Numerical Simulations of a Permeability Test on Non-Cohesive Soil Under an Increasing Water Level

Abstract

With the intensification of global climate change, extreme rainfall events are occurring more frequently. Continuous rainfall causes the debris flow gully to collect a large amount of rainwater. Under the continuous increase in the water level, the water flow has enough power to carry plenty of loose solids, thus causing debris flow disasters. The intensity of the soil is reduced with the infiltration of rainwater, which is one of the key causes of the disaster. The rise in the water level affects the infiltration behavior. There have been few previous studies on infiltration under variable head. In order to understand the infiltration behavior of soils under the action of water level rises, this paper conducted an indoor permeability test on non-cohesive soil under the condition of an increasing water level. A numerical model was established using the finite element analysis software, Abaqus 6.14, and the pore pressure was increased intermittently to simulate the intermittent increase in the water level. Thereafter, the permeability coefficient and seepage length were changed to interpret the changes in the flow velocity and rate in the permeability test of the non-cohesive soil. The results showed that the finite element numerical simulation method could not reflect the particle movement process in the soil. The test could better reflect the through passage and void plugging phenomenon in soil; when the permeability coefficient alone changed, the velocity of the measuring point with higher velocity changed more violently with the permeability coefficient; when the length of soil seepage diameter was uniformly shortened, the velocity of water flow increased faster and faster.

1. Introduction

Debris flow is a sudden geological disaster in mountainous areas and has the characteristics of a large impact range, fast speed, and wide coverage [1]. Earthquakes and other effects produce abundant loose materials in mountainous areas, which become the necessary conditions for the formation of debris flow, while rainfall becomes one of the main factors causing debris flow disasters [2]. With the impact of global climate change and human activities, the ecological environment continues to deteriorate, and the number of extreme rainfall events has increased significantly. This has led to a concomitant increase in the occurrence of debris flow disasters.

Regarding the effect of rainfall on debris flow, it is usually reflected in antecedent rainfall, rainfall intensity, and rainfall duration [3]. Antecedent rainfall, which occurs hours to days before the debris flow, plays a key role in the triggering of the debris flow, and the infiltration of rainwater in the soil reduces the shear strength and may cause a massive drift during the channel flow caused by the main convective cell [4]. In addition, when the rainfall intensity is higher than the infiltration capacity of the surface, surface runoff occurs, and the process of surface runoff scours and carries away loose material to cause debris flow [5,6]. When the rainfall intensity is low, a long period of low-intensity rainfall will lead to an increase in the underground pore water pressure, thus triggering debris flow [7]. These three factors will affect the development of debris flow. Zhang et al. [8] calculated the probability of debris flow under different antecedent effective rainfall using numerical calculation, rainfall scenario simulations, and the Monte Carlo integration method and found that the higher the antecedent effective rainfall, the higher the probability of debris flow triggered by subsequent rainfall. Some scholars have discussed the effect of rainfall intensity-duration on debris flow. In the study by Cannon et al. [9], the range of rainfall conditions associated with different magnitude classes was defined by integrating local rainfall data with the response magnitude information. Guzzetti et al. [10] plotted the rainfall intensity–duration (ID) values in logarithmic coordinates and found that the increase in rainfall duration leads to a decrease in debris flow triggering intensity. Staley et al. [11] presented a new, fully predictive approach that utilized rainfall, hydrologic response, and readily available geospatial data to predict rainfall intensity–duration thresholds for debris flow generation in recently burned locations in the western United States. These studies have mainly focused on rainfall parameters, and they provide a reference for the prediction and early warning of debris flow disasters according to rainfall conditions.

In addition, the soil state is also one of the key factors. The mechanism of rainfall influence on debris flow is actually the response of soil to rainfall, which can be reflected by parameters such as pore water pressure, moisture content, soil pressure, and cohesion [12,13,14,15,16,17]. Andrew et al. [18] found that changes in inflow can lead to large variations in simulated flow depths and velocities, and inflow conditions can potentially affect debris flow. Debris flow-prone soils usually consist of poorly graded, loose, unconsolidated material; this includes coarse particles (such as rock, gravel, and sand) and fine particles (such as silt and clay) [19]. Under seepage conditions, fine particles tend to migrate from the skeleton formed by coarse particles, causing severe internal erosion and affecting their mechanical properties [20,21,22]. Different soil pore structures, due to different coarse particle contents, affect the permeability coefficient and lead to changes in the erosion rate [23,24]. It can be seen that both the parameters of the soil and the infiltration conditions have an impact on the formation of debris flows.

In recent research, the seepage law and seepage erosion of soil have been studied a lot, but research of seepage under the variable head has not been extensive enough. Under the effect of continuous rainfall, the rising water level will cause a change in the infiltration state and thus affect the process of debris flow. Infiltration will cause changes in the size and distribution of soil strength. If the law of soil infiltration regarding the water head can be studied clearly, it is of great significance to prevent natural disasters such as landslides and debris flows. Therefore, the distribution and variation of the flow velocity and flow rate in soil under an increasing water level are studied in this paper. Based on the infiltration test, the development law of the pore pressure and flow rate in the soil is obtained. The test operation is complicated, it can be interfered with by environment and human factors, and it does not allow for observations of the development of flow velocity from the inside. In order to understand the overall change in the seepage state inside the soil, Abaqus was used to establish a finite element model and analyze the evolution of the flow velocity and flow rate under different seepage path lengths and permeability coefficients of the soil.

2. Indoor Permeation Test

The test instrument used for the indoor tests was a self-research instrument, as shown in Figure 1. The self-developed cylindrical test cylinder was made of plexiglass to facilitate the observation of soil particle movement. The length of the cylindrical test cylinder was 50 cm, the inner diameter was 14 cm, and the water pressure sensors (Leicester Co., Ltd., Qingdao, China) were set every 5 cm from the top to the bottom of the cylinder wall, so there were eight water pressure sensors in total, which were numbered 1–8 in order from bottom to top. A strip sand mesh with an aperture of 0.075 mm was pasted on the wall of the cylindrical test cylinder where the hole pressure sensors were installed to prevent the passage of soil particles while the water was able to pass through, so as to avoid direct contact between the soil and the sensors that would affect the measurement results. In this test, the soil sample loading area was 25–40 cm from the bottom of the cylindrical cylinder. The height of the soil sample could be changed by using a pedestal. Due to the smooth inner wall of the test cylinder, in order to prevent seepage channels on the side wall during the test that would cause a large number of soil particles to be transported and affect the test results, a boundary control ring was provided at 1/3 and 2/3 of the height of the soil samples (as shown in Figure 2a), with a thickness of 1 cm. The boundary control ring could be installed and removed as required.

At the bottom of the cylindrical test cylinder was a buffer zone with a height ranging from 10 to 25 cm. The height could be adjusted by the pedestal, which was used to convert the water flow into a uniform water flow so as to ensure that the upper soil sample was subjected to a stable and uniform water flow. The buffer zone was equipped with a permeable plate on the upper part, and the upper surface of the permeable plate was equipped with a stainless-steel mesh with apertures of 0.075 mm to ensure that water could pass through and the fine particles did not leak into the buffer zone. A wire for dismantling was attached to the permeable plate to facilitate the removal of the permeable plate from the cylindrical test cylinder when the instrument was adjusting the seepage diameter length of the soil sample. The cylindrical test cylinder is shown in Figure 2b.

The instruments used in this study consisted of two parts: a cylinder test tank system and a data acquisition system; a schematic diagram is shown in Figure 1. The main components of the data collector were the water pressure sensor and the flow meter (Omega Co., Ltd., Norwalk, American). The water pressure sensor used a diffused silicon pressure transmitter from Leicester Instrumentation with an accuracy of 0.2%FS and an output signal of 4–20 mA. By setting the recorder (Sinomeasure Co., Ltd., Hangzhou, China), the received current signal “4–20 mA” was converted to the corresponding water pressure “0~Range Maximum (kPa)”. Since the water pressure values of water pressure Sensor 1 and water pressure Sensor 8 differed greatly, in order to more accurately measure the real-time change in water pressure, the data acquisition system used two water pressure sensors with different ranges: PWPS-1, with a range of 0~15 kPa, and PWPS-2, with a range of 0~6 kPa. The flowmeter adopted the model FPR301 impeller flowmeter produced by American Omega Company. The flowmeter measured the flow range from 0 to 316 mL/s with an accuracy of 1%. Because the recorder could only receive 4–20 mA signals, it was necessary to configure a signal converter for the flow meter to convert the signal of the flow meter to 4–20 mA signals. Since the water pressure sensor was linearly related to the upstream head, the water pressure sensor was calibrated by adjusting the upstream head.

The composition of the soil particles used in this test was mainly silicon dioxide. The soil sample was mixed to consist of coarse particles and fine particles in a certain proportion. The coarse particles were 2.0–8.0 mm, which play the role of the skeleton in the sample; the fine particles were 0.075–0.25 mm, which fills the pores formed by the skeleton of coarse particles. According to the geotechnical test specification (SL237-1999) for coarse particles and fine particles when conducting a specific gravity test and a relative compactness test, the basic physical property parameters of the test soil sample are shown in

Table 1. Specimen parameters.

Fine Particle Content (%)	Emax	Emin	Dr	Seepage Length (cm)	Gs
15%	0.56	0.29	0.3	30	2.70

.

During the test, the initial hydraulic gradient was 0.02, and the increment of each hydraulic gradient was 0.05. The specific steps of the test are as follows: (1) Soil sample saturated for 24 h; energize the recorder, flowmeter, and water pressure sensors, and ensure the data acquisition instrument operates normally for 5 min; (2) Raise the upstream tank water level to the height corresponding to the initial hydraulic gradient of 0.02 by lifting the tank and keep it there for 30 min; (3) Continue to raise the water level of the upstream tank to the next hydraulic gradient; the increment of each hydraulic gradient is 0.05 and the duration of the action is 30 min; (4) During the test, when the flowmeter cannot work properly as the flow rate is small, it is necessary to measure the flow rate with a cylinder and measure the flow rate at least three times for each level of the hydraulic gradient; (5) When the head of the soil sample cannot be further increased, the test is completed. The change process of the pore pressure and flow rate measured in the text is shown in Figure 3. It can be seen that with the increase in the hydraulic gradient, the flow rate and the pore pressure at each sensor point increase gradually, and the pore pressure increment shows an increasing trend from top to bottom; for example, the pore pressure increment at the uppermost sensor 8 is only 0.11 kPa.

3. Numerical Simulation

3.1. Principles of Saturated-Unsaturated Seepage Analysis

Saturated-unsaturated hydraulic parameters play an important role in the accurate solution of the seepage field. In saturated soil, self-gravitational and pore water stresses cause water transfer, whereas in unsaturated soil, it is the self-gravitational stresses and surface tension that govern the transfer of pore water. Infiltration in unsaturated soils also conforms to Darcy’s law and the continuity equation [25]. The two-dimensional partial differential equation for seepage in saturated-unsaturated soil can be obtained by bringing Darcy’s law into the continuity equation and using the total head h as an unknown quantity while neglecting the change in total stress and the deformation of the soil skeleton during the infiltration process:

$${\displaystyle \frac{\partial }{\partial x}}\left(k_x{\displaystyle \frac{\partial h}{\partial x}}\right)+{\displaystyle \frac{\partial }{\partial y}}\left(k_y{\displaystyle \frac{\partial h}{\partial y}}\right)={\displaystyle \frac{\partial {\theta }_w}{\partial t}}$$

(1)

$k_x$, $k_y$ are the permeability coefficients in the x, y directions; ${\theta }_w$ is the volumetric water content; h is the total head; t is the time. Let y be the positional head, then:

$$h={\displaystyle \frac{u_w}{{\gamma }_w}}+y$$

(2)

If $m_w$ is the slope of the soil-water characteristic curve, add the $u_w$ of Equation (2) into the following formula, then:

$$\partial {\theta }_w=m_w\partial u_w=m_w{\gamma }_w\partial \left(h-y\right)$$

(3)

Since y is a constant, Equation (1) can be simplified to:

$${\displaystyle \frac{\partial }{\partial x}}\left(k_x{\displaystyle \frac{\partial h}{\partial x}}\right)+{\displaystyle \frac{\partial }{\partial y}}\left(k_y{\displaystyle \frac{\partial h}{\partial y}}\right)=m_w{\gamma }_w{\displaystyle \frac{\partial h}{\partial t}}$$

(4)

From Equation (4), it can be seen that the permeability coefficient of saturated soil can be regarded as a constant but the permeability coefficient of unsaturated soil is no longer a constant, but a function of the slope of the soil-water characteristic curve, and this function is the permeability coefficient function of unsaturated soil. Therefore, the following basic conditions are required to carry out unsteady seepage analysis: (i) the permeability coefficient function of the material, including the permeability coefficient function of the negative pore pressure zone; (ii) the rate of water level falling or rising; (iii) determining the boundary conditions. When the above conditions are known, Equation (4) can be solved to obtain the unsteady seepage field of saturated-unsaturated soil.

3.2. Finite Element Modeling

The Abaqus modeling process does not limit the units of the parameters, so in the process of use we need to stipulate the parameter units [26]. The parameters and units used in this paper are shown in

Table 2. Parameter used in the modeling.

Parameter Name	Densities	Elastic Modulus	Gravitational Acceleration	Permeability Coefficient	Poisson’s Ratio	Size
unit	kg/m3	Pa	N/kg	m/s	\	m
number	2700	3 × 106	10	0.0007	0.3	0.3 × 0.14

. Since only the seepage field is analyzed in this paper, the deformation characteristics of the soil samples are not considered. The model as a whole is constrained to avoid displacement and rotation. The two sides of the model are undrained surfaces, and the top surface is a drainage boundary. Due to the support of the pedestal, only the center hole of the pedestal has water inflow, so the center 4 cm of the bottom is set as a pore pressure boundary to simulate different head conditions. The pore fluid/stress (CPE4P) cell type was selected to mesh the model, which contained 420 cells, as shown in Figure 4.

3.3. Finite Element Model Validation

In the test, sensor 3 is located in the pore pressure boundary and the measured data are the pressure of the water inlet under the soil sample. Therefore, measuring point 3 can be set at any point on the pore pressure boundary. The locations of the remaining hole measurement points were consistent with the test locations, as shown in Figure 4. It takes the pore pressure of Sensor 3 as the boundary condition of the numerical model to make the model more consistent with the experimental situation.

Figure 5 shows a comparison of the pore pressure curve with the time obtained by the simulation and the test, in which the hollow graph shows the test results and the solid graph shows the simulation results. It can be seen that the overall trends of the test and simulation curves are the same. All the measurement points gradually increase with the increase in water pressure at the inlet, but the increment from the bottom to the top is gradually getting smaller. The pressure of measurement point 8 remains almost unchanged. Some of the measurement points have different results, of which the difference is more obvious for measurement points 5, 6, 7, and 8. The initial pore pressure difference at measurement point 6 is the largest, with a simulated value of 1.61 kPa and a tested value of 1.92 kPa, which increased by about 19.25% compared with the simulated value. In the initial stage, the numerical difference between the test and the simulation is large. Over time, the numerical difference will gradually develop and become relatively close. This is due to the fact that at the beginning of the test, the fine particles in the soil samples move, resulting in the rapid penetration of the water flow throughout the soil. As time goes on, the fine particles move to a relatively stable position, and pore blockage occurs. On the other hand, the measuring instrument is fixed on the cylindrical plexiglass surface, and once the water flow reaches the inner wall of the test cylinder through the channel created by the movement of the fine particles, it tends to form a seepage channel in the smooth side wall, which results in a higher measured pore pressure. The effect decreases as the pore is clogged.

Figure 6 shows the flow rate comparison between the numerical simulation and the test. It can be seen that the changes in the two curves are basically the same. The flow rate increases with the hydraulic gradient, and there are two obvious differences between them: The test flow rate is higher than the simulation flow rate in the first half of the time, and the deviation decreases with time. The maximum difference appears in the 33rd minute, with a difference of 1.28 mL/s. The average value of the test flow measurement data before 175 min is 5.01 mL/s, while the average value of the simulated flow is 3.98 mL/s, with a difference of 1.03 mL/s. The average difference after 175 min is only 0.07 mL/s; in the first 100 min of the model test, it is obvious that the flow rate gradually decreases under each stable head, while the numerical simulation shows that the flow remains stable under the stable head. The reason for the difference in flow rate is the same as that for the difference in the pore pressure, which is caused by the movement of soil particles. When the soil particles in the soil sample basically reach a relatively stable state, most of the seepage channels in the soil sample are blocked, so the flow rate of the two tends to be consistent in the latter half. Overall, the change trends of the numerical simulation and the test results are consistent. The deviation between them is not obvious, so it can be proved that the numerical analysis model and parameter are rational.

As can be seen from Figure 7, after the water flow enters the soil, the flow velocity is the largest in the direction of about 45° outward from the edges of the pressure boundary with a magnitude of 2.47 mm/s. On the pore pressure boundary, the water flows in the form of a sector, and the farther away from the pore pressure boundary in the sector, the slower the water flow velocity is. After leaving this sector, the water flows to the free outlet boundary with a velocity of about 0.50 mm/s. The slowest water flow velocity of 0.08 mm/s was observed at the location of the lowermost edge of the soil sample.

The distribution of pore pressure at typical time steps in the numerical simulation is shown in Figure 8. It can be seen that there is a pore pressure transition region near the water inlet, which is gradually dispersed in the shape of a sector. After two to three transition regions, the change in pore pressure roughly becomes a uniform increase with height. The topmost pore pressure is almost zero due to the drainage interface, which is consistent with the model test. In terms of the change process, the sector of the inlet is not obvious at 1 min because of the small head; the gradual increase in the pore pressure will lead to the gradual obviousness of the sector. In addition, as the pore pressure boundary increases, the pore pressure increases at any location in the soil.

4. Parametric Analysis of Simulation

4.1. Variation of Flow Rate and Velocity Under Different Permeability Coefficients

The permeability coefficient is an index that comprehensively characterizes the permeability of soil and reflects the difficulty of water permeating through soil. The permeability coefficient has a great influence on the seepage of soil [27]. In order to research the seepage of soil under different permeability coefficients, the numerical model established above is adopted to study the changes in the flow rate and velocity of soil with the permeability coefficient under the condition of a rising water level. The parameters used in this part of the simulation are shown in

Table 3. Simulation schemes under different permeability coefficients.

Number	Permeability Coefficient/(m/s)	Seepage Length/(cm)
K1	0.0001	30
K3	0.0003	30
K5	0.0005	30
K7	0.0007	30
K9	0.0009	30

. Keep the seepage length unchanged, and only change the permeability coefficient of the soil.

K7 is the test model, and its water flow velocity magnitude distribution is shown in Figure 9. The flow velocity magnitude is the largest on both sides of the pore pressure boundary. In addition to these two regions, the flow velocity spreads from the pore pressure boundary in the form of a sector. After reaching 2/5 of the height of the model, the flow continues to the drainage boundary with almost equal velocity. In addition, the flow velocity is the smallest at the bottom two corners of the model. Accordingly, the maximum flow velocity and the minimum flow velocity are selected as velocity measurement points (points 1 and 5 in Figure 9). The central position of the three sector regions is selected as velocity measurement points 2, 3, and 4. The area with almost constant velocity outside the region is selected as velocity measurement point 6. So as to compare the velocity variation rule under different working conditions, taking point 2 as the origin of the coordinate, the coordinates of other measuring points are: point 1 (2, 0), point 3 (0, 2), point 4 (0, 5), point 5 (7, 0), and point 6 (0, 15). The size of the model is 30 cm×14 cm (length × width).

Figure 10 shows the curve of the flow velocity and the permeability coefficient at 340 min. It can be seen that the curve of the flow velocity and the permeability coefficient at the same measuring point is a straight line. In comparison, $v_1>v_2>v_3>v_4>v_6>v_5$. The larger the flow velocity, the greater the slope of the curve at the measurement point. Under the same permeability coefficient, the hydraulic gradient is larger at the measurement point with a larger flow velocity. Therefore, in the actual project, it is necessary to focus on the location of the soil where the flow velocity is larger. When the soil has leakage channels and the permeability coefficient increases significantly, the increase in velocity will lead to more serious erosion of the soil. Therefore, the water flow velocity should be strictly limited to prevent the occurrence of infiltration damage.

When the inlet pore pressure is constant, the flow rate increases with the increase in the permeability coefficient, as shown in Figure 11. Figure 12 shows the relationship between the flow rate and time under different permeability coefficients. It can be seen that with the passage of time, the flow rate increases with the increase in the inlet pore pressure. Similarly, due to the proportional relationship between the flow rate and the permeability coefficient, the flow difference between adjacent permeability coefficients at the same time is basically the same. This result is based on a single permeability coefficient. In reality, when the soil is infiltrated, the soil permeability coefficient will change accordingly, as shown in the laboratory test in this paper. When the soil strength is poor, once the internal leakage channels are generated, the permeability coefficient will change more violently [18].

4.2. Variation of Flow Rate and Velocity Under Different Seepage Length

In addition to the permeability coefficient, the seepage length is also a major factor affecting the seepage of soil. The above numerical model is still used to research the change rule of the flow velocity and the flow rate by changing the seepage length, and the simulation parameters are shown in

Table 4. Simulation schemes under different seepage lengths.

Number	Permeability Coefficient/(m/s)	Seepage Length/(cm)
L15	0.0007	15
L20	0.0007	20
L25	0.0007	25
L30	0.0007	30
L50	0.0007	50

. The velocity distribution in the model is different due to the different heights of the models. Considering the flow velocity distribution results comprehensively, the selected flow velocity measurement points are shown in Figure 13. Taking point 2 as the origin of coordinate, the coordinates of other measuring points are: point 1 (2, 0), point 3 (0, 2), point 4 (7, 0), and point 5 (0, 4). The length of the model in Figure 13 is 20 cm.

From Figure 14, it can be seen that the seepage length has a great influence on the flow velocity. The longer the seepage length, the slower the flow velocity. This weakening tendency has a greater influence in the region where the flow velocity is larger. When the seepage length of the soil is uniformly shortened, the flow velocity of the water increases more and more quickly. Taking point 1 as an example, the flow velocity of 15 cm, 20 cm, 25 cm, and 30 cm seepage lengths are 6.47 mm/s, 4.52 mm/s, 3.26 mm/s, and 2.37 mm/s. Growth is 1.95 mm/s, 1.26 mm/s, and 0.89 mm/s. In the case of heavy rainfall, when the strength of the slope decreases, the fracture zone along the slope surface under the action of gravity is similar to the situation where the length of the seepage path is shortened. At this time, the acceleration of rainwater infiltration will make the slope more vulnerable to damage.

Between the flow rate and the seepage length is an inverse relationship, as shown in Figure 15; this is similar to the flow velocity relationship curve. In Figure 16, the flow rate increment of the neighboring curves at any time point shows a decreasing trend with the uniform increase in the seepage length, except for L50. The flow rate in the first half of the L50 curve is almost zero due to the fact that the soil inside the numerical model is set to be saturated, and the pore water is subjected to a pressure magnitude of 4.9 kPa at the lowermost edge under the action of gravity, which indicates that the water level is below the outflow surface and no water is flowing out; the pore pressure given by the boundary is less than 4.9 kPa before 182 min.

5. Conclusions

In this paper, the seepage test of soil under water level climb was carried out using a self-developed test instrument, and numerical simulations were conducted to analyze the test. By comparing the pore pressure and flow rate of the numerical analysis, the rationality of the model was verified. Then, through a numerical analysis of the permeability under the seepage length and permeability coefficient, the main conclusions were as follows:

The pore pressure and flow rate in the soil increased during the gradual increase in the upstream head. However, when comparing the two, the finite element numerical simulation method could not reflect the particle movement process inside the soil. The test could better respond to the phenomenon of through channels and void blockage inside the soil. The finite element method was able to show the flow velocity distribution of the model as a whole, with the maximum flow velocity at the ends of the pore pressure boundary and roughly along the 45° direction.

When the permeability coefficient changed individually, the larger the flow velocity at the measurement point, the more drastic the change in flow velocity with the permeability coefficient. The variation of flow velocity and flow rate in the actual debris flow was much higher than that in the numerical simulation. Therefore, in the actual project, it was necessary to pay attention to the location of the soil where the flow velocity was larger. When the soil produced a leakage channel and the permeability coefficient increased significantly, the increase in flow velocity led to more serious erosion of the soil. Therefore, the flow velocity at this location should be strictly limited to prevent the occurrence of infiltration damage.

When the seepage length of the soil was uniformly shortened, the flow velocity of the water increased more and more quickly. Taking point 1 as an example, the flow velocity of 15 cm, 20 cm, 25 cm, and 30 cm seepage lengths are 6.47 mm/s, 4.52 mm/s, 3.26 mm/s, and 2.37 mm/s. Growth is 1.95 mm/s, 1.26 mm/s, and 0.89 mm/s. In the case of heavy rainfall, when the strength of the slope decreased, the fracture zone along the slope surface under the action of gravity was similar to the situation where the length of the seepage path was shortened. At this time, the acceleration of rainwater infiltration would make the slope more vulnerable to damage.