effects of wetting–drying cycles on soil strength profile of a silty...

Engineering Geology 206 (2016) 60–70

Contents lists available at ScienceDirect

Engineering Geology

j ourna l homepage: www.e lsev ie r .com/ locate /enggeo

Effects of wetting–drying cycles on soil strength profile of a silty clay inmicro-penetrometer tests

De-Yin Wang a, Chao-Sheng Tang a,⁎, Yu-Jun Cui b, Bin Shi a, Jian Li a

a School of Earth Sciences and Engineering, Nanjing University, 163 Xianlin Road, Nanjing 210093, Chinab Ecole des Ponts ParisTech, Laboratoire Navier-CERMES, 6 et 8, avenue Blaise Pascal, Cité Descartes, Champs-sur-Marne, 77455, Marne-la-Vallée cedex 2, France

⁎ Corresponding author.E-mail addresses:[email protected] (D.-Y.Wa

(C.-S. Tang), [email protected] (Y.-J. Cui), [email protected](J. Li).

http://dx.doi.org/10.1016/j.enggeo.2016.04.0050013-7952/© 2016 Elsevier B.V. All rights reserved.

a b s t r a c t
a r t i c l e i n f o
Article history:Received 31 March 2015Received in revised form 30 March 2016Accepted 5 April 2016Available online 20 April 2016

A good understanding of the soil strength profile underwetting–drying (W–D) cycles is essential when studyingthe hydro-mechanical response of soils under the effects of precipitation and evaporation. In this investigation,initially saturated slurry specimens were prepared and subjected to three W–D cycles. A micro-penetrometerwas used to characterize the effect ofW–D cycles on soilmechanical behavior. Based on the obtained penetrationcurves (penetration resistance versus depth), the temporal–spatial evolution of soil strength under each dryingpath was analyzed and discussed. The results show that the developed micro-penetrometer is a simple, quick,economical and reliable tool for identification of soil strength profile. Upon drying, as the water content is rela-tively high, the strength of soil is low and uniform along the depth. The influence of evaporation is insignificantin the initial stage of drying even though a large amount of water is lost. After a low critical water content isreached, the strength of the upper soil layer increases rapidly, and the overall strength generally increases expo-nentially with decreasing water content while decreases exponentially with increasing depth. The drying in-duced strength exhibits obvious delayed effect in profile. With increasing W–D cycles, the strength tends todecrease. The penetration curves change from typical mono-peak pattern to multi-peak pattern after the thirdW–D cycle, suggesting that the W–D cycles create more defects in soil microstructure and intensify the hetero-geneity of strength in profile.

© 2016 Elsevier B.V. All rights reserved.

Keywords:Micro-penetrometerPenetration resistanceWetting–drying cycleStrength profileWater contentMicrostructure

1. Introduction

Soils in nature are subjected to climate change and undergo period-ical wetting–drying (W–D) cycles. The W–D cycles can significantlyalter soil hydro-mechanical behavior, and damage earth structures.For instance, the increase or decrease of water content in clayey soilsmay cause ground heave or subsidence. Pavements, embankments,buildings and other infrastructures are significantly influenced by thisground deformation. In UK, about one in five buildings are at risk bydrought because they are built on clays, whose engineering propertiesare very sensitive to water content change (Harrison et al., 2012). TheAssociation of British Insurers (ABI) predicts that by 2050 the annual av-erage cost of building damage claims could increase from $450 millionto $900 million, with an extreme or ‘event’ drought year costing$1800 million. Similar climate-related disasters also occur every yearin France, USA, China, Canada and other countries (Silvestri et al.,1990; Hemmati et al., 2012). Desiccation cracks would also develop onslopes, landfill covers and clay liners as they are subjected to continuous

ng), [email protected] (B. Shi), [email protected]

W–D cycles (Miller et al., 1998; Tang et al., 2011). The presence of crackscan significantlyweaken the soilmechanical performance (Velde, 1999;Chertkov and Ravina, 1999; Chertkov, 2000; Albrecht and Benson, 2001;Tang et al., 2010). Thus, it is vital towell understand the soil response toclimate-related loading and particularly to W–D cycles. Besides the cli-mate change, big trees adjacent to a building can often be the cause ofstructural problems such as cracking and sometimes movement oreven cause damage to drains (Biddle, 1998), because the water-absorbing effect of tree roots can significantly change soil moisture con-tent and its distribution, affecting the geotechnical behavior of founda-tion soil (Hemmati et al., 2010).

In the past decades,many testswere conducted to investigate the ef-fects ofW–D cycles on soil properties. It was found that substantial irre-versible accumulation of swelling or shrinkage with significant changesin soil fabric may occur uponW–D cycles (Alonso et al., 2005; Cuisinierand Masrouri, 2005; Tripathy et al., 2009). Moreover, the irreversiblevolumetric deformation during W–D cycles was found to be functionof compaction conditions and the subsequent variation of stress/hydra-tion paths (Cui et al., 2002; Nowamooz and Masrouri, 2009). Goh et al.(2010) showed that the shear strength characteristics of soils underW–D cycles are different. Generally, the shear strength on drying pathis higher than that on wetting path due to the hysteresis effects(Nishimura and Fredlund, 2002; Tse and Ng, 2008). The differences

http://crossmark.crossref.org/dialog/?doi=10.1016/j.enggeo.2016.04.005&domain=pdf

mailto:[email protected]

http://dx.doi.org/10.1016/j.enggeo.2016.04.005

www.elsevier.com/locate/enggeo

Table 1Physical properties of the used soils.

Soil properties Values

Specific gravity 2.73Consistency limit

Liquid limit (%) 36.5Plastic limit (%) 19.5Plasticity index (%) 17.0

USCS classification CLCompaction study

Optimum moisture content (%) 16.5Maximum dry density (Mg/m3) 1.7

Grain size analysisSand (%) 2Silt (%) 76Clay (%) 22Uniformity coefficient 18Shrinkage limit (%) 8.3

Note: CL, low-plasticity clay; USCS, Unified Soil Classification System (ASTM,2006a).

61D.-Y. Wang et al. / Engineering Geology 206 (2016) 60–70

between the shear strength on the drying andwetting pathswere foundto be more significant for the first cycles (Goh et al., 2014). Sivakumaret al. (2006); Nowamooz andMasrouri (2008) found that at a given suc-tion the pre-consolidation stress on the drying path is lower than on thewetting path, and they also attributed this phenomenon to the hydrau-lic hysteresis duringW–D cycles. Zha et al. (2013) showed that, with in-creasing W–D cycles, the unconfined compressive strength of soilincreased first and then declined. Moayed et al. (2013) investigatedthe effect of W–D cycles on bearing capacity of lime-silica fume treatedsoil. They observed that the California Bearing Ratio (CBR) increasedafter the first W–D cycle, while it started to decrease during the subse-quent cycles. The increase of CBR during the first cycle was attributed tothe fact that enoughwaterwas provided during thefirst wetting path toensure the reaction between lime, silica fume and soil. Chen and Ng(2013) performed a series of isotropic compression tests on acompacted clay subjected to several W–D cycles. They concluded thatthe hydro-mechanical behavior of soil was significantly influenced bythe W–D history. Moreover, the desiccation cracking behavior of soil isalso significantly related to W–D cycles. Tang et al. (2011) performeddesiccation tests and observed that the measured cracking water con-tent, surface crack ratio and final thickness of soil specimen increasedmonotonously in the first three W–D cycles and then tended to reachstabilization during the subsequent cycles.

The above-mentioned studies provided a rich database for analyzingboth the volume change and mechanical behavior of soils upon W–Dcycles. But most experiments were performed in a macroscopic waythrough oedometer tests, triaxial tests, direct shear tests, etc. These con-ventional macroscopic tests were usually conducted on small speci-mens with relative homogenous moisture distribution. Moreover, thetest conditions are not compatible with field conditions where soilmoisture distribution is always non-uniformdue to the continuouswet-ting/drying processes induced by varying atmospheric conditions. Inother words, the mechanical behavior of soil in field must be depthand time dependent. On the other hand, if the temporal–spatial evolu-tion of soil strength during infiltration and evaporation is well known,the soil geotechnical response to rainfall and drought can be evaluatedand predicted. It is therefore of primary importance to find or developproper testmethodologies for examining soil strength profile under dy-namicwater content conditions. In this regard, the conepenetration test(CPT) is regarded as an easy, rapid and relatively economical field testmethod to investigate the mechanical behavior of soils (Shin and Kim,2011). Nevertheless, the equipment used is of large size and thus cannotbe used in laboratorial testing.

In this investigation, a micro-penetrometer was used to characterizethe temporal–spatial evolution of soil strength duringW–D cycles. A se-ries of penetration tests were conducted during each drying path. Theobtained results allowed the effect ofW–D cycles on soil strength profileto be analyzed.

2. Materials and methods

2.1. Materials

The tested soil was collectedwith spade fromNanjing area, China, in adepth about 0.5–1m (water table is about−3m). It is widely distributedin themiddle and lower reaches of Yangtze River, and is a very importantfoundation soil involved in numerous construction projects. In some re-gions, the thickness of the tested soil layer is larger than 60m. Its physicalproperties are summarized inTable 1. According to theUSCS classification(ASTM, 2006a), it is a lowplasticity clay soil (CL). The clay fraction is dom-inated by illite and interstratified illite-smectite (Liu and Lv, 1996).

2.2. Apparatus

In order to quantitatively characterize the evolution of soil strengthprofile during W–D cycles, a specific micro-penetrometer equipment

was used (Liu et al., 2006; Gu et al., 2014), as illustrated in Fig. 1. Itmainly consists of three parts: (i) load/displacement controlling andmeasuring system (control box, control panel, load transducer anddisplacement transducer), (ii) platen positioning system (platen andposition adjuster), and (iii) data collecting system (data logger andcomputer). The load frame was driven by a motor that can apply con-stant displacement rate during the penetration test. The soil samplewith the container was placed on the platen. A penetration probe wasconnected to the load transducer (with a capacity of 100 N to an accura-cy of 0.01N), and fixed on the center of the crossbeamof the load frame.A displacement transducer with a capacity of 50 mm (accuracy of0.01mm)wasmounted on the platen. Both the load transducer and dis-placement transducer were calibrated, as indicated in Gu et al. (2014).

In order to reduce the side friction between themicro-penetrometerand the soil matrix, also considering both the load capacity and the test-ed soil specimen size, a special mini-probe with an enlarged conicalhead was designed (Fig. 2). The probe has a length of 60 mm. Thecone at the probe head has an apex angle of 60°, and the base diameteris 2.0 mm which is slightly larger than the rod (1.5 mm in diameter).The stiffness of the probe was improved by quenching treatment toreach the requirements of the penetration test. During penetration,the force (N)measured by the load transducer corresponds to the tip re-sistance. The measured displacement corresponds to the penetrationdepth (dp). Generally, when the probe penetrates into a soil verticallyat a constant rate, the higher the soil strength, the greater themeasuredpenetration resistance (Rp). The resistance of soil matrix can thereforebe considered as an indicator of strength (Liu et al., 2006; Gu et al.,2014). According to the obtained penetration resistance and displace-ment, the penetration curve (Rp versus dp) can be plotted, and the var-iation of strength along penetration depth is characterized.

2.3. Preparation of specimens

The collected soil samples were air-dried, crushed and passedthrough 2 mm sieve. The required mass of soil powder was mixedwith deionized water by hand to prepare saturated and homogeneousslurry with an initial water content of 50%, which is about 1.5 timesthe liquid limit (Table 1). A desired quantity of slurry was poured intoa stainless steel container, which are 105 mm in height and 99.2 mmin inner diameter, as schematically shown in Fig. 3 (a). Three long fixingbars were used to fix the baseplate. The container with slurry wasplaced on a vibration (with amplitude of 0.3 mm and frequency of50 Hz) device for 5 min to remove entrapped air bubbles. Afterwards,the container was sealed with plastic membrane for at least 3 days toallow slurry deposition. The final settled slurry thickness was about60 mm. In this investigation, the uniformity of the prepared slurrywas checked and ensured by analyzing the grain size distribution of

Fig. 1. Schematic drawing of the micro-penetrometer equipment.

62 D.-Y. Wang et al. / Engineering Geology 206 (2016) 60–70

the samples taken from the top, middle and bottom layer. Because onespecimen is not enough to run all the penetration tests, two identicalspecimens were therefore prepared following this protocol.

2.4. Test procedures

After the specimens were prepared, they were subjected to threeW–D cycles. As mentioned above, drought can results in soil crackingand significant damage to infrastructures, while the mechanical re-sponse of soil to drought has not receive the deserved concern whencompared with that to rainfall. In the present investigation, as the firststep, the penetration test was only performed at each drying path. The

Probe

Soil

NPenetration load

Tip resistance

1.5 mm

1.73 mm

2 mm

Fig. 2. Schematic drawing of probe penetration process.Fig. 3. Schematic drawing of the specimen container and the porous metal cover.(a) Profile of the container. (b) Top view of the metal cover.

0 5 10 15 20 25 30 35 400

20

40

60

80

100

Average water content29.75%27.07%24.27%20.93%19.79%16.61%14.60%11.64%

(a)

0 5 10 15 20 25 30 35 400

20

40

60

80

100(b)

Average water content29.75%27.07%24.27%20.93%19.79%16.61%14.60%11.64%

Penetration depth dp, mm


Pene

trat

ion

resi

stan

ce R

p, N

Pene

trat

ion

resi

stan

ce R

p, N

p p*exp( / )R a d b c= +−

Fig. 4. The typical penetration curves during the first drying path. (a) For 0 ≤ dp ≤ 35 mm.(b) For 5 ≤ dp ≤ 35 mm.


first drying path started from the initial slurry state. The specimenswere exposed to ambient condition (temperature is 22 ± 2 °C and rel-ative humidity is 55 ± 5%) and moved to the platen of load frame forperforming penetration test at different time intervals correspondingto different water contents. The test ended when the penetration resis-tance reached the maximum capacity of the load transducer (100 N).However, the specimenswere still subjected to drying until the residualwater content was reached. During drying, if the mass change of thespecimen is less than 0.1 g in four hours, the residual water contentwas assumed to be reached. It should be noted that, due to the limita-tion of the measurement capacity of the load transducer, the penetra-tion tests ended before the specimens reached the residual watercontent. The subsequent wetting path was followed by adding deion-izedwater into the container.Water was carefully sprayed on specimensurface until the initial water content was reached (about 50%, corre-sponding to 1.5 times the liquid limit). The container was also sealedwith plastic membrane to prevent moisture loss by evaporation. Sevendays were allowed for this saturation process. Then, the specimenswere exposed to ambient condition and underwent penetration testsagain. By repeating the above procedure, a total of three W–D cycleswere applied to each specimen.

It should be noted that, during the penetration test, a porous metalcover of about 1mm thick (Fig. 3 (b))was placed on the top of the spec-imen to preventmoisture evaporation. In themetal cover, small holes of3mm in diameter were drilled for positioning the penetration tests. Thepenetration positions were carefully selected andmarked in the centralarea of the specimens (about 15–20mmfar away from the boundary) toreduce the boundary effect. About 10–13 penetration tests were per-formed one after another during each drying path. The distance be-tween each neighboring penetration point was kept 20 mm orlarger, which is 10 times larger than the diameter of the penetrationcone (i.e. 2 mm). After performing a series of verification tests thisdistance was proved to be large enough to eliminate any interferencebetween the penetration tests. In order to avoid the probe penetrat-ing in the same hole of the last drying path, the porous metal coverwas rotated slightly by about 20 degree for the subsequent dryingpath. Moreover, only one hole in the top cover was opened for eachpenetration test, the others being sealed by adhesive tape to reducemoisture evaporation. The penetration rate was fixed at 5 mm/minaccording to Gu et al. (2014). After each penetration, the mass ofspecimen was determined by a balance to an accuracy of 0.01 g.The value was used to back-calculate the average water content(wa), and describe the drying process of the specimen. At the endof the third drying path, small soil samples were taken at differentdepths of each specimen, i.e. 0–5, 5–10, 10–15, 15–20, 20–25, 25–30, 30–35 and 35–40 mm, to determine the water content profileby oven drying at 110 °C (ASTM, 2006b).

3. Results

3.1. The first drying path

Fig. 4 (a) shows the typical penetration curves (penetration resis-tance Rp with penetration depth dp) at different wa values during thefirst drying path. All the curves show similar pattern at the initialstage of the test; the Rp increases with increasing dp before a peakvalue is reached. Further examination of the results in Fig. 4 (a) showsthat the critical penetration depth corresponding to the peak Rp gener-ally falls in the range of 2.0–3.0 mm, and increases slightly with the de-crease of wa. Apparently, it is little higher than the height (1.73 mm) ofthe conical probe head. This phenomenon seems more pronounced atlower water contents.

For simplicity and clarity, only the penetration data after dp ≥ 5 mmare plotted for further analysis, as shown in Fig. 4 (b). It can be seen thatthe overall Rp is quite small in the relatively high wa range (i.e. 29.75–19.79%). The Rp increases slightly from about 2 to 12 N as the wa

decreases from 29.75% to 19.79%, and the Rp generally remains roughlyconstant (ormore exactly slightly decrease) with increasing dp. Howev-er, as wa reaches the relatively low range (16.61–11.64%), the penetra-tion curves are different from that observed in the high wa range. TheRp increases rapidly at shallow depth (Fig. 4 (a)) and thereafter de-creases exponentially with increasing dp (Fig. 4. (b)). For instance, at5 mm depth, the Rp increases by 342% from approximately 19 to 84 Nas wa decreases from 16.61 to 11.64%. This suggests that drying resultsin not only significant increase in soil strength but also the increase ofsoil strength heterogeneity in profile. The strength in upper soil layeris higher than that in lower soil layer.

Fig. 5 shows the development of Rp at different depths (5, 10, 15, 20,25, 35 and 40 mm) with the decrease of wa. When wa is higher than19.79%, the Rp at different depths are almost the same even though wa

decreases by about 10%. During this drying period, the overall Rp in pro-file increases slightly, and the corresponding increase rate (defined asincrement of Rp induced by unit decrease of wa — 1%) is approximately0.7 N/%. Whenwa is lower than 19.79%, the Rp at shallow depths (5 and10 mm) starts to increase rapidly first. When wa reaches 16.61%, the Rpat 15 mm starts to increase rapidly. However, the Rp at other deeperdepths (dp ≥ 20 mm) generally increases slightly with decreasing wa.It can be also observed that the increase rate of Rp at each depth in-creases with the decrease ofwa. These results suggest that the develop-ment of strength in soil profile by drying exhibits obvious delayed effectwith increasing depth.

30 25 20 15 10

0

20

40

60

80

100

Depth5mm10mm15mm20mm25mm30mm35mm

Pene

trat

ion

resi

stan

ce R

p, N

wa= 19.79% wa= 16.61%

Average water content wa, %

Fig. 5. Change of Rp at specific depth with wa during the first drying path.

0 5 10 15 20 25 30 35 400

20

40

60

80

100(a)

Average water content29.51%26.30%21.03%19.40%18.37%16.74%14.39%10.16%8.30%

0 5 10 15 20 25 30 35 400

20

40

60

80

100(b)


Pene

trat

ion

resi

stan

ce R

p, N

Pene

trat

ion

resi

stan

ce R

p, N




Fig. 6. The typical penetration curves during second drying path. (a) For 0 ≤ dp ≤ 35 mm.(b) For 5 ≤ dp ≤ 35 mm.


3.2. The second drying path

After thefirst drying pathwasfinished, the specimenswere fully sat-urated and subjected to a second drying and penetration test. The ob-tained penetration curves at different wa values are shown in Fig. 6 (a)for the whole penetration depth, and Fig. 6 (b) for dp ≥ 5 mm. The pat-terns of penetration curves are similar to those of the first drying path(Fig. 4). It can also be observed that, in the relatively high wa range(29.51–18.37%), the overall Rp is relatively small (about 2–10N) and re-mains constant. Upon further drying, the Rp at shallow depth starts toincrease rapidly. For instance, as wa decreases from 16.74 to 8.30%, theRp at 5 mm depth increases from 12.4 to 81.3 N. The Rp also decreasesexponentially with increasing dp (Fig. 6 (b)).

Fig. 7 shows the development of Rp at different depths (5, 10, 15, 20,25, 35 and 40 mm) with decreasingwa. Whenwa is higher than 21.03%,the Rp at different depths are almost the same. The effect of drying onstrength profile is then negligible. The Rp at 5 mm depth starts to in-crease rapidly. As wa decreases to 16.74 and 10.16%, obvious increaseof Rp can be observed at 10 and 15 mm depths. The Rp at other deeperdepths (dp ≥ 20 mm) generally increases slightly. This is consistentwith the observation of the first drying path (Fig. 5).

30 25 20 15 10 5

0

20

40

60

80

100


Pene

trat

ion

resi

stan

ce R

p, N

wa= 21.03% wa= 16.74% wa= 10.16%


Fig. 7. Change of Rp at specific depth with wa during the second drying path.

3.3. The third drying path

The obtained typical penetration curves during the third drying pathare shown in Fig. 8 (a) for the whole penetration depth, and in Fig. 8(b) for dp ≥ 5 mm. Unlike the penetration curves obtained from thefirst and second drying paths, the penetration curves during the thirddrying path show some fluctuations. Moreover, some curves showmulti-peak pattern instead of mono-peak pattern, especially at low wa

values. For instance, at wa = 9.33%, the Rp reaches the first peak atdp = 3.1 mm; after that it decreases and increases to the second peckat dp = 9.5 mm; finally, it reaches the third peak at dp = 29.1 mm(Fig. 8 (a)). This indicates the more pronounced spatial difference ofstrength in profile after three W–D cycles.

During the first and second drying paths, the corresponding criticaldepth dp at the maximum Rp usually falls in the range of 2.0–3.0 mm(Figs. 4 (a) and 6 (a)). The Rp generally decreases monotonously withincreasing dp after 5 mm depth, and the corresponding penetrationcurves can be best fitted by a declined exponential function (Figs. 4(b) and 6 (b)). However, during the third drying path, the criticaldepth dp shifts rightward to larger values (2.5–9.5 mm), as shown inFig. 8 (a), the largest value of dp (9.5 mm) being at wa = 9.33%.

Fig. 9 shows the development of Rp at the specific depths (5, 10, 15,20, 25, 35 and 40 mm) with decreasing wa. Similar observation can be

made as in the first and second drying paths. When wa is higher than21.12%, the effect of drying on the development of strength profile islow and negligible. After that, the Rp at dp = 5 mm starts to increase.As wa reaches 17.72 and 14.54%, the Rp values at dp = 5, 10 and15 mm start to increase rapidly, with evident delayed effect. In particu-lar, unlike in the former twodryingpaths, theRp at 5mmdepth does not

0 5 10 15 20 25 30 35 400

20

40

60

80(a)


0 5 10 15 20 25 30 35 400

20

40

60

80


(b)


Pene

trat

ion

resi

stan

ce R

p, N

Pene

trat

ion

resi

stan

ce R

p, N



Fig. 8. The typical penetration curves during the third drying path. (a) For 0 ≤ dp ≤ 35mm.(b) For 5 ≤ dp ≤ 35 mm.


increase monotonously with further decrease of wa. It reaches a peakvalue of 43.9 N at wa = 11.47%, and then decreases to 39.1 N at wa =9.33%.

At the end of the third drying path, the actual water content wasmeasured at 5 mm spacing and presented in Fig. 10, the value of wa

30 25 20 15 10 5

0

20

40

60

80


Pene

trat

ion

resi

stan

ce R

p, N


wa= 17.72%wa= 21.12% wa= 14.54%

Fig. 9. Change of Rp at specific depth with wa during the third drying path.

being 9.33%. The corresponding penetration curve is also plotted. Thewater content generally increases monotonously over depth as expect-ed, while the corresponding penetration curve shows significant fluctu-ation and apparently has no direct connection with the variation ofwater content.

3.4. Comparison of the results

In order to better understand the effect of W–D cycles on the soilstrength development during drying, the penetration curves at similarwa for each drying path are shown in Fig. 11 (a) for wa about 21.0%,and in Fig. 11 (b) for wa about 14.5%. It is observed that the overall Rpand the maximum Rp generally decrease with increasing W–D cycles.For instance, at wa = 21.0 and 14.5%, the maximum Rp during thethird drying path decreased by 72.7% and 72.3% respectively as com-pared with the values in the first drying path. Moreover, the fluctuationof Rp is intensified after the third W–D cycle.

In addition to the averagewater content, another parameter namelythe average penetration resistance Rap was employed to describe theevolution of soil strength during drying. The Rap is defined as the ratioof the area under the penetration curve to the penetration depth.Based on the penetration curves shown in Figs. 4, 6 and 8, the Rap ateach wa can be determined. Fig. 12 shows the variation of Rap with de-creasing wa during the three drying paths. The Rap generally increasesexponentially with decreasingwa. At a given wa, the Rap during the for-mer drying path is always higher than that during the subsequent dry-ing paths.

4. Discussion

4.1. The development of strength profile during drying path

As observed in Figs. 4, 6 and 8, at given water content, the soilstrength decreases exponentially with increasing depth. This is mainlydue to the suction gradient induced in soil during drying. For an initiallysaturated soil, evaporation generally takes place in three stages(Fig. 13), and starts from the soil surface (Hillel, 1982). In stage I(Fig. 13 (a)), soil is fully saturated. The moisture in deeper soil layersis drawn to the surface for maintaining the evaporation by means ofcapillary forces which form capillary water columns passing throughthe air/water menisci. When the air-entry value is reached, air bubblesenter, and the top layer changes from saturated to unsaturated state.The evaporation process transits to stage II (Fig. 13 (b)). After that, suc-tion in the top unsaturated zone tends to be very sensitive to further

0 5 10 15 20 25 30 35 400

20

40

60

80 Penetration curve at w

a = 9.33%

4

6

8

10

12

14 Actual water content profile

Act

ual w

ater

con

tent

w, %

Pene

trat

ion

resi

stan

ce R

p, N


Fig. 10. Penetration curve at wa = 9.33% and the corresponding profile of actual watercontent.

0 5 10 15 20 25 30 35 400

5

10

15

20(a)

The first drying path The second drying path The third drying path

wa is about 21.0%

0 5 10 15 20 25 30 35 400

20

40

60

80

The first drying path The second drying path The third drying path

(b)

Pene

trat

ion

resi

stan

ce R

p, N

Pene

trat

ion

resi

stan

ce R

p, N



wa is about 14.5%

Fig. 11. Typical penetration curves at similar average water content during each dryingpath. (a) wa is about 21.0%. (b) wa is about 14.5%.


water content change — a little decrease in water content would resultin significant increase in suction.With continuous drying, the top unsat-urated zone is enlarged, and pushes the evaporation surface deeper.Below the evaporation surface, soil is usually saturated or with highdegree of saturation, and the deep pore water is still drawn to the

30 25 20 15 10 50

10

20

30

40

50

The first drying path

The second drying path

The third drying path

Ave

rage

pen

etra

tion

resi

stan

ce R

ap, N


2ap 357.22*exp( / 4.87) 1.49, 0.98aR w R= =+−

2ap 231.30*exp( / 4.53) 1.47, 0.99aR w R= =+−

2ap 216.22*exp( / 4.43) 1.43, 0.99aR w R= =+−

Fig. 12. The relationship between Rap and wa during each drying path.

evaporation surface by capillary force. Above the evaporation surface,soil is unsaturated and the evaporation process is dominated by vapordiffusion. Because the pore water that is closer to the soil surface hasshorter transfer distance to the ambient environment, it evaporatesfaster than that in deeper pores. As a result, significant water contentgradient as well as suction gradient is established in the upper unsatu-rated zone. Indeed, Song et al. (2014) performed evaporation testusing an environmental chamber and measured suction during drying,their results indicated that suction decreases exponentially with in-creasing depth. In addition, the measured peak penetration resistancein the near surface zone and the calculated average penetration resis-tance increase with the decrease of average water content. This phe-nomenon can also be attributed to the increase of suction in the soildue to drying. With further drying, when the vapor pressure insidesoil pore spaces is close to that in the ambient air, the evaporation pro-cess transits to stage III (Fig. 13(c)) and the corresponding evaporationrate is near zero. Finally, the residual water content of soil is reached.

It is found that, before the averagewater contentwa reaches a criticalvalue, the strength in profile is generally similar and slightly influencedby the decreasingwater content. Based on the results shown in Figs. 5, 7and 9, this critical value ofwa is determined as about 20 ± 1%. By refer-ring to the above-mentioned evaporation process, it can be deducedthat the observed critical average water content is probably related tothe soil air-entry value. Before the air-entry value is reached, the soil issaturated, and the moisture distribution is almost uniform, the effectof evaporation on soil suction being insignificant. After the criticalvalue of wa is reached, the penetration resistance in the surface layerstarts to increase rapidly. This is because the surface layer becomes un-saturated, and suction increases rapidly even though the water contentdecreases slightly. With further drying, the evaporation surface movesdown gradually, explaining the observed delayed effect of strength pro-file. Therefore, according to the penetration results, it is possible to eval-uate the development of unsaturated zone in profile, or to assess theevolution of the influence depth of drying. The development of Rpshown in Figs. 5, 7 and 9 suggests that the variation of soil strength atdeeper depths (dp ≥ 20 mm) is insignificant for each drying path. Itcan be deduced that the maximum influence depth of drying is almostlimited to 20 mm for the tested specimens. Note that if accurate rela-tionship among penetration resistance, water content and suction isavailable, the introduced penetration testmethodwould be an econom-ic and rapid way to detect soil mechanical response to climate changes,especially during drought periods.

In addition to the contribution of suction, volumetric shrinkagewouldbe another important factor affecting the development of strength duringdrying. Because clayey soil was tested, volumetric shrinkage must occurduring drying, resulting in denser fabric. The number of particle contactsin unit volume therefore increaseswith decreasingwater content, and thestrength increases.

4.2. The effect of W–D cycles on strength profile

It is observed that the soil strength profile is significantly influencedby the W–D cycles. For the initially saturated specimens, the overallstrength at given water content decreases with increasing W–D cycles,as shown in Figs. 11 and 12. One of the possible reasons is that the soilwater retention capacity is decreased after the specimens undergomul-tiple W–D cycles. Goh et al. (2014) investigated the W–D cycles on soilwater retention curve (SWRC) at zero confining pressure, and theyfound that, at given water content, the soil suction in the drying pathis decreasing with the W–D cycles.

Another reasonmay be the soil texture alteration or the structure re-arrangement duringW–D cycles.When a clayey soil is subjected towet-ting, the active clay minerals may adsorb water and swell. Generally,there are two levels of swelling: interlayer swelling at the microscopicscale and bulk volume increase at the macroscopic scale (Likos and Lu,2006). The interlayer swelling is generally reversible when the soil is

Soil surface/Evaporation surface

Curvature of water surface

Soil particle

Pore water

Unsaturatedzone

Saturatedzone

Evaporation surface

Moves down as evaporationproceeds

Evaporation Residual water between soil particles

Pore air

t0t1t2t3 t0 t2 t3t1Water content

Dep

th

Dep

th

Suction

Unsaturated zone

Saturated zone

The changing water content/suction profile during drying

Stage I, t=t0 Stage II, t=tn, n=1, 2, 3… Stage III, t t∞

(a) (b) (c)→

Fig. 13. Schematic drawing of soil water evaporation process. (a) Stage I. (b) Stage II. (c) Stage III.


subjected to subsequent drying. However, the bulk volume increase isgenerally featured with irreversibility and hysteresis. Many researchersfound that accumulated irreversible swelling or volume increase oc-curred under the effect of W–D cycles (Basma et al., 1996; Cui et al.,2002; Wang et al., 2008; Tripathy et al., 2009; Nowamooz andMasrouri, 2009; Chen and Ng, 2013).

Furthermore,W–D cycles can create larger pores in soil. Sartori et al.(1985) found a strong increase in the intra-aggregate porosity aftersome W–D cycles for a clayey soil. Pires et al. (2008) investigated thechanges in soil structure induced by W–D cycles, and also found thatthe total pore area increased significantly after repeatedW–D cycles, es-pecially for the large pores. Zemenu et al. (2009) performed scanningelectron microscopy (SEM) and mercury intrusion porosimetry (MIP)tests on a clayey soil which was subjected to multiple W–D cycles. Theresults indicated that W–D cycles causes significant changes of soilstructure— the pore volume and the average diameter of pores increasewith W–D cycles.

W–D cycles can also result in crack development in soil. Tang et al.(2011) investigated the effect ofW–D cycles on the desiccation crackingbehavior of a clay slurry. They found that the extent of soil crackingwassignificantly enhanced by increasing W–D cycles. They also found that,after the first W–D cycle, the subsequent W–D cycles change the ho-mogenous slurry structure to an aggregated-structure. A large amountof macroscopic inter-aggregate pores were created, significantly in-creasing the soil heterogeneity. Lu et al. (2002) studied the effect ofW–D cycles on soil microstructure through CT tests. They found thatW–D cycles gave rise to the development of soil cracks. The numberand connectivity of cracks increased with increasing W–D cycles.

With the increase of total bulk volume, the creation of large poresand the development of cracks by the W–D cycles, the soil overall den-sity and particle contacts per unit volume decrease. Consequently, theoverall strength tends to decrease (Figs. 11 and 12). The presence oflarge pores or possible cracks in soil creates weak zones in soil and sig-nificantly intensifies the spatial heterogeneity of structure. This explainsthe observed fluctuation of penetration curves during the third dryingpath (Figs. 8 and 11). The significance of these observations is that the

penetration curves can be used to rapidly detect soil inner structure de-fects with minimum disturbance. It can be also concluded that the soilstrength is not only conditioned by the water content/suction distribu-tion, but also significantly linked tomicrostructure features, as shown inFig. 10.

In geotechnical practice, themeasured penetration resistance is usu-ally used to evaluate the in-situ overconsolidation ratio (OCR) of soil(Mayne and Kemper, 1988; Jian et al., 2005). Generally, for a givensoil, the penetration resistance increases with increasing OCR. For thetested soil specimens in this study, the increase of suction during dryingpath could result in the increase of OCR. According to the measuredwater content profile shown in Fig. 10, it can be assumed that OCR de-creases monotonously over depth as well as the penetration resistance.However, this is not the case if the soil has undergonemultipleW–D cy-cles. As can be seen in Fig. 10, the measured penetration resistance pro-file shows significant fluctuation along the depth, indicating that thecorrelations between penetration resistance and OCR is not uniqueeven for a given soil. It confirms again that the soil mechanical behavioris conditioned by both the water content/suction and fabric features.

4.3. The critical penetration depth

For all the penetration curves, the penetration resistance increasesmonotonously with increased dp at the initial stage of the test. This phe-nomenon can be attributed to the used conical probe head (Fig. 2). Be-fore the conical probe head was fully penetrated into the soil, thecontact area between the probe and the soil gradually increased. How-ever, it is observed that the critical penetration depth that correspondsto the maximum penetration resistance is higher than the height ofthe cone head (Fig. 2), and generally increases with decreasing watercontent. It means that, even though the cone head fully penetratesinto the soil and has the maximum contact area with the soil, the pene-tration resistance does not still reach the maximum value. In otherwords, upon drying, the maximum strength moves down to deeper lo-cations and this is in conflict with the suction development as discussedabove. Usually, the suction of upper soil layer should be always higher

30 25 20 15 10 5

0

2

4

6

8

10

12

The third drying path

The second drying path

R2=0.98qc = 65.01*exp(-0.22*wa)+0.26

R2=0.99qc = 80.82*exp(-0.23*wa)+0.48

qc = 107.12*exp(-0.2*wa)+0.37 R2=0.98The first drying path


Ave

rage

uni

t con

e tip

res

ista

nce

q c, M

Pa

Fig. 14. The relationship between qc and wa during each drying path.


than that of lower soil layer, because the water content of upper layer isalways lower before the drying comes to the end. Therefore, in theory,the critical penetration depth should be lower than the height of thecone head. Actually, the observed intriguing phenomenon may beinterpreted by referring to the typical shear failure modes of shallowfoundation soil under footing (Chen and Abu-Farsakh, 2015; Lambeand Whitman, 1969). At high water content condition, the specimensare loose and soft, and present high compressibility. During penetration,punching shear failure of upper soil layer may occur. The soil under thecone head is compressed with large strain, resulting in the increase ofthe strength of lower soil. For given diameter of the probe, the compres-sion zone increaseswith increasing contact area between conehead andsoil. After the cone head is fully penetrated into the soil, the compres-sion zone may still develop until the critical penetration depth isreached, where the penetration resistance reaches themaximum value.

With water content loss, the developed suction leads to soil volumeshrinkage, increasing the density and the stiffness, while decreasing thecompressibility. During penetration, generally shear failure may occurat shallow depth around the probe. Consequently, the maximum pene-tration resistance can be only obtained at a deeper position, which isusually several times of the cone head height/diameter for stiff soils(Chen and Abu-Farsakh, 2015; Lambe and Whitman, 1969), as shownin Figs. 4, 6 and 8.

It is also found that the critical penetration depth increases with in-creasing W–D cycles, and this phenomenon is more pronounced afterthe third cycle. This is because theW–D cycles result in irreversible vol-ume change and structure rearrangement as discussed above. The re-duction of overall density and the presence of large pores or crackscan significantly increase the soil compressibility.

4.4. Effect of water content/W–D cycles on the bearing capacity of pile

It is well known that the cone penetration test “CPT” is one of themost commonly used field investigation technique for evaluating thegeotechnical properties of soils. In particular, thanks to the similarity be-tween cone penetrometer and pile, the measured penetration resis-tance is widely applied to the estimation of the bearing capacity ofpile. In the past several decades, numerous empirical relationships be-tween cone penetration resistance and pile bearing capacity were pro-posed (Eslami and Fellenius, 1997; Cai et al., 2009). However, none ofthem took into account the effect of water content or W–D cycles. Asdiscussed above, the penetration resistance measured by the micro-penetrometer reflects the mechanical behavior of soil along depth. Theobtained results may therefore be useful for assessing the effect ofwater content/W–D cycles on bearing capacity of pile.

As shown in Figs. 4, 6, 8 and 12, the penetration resistance de-pends significantly on soil water content, and generally increases ex-ponentially with the decrease of water content. Therefore, if thevariation of soil moisture is not properly accounted, the determinedpile bearing capacity based on one in-situ CPT data would beoverestimated or underestimated. More importantly, if the field con-sidered underwent a long-term heavy drought before the CPT tests,the determined pile bearing capacity would be significantlyoverestimated. Consequently, large settlement of foundation evendamage of building may occur during the subsequent wet weather.In arid and semi-arid regions where the field is usually composedof thick unsaturated soil, it is recommended to consider soil watercontent change induced by extreme climate events (i.e. heavydrought and precipitation) as well as its effect on the bearing capac-ity of foundation, especially shallow foundation.

Based on the determined average penetration resistance Rap and thebase area (0.03 cm2) of the used cone head (Fig. 2), the average unitcone tip resistance (qc) is calculated for each drying path, and the resultsare presented in Fig. 14. Taking thefirst drying path as an example, aswa

decreases from 29.75 to 11.64%, qc increases dramatically by 23.5 timesfrom 0.47 to 11.05 MPa. This means that per unit change of water

content would result in about 0.6 MPa variation of the unit cone tip re-sistance. Based on the best fitting curve of the first drying path, the rela-tionship between qc and wa can be derived, as follows:

qc ¼ 107:12� e−0:2�wa þ 0:37: ð1Þ

It should be noted that the above equation cannot be directly used toestimate the end-bearing capacity of pile, because the employedmicro-penetrometer is not a real CPT device, and cannot also be simply consid-ered as amini-CPT (MCPT) device.We assume theremust be significantscale effects as compared with the standard CPT test. Furthermore, therelationship is deduced from the average penetration resistance andcharacterized by the average water content. Comprehensive calibrationtests should be performed to refine the relationship between qc and CPTcone resistance. Systematic theoretical work on themechanical interac-tion between the probe and soil particle and sufficientfield tests are alsorequired before the application of the proposed micro-penetrometer topractice.

In addition to water content, theW–D cycle is another factor affect-ing pile bearing capacity. From the results shown in Fig. 14, it can beseen that the pile bearing capacity decreases with increasing W–D cy-cles. Based on the fitting equations of the three drying paths, the de-crease of average unit cone tip resistance Δqc from the first to thesecond drying path Δqc, 1–2, and from the second to the third dryingpath Δqc, 2–3 at different given water contents are calculated andshown in Fig. 15. It can be seen that, with increasingW–D cycles, the re-duction of unit cone tip resistance is more pronounced at lower watercontent. On the whole, Δqc increases exponentially with decreasingwater content. The reduction of qc from the first to the second W–Dcycle Δqc, 1–2 is more pronounced than that from the second to thethirdW–DcycleΔqc, 2–3.Many researchers studied the effect ofW–D cy-cles on soil shear strength (Liu and Yin, 2010; Yang et al., 2014), and themaximum reduction of cohesion was also usually observed from thefirst to the second cycle. This is consistent with the results shown inFig. 15, and can be attributed to the weakening of soil structure (Tanget al., 2011). Moreover, it was found that soil hydro-mechanical behav-ior generally reaches stabilization after three to five cycles (Basma et al.,1996; Liu and Yin, 2010; Tang and Shi, 2011; Yang et al., 2014).

5. Conclusions

A micro-penetrometer was used to investigate the evolution of thestrength profile of a silty clay subjected to multiple W–D cycles. The

30 25 20 15 10 5

0

2

4

6

8

10

12

14

16

18D

ecre

men

t of

the

aver

age

unit

cone

tip r

esis

tanc

e

q c, M

Pa

qc, 1-2

qc, 2-3


Fig. 15. Change of Δqc with decreasing wa.


effects of drying andW–D cycles on soil strength profile were analyzed.The following conclusions can be drawn:

(1) The proposed micro-penetration test provides a simple, quick,economic and reliable laboratory method to investigate the soilstrength profile. The test shows a potential in evaluating the tem-poral–spatial evolution of soil strength profile.

(2) When the initially saturated soil is subjected to drying, thestrength along the profile is uniform and is slightly affected bywater evaporation before a relatively low critical water contentis reached. After that, the strength generally increases exponen-tially with the decrease of water content, and decreases expo-nentially over depth. The observed critical water content isprobably related to the air-entry value of soil. Moreover, the me-chanical response of the upper soil zone is more sensitive to dry-ing as compared to the deeper soil zone. The drying induceddevelopment of strength presents evident delayed effect alongthe depth.

(3) The temporal–spatial evolution of soil strength profile is signifi-cantly influenced by the W–D cycles. Generally, the averagestrength at a given average water content decreases with in-creasing W–D cycles, and this reduction is more pronounced atlower water contents.

(4) During the first two W–D cycles, the penetration resistance ver-sus average water content shows similar behavior and the effectofW–D cycles is small. However, after the specimens are subject-ed to the third W–D cycle, the penetration resistance with pene-tration depth shows obvious fluctuation, and the penetrationcurves change from typical mono-peak pattern to multi-peakpattern. The heterogeneity of strength distribution is significant-ly intensified after the thirdW–D cycles. This phenomenon is alsomore significant at lower water content.

(5) The development of strength profile during drying is related tothe soil water evaporation process. In addition to suction, thestrength profile also significantly depends on soil structure fea-tures, i.e. pore sizes, heterogeneous zones and cracks.

It should be noted that the present study provides primary dataabout soil strength profile and the effect ofW–D cycles. The lack of actu-al water content and suction does not allowmore quantitative analysis.In the futurework, calibration testswill be performed to establish quan-titative relationships among penetration resistance, water content andsuction. Some MIP and SEM tests will be also carried out to clarify themicrostructure changes.

Acknowledgments

This work was supported by the National Natural Science Foundationof China for Excellent Young Scholars (Grant No. 41322019), NationalNatural Science Foundation of China (Grant No. 41572246), Key Projectof National Natural Science Foundation of China (Grant No. 41230636),and the Fundamental Research Funds for the Central Universities (GrantNo. 020614380013). The authors want to address their thanks to the 6anonymous reviewers for their constructive comments, and Mr. Xiao-Wei Luo at Nanjing University for his help in performing the tests. The au-thors also wish to acknowledge the support of the European Commissionvia the Marie Curie IRSES project GREAT-Geotechnical and geological Re-sponses to climate change: Exchanging Approaches and Technologies ona world-wide scale (FP7-PEOPLE-2013- IRSES-612665).

References

Albrecht, B.A., Benson, C.H., 2001. Effect of desiccation on compacted natural clays.J. Geotech. Geoenviron. 127, 67–75.

Alonso, E.E., Romero, E., Hoffmann, C., García-Escudero, E., 2005. Expansive bentonite-sand mixtures in cyclic controlled-suction drying and wetting. Eng. Geol. 81,213–226.

ASTM. D2216-10, 2006b. Standard Test Methods for Laboratory Determination of Water(Moisture) Content of Soil and Rock by Mass. American Society for Testing andMate-rials, West Conshohocken, Pa.

ASTM. D2487-11, 2006a. Standard Practice for Classification of Soils for Engineering Pur-poses (Unified Soil Classification System). American Society for Testing andMaterials,West Conshohocken, Pa.

Basma, A.A., AlHomoud, A.S., Malkawi, A.I.H., AlBashabsheh, M.A., 1996. Swelling-shrinkage behaviour of natural expansive clays. Appl. Clay Sci. 11, 211–227.

Biddle, P.G., 1998. Tree Root Damage to Buildings. Patterns of Soil Drying in Proximity toTrees on Clay Soils vol. 2. Willowmead Publishing, Wantage (300 pp.).

Cai, G., Liu, S., Tong, L., Du, G., 2009. Assessment of direct CPT and CPTU methods forpredicting the ultimate bearing capacity of single piles. Eng. Geol. 104, 211–222.

Chen, Q., Abu-Farsakh, M., 2015. Ultimate bearing capacity analysis of strip footings on re-inforced soil foundation. Soils Found. 55 (1), 75–85.

Chen, R., Ng, C.W.W., 2013. Impact of wetting–drying cycles on hydro-mechanical behav-iour of an unsaturated compacted clay. Appl. Clay Sci. 86, 38–46.

Chertkov, V.Y., 2000. Using surface crack spacing to predict crack network geometry inswelling soils. Soil Sci. Soc. Am. J. 64, 1918–1921.

Chertkov, V.Y., Ravina, I., 1999. Tortuosity of crack networks in swelling clay soils. Soil Sci.Soc. Am. J. 63, 1523–1530.

Cui, Y.J., Yahia-Aissa, M., Delage, P., 2002. A model for the volume change behaviour ofheavily compacted swelling clays. Eng. Geol. 64, 233–250.

Cuisinier, O., Masrouri, F., 2005. Hydromechanical behaviour of a compacted swelling soilover a wide suction range. Eng. Geol. 81, 204–212.

Eslami, A., Fellenius, B.H., 1997. Pile capacity by direct CPT and CPTu methods applied to102 case histories. Can. Geotech. J. 6, 886–904.

Goh, S.G., Rahardjo, H., Leong, E.C., 2010. Shear strength equations for unsaturated soilunder drying and wetting. J. Geotech. Geoenviron. Eng. 594–606 http://dx.doi.org/10.1061/(ASCE)GT.1943-5606.0000261.

Goh, S.G., Rahardjo, H., Leong, E.C., 2014. Shear strength of unsaturated soils under mul-tiple drying–wetting cycles. J. Geotech. Geoenviron. 140, 06013001.

Gu, K., Tang, C., Shi, B., Hong, J., Jin, F., 2014. A study of the effect of temperature on thestructural strength of a clayey soil using a micropenetrometer. Bull. Eng. Geol. Envi-ron. 73, 747–758.

Harrison, A.M., Plim, J.F.M., Harrison, M., Jones, L.D., Culshaw, M.G., 2012. The relationshipbetween shrink-swell ovvurrence and limate in south-East England. Proc. Geol.Assoc. 123, 556–575.

Hemmati, S., Gatmiri, B., Cui, Y.J., Vincent, M., 2010. Soil-vegetation-atmosphere interac-tion by a Multiphysics approach. J. Multiscale Model. 2 (3 & 4), 163–184.

Hemmati, S., Gatmiri, B., Cui, Y.J., Vincent, M., 2012. Thermo-dydro-mechanical modellingof soil settlements induced by soil–vegetation–atmosphere interactions. Eng. Geol.139-149, 1–16.

Hillel, D., 1982. Introduction to Soil Physics. Academic Press, New York.Jian, W.B., Wu, Z.X., Tong, W.D., Liu, H.M., Zhang, M.X., 2005. Consolidation state of soft

soil differentiated by static cone sounding. Chin. J. Rock Mech. Eng. 24 (12),2166–2169.

Lambe, T.W., Whitman, R.V., 1969. Soil Mechanics. John Wiley and Sons, Inc., New York,USA.

Likos, W.J., Lu, N., 2006. Pore-scale analysis of bulk volume change from crystalline swell-ing in Na+-and Ca2+-smectite. Clay Clay Miner. 54 (4), 515–528.

Liu, J.G., Lv, M.K., 1996. The engineering geological properties of Xiashu loess in Nanjing.Investig. Sci. Technol. 6, 36–38.

Liu, H.Q., Yin, Z.Z., 2010. Test study of influence of crack evolution on strength parametersof expansive soil. Rock Soil Mech. 31 (3), 727–731.

Liu, Z.B., Shi, B., Sheng, D.C., 2006. A micropenetrometer for detecting structural strengthinside soft soils. Geotech. Test. J. 29, 443–450.

Lu, Z.H., Chen, Z.H., Pu, Y.B., 2002. A CT study on the crack evolution of expansive soil dur-ing drying and wetting cycles. Rock Soil Mech. 23 (4), 417–422.


Mayne, P.W., Kemper, J.B., 1988. Profiling OCR in stiff clays by CPT and SPT. Geotech. Test.J. 11 (2), 139–147.

Miller, C.J., Mi, H., Yesiller, N., 1998. Experimental analysis of desiccation crack propaga-tion in clay liners. J. Am. Water Resour. Assoc. 34 (3), 677–686.

Moayed, R.Z., Lahiji, B.P., Daghigh, Y., 2013. Effect ofWetting–Drying Cycles on CBR Valuesof Silty Subgrade Soil of Karaj Railway. Proceedings of the 18th International Confer-ence on Soil Mechanics and Geotechnical Engineering, Paris, pp. 1321–1324.

Nishimura, T., Fredlund, D.G., 2002. Hysteresis Effects Resulting from Drying andWettingunder relatively Dry Conditions. Proc. 3rd Int.Conf. On Unsaturated Soils, UNSAT2002. Swets & Zeitlinger, Lissie, Netherlands, pp. 301–306.

Nowamooz, H., Masrouri, F., 2008. Hydromechanical behaviour of an expansive benton-ite/silt mixture in cyclic suction-controlled drying and wetting tests. Eng. Geol. 101,154–164.

Nowamooz, H., Masrouri, F., 2009. Density-dependent hydromechanical behaviour of acompacted expansive soil. Eng. Geol. 106, 105–115.

Pires, L.F., Cooper, M., Cássaro, F.A.M., Reichardt, K., Bacchi, O.O.S., Dias, N.M.P., 2008. Mi-cromorphological analysis to characterize structure modifications of soil samplessubmitted to wetting and drying cycles. Catena 72, 297–304.

Sartori, G., Ferrari, G.A., Pagliai, M., 1985. Changes in soil porosity and surface shrinkage ina remolded, saline clay soil treated with compost. Soil Sci. 139, 523–530.

Shin, Y.J., Kim, D., 2011. Assessment of undrained shear strength based on cone penetra-tion test (CPT) for clayey soils. KSCE J. Civ. Eng. 15, 1161–1166.

Silvestri, V., Soulie, M., Lafleur, J., Sarkis, G., Bekkouche, N., 1990. Foundation problems inChamplain clays during droughts. I: rainfall deficits in Montreal (1930-1988). Can.Geotech. J. 27, 285–293.

Sivakumar, V., Tan, W.C., Murray, E.J., Mckinley, J.D., 2006. Wetting, drying and compres-sion characteristics of compacted clay. Geotechnique 56, 57–62.

Song, W.K., Cui, Y.J., Tang, A.M., Ding, W.Q., Tran, T.D., 2014. Experimental study on waterevaporation from sand using environmental chamber. Can. Geotech. J. 51, 115–128.

Tang, C.S., Shi, B., 2011. Swelling and shrinkage behaviour of expansive soil during wet-ting–drying cycles. Chin. J. Geotech. Eng. 33 (9), 1376–1384.

Tang, C.S., Cui, Y.J., Tang, A.M., Shi, B., 2010. Experiment evidence on the temperature de-pendence of desiccation cracking behaviour of clayey soils. Eng. Geol. 114 (3–4),261–266.

Tang, C.S., Cui, Y.J., Shi, B., Tang, A.M., Liu, C., 2011. Desiccation and cracking behaviour ofclay layer from slurry state under wetting–drying cycles. Geoderma 166, 111–118.

Tripathy, S., Kanakapura, S., Rao, S., 2009. Cyclic swell–shrink behaviour of a compactedexpansive soil. Geotech. Geol. Eng. 27, 89–103.

Tse, E.Y.M., Ng, C.W.W., 2008. Effects of Drying and Wetting Cycles on Unsaturated ShearStrength. Proc., 1st European Conf. On Unsaturated Soils. CRC Press/Balkema, Leiden,Netherlands, pp. 481–486 (Geol. Eng. 27, 89–103).

Velde, B., 1999. Structure of surface cracks in soil and muds. Geoderma 93, 101–124.Wang, W., Kravchenko, A.N., Smucker, A.J.M., Rivers, M.L., 2008. Effect of Wetting and

Drying Cycles on the Spatial Variability Characteristics of the Internal Structures ofSoil Aggregates Vol. 10.

Yang, H.P., Wang, X.Z., Xiao, J., 2014. Influence of wetting–drying cycles on strength char-acteristics of Nanjing expansive soils. Chin. J. Geotech. Eng. 36 (5), 951–954.

Zemenu, G., Martine, A., Roger, C., 2009. Analysis of the behaviour of a natural expansivesoil under cyclic drying and wetting. Bull. Eng. Geol. Environ. 68, 421–436.

Zha, F.S., Liu, J.J., Xu, L., Cui, K.R., 2013. Effect of cyclic drying and wetting on engineeringproperties of heavymetal contaminated soils solidified/stabilized with fly ash. J. Cent.South Univ. 20, 1947–1952.

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页文献云下载图书馆入口外文数据库大全疑难文献辅助工具

http://www.xuebalib.com/cloud/

http://www.xuebalib.com/

http://www.xuebalib.com/cloud/


http://www.xuebalib.com/vip.html

http://www.xuebalib.com/db.php

http://www.xuebalib.com/zixun/2014-08-15/44.html


effects of wetting–drying cycles on soil strength profile of a silty...

Documents