Instability Analysis of a Low-Angle Low-Expansive Soil Slope under Seasonal Wet-Dry Cycles and River-Level Variations

-ere were a small amount of obvious offsets at the bearing of bridge piers built on an artificial gentle canal bank terrace andmany tensile cracks visible at the surface of the mortar block stones covering the terrace soil in several years following construction. To determine these reasons, a comprehensive site investigation and a wide variety of tests were implemented, which included geophysical tests, in situ tests, laboratory tests, pile integrity detection, and numerical analysis with the finite element method (FEM). -e results revealed that the soil of the low-angle slope was the potentially low-expansive clay soil. -e reduction in soil shear strength deriving from seasonal wet-dry cycles and river-level variations led to the instability and failure of the low-angle low-expansive soil slope, which triggered the collapses of the soil slope and lots of fractures in the piles of the bridge foundation. -e typical characteristics of the instability and failure of the low-angle low-expansive soil slope were tractional detachment and slow sliding.


Introduction
Expansive soil failure events are not as dramatic as earthquakes or hurricanes. e failure soil occurs more slowly than those in other disasters, and the damage may not be concentrated in a small location but instead spread over wide areas. At times, expansive soil damage may considerably worsen and cause very large disasters despite minor maintenance and visual inspections [1]. Expansive soil is the special clay that undergoes expansion, shrinkage, fissuring, and superconsolidation in natural geological processes. Expansive soil is sensitive to climatic and hydrological conditions. Additionally, it exhibits many consistent and long-term potential characteristics. Expansive soils can be found on almost all continents. e destructive effects of these clay soils appear all over the world, such as China, India, Australia, Canada, the United States, Israel, and South Africa [2]. According to the statistics provided by Steinberg, the annual loss caused by expansive soil damage is US$10 billion in the United States and up to ¥100 million in China. Because the expansive soil expands, cracks, and overconsolidates, it is called "engineering cancer," and a sliding failure of a slope with a slope angle only between 16.7% and 25% is usually caused by the presence of highly expansive soil [3]. When this type of soil dries, the remaining moisture concentrates near the particle interface, forming menisci, and the particles are pulled closer together by the resulting surface tension forces, which cause the soil to shrink [1].
Regarding the instability mechanism of expansive soil slopes, it is generally believed that the instability of expansive soil slopes is characterized by shallow soil layers (normally less than 3 m), step-by-step traction, and low-angle slope sliding. e wet and dry cycles (rainfall and evaporation) easily trigger swelling and shrinkage cracking of the shallow expansive soil. e fracturing accelerates water infiltration, destroys the soil structure, and thus leads to a decrease in soil strength [4]. Shallow slope failure may take place during the wet season following construction [5,6]. e effect of precipitation on slope instability was analyzed based on Liu and Li in the case of an actual landslide [7]. e instability of the soil slope can be severely impacted by precipitation, infiltration, and variation in water level. e wet-dry cycle leads to cracking of the expansive soils in the slope. e widths and depths of the fractures will increase with the number of wet-dry cycles. e fractures not only weaken the soil structure of the slope but also provide conduits for water penetration. is further leads to softening soil and decreasing strength. Each wet-dry cycle results in the accumulation of the downward settlement and horizontal displacement of the slopes. With the increase in the number of wet-dry cycles, the expansive soil slopes gradually became damaged, regardless of the soil density [8].
Based on the results of previous studies [9,10], precipitation might cause shallow slope composed of plastic clay soil to failure due to reduction in its shear strength caused by soil softening. e rainwater easily intrudes the soil due to existing desiccation cracks that formed during the dry period. Low-permeability soils stop the downward intrusion of rainwater. Consequently, both excess pore water pressure and reduced shear strength have led to slope sliding several years after its completion [11]. e shear strength of expansive soil slope may decrease on account of swell-shrink behavior following construction, and slope failure often occurs in shallow regions of slope. e sliding interface is normally parallel to the slope surface [12]. Instability of the slope can take place a few years or even decades following construction. e water level fluctuation caused by tides, irrigation, freshwater recharge, and hydropower production also influences the stability of soil slope. Cyclical fluctuation of the water level in reservoir often results in the failure of its bank slope. Water-level fluctuation can cause landslide to reactivate. Slope instability will occur during not only water level drawdown but also water level rise [13]. Slope instability is related to fluctuation of water level [14]. Slope soil is susceptible to variations in water level, and shear strength of fully softened soil will change accordingly. e safety factor of soil slope will alter due to variation in water level [15]. Consequently, the relationship of slope instability and water level fluctuation should be researched and evaluated. e slope instability of expansive soil slope of a canal located in Nanjing, East China, was analyzed in this paper using cone penetration tests (CPTs), standard penetration tests (SPTs), standard soil tests, soil potential expansion tests, and finite element method (FEM) numerical analysis. e soil shear strength was attained by using various shear strength tests, such as direct shear tests and triaxial tests. e analysis of slope instability in ABAQUS [16] was conducted according to the parameters determined from the soil tests.

Background
e slope investigated during this study is an artificial canal slope of accumulated canal bank terrace and is located in the southeast corner of Nanjing, East China. e slope soil attained by dredged river clay across the site was fairly uniform.
e slope was relatively gentle, as is presented schematically in Figure 1. e vertical height of the slope is approximately 16.6 m with slope angles ranging from 8.3% to 15.5%. e highest point of the slope is in the south of the study area, and the lowest point is in the north. ere are 3 steps with slope covered with stone masonry, as shown in Figures 2(a)-2(c). e slope protection is approximately 50 m wide along the canal and approximately 80 m long perpendicular to the canal.
It was observed in the seventh year after its construction that a small amount of offset had occurred at the bearing of bridge piers #1∼#4 built on the canal slope, and there were many tensile cracks visible at the surface of the stone masonry covering above the slope. It was speculated that instability of the canal slope might be triggered due to several seasonal wet-dry cycles and the variation in river level. When the bridge was examined in the ninth year after its construction, the magnitude of the bearing offset at the top of piers #1∼#4 was larger than ever before, and the offset was accelerating. A comprehensive inspection of the slope and the bridge built on it was carried out ten years after its construction.
e investigation results revealed that the bearing offset of pier #2 was 23 cm and that the bearing offset of pier #1 was 18 cm. ere were many wider tensile cracks at the surface of the stone masonry covering above the slope, as shown in Figure 2(c).
ere was a thoroughgoing cross fracture in the asphalt concrete road approximately 20 m away from the crest of the slope, as shown in Figure 2(d). e piers of the bridge built on the slope had clearly moved in the direction of the riverbed, northwardly. e maximum offset of pier #3 was 44.5 cm, and the vertical tilt of pier #4 was the largest, 1.52%, as shown in Table 1. e ground-water level is dominantly controlled by precipitation and the lateral recharge of rivers. Rainfall is mainly produced by natural evaporation and discharge of rivers. e lowest river water level of the canal is 6 m and occurs in winter. e river water level rises slowly from April to June and rapidly during July and August to the highest river water level, which is absolute height of 8.9 m above mean sea level. e river water level drops rapidly from September to October and then gradually decreases to the lowest river water level in January. e absolute height of the riverbed above mean sea level is 3.6 m.

Ground Investigation
e ground investigation included collecting soil samples from six groups of test boreholes, named BH1, BH2, BH3, BH4, BH5, and BH6. BH1 was situated adjacent to the crest of the slope, BH2, BH3, BH4, and BH5 were situated in the medium of the slope, and BH6 was situated at the toe of the slope. e soil samples collected in situ from the site were tested to evaluate the index properties and shear strength parameters of the slope soil. e test results were then utilized in FEM simulations. e ground investigation was implemented by using SPTs and CPTs. e layout of the soil test borings is presented in Figure 3.    and BH6 were 20 m. All the test borings were 110 mm in diameter.

Laboratory Tests.
Index properties of the slope soil, including the water content (WC), plastic limit (PL), plasticity index (PI), and liquid limit (LL), determined from the laboratory tests of the soil samples are shown in Table 2 and Figure 4. According to the geological age and soil properties, the slope soils were divided into three soil layers, whose soil profile is shown in Figure 5(a). Soil layer 1, namely, "made ground," was loose yellowish grey clay mixed with macadam and plant roots. Soil layer 2 was greyish yellow clay interlayered with thin grey silty clay layers. Soil layer 3 was stiff yellowish brown clay with sparse iron manganese concretions and was partially mixed with gravel.
According to the results of the PI and LL tests, the PI and LL values of the slope soil ranged from 34.8 to 42.8 and 20.6 to 23.1, respectively. e magnitude of the water content of the slope soil is presented in Table 2, which shows that the average water contents of soil layer 1 and soil layer 2 were 25.5% and 23.6%, respectively, while soil layer 3 had moisture content of 21.7%. e results of the soil tests were evaluated, as shown in Figure 4. e slope soil was categorized as low-plasticity clay (CL). e coefficients of permeability of soil layer 1 and layer 2 were 4.5 × 10 −5 cm/s and 7.8 × 10 −6 cm/s, respectively, while soil layer 3 had coefficient of permeability of 2.7 × 10 −7 cm/s, which indicates that the permeability of soil layer 1 and soil layer 2 was only slightly permeable and that of soil layer 3 was nearly impermeable.
ere was ground-water in soil layer 1 and layer 2 in the vicinity of the canal, and the ground-water level varied with the rise and fall of the river water level. Figure 3(a), were drilled on the slope. Each group comprised two bore holes spaced 1 m apart. One borehole was drilled for the SPT and soil sampling; the other was drilled for the CPT. e soil CPT profile and soil parameters from the SPT values determined from BH1 to BH6 are shown in Figure 5. e variations in CPT value and SPT value with the depth are shown in Figures 5(a) and 5(b), respectively.

In Situ Tests. Six groups of boreholes, as shown in
As shown in Figure 5(a), the variation in the CPT value with the depth of each borehole indicates that the CPT values of soil layer 1 were within 0.715 to 2.484 MPa and those of soil layer 2 were within 0.835 to 3.358 MPa, while the CPT values of layer 3 ranged between 2.456 and 7.778 MPa, which were much higher than those of the upper two strata. Figure 5(b) shows that the SPT values ranged from 4 to 11, 6 to 21, and 17 to 67 for soil layers 1, 2, and 3, respectively. e SPT values increased with depth, like the results of CPTs. e SPT values of soil layer 3 were much higher than those of the tow soil layers above, clearly indicating that the lower stratum had higher strength.

Estimating Potential Expansion.
Expansive soil has the conspicuous clay content. Many types of testing and determination methods have been proposed, but none of them are generally applicable. In this paper, the magnitude of swelling is determined using the free swelling ratio, δ ef , which is the percentage equal to the volume increment to the original volume of dry soil after steady expansion in water. e free swelling ratio tests were conducted in accordance with the GB50112-2013 guidelines. During the test, a dry soil specimen is remoulded into a 50 ml graduated cylinder. e specimen is immersed in water to be fully saturated and is permitted to stand freely until the rate of swelling reaches a critical value. e free swelling ratio, δ ef , is expressed as where v w is the final volume of the soil after expansion and v 0 is the initial volume of the specimen. According to the typical classification of soil expansiveness based on free swelling ratio test results, the potential expansion of soil is very low, low, medium, and high when the free swelling ratio values range from 0 to 40, 40 to 65, 65 to 90, and more than 90, respectively. e free swelling ratios of the soil samples from the study site at borehole BH1 are shown in Figure 6, which indicates that the potential expansion of soil layer 1 was very low and that the potential expansion of soil layers 2 and 3 was low.
As shown in Figure 6, the slope soil is potentially low expansive clay, which could result in many problems such as slope instability and foundation cracking [17]. [18] data of the project site in Nanjing, East China, were evaluated. ese precipitation and evaporation data of Nanjing are used and are shown in Figure 7. e total annual precipitation and evaporation are very high and the highest occur during summer. In contrast, there are less precipitation and evaporation during the winter. at is, the wet summer is followed by the dry winter. e clay soil might have a high expansion due to the variation in water content, according to the results of Holtz and Kovacs [19]. Obvious shrinkage cracks might form due to wet-dry cycling, which would probably be used as the preferential flow path for water during precipitation. Because low-plasticity clay soil has weak osmosis property, rain water is likely to remain below the active area. erefore, the instability of the slope due to soil strength reduction caused by precipitation requires analysis. e depth of the active zone is a crucial parameter when analyzing the expansion of soil. e active zone [20] by the design can be taken as the zone of moisture content fluctuation. e engineering design also needs to consider how human activities affect soil expansion at the given site. e water condition may change due to human activities at the particular place; hence, the depth of the active zone is influenced as well. Because the fissures transport water to greater depth, heavily fissured soils will generally have deeper active zone. Field studies conducted in Colorado indicated that the studied active zone extended to 16 m below the ground surface in certain areas [21].

Precipitation and Evaporation. Precipitation and evaporation
If the site is in its natural state under the influence of climatic changes, without any influence of human activities, the water content of the soil underlying the site will vary more near the slope surface than deep underground. e reason for this trend is that the variations in precipitation and evaporation change more rapidly within these upper soils. It is assumed that the water content is approximately constant below this zone. Hence, no expansion occurs below this zone. It was determined from the precipitation and evaporation data that an approximate average depth of natural active zone in Nanjing, East China, is 3.0 m below the ground surface.

Pile Integrity Detection
e drilling core measurement method (DCM) is broadly utilized to detect the continuity of pile integrity and concrete quality [22]. e method is very simple and easily applied, whereas the discontinuities, especially fractures, in pile are very difficult to determine precisely. Fractures in pile are crucial to identify the evaluation of the safety of the whole structure. A borehole optical imaging system is an effective, convenient, and practical drilling investigation technology. e interior status of the borehole can be directly observed from optical images [23,24]. To establish the spatial coordinate system of the drilling hole, the diameter of the borehole must be known, and the shape of the borehole should be a standard cylinder. erefore, the position and orientation of the panoramic image can be determined by utilizing a location method. A continuous orientated digital 360°colour image of the borehole wall can be attained by borehole optical imaging technology. us, the continuous change in the borehole wall and the fractures in the pile can be directly surveyed [25]. In this paper, borehole optical imaging technology is used to obtain an image along the borehole wall in pile.
In the space beneath the bridge, piles 2 and 4 of pier #3 and pile 1 of pier #4, which was 1200 mm in diameter and 33 m in depth, were selected, as shown in Figure 8. e drilling hole in each pile is 35 m deep and 91 mm in diameter. All of the drilling holes were flushed after drilling to remove debris and broken stones and then swilled using clean river water. Finally, the selected piles were imaged with borehole optical imaging technology to identify fractures in the piles. e typical borehole wall images of the fractures obtained at the site are presented with the depth of each pile drilling hole, in Figure 9, which shows the emerging fractures of piles 2 and 4 of pier #3 and pile 1 of pier #4. Figure 9 shows that there were 6 irregular fractures at a depth of 10.0 m∼14.0 m in pile 2 of pier #3 and 10 irregular fractures at a depth of 10.0 m∼15.0 m in pile 2 of pier #3, while there were 12 irregular fractures at a depth of 5.0 m∼12.0 m in pile 1 of pier #4. From the results of the detected piles, it could be speculated that there were also many fractures in other unselected piles of the piers, which shows that the soil slope had undergone potential instability and failure. Additionally, as shown in Figure 9, the depth of the fractures observed in the piles of pier #4 was shallower than that of pier #3; at the same time, the depth range of the fractures observed in the piles of pier #4 was greater. e depth of the potential slide interface in the soil slope could be derived from the depth of the fractures in the piles.

Slope Instability Analyses
e slope instability analyses were conducted by using the FEM program ABAQUS, which utilized the elastic-perfectly plastic Mohr-Coulomb soil model. During this study,    standard fixities were used as the boundary condition in the ABAQUS models. e shear strength reduction method was used to analyze the slope instability. e strength reduction coefficient of the soil slope reflects that the original shear strength parameters may be reduced enough to trigger the failure of the soil slope [26][27][28]. e results of the current research reveal that the shear strength of expansive soils will decrease with an increasing number of wet-dry cycles. e cohesion of the expansive soil clearly decreases, but the friction angle is relatively stable with the increase in the number of wet-dry cycles [29]. e reduction in cohesion with the increase in the number of wet-dry cycles is similar to an exponential function relationship, while the friction angle has a decreasing trend with the increase in wet-dry cycles, but the rate of decrease decreases [30]. e soil parameters, as shown in Table 3, were utilized to evaluate the instability of the slope by using ABAQUS. e original model of the soil profile is presented in Figure 10. To simplify, Poisson's ratio of all the soils of the slope was 0.3. e two-dimensional model was 190 m long in the horizontal direction and has a maximum depth of 38 m; a schematic of the original model is shown in Figure 10.
Considering all factors, the Mohr-Coulomb strength criterion was chosen as the failure criterion of the material in this calculation. e soil strength theory of the Mohr-Coulomb strength criterion is very important in soil mechanics and can reflect the plastic failure characteristics of soil. Because the Mohr-Coulomb strength criterion is simple and practical and the strength parameters of soil are easily measured, this criterion is widely applied in soil mechanics and plasticity theory. At present, many traditional analytical methods, such as slippery line theory, ultimate bearing calculation theory, Earth pressure calculation theory, and structural strength theory, are based on this criterion; its frequent use in practical applications has resulted in the accumulation of experimental data [31][32][33].
e ABAQUS model was constructed by using a large deformation model. e results of the numerical simulated calculation are presented in Figure 11. e shear strain increment contour plots are mainly used to analyze the trend of slope instability.
As shown in Figure 11, soil layer 2 of the slope is clearly situated in the plastic strain zone, indicating that soil layer 2 has the greatest influence on the slope instability in the FEM models under fully softened conditions. As mentioned above, many tensile cracks happened at the surface of the stone masonry covering the slope soil. e shear strength of the slope soil decreased greatly after the river water level rose and the precipitation intruded in summer, which led to the instability of the soil slope. e desiccation cracks caused by soil swelling-shrinking cycles carried rainwater and river water into the deep soil. e deep low-plasticity clay had a low water conductivity, which prevented the deep intrusion of the water. Consequently, the reduction in fully softened soil shear strength caused by seasonal wet-dry alternation and river level fluctuation conditions in the soil on the slope leads to low-angle low expansive soil slope instability for several years following construction. e FEM simulation was utilized to determine the safety factor of this soil slope. e factors of safety of the soil slope were calculated to be 1.38 in the first winter and 1.04 in the first summer, which revealed that this soil slope was stable in the first year. In this analysis, slope instability would happen if the factor of safety was equal to 1.0. e factors of safety of the soil slope were calculated to be 0.80 in the third summer, 0.70 in the sixth summer, and 0.65 in the ninth summer after its construction, as presented in Table 4, which indicated that the slope began to be unstable and damaged 3 years following construction due to the drywet cycle and the rise and fall of water level. erefore, this soil slope might have failed because of the reduction in soil strength resulting from seasonal wet-dry alternation and river level variation events.      Figure 10: Schematic of the original model. Advances in Civil Engineering

Conclusions and Discussion
e instability of a gently sloping low expansive soil slope resulted in many tensile cracks at the surface of the mortar block stones covering the slope soil and lots of fractures in the piles of the bridge foundation. A comprehensive site investigation and a wide variety of tests were implemented, which included geophysical tests, in situ tests, laboratory tests, pile integrity detection, and numerical analysis with finite element method (FEM). e results revealed that this gently sloping expansive soil slope has become unstable due to the events of several seasonal wet-dry alternations and river level variations after its construction. e low-plasticity clay caused dry shrinkage, which caused the mortar block stones covering the slope soil to undergo tensile cracking during the dry season and develop paths for water to flow into the slope soil during the wet season. e hydraulic conductivity of the slope soil would be changed significantly owing to the tensile cracks in the mortar block stones and the shrinkage cracks in the slope soil, which brought water from the topsoil of the slope to the active zone. According to the study conducted   [34]. According to the pile integrity detection results, it could be speculated from the many irregular fractures observed in the pile that the soil slope had undergone instability. e data from the FEM numerical analysis indicated that the soil slope began to be unstable and damaged 3 years after its construction. Both seasonal wet-dry alternation and river level fluctuation were responsible for the slope instability, and the typical characteristics of the instability and failure of the gently sloping low expansive soil slope were tractional detachment and slow sliding.
Findings from the study demonstrated that the reduction in expansive soil shear strength with wet-dry cycling plays a very important role in the evaluation of slope instability. Furthermore, the slope stability and the factor of safety decrease with the increase in the number of wet-dry cycles. In this paper, the reduction in soil shear strength resulting from seasonal wet-dry cycles and river level variations led to the instability and failure of the low-angle low expansive soil slope.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that they have no conflicts of interest.