Study On Seepage Laws Of Completely Weathered Phyllite Slope Under Rainfall

: According to the available geological data and monitoring data, the completely weathered phyllite slope dilates and softens under the condition of continuous rainfall, which is then prone to instability failure. The indoor artificial rainfall test was carried out through the construction of the slope model, and the soil moisture sensor, pore water pressure sensor and matric suction sensor were used to study the variation laws of moisture content, pore water pressure and infiltration line at the back edge, slope body and the foot of the slope under continuous heavy rainfall. According to the sensor data and recorded information, with the influence of heavy rainfall over a long period of time, the water content and pore water pressure increased firstly, then decreased, and finally stabilized. The infiltration line moved from the top, the surface and the foot of the slope to the slope body, and shallow slip failure occurred in the shallow layer of the slope body.


INTRODUCTION
The Tonggu-Wanzai expressway in JiangXi province runs through the Jiuling mountainous area, along which a large number of completely weathered phyllite slopes are distributed, and this area is also the storm center of JiangXi province. Completely weathered phyllite dilates and softens in case of water, which is prone to slope instability, seriously threatening the safe operation of the expressway and the safety of people's lives and properties.
Heavy rainfall is the main cause of such slope failure [1][2][3] . Simplified slope model was built up with site samples [4][5][6] and the change rules of moisture content, matrix suction, saturation line and slope position shift were studied under continuous heavy rain by setting volumetric moisture content sensor, matrix suction sensor and pore water pressure sensor. All the results measured above provided references for further slope failure and reinforcement mechanism research.

Configuration of Slope Soil Samples
The soil used in the test was completely weathered phyllite soil collected from a slope of B5 tender of Tongwan expressway. The strength of this kind of rock and soil mass was low, easy to be softened in water, and it was like mud. After site sampling, the soil samples were prepared following the sequence of drying, crushing and sieving. Before the test, moisture content of the soil samples was measured. Soil with the same moisture content was configured and poured into the model box.

Model Device
The model box was made of plexiglass and steel and the size was 2m×1.9m×1.2m. To reduce the boundary effect, protective film was applied to the inner side of the plexiglass.

Artificial Rainfall Equipment
The artificial rainfall simulation system consisted of three parts: reservoir, artificial rainfall pipeline, sprinkler head and rainfall control system. The effective rainfall area was 2m 2 , the rainfall height being 4m and the rainfall intensity range being 6-240mm/h, while the rainfall uniformity coefficient was greater than 0.86 and the rainfall adjustment accuracy was 7mm/h. The composition of rainfall system is shown in Fig 3. (a) diagrammatic drawing (b) entitative graph

Model Construction
In order to determine the location of soil easily and facilitate the observation of soil deformation, one layer of soil in the model box was set to be 10cm. The configured soil sample was poured into the model box to spread out and compacted in layers with a gradient of 1:1.

Sensor Layout
Seven moisture content sensors and seven matric suction sensors were placed in the slope at different depths and heights. Four pore water pressure sensors were placed at the foot of the slope and inside the slope body respectively, and the specific arrangement is shown in

Process of Rainfall Test
Weili [7] made a statistical summary of the rainfall situation in JiangXi province since 2003. The maximum rainfall in an hour in JiangXi province ranged from 56.2 to 132.5mm. The rainfall intensity of this experiment was set at 70mm/h, and the sensor was embedded and then set for 7 days. After the initial values of the sensors were stabilized, artificial rainfall began. The duration and cumulative time of rainfall are shown in Table 1.

Change law of volumetric moisture content
With the progress of rainfall, the readings of the 7 volumetric moisture content sensors increased successively in the order of SM1, SM2, SM3, SM4, SM7, SM6 and SM5. The volumetric moisture content of SM1 and SM2 on the top of the slope increased firstly, then SM4 at the foot of the slope, and lastly SM3 on the bottom slope. As the rainfall time increased, the volumetric moisture content of SM7 on the bottom of the slope, SM6 on the middle and upper part of the slope, and SM5 on the middle and lower part of the slope increased successively. After the rainfall was stable, the volume moisture content inside the slope body was stabilized at about 34.4%. The shallow sensors were

Variation Law of Pore Water Pressure
Readings of the four sensors began to increase on the second day of rain. As the rain water flowed along the slope and the stratum at the foot of the slope was below the water surface, HC1 sensor value increased firstly, and after all the sensors reached stability, its pore water pressure was the highest. As the rain water penetrated, HC2 numerical reading increased, then HC3. During the rainfall observation period, because the saturation line did not enter the right lower slope, HC4 value showed no obvious changes, as shown in Fig 7.

Variation Law of Matric Suction
With the continuing rainfall, the top of the slope reached saturation state firstly, with the readings of sensor WP1, WP2 and WP7 started to reduce, then near the foot of the slope, and then in the middle of the slope. The sequence of the sensor value reduction was WP4, WP3, WP6 and WP5. The variation trend is shown in

Laws of Infiltration Line and Slope Deformation
After the first two days of rainfall, infiltration area appeared parallel to the shallow surface, and small gully appeared near the top area of slope under rainfall erosion. The invasion zone width of the bottom appeared greater than those of the slope body and the top area of the slope after the next-two-day rainfall completed. At the same time, the left ditch deepened and shallow sliding happened.
(a) before the rain (b) the first day of rain (c)the second day of rain

Model Parameters
In the analysis of this paper, the calculation formula of permeability coefficient of water in soil was cited to be defined based on the relationship between permeability coefficient and matric suction [8] : Abaqus was used for numerical simulation. The characteristic curves of hydraulic permeability coefficient and soil-water characteristic curves adopted are shown in Fig 12. The basic mechanical parameters of the slope are shown in Table 2.

Result Analysis
In the process of numerical simulation, the humidity results were almost the same as in model test, while the negative pore water pressure showed a greater difference from the measured values. The main reason is that the soil water characteristic curve is calculated by empirical formula in the numerical simulation, rather than referring to the practical data acquired from soil samples. For different types of rock and soil, the soil water characteristic curve is quite different, which leads to the huge difference of permeability coefficient in unsaturated zone. Therefore, this paper only makes qualitative comparative analysis on the experimental and numerical simulation results. Fig 13 showed the distribution of humidity and negative pore pressure inside the slope before rainfall. It can be seen that the distribution is completely consistent with the initial setting conditions because there is no influence of rainfall. Fig 14 showed the distribution of saturation and pore pressure in the slope body when the slope toe reach saturation. It can be seen that the front portion of the slope toe reach saturation firstly, with positive pore water pressure which is consistent with the model test results. The main reason is that in the process of rainfall, water accumulated near the slope toe, and the isolines of saturation and pore water pressure in the body are in arc shape and don't overlap. Fig. 15 showed the distribution of saturation and pore pressure in the slope body when the saturated zone appears at the top of the slope. It can be seen that when the foot of the slope is fully saturated, the saturated zone and positive pore water pressure begin to appear directly in the front portion of the top of the slope. With the development of rainfall, the saturated area and the negative pore water pressure area decrease gradually, and shrink to the inside in the form of arc. Fig 16 showed the distribution of saturation and pore pressure inside the slope when 320 points in the middle of the slope reached saturation. It can be seen that the saturated regions at the top of the slope are connected with that at the foot of the slope. The unsaturated region further decreases, and the positive pore water pressure in the saturated region at the top of the slope gradually increases and is larger than that at the foot of the slope.   17 showed the calculated results when the seepage inside the slope reaches a stable state under the effect of rainfall. It can be seen from figure (a) that there is a large positive pore water pressure in the shallow layer of the slope, and since the right boundary is the drainage boundary at this time, the pore water pressure at the top of the slope dissipates. Figure (b) shows that stress concentration occurs near the foot of the slope, which is prone to plastic deformation. Figure (c, d) shows respectively horizontal displacement and vertical displacement of the slope at stable seepage state. It can be seen that the peak horizontal displacement is less than half of the vertical displacement, and the maximum horizontal displacement occurs in the middle of the slope body, whereas the maximum vertical displacement in the front of the slope top. The displacement was produced mainly by the infiltration of rainwater which is in accordance with the model test results. With continuous rainfall, the pore pressure, Mises stress and displacement in the slope remained unchanged, and the rainwater infiltrated into the slope was discharged through the boundary on both sides.

CONCLUSION
In this paper, the variation law of volumetric moisture content, pore water pressure, matric suction and slope deformation of the top, the surface and the foot of the completely weathered phyllite slope were studied under the condition of continuous heavy rainfall by means of artificial rainfall model test.
(1) in this test, the volume moisture contents of the top, foot and slope body increased successively, while the matric suction decreased under continuous heavy rainfall. When the saturation state was reached, the highest volumetric water content was observed below the foot of the slope and the top of the slope.
(2) the persistent heavy rains caused the pore water pressure to increase obviously under the middle of the slope and the slope toe, and the saturation line moves from the surface to the slope body. Besides, slope surface erosion and shallow sliding failure showed that the completely weathered phyllite slope was prone to shallow sliding failure under the continuous heavy rainfall. In consequence, slope surface water erosion, waterproof and drainage measures in the body should be fully considered in this kind of slope control.
(3) the numerical simulation analysis results were basically consistent with the variation laws of the parameters measured in the test. As a result, numerical analysis can be fully used to simulate the variation laws of the completely weathered phyllite slope under various rainfall conditions, so as to optimize the model test.