Experimental study on comprehensive improvement of shear strength and erosion resistance of yellow mud in Qiang Village

2.2.1 Preparation of samples

The specimens were assigned into 17 groups, including plain soil control group (group A), 0.25% straw (group B1), 0.5% straw (group B2), 0.75% straw (group B3), 1% straw (group B4), 1% starch (group C1), 3% starch (group C2), 5% starch (group C3), 7% starch (group C4), 2% cement (group D1), 5% cement (group D2), 8% cement (group D3), 11% cement (group D4), 3% epoxy resin (group E1), 5% epoxy resin (group E2), 7% epoxy resin (group E3), as well as 11% epoxy resin (group E4). In accordance with the “Standards for Geotechnical Test Methods” (GB/T50123-2019), the round cake-shaped raw soil specimens with a diameter of 61.8 mm and a height of 20 mm were employed for direct shearing test, and there were a total number of 68 specimens. A mold with a size of 100 mm × 100 mm × 100 mm was used to make cube-shaped raw soil specimens for the anti-scour test, and the total number of specimens was 51. After the specimens were prepared, they were cured in a natural state (Figure 7). Table 3 lists the material ratio of the specimen, and Table 4 depicts the main equipment and instruments of the test.

Figure 7

Forming sample.

2.2.2 Direct shear test

The direct shear test was performed using a quadruple constant strain direct shear tester (Figure 8). In this test, a sample with a diameter of 61.8 mm and a height of 20 mm was applied, and the shear rate was set at 0.8 mm·min⁻¹. Under no lateral deformation, the vertical pressures including 100, 200, 300, and 400 kPa were applied to the 4 raw soil specimens, respectively. Before the test was performed, the balance weight of the direct shearing instrument should be adjusted to level the pressure lever, and Vaseline was applied evenly on the inner wall of the shearing box to reduce the effect of resistance. Before the test piece was installed, the upper and lower boxes of the shearing container should be aligned before the fixing pin was inserted. The permeable plate and filter paper were placed in the box. Subsequently, the test piece was gently placed in the box, and the filter paper and the permeable plate were placed on the test piece. Lastly, the pin was removed, and the cutting began. During the shearing process, the test block underwent shear failure along the horizontal direction of the shear box, thus resulting in displacement (Figure 9). Accordingly, the shear stress τ of the specimen on the shear plane when shear failure occurred, and the shear strength parameters, cohesion c and internal friction angle φ, were calculated according to Coulomb’s law.

Figure 8

Quadruple constant strain direct shear apparatus.

Figure 9

Shear failure of test block.

2.2.3 Anti-scour test

In this test, 100 mm × 100 mm × 100 mm cube-shaped raw soil specimens were used. To be specific, a total of 17 groups were involved, of which each was tested with 3 specimens. The average value was taken as the final value. During the test, a sprinkler head, a rubber hose, a pressure gauge, tap water, and others were used as a set of sprinkler devices (Figure 10). Based on the rainfall in Lixian County over the past few years, the parameters applied in this test to simulate rainfall included: the distance from the nozzle to the specimen was 30 cm, the diameter of the nozzle was 8 cm, there were 52 nozzle holes with a diameter of 1 mm, and the water pressure ranged from 0.05 to 0.07 MPa. The test pieces were put on the iron net one by one, and the nozzles were fixed to ensure that the water flow was sprayed down to the test pieces, the flow of the water was adjusted, and the water pressure was controlled. Natural precipitation was simulated to perform a spray test on the surface of the raw soil specimen, the spraying time and surface changes were observed, and the size of the specimen was measured every 10 min. Lastly, the volume loss rate of the specimen was calculated after 60 min, and the average value of the respective group was taken. On that basis, the anti-scour performance of the test piece in the spray test was examined.

Figure 10

Self-made sprinkler.

2.2.4 Microstructure

After the direct shear test and the anti-scour test were performed, the improvement effect of different contents of the respective material was analyzed, and the sample mixed with straw, starch, cement, and epoxy resin was selected as the test sample. It was cut into soil sample of 8 mm × 8 mm × 15 mm. Subsequently, it was sprayed with metal coating and observed by SEM (Thermo Fisher Scientific’s Prisma-E) (Figure 11).

Figure 11

Electron microscope scanner.

3.1.1 Analysis of shear strength index

Figure 12 presents the change in shear strength of specimens with different material dosages, and Figure 13 illustrates the shear strength index of specimens with different material dosages.

Figure 12

Change in shear strength of yellow mud with different material contents: (a) straw; (b) starch; (c) cement; and (d) epoxy resin.

Figure 13

Shear strength indexes of different materials: (a) straw; (b) starch; (c) cement; and (d) epoxy resin.

Figure 12a shows the strength change in the straw-modified yellow mud specimen. As depicted in Figure 12a, when the straw content was 0.5%, the shear strength of the specimen reached its peak value. Under different vertical pressures, the corresponding shear strengths were 76.8, 123.69, 176.6, and 225.4 kPa, 12.64–20% higher than those of the pure yellow mud specimens. As depicted in Figure 13a, the shear strength index of the specimen changed with the increase in the straw content. When the straw content was 0.5%, both the cohesion and internal friction angle of the specimen reached their peaks, including 25.95 kPa and 26.51°, respectively, 26.13 and 13.9% higher than those of the pure yellow mud specimen. This was due to the tie effect of the straw, which increased the cohesion of the specimen; the surface of the straw had irregular grooves, thus increasing the occlusal force between the soil and the material and increasing the internal friction angle of the soil. When the straw content was higher than 0.5%, the shear strength index of the specimens decreased. This was because the straw was hollow inside (SEM figure). When the doping amount was too high, the pores inside the straw were connected in the specimen, thus increasing the weak points inside the specimen and ultimately resulting in reduction in the shear strength, cohesion, and internal friction angle of the specimen. Thus, straw dosage is recommended to be controlled at 0.5%

Figure 12b shows strength changes of starch-modified yellow mud. As depicted in Figure 12b, when the starch content was 5%, the shear strength of the specimen reached the peak. Under different vertical pressures, the corresponding shear strength was 72.9, 141.5, 184.4, and 210.8 kPa, respectively, 9.83–37.28% higher than that of the pure yellow mud specimens. As depicted in Figure 13b, the shear strength index of the specimen increased first and then decreased with the increase in the straw content. When the straw content was 5%, the cohesion and internal friction angle of the specimen reached the peaks of 38.2 kPa and 24.5°, respectively, 85.71 and 5.29% higher than those of the pure yellow mud specimen. This was because the gelation of starch made the soil better bond together, thus improving the cohesion and internal friction angle of the specimen. When the dosage was too high, the shear strength index of the specimens decreased. This was because when the starch content was too much, the starch tended to cluster (SEM figure). The above starch groups increased the weak area inside the specimen, and the soil could not be wrapped with enough starch, thus resulting in reduction in the shear strength, cohesion, and internal friction angle of the specimens. Accordingly, the starch content is recommended to be controlled at 5%.

Figure 12c illustrates the strength changes in cement improved yellow mud. As depicted in Figure 12c, the shear strength of specimens increased with the increase in the cement content. When the vertical pressure was 400 kPa, the shear strength corresponding to different cement contents was 227.64, 301.4, 387.46, and 414.074 kPa, respectively, which increased by 18.58–115.74% compared with pure yellow mud specimens. As depicted in Figure 13c, the higher the cement content, the higher the shear strength index of the specimen would be. The cohesive force corresponding to different cement contents was 32, 36.05, 42.61, and 59.47 kPa, marking the increases by 55.57–189.11% compared with yellow mud samples. The internal friction angles corresponding to different cement contents were 27.02°, 34.22°, 39.69°, and 41.02°, 16.12–76.28% higher than those of yellow mud specimens. The reason was that after the yellow mud was mixed with cement, the fine cement particles and hydration products played a significant filling effect and decreased the number of pores. Moreover, the hydration of cement produced hydraulic gel products, which increased the cohesion between loose yellow mud particles, thus improving the shear strength of the samples. As revealed by the situation of Taoping Qiang Village in Li County, the newly built village over the past few years has completely replaced the yellow mud with cement, which has destroyed the style of the Qiang Village. Some scholars have shown that when the cement content is 11%, the raw soil material is stable enough [22,23]. Furthermore, according to “Code for Design of Masonry Structure” GB50003-2011, the shear strength standard of stone masonry mortar is 170 kPa. According to the local construction experience multiplied by the safety factor of 1.77, 300.9 kPa already satisfies the local strength requirements. When the cement content is 5%, the shear strength is 301.4 kPa, which has exceeded the standard requirement. Combined with the current market price of cement, the cost of different cement dosages has been rising (Figure 14). Based on the above two indicators, for the shear resistance of the specimen, the cement content is recommended to be controlled at 5%.

Figure 14

Influence of strength and cost on cement content.

Figure 12d illustrates the strength change in epoxy resin-modified yellow clay. As depicted in this figure, the shear strength first increased and then decreased with the increase in epoxy resin content; it reached a peak when the content of epoxy resin was 7%. Under different vertical pressures, the corresponding shear strengths were 74.5, 155.6, 202.6, and 233.6 kPa, marking the increases by 14.61–41.26%. As depicted in Figure 13d, the cohesion and internal friction angle of the specimens were consistent, and peaks were reached when the dosage was 7%, which were 32 kPa and 27.92°, respectively, marking the increases by 55.57 and 19.98%, respectively. Since the three-dimensional (3D) network structure formed by epoxy resin in the modification process wrapped the soil in it, the internal connection of the soil was more tight, and the network structure would increase the surface friction and occlusal force between the particles. Thus, with the increase in epoxy resin content, the cohesion and internal friction angle of the soil would be larger. When the dosage was 7%, the pores inside the specimen were rare, and the soil was well wrapped. Accordingly, when the dosage increased again, the shear strength indexes did not change much. Thus, the dosage of epoxy resin is recommended to be controlled at 7%.

3.1.2 Shear stress–shear displacement curve analysis

To analyze the deformation characteristics of the samples at different stages, the shear stress–shear displacement curves of 0.5% straw, 5% starch, 11% cement, and 7% epoxy resin were plotted for the specimens characterized by the maximum shear strength. The deformation of the sample could fall into three main stages, including elastic deformation, plastic deformation, and shear failure, in accordance with the change rule of the above curves. At the elastic deformation stage, the shear stress increased rapidly and linearly with the shear displacement. At this stage, the specimen was primarily in the elastic deformation, and the corresponding shear stress at the end of the curve was proportional strength τ _e. At the plastic deformation stage, the shear stress increased slowly, and the corresponding shear stress at the end of the curve was the shear strength τ _f. At the shear failure stage, the shear stress will reach a peak point, after which, as the shear strain continues to grow, the shear stress will form a descending section until the specimen is completely destroyed. Finally, there was a relatively stable residual shear strength τ _p. In this study, the end of the curve of shear stress and shear displacement with approximate linear growth was considered the end of the first stage, and the point corresponding to the shear strength displayed on the device of geotechnical test data acquisition system was recognized as the end of the second stage. Table 5 lists the shear displacements corresponding to the ends of the elastic and plastic stages.

Figure 15 presents the shear stress–shear displacement curve of pure yellow mud, belonging to the strain hardening type. Figure 15 and Table 5 depict that the elastic deformation stage of the pure yellow mud specimen under different vertical pressures ends when the shear displacement reached 1.17, 1.06, 1.66, and 2.19 mm, and the plastic deformation stage ended at 3.9, 4.01, 4, and 3.93 mm. By averaging the elastic deformation and plastic deformation under different vertical pressures, it was found that the corresponding elastic deformation and plastic deformation of the pure yellow mud specimen were 1.52 and 2.44 mm.

Figure 15

Shear stress–shear displacement curve of yellow mud.

Figure 16 illustrates the strength characteristic curve of yellow mud treated with different materials. As depicted in this figure, the curves of adding 0.5% straw, 5% starch, and 7% epoxy resin were the same as those of pure yellow mud. Figure 16c presents the strength characteristic curve of the specimen after the cement treatment, belonging to the strain softening type, thus revealing that the shear failure was brittle and had a significant peak value. After exceeding the peak value, the shear stress tended to decrease with the increase in the shear displacement, high residual strength was maintained, and the strength enhancement effect was obvious.

Figure 16

Shear stress–shear displacement curves of improved specimens of different materials: (a) straw; (b) starch; (c) cement; and (d) epoxy resin.

Table 5 lists the corresponding shear displacement at the end of the respective stage. The table indicates that the average elastic deformation of the test piece with 0.5% straw under different vertical pressures was obtained as 0.67 mm, 55.92% less than that of the pure yellow mud test piece, and the average plastic deformation was measured as 3.28 mm, 34.43% more than that of the pure yellow mud test piece. The average elastic deformation of the specimen with 5% starch under different vertical pressures was obtained as 1.63 mm, 7.24% higher than that of the pure yellow mud specimen, and the average plastic deformation was 2.39 mm, 12.45% lower than that of the pure yellow mud specimen. The average elastic deformation of the specimen with 11% cement under different vertical pressures measured 0.93 mm, 38.82% less than that of the pure yellow mud specimen, and the average plastic deformation was found as 1.94 mm, 28.93% less than that of the pure yellow mud specimen. The average elastic deformation of the specimen with 7% epoxy resin under different vertical pressures was found as 1.55 mm, 1.97% higher than that of the pure yellow clay specimen, and the average plastic deformation was obtained as 2.89 mm, 5.86% higher than that of the pure yellow clay specimen.

3.4.1 Qualitative description of SEM images

To explain the modification effect of yellow mud of Qiang Village mixed with straw, starch, cement, and epoxy resin from the micromorphology, the pure yellow mud and the samples mixed with different materials were analyzed by the SEM. The results are shown in Figure 20.

Figure 20

SEM images of specimens with different materials: (1) SEM of pure yellow mud; (2) SEM of straw-improved yellow mud; (3) SEM of starch-modified yellow mud; (4) SEM of cement-modified yellow mud; and (5) SEM of epoxy resin-modified yellow mud.

Figure 20(1) depicts the SEM figure of the yellow mud. As can be seen from the figure, the pores inside the pure yellow mud specimen were large and numerous, thus increasing the weak areas inside the specimen. Figure 20(1b) depicts that the soil mass of the specimen without materials was loose, and the soil particles were stacked together but not closely connected, which would be easy to be destroyed under the action of external forces. Accordingly, the strength of the pure yellow mud specimen was low, and the anti-erosion ability was poor.

Figure 20(2) depicts the SEM diagram of the sample mixed with straw. Straw acted as aggregate in the test piece and played a role of “reinforcement.” The unsmooth surface of straw laid good mechanical bite conditions between the straw and yellow mud, as presented in Figure 20(2a). There were small grooves and spikes on the surface of straw, and soil particles were well attached to the surface of straw, which suggested that straw could increase the internal friction of raw soil test piece and pull the soil, so as to improve the shear strength and scouring resistance of the yellow mud. However, due to the hollow shape of straw (Figure 20(2b)), if the amount of straw was too high, these hollow areas would pass through the interior of the test piece, thus increasing the weak points of the test piece. In the modification process, the soil will not be closely combined, the moisture in the specimen will make the straw rotten and moldy in the curing process, resulting in a soft filling medium and weakened mechanical properties of the specimen, so the amount of straw should not be too high.

Figure 20(3) depicts the SEM diagram of starch-modified yellow mud. As indicated by Figure 20(3a), starch could be bonded into a film to wrap the soil and bond it together, which played a certain cementation role and made the soil bond more closely, so as to improve the strength and anti-scouring ability of the soil. However, when the starch content was too high, the starch would bond into a mass (Figure 20(3b)). The more these starch clusters inside the specimen, the easier it would be to destroy the specimen when receiving external force. Thus, the amount of starch should not be too high.

Figure 20(4) depicts the SEM figure of cement-modified yellow mud. As depicted in Figure 20(4b), hydration reaction and hydrolysis reaction would occur during cement modification of the yellow mud, and fibrous calcium silicate hydrate, CAH gel, calcium hydroxide, needle-like ettringite, and other substances would be generated. As depicted in Figure 20(4a), the pores of the specimens were significantly reduced, since these cementing substances and hydration products would adhere to the inner wall and surface of the pores between the soil particles, and also condense to form soil skeleton and reduce the pores between the soil particles. When the calcium, sodium, magnesium, and aluminum ions reacted with the ions adsorbed on the surface of soil particles, the ion exchange reaction could reduce the thickness of water film on soil particles, facilitate the coagulation of soil particles, enhance the strength of soil, and also improve the anti-scouring ability of soil to a certain extent.

Figure 20(5) depicts the SEM diagram of yellow mud modified by epoxy resin. As depicted in Figure 20(5a), the pores of specimens were significantly smaller than those of pure yellow mud. This was because epoxy resin was stable and had strong adhesion. Epoxy resin E-51 did not damage the soil. When it was mixed with curing agent, a cross-linking reaction occurred, as presented in Figure 20(5b). The 3D network structure generated by the cross-linking reaction connects the soil and forms a new skeleton to fill the void of the soil, thus improving the cohesion of the soil and enhancing its shear strength and anti-erosion ability. As indicated by Figure 20(5a), when epoxy resin content was 7%, the soil mass and epoxy resin were well combined and tended to be saturated. Accordingly, when the dosage was higher than 7%, the modification effect of epoxy resin on soil is no longer improved.

3.4.2 Quantitative analysis of PCAS operation

PCAS software was used to analyze the 500 times magnified SEM images of the specimens mixed with different materials. This software binarizes the images and obtains the binarized images and particle recognition results primarily based on the threshold segmentation method (Figure 21). After processing, the software automatically segmented and identified pores and particles, while eliminating impurities. After image processing, geometric parameters were automatically output, which consisted of porosity, area, perimeter, shape coefficient, probability entropy, and fractal dimension. The specimens with and without materials were quantitatively compared and analyzed to explain the structural characteristics of the specimens with different materials and the effects of shear strength and scouring resistance from a microscopic point of view. It is noteworthy that the length and area obtained by PCAS software were in pixels and should be converted into real parameters through equations (2) and (3).

where S _t and C _t represent the actual area and actual perimeter; S and C represent the pixel area and pixel perimeter; R is the resolution of the image (in this study, the sub-change in the 500 times magnified image is 0.54 μm per pixel)

Figure 21

SEM images of specimens with different modified materials: (a) binarization graphics and particle recognition results of pure yellow mud specimens; (b) binary graphics and particle recognition results of specimens with straw; (c) binary graphics and particle recognition results of specimens with starch; (d) binary graphics and particle recognition results of specimens with cement; and (e) binary graphics and particle recognition results of specimens with epoxy resin.

In accordance with the classification criteria in Table 6, the number of pores and pore area of samples of different materials were statistically analyzed in the histogram. Figure 22 presents the statistical results. As depicted in Figure 22a, when the modified material was added to the yellow mud, the total number of macropores of the sample increased or changed insignificantly, and the micro pores increased. The modified yellow mud specimen primarily had micro pores. The total pore areas of the pure yellow mud specimen, the straw specimen, the starch specimen, the cement specimen, and epoxy resin specimen were 50553.5, 31034.6, 33209.2, 37866.3, and 39029.6 μm². As depicted in Figure 22b, the areas of micropores, small pores, and medium pores of the modified yellow mud test block were increasing, while the area of large pores was found as the largest of the pure yellow mud test piece. The pore area of the specimen was primarily dependent on the area of large and medium pores. Pure yellow mud specimens had large pore areas and low connection strength between the particles. When rainwater was immersed in the soil, or the soil was sheared by external force, the soil particles fell into the pores and would be more likely to be damaged. The pore area of the specimens with straw, starch, cement, and epoxy resin was relatively low and stable as compared with that of the yellow mud, so the shear strength and scouring resistance were found to be strong.

Figure 22

Pore area distribution of different materials: (a) pore number distribution and (b) pore area distribution.

Probability entropy is capable of characterizing the orientation of particles and pores. When the value approached 1, the order of the specimen was worse, and the soil was more stable. The separation coefficient represents the uniformity of the particles. The larger the value, the worse the separation would be, the smaller the porosity and the more stable the soil would be. The average shape coefficient of particles represents the difference between the particle shape and sphere. The larger the fractal dimension, the more complex and stable the soil shape would be [39,40,41].

As depicted in Figure 23a, when the modified material was added to the yellow mud, the probability entropy of the specimen increased and approached 1 compared with the pure yellow mud, thus indicating that the pores of the modified specimen were randomly arranged. The order was poor, and the soil was more stable. Compared with pure yellow mud, the sorting coefficient of the samples with materials increased, which revealed that the modified yellow mud significantly changed the particle shape and poor sorting performance due to the binding effect of straw, the gelation of starch, the hydration and hydrolysate of cement, and the 3D network structure generated by the cross-linking reaction of epoxy resin. After starch, cement, and epoxy resin were added, the fractal dimension of the specimen increased, which revealed that soil particles in the specimen had a complicated morphology. There was little difference between the fractal dimension of straw and pure clay, since there was no significant difference between straw and pure clay.

Figure 23

Parameter statistics: (a) fractal dimension, sorting factor, probability entropy; (b) porosity; average circumference; average area.

The average pore area and average pore perimeter could reflect the pore size. The larger the value, the more the macropores would account for. Porosity is a vital parameter affecting water flow in pores. As depicted in Figure 23b, the average pore area of the test piece added with straw, starch, cement, and epoxy resin decreased by 52.08, 41.36, 32.92, and 40.42% compared with those of the yellow mud test piece, the average perimeter was reduced by 14.56, 3.05, 5.23, and 11.07% as compared with that of the yellow mud test piece, and the porosity decreased by 35.14, 35.05, 24.12, and 24.28% compared with that of the yellow mud test piece. The three indexes of the pure yellow mud specimen were found to be large, which suggested that the pore size was large, the proportion of macropores was large, the soil structure was loose, and it would be easy to be damaged under the action of external force. In contrast, the specimen with material exhibited low porosity, inhibited the flow of water in the pores, and improved its shear strength and scouring resistance.