 38:32:20:10 with
SG. Three contents (17%, 19%, and 21%) of lime and fly ash for each gradation were
aggregates during VTM compaction. Given that the VTM decreases
selected.
aggregate breakage and can be used to simulate vibrational com-
paction in practical engineering [27,28], the mechanical property
2.2. Specimen preparation methods
that results from compaction by this method is representative, reli-
able, and accurate. The correlation between specimens prepared in In this work, the fatigue properties of LFASM were studied using specimens cast
the laboratory using the VTM with in-field specimens reaches up to through the VTM.
90% [22,23]. The mechanism of the VTM is similar to that of gyra- The vertical vibration test equipment (VVTE) is the core of the VTM. The VVTE
used in the VTM was manufactured in accordance with the principle of a directional
tory compaction initially developed by the Texas Highway Depart-
vibratory roller. The VVTE for indoor testing was designed to imitate the working
ment in 1939 to aid the design and control of asphalt mixtures
[29]. In contrast to gyratory compaction, however, the VTM is
aimed at improving the simulation of practical field compaction Table 1
[13–15]. Test results of lime character.
Most studies on LFASM are based on the QSTM; thus, informa- Test Fineness/% Mass fraction of CaO
tion about the fatigue performance of LFASM designed with the index and MgO/%
Residue of the sieve Residue of the sieve
VTM is limited. However, previous studies have also indicated that of 0.71 mm of 0.125 mm
LFASM behaves differently under different compaction methods
Results 0 2.3 61.0
even at the same compaction level [23]. Therefore, the fatigue per-
 J. Xue, Y. Jiang / Construction and Building Materials 155 (2017) 531–541 533

Table 2
Grain compositions of Fly ash.

Sieve size/mm 4.75 2.36 0.6 0.075

Percentage passing (by mass)/% 100 97.1 88.6 81.02

Table 3 (u 150 mm  h 150 mm) were compacted with the different compaction methods
Chemical compositions of fly ash. on the basis of the maximum dry density and optimum content of the specimens.
The compaction degree was set to 98%. Then, the specimens were placed in plastic
Test index Content of oxides/% Loss on ignition/% bags and stored in the curing room, where the temperature was maintained at
SiO2 Al2O3 Fe2O3 20 ± 2 °C and humidity at more than 95%. VCTM was performed for 120 s, and VPSM
was performed for 80 s [22,23].
Results 51.6 19.8 8.0 12.1

2.3. Splitting strength test

principle and structural properties of a directional vibratory roller. The structure of
the VVTE is shown in Fig. 1. It was installed with two symmetrical (Fig. 2) vibrators
that inversely rotate to counteract the horizontal component of the centrifugal force
and to multiply the vertical component such that the VTM remains stable and effec-
tive. The basic parameters of the VVTE are presented in Table 6.
The VTM includes the vibration compaction testing method (VCTM) and vibra-
tion pressure producing specimen method (VPSM). In this study, the maximum dry
density and optimal water content of LFASM were determined using the VCTM, and
specimens were cast in accordance with the VPSM. Cylindrical specimens
The limit splitting strength is the basic parameter for the establishment of split-
ting fatigue equation. In the current work, the limit splitting strength was utilized
to test the dynamic splitting strength of LFASM to ensure the accuracy and reliabil-
ity of the test results. The principle and composition of the Dynamic Material Test-
ing System (MTS) are shown in Fig. 3.
The cylindrical LFASM specimens were formed through the VTM, and the split-
ting test was conducted on specimens that were saturated for a day after a certain
curing period. Mechanical tests were repeated six times [32,33], and the results
were examined and averaged through Grubbs’ method. The representative values
were then calculated.

Table 4
Technical Performances of Aggregates.

Aggregate Size Crushing Value/% Needle Flake Content/% Apparent Relative Density Water Absorption/%
19–37.5 mm 17.9 12.4 2.73 0.40
9.5–19 mm 21.0 15.5 2.74 0.40
4.75–9.5 mm – 15.2 2.74 1.50
54.75 mm – – 2.76 –

Table 5
Gradation of mineral aggregate.

Gradations Percentage passing/% of sieve sizes/mm

37.5 31.5 19.0 9.5 4.75 2.36 0.6 0.075
SD 100 97.5 51.5 26.0 9.5 5.5 1.5 0.5
SG 100 100 58.0 29.0 12.0 7.0 3.0 1.5

Fig. 1. The structure of VVTE.

534 J. Xue, Y. Jiang / Construction and Building Materials 155 (2017) 531–541

Fig. 2. Eccentric blocks.

Table 6
Vibration parameters of VVTE.

Vibration force/kN Load of upper system/kN Load of nether system/kN Nominal amplitude/mm Vibration frequency/Hz
7.6 1.8 1.2 1.2 30

The respective representative values of the splitting strength of the VTM spec-
imens with a 95% guarantee rate were defined as Ri0.95. Ri0.95 was calculated with Eq.
(1) [21]:

 1:645S
Ri0:95 ¼ R ð1Þ

is the average of splitting strength, MPa; S is the standard deviation.
where R
2.4. Fatigue test method
The fatigue test methods for road materials are classified into the small-beam
bending test and cylindrical-specimen-splitting test. The bending test better simu-
lates the stress conditions of actual pavement structures than the splitting test but
is more complicated to use for forming specimens. By contrast, cylindrical
specimens easily form in the splitting test given the similarity between the stress
conditions of the test and those of actual roads [34]. Therefore, the splitting fatigue
test was adopted in this work.
The MTS was adopted to test the fatigue performance of the specimens.
The strain control mode was used. Five stress levels were set: 0.85, 0.80,
0.75, 0.70, and 0.65. The loading frequency was 10 Hz and the peak-to-valley
stress ratio (circulation characteristic value) under a sinusoidal wave load
was 0.1. The failure of the specimens was used as the judging criterion for
fatigue failure. Testing was terminated if the specimens did not fail after
106 loading times.
The fastener and installment of the specimens are shown in Fig. 4.

Fig. 3. The principle and composition of MTS.

 J. Xue, Y. Jiang / Construction and Building Materials 155 (2017) 531–541
3. The fatigue test results and analysis
3.1. Splitting strength
Given that LFASM grows stronger with age, the appropriate cur-
ing duration should be selected to reflect the strength of LFASM.
Fig. 5 shows the influence of curing time and LFA content on the
splitting strength of LFASM.
Fig. 5 shows that splitting strength increases sharply at the ini-
tial stage, which involves a curing time of 0–90 days. Subse-
quently, the rate of the increase in strength decreased, and the
splitting strength of LFASM remained stable after 180 days of cur-
ing. Thus, the splitting strengths of LFASM with a curing time of
180 days were chosen in the current work.Table 7 shows the
splitting strength of LFASM with SD gradation and SG gradation.
The contents of lime and fly ash for each gradation were 17%,
19%, and 21%. The average of the splitting strength was denoted

and the variation coefficient of the splitting strength was
as R,
defined as Cv.
As shown in Table 7, splitting strength increased as LFA content
increased. The rates of increase, however, gradually decelerated. As
inferred from the data presented in Table 6, the splitting strength
of LFASM was influenced by gradation. The splitting strength of
LFASM with SD gradation was greater than that of LFASM with
SG gradation.
Fig. 4. The fastener and installment of the specimen.
Fig. 5. Variation in the splitting strength of LFASM with curing duration and different LFA contents.
535
3.2. The fatigue test results
The maximum and minimum load stresses were calculated on
the basis of the results shown in Table 7. The fatigue lives of the
LFASM specimens with different gradations and LFA contents were
tested. Table 8 shows the test results for the fatigue lives of the
LFASM specimens.
As shown in Table 8, the results of the fatigue test were discrete.
Given that the fatigue lives of specimens with the same gradation,
LFA content, and stress level were markedly different, the data
were not easily analyzed. Therefore, the Weibull distribution was
used to analyze the fatigue lives of the LFASM specimens [35,36].
The Weibull distribution, which is a minimum asymmetric dis-
tribution, is particularly suitable for the study of fatigue life. This
type of distribution closely reflects the regularities of the distribu-
tion of the design parameters of LFASM. The statistical model of a
two-parameter Weibull distribution is widely used in structural
design and fatigue life analysis [37].Assuming that the equivalent

N
fatigue life Nð
¼ N 1n Þ obeys a two-parameter distribution, the
failure probability P should conform to Eq. (2).


m !

N
P ¼ FðNÞ ¼ 1  exp  ð2Þ
g
Eq. (2) can be logarithmically transformed to Eq. (3).
536 J. Xue, Y. Jiang / Construction and Building Materials 155 (2017) 531–541

Table 7 lg S ¼ lg a  b lg N ð6Þ
Splitting strength of lime and fly-ash stabilized macadam.
where a and b are the regression coefficients of the fatigue equation.
Gradation Content of lime and fly-ash/% Splitting strength
Regression analysis was used to establish the LFASM fatigue


R/MPa Cv/% Ri0.95/MPa equations on the basis of the fatigue data shown in Table 10. The
SD 17 1.52 7.8 1.32 regression coefficients are listed in Table 11. The coefficients a
19 1.61 6.5 1.41 and b denote the intercept of the equation curve on the vertical
21 1.58 4.2 1.44
axis and the slope of the equation curve, respectively. The fatigue
SG 17 1.33 7.3 1.14 performance of LFASM under the action of high stress levels
19 1.36 5.6 1.18
increased as indicated by the increase in coefficient a and decrease
21 1.37 3.1 1.21
in coefficient b.
The following fatigue equations under the failure probability of
1
 ln t 50% were established in accordance with the coefficients shown in
ln ln ¼ m ln N ð3Þ
1P Table 11. The LFASM sample with SD and LFA content of 17% was
where m is the shape parameter; t is the scale parameter. denoted as SD17. Other samples with different gradations and
The regression coefficients m and lnt and the correlation coeffi- LFA contents were denoted similarly.
cient r of the Weibull distribution model for the equivalent fatigue
SD17lgS ¼lg1:278  0:058lgN ð7Þ

of LFASM are shown in Table 9. The type of coefficients is
life N
SD19lgS ¼lg1:206  0:051lgN ð8Þ
denoted as X.
As shown in Table 8, the correlation coefficient r of the Weibull SD21lgS ¼lg1:201  0:052lgN ð9Þ
distribution model exceeded 0.9. This result indicated that the SG17lgS ¼lg1:236  0:057lgN ð10Þ
equivalent fatigue life (N) conformed to the two-parameter SG19lgS ¼lg1:198  0:053lgN ð11Þ
Weibull distribution. SG21lgS ¼lg1:199  0:059lgN ð12Þ

3.3. Fatigue equation

The equivalent fatigue life that corresponds to different stress
levels (S) and failure probabilities (P) shown in Table 10 were
obtained by substituting the value of the regression coefficients
in Eq. (3). Table 10 shows the equivalent fatigue lives under differ-
ent failure probabilities.
Assuming that LFASM has no fatigue limit, the fatigue equation
should satisfy the boundary conditions of Eqs. (4) and (5).
If S ¼ 1; N ¼ 1
If S ¼ 0; N ! 1
The fatigue equation adopted in the study given the apparent single/
double logarithmic relationship between N and S is shown as Eq. (6).
3.4. Influence of material composition on LFASM fatigue performance
3.4.1. Influence of gradation
Fig. 6 shows the curves of fatigue properties with different gra-
dations. The curves were derived using Eqs. (7)(12) from the data
presented in Table 11.
Fig. 6 shows that gradation exerted a remarkable effect on the
fatigue performance of LFASM. Under the same LFA content, LFASM
with SD had a higher value for coefficient a and a lower value for
ð4Þ
coefficient b than LFASM with SG. Hence, LFASM with SD exhibits
better anti-fatigue performance than LFASM with SG.
ð5Þ
Thus, the material composition of LFASM dominates its perfor-
mance. The SG gradation belongs to a suspended dense gradation,
which contain more fine aggregates and less coarse aggregates, so

Table 8
Splitting fatigue life of LFASM.

Gradation Content of LFA/% Fatigue life (loading times) with different stress levels S
0.85 0.8 0.75 0.70 0.65
SD 17 486 905 3182 10,145 44,858
843 2601 6925 26,388 64,300
2453 4412 13,814 29,445 122,768
– 6883 22,814 72,245 259,807
19 428 1371 4979 18,936 76,935
669 3056 8108 33,879 163,633
2008 3922 16,719 49,253 243,120
– 7941 24,710 92,686 316,153
21 274 850 2682 9860 39,687
702 2049 8467 27,215 70,315
1078 4434 12,854 38,819 188,724
– 6758 22,197 83,539 256,970
SG 17 312 850 2682 9860 39,687
880 2449 8967 29,215 77,315
1896 4434 12,854 38,819 188,724
– 6758 22,197 83,539 256,970
19 261 696 2108 9208 48,131
769 1653 5788 19,931 71,518
1314 3140 7973 28,332 178,785
– 4981 14,705 59,904 218,382
21 331 623 1824 6675 20,987
1372 2708 4349 15,406 46,390
1898 4176 11,776 30,091 124,280
– 5961 16,936 71,068 189,612
 J. Xue, Y. Jiang / Construction and Building Materials 155 (2017) 531–541 537

Table 9
Result of test data examined by Weibull distribution.

Gradations Content of LFA/% Coefficient type Coefficient of the Weibull distribution model under the following stress level S
0.85 0.80 0.75 0.70 0.65
SD 17 m 1.030 1.057 1.095 1.135 1.179
lnt 6.804 8.101 9.438 10.888 12.636
r 0.938 0.986 0.998 0.943 0.941
19 m 1.237 1.279 1.394 1.399 1.474
lnt 7.876 9.801 11.622 13.883 16.404
r 0.981 0.976 0.985 0.972 0.986
21 m 1.042 1.024 1.057 1.050 1.063
lnt 6.565 7.227 8.938 10.203 11.529
r 0.903 0.992 0.981 0.985 0.961
SG 17 m 0.907 0.931 0.968 1.010 1.064
lnt 5.716 6.823 8.057 9.600 11.224
r 0.918 0.972 0.995 0.972 0.950
19 m 1.055 1.099 1.146 1.196 1.254
lnt 6.568 7.990 9.441 11.286 13.519
r 0.985 0.997 0.982 0.983 0.936
21 m 0.906 0.915 0.922 0.931 0.939
lnt 5.079 5.962 6.976 8.034 9.035
r 0.924 0.935 0.980 0.990 0.982

Table 10
Equivalent fatigue life of different stress levels under different failure probabilities.

Gradations Content of LFA/% Failure probability/% Equivalent fatigue life under the following stress level S
0.85 0.8 0.75 0.7 0.65
SD 17 50 517 1410 3956 10,611 33,184
5 41 123 367 1071 3646
19 50 445 1595 4603 15,748 53,099
5 54 208 654 2448 9081
21 50 458 802 3897 11,724 36,440
5 35 64 316 982 3146
SG 17 50 381 1003 2884 9021 26,324
5 22 63 197 702 2320
19 50 356 1032 2754 9217 36,004
5 33 99 284 1046 4514
21 50 182 454 1295 3785 10,201
5 10 26 77 231 638

Table 11
Regression coefficients of fatigue equation.

Gradations Content of LFA/% Coefficient Coefficient of fatigue equation under the

following failure probability/%
50 5
SD 17 a 1.278 1.067
b 0.058 0.054
R 0.989 0.994
19 a 1.206 1.054
b 0.051 0.047
R 0.981 0.992
21 a 1.201 1.029
b 0.052 0.051
R 0.989 0.979
SG 17 a 1.236 0.987
b 0.057 0.051
R 0.991 0.992
19 a 1.198 1.021
b 0.053 0.049
R 0.989 0.983
21 a 1.199 0.999
b 0.059 0.058
R 0.987 0.986

the coarse aggregates are suspended in the fine aggregates. The certain coarse aggregates to formed the skeleton structure that
LFASM with SG gradation has high compactness and poor stability, having stronger inherent resistance and the gap of coarse aggre-
so it has relatively low anti-fatigue performance. While the SD gra- gates filled by the dense fine aggregates, so its form skeleton-
dation belongs to a skeleton dense gradation, which contain the density frame after combining coarse aggregate and fine aggregate.
538 J. Xue, Y. Jiang / Construction and Building Materials 155 (2017) 531–541
The LFASM with SD gradation maintains high strength, good crack-
resistance property, and good fatigue-resistance property.
Fig. 7 shows the fatigue–fracture surface of the LFASM speci-
mens. Fig. 6 shows that large aggregates, e.g., 19–37.5 mm and
9.5–19 mm, snapped when fatigue failure occurred. Moreover,
the amount of large aggregates was correlated with the increase
in the number of snapped aggregates, leading to the prolonged fati-
gue lives of specimens. This condition explained why LFASM with
SD exhibit better anti-fatigue performance than LFASM with SG.
3.4.3. Influence of LFA content
The curves of the fatigue properties of the samples with differ-
ent LFA contents are shown in Fig. 8.
As shown in Eqs. (7)(12) and Fig. 8, the fitted lgS–lgN lines indi-
cated that LFASM with 19% LFA content showed the highest fatigue
performance, followed with LFASM 17% and 21% LFA content. As
mentioned previously, when the voids between coarse aggregates
are filled with the appropriate amount of fine aggregates (LFA), a
dense skeleton framework that combines coarse and fine aggre-
gates is formed and enhances the anti-fatigue performance of
LFASM. When LFA content was less than 19%, the voids between
coarse aggregates cannot be completely filled with LFA. Thus, the
LFASM specimens contained numerous voids that easily led to fati-
Fig. 6. The influence of gradation on LFASM fatigue performance.
gue failure. However, when the LFA content exceeded 19%, the LFA
propped open the skeleton structure of the coarse aggregates. The
coarse aggregates thus could not be compacted into a rigid frame-
work. This condition also easily led to fatigue failure. The analytical
results of the fatigue performance tests revealed that the optimal
LFA content of LFASM is approximately 19%.
3.5. Fatigue failure mechanism
LFASM is an inhomogeneous multiphase material with numer-
ous voids. Therefore, its fatigue failure is caused mainly by the
long-term accumulation of fatigue damage. The failure process of
the LFASM specimens in the splitting fatigue test could be divided
into six stages: (1) full stress stage, in which microcracks and par-
tial damage do not appear and the specimen remains fully intact;
(2) crack initiation stage, in which microcracks gradually initiate
at the bottom of the specimen under continuous loading and visi-
bly propagate in the vertical direction; (3) slow crack propagation
stage, in which cracks slowly propagate in the vertical direction;
(4) rapid crack propagation stage, in which cracks quickly propa-
gate in the vertical direction and the widths of the cracks rapidly
expand; (5) crack cut-through stage, in which all cracks vertically
cut through the specimen; and (6) buckling failure stage, in which
 J. Xue, Y. Jiang / Construction and Building Materials 155 (2017) 531–541 539

Fig. 7. Fatigue fracture surface of the LFASM specimens.

Fig. 8. The influence of LFA content on LFASM fatigue performance.

Table 12
The recommended tensile strength structure coefficient (KS) of LFASM based on different design method.

KS Cumulative equivalent axle loads Ne (timelane1)

Light traffic Medium traffic (3.0  106–1.2  107) Heavy traffic (1.2  107–2.5  107) Special heavy traffic (>2.5  107)
(<3.0  106)
KS(VTM) <1.69 1.69–1.84 1.84–1.92 >1.92
KS(SD) <2.12 2.12–2.47 2.47–2.68 >2.68

all cracks cut through the whole specimen and the specimen loses definition of the tensile strength structure coefficient (KS) and fati-
its bearing capacity. gue equation, is shown in Eq. (14).

3.6. Tensile strength structure coefficient
In Chinese pavement design specifications, flexural tensile
stress is utilized as a test criterion for the tensile stress of the bot-
tom layer of a semi-rigid base course and is used as the design
index in the control of the fatigue cracking of structural layers [38].
To control the fatigue distress of asphalt pavement, the tensile
stress rm of each layer should be lower than or equal to the allow-
able tensile strength rR, as shown in Eq. (13):
rR
rm 6 rR ¼
KS
where rm is the tensile stress of each layer under the standard load
in MPa; rR is the allowed tensile strength in MPa; rS is the limit
splitting strength in MPa; and KS is the tensile strength structure
coefficient.
KS, a key parameter for pavement structure design, can be
derived from the fatigue life of pavement materials. The simplified
calculation equation of KS, which is defined in accordance with the
1
Ks ¼  Nb ð14Þ
lg a
where N is the fatigue life; a and b are the regression coefficients
whose specific values are shown in Table 11.
KS was calculated on the basis of the results of the laboratory
fatigue test. However, actual pavement conditions are influenced
by many factors, including interval time, worst season, transverse
distribution of the wheel track, and stress state. Thus, KS must be
corrected in actual pavement design. The relationship between
indoor and actual pavement fatigue life is shown in Eq. (15).
ð13Þ
N ¼ A1  A2  A3  A4  N e ð15Þ
where Ne is the cumulative equivalent axle loads; N is the fatigue
loading times of the fatigue test; A1 is the coefficient of interval time
(0.2); A2 is the coefficient considering the transverse distribution of
the wheel track and the maximum probability and frequency of the
wheels (the value of A2 is 0.5); A3 is the coefficient of stress state,
which is 2/3; and A4 is the coefficient related to the days of the
540 J. Xue, Y. Jiang / Construction and Building Materials 155 (2017) 531–541
worst season. Given that the worst season comprises 60 days, the
value of A4 is 60/365.
Considering the above factors, Eq. (14) was reduced to Eq. (16).
Coefficients a and b are the average values of a and b when P is 50%
and were selected from Table 10. Thus, the corrected KS was calcu-
lated using Eq. (16).
Ks ¼ 0:642N0:054
e Ac ð16Þ
where Ac is the coefficient of highway classification, the Ac of
expressway, class I highway, class II highway, and class III or class
IV highway are fixed at 0.85, 1.0, 1.1, and 1.2, respectively.
The recommended values of KS in accordance with the cumula-
tive equivalent axle loads of different traffic levels are shown in
Table 12. Here, KS(VTM) and KS(SD) are the recommended values of
the KS of the VTM and Chinese pavement design specifications.
In designing asphalt pavement structure, the tensile strength
structure coefficient can be used as the basis of the tensile stress
check of the road base, thus providing guidance for pavement
design. As shown in Table 12, the KS(VTM) values are lower than
KS(SD), indicating that the KS(SD) used in pavement design is too con-
servative and can lead to the adoption of LFASM with high strength
and high LFA content, thus decreasing the crack resistance of the
pavement base course. The KS(VTM) value obtained in this research
was applied in pavement design to fully meet the required strength
of LFASM, to decrease LFA content, and to improve the crack resis-
tance of the pavement base course.
The value of Ks obtained in this research was applied in the
reconstruction of a provincial Yaozhou highway into a class II
Xunyi highway in Tongchuan, Shanxi Province. The pavement
designed with this method exhibited decreased road thickness,
and the LFA content used in the VTM design method was lower
than that used in the traditional design method. Thus, the crack
resistance of the pavement base course of the reconstructed high-
way was improved.
4. Conclusions
(1) The splitting strength of LFASM increases with increasing
LFA content. The rate of increase in the splitting strength
of LFASM, however, gradually declines. The splitting
strength of LFASM with SD is greater than that of LFASM
with SG.
(2) The fatigue life of LFASM conforms to the two-parameter
Weibull distribution. The established fatigue equations
accurately reflect the fatigue life of CRPMs.
(3) Gradation significantly influences the fatigue performance of
LFASM. LFASM with SD have potentially better fatigue resis-
tance than LFASM with SG. As LFA content increases, the
effect of gradation on the fatigue performance of LFASM
remains negligible.
(4) The coefficients of tensile strength structure are used as a
reference for fatigue equations and the cumulative equiva-
lent axle loads of the different traffic levels. The value of Ks
obtained in this study can effectively serve as a guideline
for pavement structure design. The pavement designed with
the VTM has improved crack resistance and durability.
The obtained results have considerable practical significance in
the promotion and application of LFASM in overhauling the asphalt
pavement of major highways and in the implementation of sus-
tainable development strategies for highway construction. How-
ever, given that aggregate type also affects fatigue performance,
our future research will investigate the influence of different
aggregate types on LFASM.
Acknowledgements
This research is supported by the scientific project from Henan
Provincial Communication (No. 2014K49-1), the National Natural
Science Foundation of China under project No. 51408044, the
China Postdoctoral Science Foundation - China under Project No.
2015T80999, and the Scientific Research of Central Colleges of
China for Chang’an University under project No. 310821171113.
The author gratefully acknowledges its financial support.
References
[1] A.H. Lav, P.J. Kenny, Fly ash production and its utilization in road construction,
Road Transport Res. 6 (3) (1997) 4–15.
[2] M.A. Kumar, V.R.P. Raju, Use of lime cement stabilized pavement construction,
Indian J. Eng. Mater. Sci. 16 (4) (2009) 269–276.
[3] V. Sahu, V. Gayathri, The use of stabilized fly ash as a green material in
pavement substructure: a review, Int. J. Civ. Struct. Eng. 4 (3) (2014) 306–314.
[4] M.J. Mccarthy, L.J. Csetenyi, A. Sachdeva, et al., Engineering and durability
properties of fly ash treated lime-stabilised sulphate-bearing soils, Eng. Geol.
174 (1) (2014) 139–148.
[5] S. Antiohos, S. Tsimas, Activation of fly ash cementations’ systems in the
presence of quicklime: part I. compressive strength and pozzolanic reaction
rate, Cem. Concr. Res. 34 (5) (2004) 769–779.
[6] J.M. Hoover, R.T. Huffman, D.T. Davidson, Soil stabilization field trials, Primary
Highway 117, Jasper County, Iowa, in: 41st Annual meeting of the highway
research board, Washington DC, USA, 1962.
[7] S. Arora, A.H. Aydilek, Class F fly-ash-amended soils as highway base materials,
J. Mater. Civ. Eng. 17 (6) (2005) 640–649.
[8] R.L. Parsons, Field performance of fly ash stabilized sub grades, Proc. Inst. Civ.
Eng. Ground Improv. 9 (1) (2005) 33–38.
[9] M.E. Kalinski, B.T. Hippley, The effect of water content and cement content on
the strength of Portland cement-stabilized compacted fly ash, Fuel 84 (14–15)
(2005) 1812–1819.
[10] A. Ghosh, C. Subbarao, Strength characteristics of class F fly ash modified with
lime and gypsum, J. Geotech. Geoenviron. Eng. 133 (7) (2007) 757–766.
[11] D.R. Pokhrel, P.E.A. Godiwalla, Performance evaluation of lime fly-ash
stabilization as a pozzolanic stabilizer for airfield pavement used at houston
airport systems, Electron. J. Geotech. Eng. 17 (2012) 1391–1406.
[12] P. Solanki, M.M. Zaman, J. Dean, Resilient modulus of clay subgrades stabilized
with lime, class C fly ash, and cement kiln dust for pavement design, Transp.
Res. Rec. J. Transp. Res. Board 2186 (1) (2010) 101–110.
[13] F.G. Bell, An examination of the use of lime and pulverized fly ash to stabilize
clay materials, Bull. Assoc. Eng. Geol. 4 (1993) 469–479.
[14] J. Mallela, P.E. Quintus, K.L. Smith, Consideration of lime-stabilized layers in
mechanistic- empirical pavement design, Nat. Lime Assoc. (2004) 17–29.
[15] J. Panda, Design of flexible pavement widening and strengthening of bolangir-
sonepur road, Indian Highways 31 (2003).
[16] J. Judycki, Comparison of fatigue criteria for flexible and semi-rigid pavement,
in: Conference on Asphalt Pavements - International Society of Asphalt
Pavements, 1997, pp. 919–937.
[17] A.H. Lav, M.A. Lav, A.B. Goktepe, Analysis and design of a stabilized fly ash as
pavement base material, Fuel 85 (16) (2006) 2359–2370.
[18] A.M. Sha, K. Jia, X.G. Li, Fatigue performances of semi-rigid base course
materials, J. Traffic Transp. Eng. 9 (3) (2009) 29–33.
[19] X.G. Li. Research of fatigue damage forecast model of stable inorganic binder
class at the grass-roots level (Ph.D. thesis), Chang’an University, 2006. (in
Chinese).
[20] Q. Gao, The fatigue performance of two ash crushed stone mixture and its
reliability analysis (Ph.D. thesis), Hunan University, 2007 (in Chinese).
[21] Rioh, Testing methods of material stabilized with inorganic binders for
highway engineering (JTG E51-2009), China Communications Press, 2009 (in
Chinese).
[22] Y.J. Jiang, L.F. Fan, An investigation of mechanical behavior of cement-
stabilized crushed rock material using different compaction methods, Constr.
Build. Mater. 48 (11) (2013) 508–515.
[23] Y.J. Jiang, X.R. Yang, N.F. Li, et al., Evaluation of vertical vibration test method of
lime-fly ash-stabilized aggregate, J. Build. Mater. 18 (2) (2015) 351–356 (in
Chinese).
[24] Y.J. Jiang, L.F. Fan, An experimental investigation of optimal asphalt-aggregate
ratio for different compaction methods, Constr. Build. Mater. 91 (2015) 111–
115.
[25] M.J. Browne, Feasibility of using a gyratory compactor to determine
compaction characteristics of soil, Soil Compaction, 2006.
[26] R.C. Chaney, K.R. Demars, P. Kouassi, et al., A new technique of kneading
compaction in the laboratory, Geotech. Test. J. 23 (01) (2000) 72–82.
[27] F.A. Crispim, D.C.D. Lima, C.E.G.R. Schaefer, et al., The influence of laboratory
compaction methods on soil structure: mechanical and micromorphological
analyses, Soils Rocks 34 (01) (2011) 91–98.
[28] E. Yaghoubi, M.M. Disfani, A. Arulrajah, et al., Impact of compaction method on
mechanical characteristics of unbound granular recycled materials, Road
Mater. Pavement Des. 18 (02) (2017) 1–23.
 J. Xue, Y. Jiang / Construction and Building Materials 155 (2017) 531–541
[29] T. Harman, J.R. Bukowski, F. Moutier, et al., History and future challenges of
gyratory compaction: 1939 to 2001, Transp. Res. Rec. J. Transp. Res. Board
1789 (01) (2002) 200–207.
[30] Rioh, Specification for design of highway asphalt pavement (JTG D50-2006),
China Communications Press, 2006 (in Chinese).
[31] Y.J. Jiang, Z.P. Fu, N.F. Li, Design method of LSCR based on mortar theory 33 (in
Chinese), J. Traffic Transport. Eng. 15 (5) (2015) 8–14.
[32] L.F. Fan, L.N.Y. Wong, G.W. Ma, Experimental investigation and modeling of
viscoelastic behavior of concrete 33 (in Chinese), Constr. Build. Mater. 48
(2013) 814–821.
[33] L.F. Fan, Z.J. Wu, Z. Wan, et al., Experimental investigation of thermal
effects on dynamic behavior of granite, Appl. Therm. Eng. 125 (2017) 94–
103.
541
[34] G. Valdes, F. Perez-jimenez, R. Miro, et al., Experimental study of recycled
asphalt mixtures with high percentages of reclaimed asphalt pavement,
Constr. Build. Mater. 25 (3) (2011) 1289–1297.
_
[35] R. Sakin, A. Irfan, Statistical analysis of bending fatigue life data using weibull
distribution in glass-fiber reinforced polyester composites, Mater. Des. 29 (6)
(2008) 1170–1181.
[36] J. Schijve, A normal distribution or a weibull distribution for fatigue lives,
Fatigue Fract. Eng. Mater. Struct. 16 (08) (1993) 851–859.
[37] S.P. Singh, S.K. Kaushik, Fatigue strength of steel fiber reinforced concrete in
flexure, Cement Concr. Compos. 25 (7) (2003) 779–786.
[38] S. Zhao, B. Huang, X. Shu, et al., Laboratory performance evaluation of warm–
mix asphalt containing high percentages of reclaimed asphalt pavement, J.
Transp. Res. Board 44 (03) (2012) 98–105.