Experimental Study on Destruction Mode and Influence Factors of the Gridded Hard Crust Using Transparent Soil

Abstract

In this study, 12 groups of plane strain model tests of gridded hard crust with different cement contents and structures were conducted with a transparent soil experimental technique. The destruction mode and influence factors in the ultimate state were investigated by analyzing the law of soil displacement and ultimate load change around the structure. The test results revealed that the destruction mode of gridded hard crust under 8% cement content was mainly the destruction of the upper hard crust. Under the condition of small spacing, the grid structure was destructed when the thickness of the hard crust increased. The destruction mode of the gridded hard crust was majorly the destruction of the lower grid structure when the cement content was 15%, and the thinner hard crust was destroyed when the space between grid structures enlarged.

1. Introduction

In China, marine silt is widely distributed in coastal areas, and its physical and engineering properties are close to those of soft clay. It is difficult to use in engineering due to characteristics such as high water content, high compressibility, low shear strength, and high sensitivity. Therefore, marine silt must be treated to meet engineering needs [1]. Since marine silt contains humus and many kinds of ions, it is frequently added to cement and other curing agents for treatment [2,3]. In situ solidification technology can quickly form an artificial hard crust of a certain strength and thickness on the surface of soft soil by adding a curing agent to in situ soft soil and stirring it evenly [4]. Currently, in situ solidification mostly adopts a single and equal thickness solidification method, which is inflexible, for soft foundation treatment. Meanwhile, its bearing capacity may be much higher compared to the actual needs, resulting in a waste of resources. On the basis of in situ solidification, the gridded hard crust with different thicknesses is formed considering that the project should solidify the soil with different depths according to local conditions (Figure 1). Then, it is utilized to lower the project cost, presenting a good application prospect. Therefore, it is of great significance to investigate the destruction mode and influence factors of the gridded hard crust for its popularization and application. Gridded hard crust is mainly used for shallow in situ curing, where the loads are transferred to the bottom bearing layer. This type of hard-shell layer is known as a gridded hard crust because of the gridded structure it forms after construction and depending on the shape the structure is divided into two parts: the hard crust and the grid structure. The addition of this special structure to a soil that is inherently multi-phase and complex makes the interaction between them extremely complex and the deformation and damage limit states are highly uncertain.

Physical simulation test technology takes the similarity principle as a basis. Through the establishment of an effective physical model, the prototype has been deeply studied, has been widely used in many fields of geotechnical engineering, and has become one of the important means of scientific research and technical problem-solving. However, the internal deformation law of soil for traditionally common similar materials cannot be directly observed as its interior is invisible. Transparent soil material is a kind of transparent material similar to soil. This is a general name for synthetic transparent simulation materials with natural soil engineering properties. They are composed of aggregate and pore fluid with the same or similar refractive indices. Owing to the similar refractive indices of aggregate and pore fluid, there is no significant refraction or scattering when light passes through transparent soil. Hence, this kind of “soil” is transparent, and the appearance of similar materials of transparent soil makes it possible to “directly observe” the internal deformation of the model [5,6]. At present, it has been widely adopted in foundation engineering, tunnel engineering, foundation pit engineering, slope engineering, and other related scientific and technological fields [7,8,9,10,11]. On the basis of verifying that transparent clay can be applied to the experimental study of shield tunnel excavation face instability, Lei [12] analyzed the evolution process of excavation face instability under different buried depth ratios from the perspective of soil movement and studied the characteristics of instability stages, zoning, and morphology, as well as the change law of surface settlement. With transparent sand and PIV technology, Wang [13] performed a series of centrifugal model tests on anchor piles under the oblique uplift load and explored the relationship between the soil displacement field and pile foundation bearing capacity. Xu [14] discussed the influence of shallow neck on the vertical bearing capacity of a single pile with a cap by transparent soil test technology and explored the reasons for the loss of the vertical bearing capacity of the shallow neck pile using the displacement field of the soil around the pile. In recent years, similar materials of transparent soil have been constantly updated, and materials with closer properties and better transparency are being continuously developed [15,16,17]. The transparent clay made of Carbopol^® Ultrez10 polymer (U10), a new material studied by Kong [18,19], demonstrates the properties of medium- and low-sensitivity clay and presents noticeably reinforcing strength over time. Its strength and compression consolidation characteristics are similar to those of marine silt, which is broadly employed in studies related to the interaction between structure and marine silt.

Image processing and results analysis are the key to physical simulation test technology for transparent soils and are an effective means of obtaining deformation data within the model [20,21]. Currently, the PIV technique is used to analyze the results of the experimental images. The image is segmented into numerous regions with unique features to extract the desired information. The image feature obtained from the transparent soil test is the grey scale. The basic principle of image correlation matching is to calculate the correlation between the grey scale values of an image by means of a computer program, which in turn allows the velocity and displacement of the specific point required to be obtained. The process is as follows: a cell is selected for analysis at time t₀ and a corresponding area larger than the cell size is selected for retrieval at time t₁. The cell with the largest coefficient is determined to be the best match for the cell at time t₀, and the coordinates of the cell are the position after displacement at time t₀. The image correlation function is shown in Equation (1).

$$C\left(\Delta x,\Delta y\right)={{\displaystyle \iint }}_A{I}_0\left(x,y\right)I_1\left(x+\Delta x,y+\Delta y\right)dxdy$$

(1)

$C$ is the correlation coefficient. $x,y$ are the pixel coordinates of the image. $\Delta x,\Delta y$ is the displacement of the comparison image. $I_0,I_1$ is the grey scale value of the pixels of the image and the comparison image, respectively.

Wang [22] combined the PIV technique with two-dimensional model tests to investigate the damage behavior of strip foundations of geosynthetically reinforced granular soil foundations at the mesoscale and obtained three damage modes. Kwak [23] captures digital images through the transparent sidewalls for particle image velocimetry (PIV) analysis and to assess the displacement and deformation of soil particles. Sun [24] uses particle image velocimetry (PIV) and 32 cluster strain gauges to monitor the deformation of the tunnel structure and the landslide soil and sliding surface, respectively, and discusses in detail the effect of cyclic loading on the mechanical behavior and displacement of the tunnel and sliding surface.

A model test of gridded hard crust formed by silt solidification was conducted in this study to explore the destruction mode of gridded hard crust under uniform load and the influence of cement content and structure on it. Besides, the displacement field diagram of soil and the load-displacement curve of the model were obtained based on transparent soil technology. Following the deformation of soil, the destruction mode of the gridded hard crust was obtained, and the reasons for the influence of the cement content and structure on its destruction mode were analyzed. The results of the transparent soil test were verified by the law of displacement-load curve. The results of this study lay a foundation for the application of gridded hard crust and the subsequent theoretical derivation of ultimate bearing capacity.

2. Materials and Methods

2.1. Parameters and Preparation of Transparent Soil Materials

The transparent soil in this experiment was made of Carbopol^® Ultrez10 polymer (U10). The U10 was white powder at room temperature, as illustrated in Figure 2a. The particle size ranged from 50 nm to 100 nm, the apparent density was 0.21 g/cm, and the relative density was 1.4543. Since the transparent soil prepared by U10 was transparent and uniform, it could not produce a speckle field under laser irradiation. Thus, titanium oxide was used as the tracking formula in the experiment. Titanium oxide has stable chemical properties and does not react with U10. Moreover, its particle diameter was 5 μm, demonstrating a white powder shape, as exhibited in Figure 2b. First, the U10 powder was slowly poured into purified water at 70 °C, stirred with a low-power stirrer for 30–40 min, and then sealed and left to stand for 6–8 h. Afterward, the NaOH solution was poured into the mixed solution containing U10 at the ratio of U10: NaOH: water =1: 0.4: 98.6 and stirred with a high-power stirrer for 30–40 min until it was uniform and viscous. Second, titanium oxide was added according to 0.2 ‰ of the mass of transparent soil for stirring. Subsequently, the transparent soil was vacuum-degassed for 6–8 h. Finally, the transparent soil was put into a clean plastic sealed bucket for later use.

2.2. Model Materials and Design

Currently, there are some studies on cemented transparent soil [25,26]. Nonetheless, cemented transparent soil and cement soil are different, and the sample preparation is complicated. Therefore, in the model of this test, washed kaolin was taken as the curing object [27], and 42.5 ordinary Portland cement was regarded as the curing agent. The plastic limit of kaolin was 33%, the liquid limit was 60%, and the pH value was 4.14. Soft soil with high water content can be prepared by mixing purified water with the kaolin liquid limit. The gridded hard crust model cannot be cast in transparent soil and hence was prefabricated in advance. Then, a mold made of 6 mm thick PVC cardboard splicing was designed to establish the model. Simultaneously, purified water was first added with 60% dry soil mass to kaolin, stirred into a flowing state, and then added with a cement of 8% and 15% wet soil mass, stirred evenly, and poured into a mold. After curing indoors for 7 days, all templates were removed, and finally the model was formed, as exhibited in Figure 3.

For illustration purposes, the dimension symbol of the gridded hard crust is defined in Figure 4. A model with the ratio of hard crust thickness (t) to grid structure height (h) of 0.5 and the ratio of grid structure spacing (s) to grid structure width (d) of 2 was established to explore the destruction mode of gridded hard crust based on previous engineering experience. Besides, 2 cases were considered when the cement content was designed.

First, when the cement content was small (namely the modulus of the gridded hard crust was low), the deformation of the hard crust was large and coordinated with the deformation of the foundation, and the soil exerted more bearing capacity. Second, when the cement content was large (namely, when the modulus of the gridded hard crust was large), the deformation of the hard crust was small, the soil did not exert its bearing capacity, and the hard crust was destroyed. Then, two kinds of cement content (8% and 15%) were designed. Control groups with t/h = 0.5, 1, and s/d = 1 were placed to research the influence of structural differences on destruction modes. The test scheme is presented in

Table 1. Test scheme.

Model Type	Cement Content	t/h	s/d
SN-8	8%	0.5	1
SP-8	8%	1	1
MN-8	8%	0.5	2
MP-8	8%	1	2
LN-8	8%	0.5	3
LP-8	8%	1	3
SN-15	15%	0.5	1
SP-15	15%	1	1
MN-15	15%	0.5	2
MP-15	15%	1	2
LN-15	15%	0.5	3
LP-15	15%	1	3

. Among them, S, M, and L represent the ratio of s to d as 1, 2, and 3, respectively; N and P denote the ratio of t to h of 0.5 and 1, respectively; 8 and 15 indicate cement content. The width of the model is aligned with the model box at 20 cm.

2.3. Test Equipment and Design

Figure 5 illustrates transparent soil model test equipment composed of a loading system, CCD industrial camera, laser, model box, optical platform, and computer. The loading system was a microcomputer control electron universal testing machine WDW-100D, the measuring force range was 0–100 kN, and the displacement measurement accuracy was 0.01 mm. Additionally, the accuracy of force measurement was 0.03 F.S., the resolution of the CCD industrial camera was 1280 × 960, and the maximum frame rate was 40 fps. The laser model was YN-532-3W, the wavelength was 532 nm, the maximum power was 5 w, and the working distance was 1 m. The length, width, and height of the model box were 30 cm, 20 cm, and 15 cm, respectively, and the box was bonded by acrylic plates with a thickness of 15 mm.

The gridded hard crust model studied in this experiment was different from the previous study of piles or other small components. In the past, the model box could be directly placed on a professional optical test platform, and the laser was irradiated from the side to form a speckle field for testing. However, the laser could not penetrate the transparent soil at the interval of the grid structure in this test owing to there being no gap between the side of the model and the bottom of the model box. Thus, an optical test table was designed for testing. The optical test table consisted of welded steel plates, and the test table was provided with a square hole with a width of 20 cm in the vertical direction facing the load plate. In this way, the fan-shaped laser surface generated by the laser could pass through the model box from bottom to top.

Before the test, the transparent soil was put into the model box, and the model box with transparent soil was put into a vacuum bucket for vacuum degassing for 2–3 h. This step aimed to discharge bubbles caused by disturbance during sample loading. After the air pumping was finished, the model was embedded into the transparent soil of the model box according to the calibrated position. Subsequently, the transparent soil was added to the positions on both sides of the model and put into the vacuum bucket again for air pumping for 2–3 h. After standing for 12 h, the model box was placed at the center position of the desktop of the optical platform aligned with the gap opening of the desktop. A laser emitter was placed at the bottom of the model box to make the laser spot, and the position of the laser was adjusted to ensure that the sheet laser emitted by the laser faced the middle of the transparent model box. Besides, the industrial camera was installed on the tripod and fixed in front of the model box. It was focused manually, and the position of the camera was constantly adjusted until a clear speckle surface could be photographed. The image acquisition software was debugged and operated to ensure that the pictures could be collected normally and continuously. The photo storage format was BMP. Normal and continuous data acquisition was performed by data acquisition software. Before loading, three load plates with lengths of 10.5 cm, 15 cm, and 19.5 cm were prepared following models of different sizes. Each load plate was 20 cm in width and 2 cm in thickness, and threaded holes were drilled in the middle to connect with the loading equipment. The concentrated load applied by the universal testing machine was transformed into a uniform load through the load plate. Furthermore, the loading speed was set to 0.5 mm/min, so as to achieve a more harmonious deformation of the model in the test.

3. Results

3.1. Displacement Vectors and Contour Diagram

With the load-displacement diagram, the displacement of the model during destruction was obtained, and the corresponding image was found, processed, and analyzed by PIVview2C software. Considering that the load and structure were symmetrical, half of the image (Left side) was selected for analysis. The vector diagram and displacement nephogram of the soil are illustrated in Figure 6, Figure 7, Figure 8, Figure 9, Figure 10, Figure 11, Figure 12, Figure 13, Figure 14 and Figure 15.

The soil displacement of the MN-8 model was mainly concentrated on the bottom (area 1 and area 2) of the first and second spans and the outer side (area 3) of the hard crust when the cement content was 8% (Figure 6). Meanwhile, the vertical downward displacement of the soil in area 2 increased. In other words, the hard crust was destroyed and extruded the soil downward, contributing to the production of a large displacement downward and a large number of upward displacements. The occurrence of massive soil displacements in Area 3 demonstrated that the connection between the hard crust and the grid structure was destroyed, and the connection extruded the soil outward when subjected to vertical load, enabling the soil to displace upward. The X-direction displacement contour diagram (Figure 7a) revealed that the horizontal displacement of soil was only large at area 1 and area 2, with a maximum of about 0.7 mm. Nonetheless, the displacement near the grid structure slightly changed, with a maximum displacement of only 0.13 mm, indicating that the deformation of the grid structure was small during destruction. In the Y-direction contour diagram (Figure 7b), the downward displacement at the bottom of the hard crust increased significantly to 1.4 mm. This suggested that the soil under the hard crust bore part of the load when the structure was destroyed. The displacement vector diagram of the MP-8 model soil with the increasing thickness of the hard crust is exhibited in Figure 8. Different from the MN-8 model, there was no large displacement of soil in area 3, reflecting that the connection between the hard crust and grid structure was relatively complete during destruction. Nevertheless, a large amount of downward displacement of soil was observed in area 1, implying that the destruction point was still in the hard crust when the structure reached the ultimate bearing capacity. Besides, the distribution law of the contour diagram (Figure 9) was consistent with the MN-8 model.

The displacement vector diagram of other structures during destruction in Figure 10 revealed that the soil displacement was still mainly concentrated at the bottom of the hard crust, while the soil displacement near the grid structure was small. When the load was applied to the structure, it was similar to the beam-slab structure, and the hard crust bore tensile force at the bottom during bending. Since the cohesion of cement soil was poor when the cement content was small, its tensile capacity was weak, and it was easier to destroy when stretched. Therefore, in the limit state, the hard crust with weak flexural capacity caused the whole structure to be destroyed during bending. In practice, when the cement content is small, it is recommended to add fibers and other materials to the bottom of the hard crust to improve its tensile strength and increase the overall load bearing capacity of the gridded hard crust.

Under the same structure, the position of the soil displacement concentration changed with the increase in the cement content. Figure 11 reflects that the soil of the MN-15 model was mainly distributed on both sides of the bottom of the grid structure (area 4 and area 5), except for area 3, and both area 4 and area 5 were oblique upward displacements. In the X-direction contour diagram (Figure 12a), a large area of lateral displacement appeared only at the bottom of the outer grid structure, and the maximum displacement was about 3 mm. This uncovered that the grid structure possessed a large lateral deformation. In the Y-direction contour diagram (Figure 12b), the main displacement changes occurred around the outer grid structure with an upward direction. The maximum was about 2.8 mm, implying that the bottom of the grid structure was damaged and deformed upwards by a downward load. As a result, the surrounding soil moved. The maximum downward displacement at the bottom of the hard crust was only 0.2 mm. Thus, the soil bore less load when it was damaged. As h increased, the larger soil displacement only occurred at area 4 and area 5 (Figure 13 and Figure 14).

It can be observed in Figure 15 that the soil displacement law of other structures similar to the former was obtained. With the increase in cement content, the flexural capacity of the upper hard crust was enhanced, enabling the structure to keep intact before the whole destruction. With the increasing load, the pressure on the grid structure gradually increased. The material destruction emerged when the shear strength of the cement soil was reached. Therefore, under the premise of meeting the bending resistance achieved by the hard crust, its cement content can be appropriately reduced and the cement content of the grid structure can be increased to enhance the overall bearing capacity of the structure while making reasonable use of resources. In Figure 15c, the hard crust was also destroyed because the modulus of the structure and soil was relatively large, and the gridded hard crust bore most of the load. Nevertheless, the increase in s made the bending moment of the hard crust increase, and the hard crust with smaller t was destroyed by bending when the structure reached the limit state. Therefore, when this type of structure is chosen for construction, the flexural resistance of the hard crust needs to be improved by, for example, upgrading the cement content or adding fibers.

3.2. Load-Displacement Curve

Figure 16 presents the load-displacement curve of the gridded hard crust structure. Considering that the destruction mode of the gridded hard crust structure under the limit state was investigated in this study, the ultimate load of the structure, namely, the peak load, was emphasized in the load-displacement diagram. The relationship between the ultimate load and cement content and model construction is demonstrated in Figure 17. As observed in the figure, the cement content exerted a great influence on the law of destruction load change. Under the same thickness of hard crust, the ultimate load increased with the increase in s when the cement content was low. At t/h = 0.5, the ultimate loads of MN-8 and LN-8 increased by 24.1% and 55.7% compared to SN-8, and at t/h = 1, the ultimate loads of MP-8 and LP-8 increased by 12.5% and 61.8% compared to SP-8, respectively. However, the opposite situation appeared when the cement content was high. At t/h = 0.5, MN-15 and LN-15 each lost 27.9% and 53.3% of their ultimate load compared to SN-8; at t/h = 1; MP-15 and LP-15 each lost 21.7% and 30.5% of their ultimate load compared to SP-15.

Different cement contents obtained different ultimate load variation laws corresponding to destruction modes. Less cement content led to less structural stiffness. Therefore, when the load was applied, the structure and soil bore the load together, and the destruction of the structure appeared at the hard crust. As s increased, the contact area between the hard crust and the lower soil enlarged, and the soil bore more load. Therefore, the destruction load increased with the increase in s/d in the load-displacement diagram of the 8% cement content. The destruction mode with large cement content was majorly the destruction of the grid structure. At this time, the ultimate load mainly depended on the strength of the grid structure. As s increased, the load area of the hard crust enlarged, and the load transferred to a single bottom grid structure also grew. Therefore, the structure with the larger s was more likely to be destroyed compared with the structure with the smaller s. Hence, the ultimate load in the displacement-load diagram with 15% cement content decreased with the increase in s/d. The increase in t has a positive effect on the increase in ultimate load, especially at high cement admixtures. When the cement content is 8% and t/h is increased to 1, the maximum increase in the ultimate load of the model is only 39.1%, whereas when the cement content is 15% and increased to the same thickness, the maximum increase in the ultimate load of the model reaches 67.2%. On the one hand, the increase in t changes the distribution of stresses and increases the flexural stiffness of the hard crust, thus increasing the load-carrying capacity of the low cementitious structure; on the other hand, as the flexural stiffness of the hard crust increases, more loads are assumed and the loads distributed in the lower grid structure will be reduced; thus, the ultimate load-carrying capacity of the high cementitious structure controlled by the strength of the grid increases with the increase in t.

4. Conclusions

In this paper, the soil displacement around 12 groups of gridded hard crust models with different cement content and structure were analyzed through the transparent soil test. The conclusions are drawn as follows.

Since the cement content impacted the tensile strength of the cement soil and then the flexural stiffness of the hard crust, there were two main destruction modes of the gridded hard crust in the limit state: the destruction of the upper hard crust and the destruction of the lower grid structure. The destruction of the upper hard crust mostly occurred in the working condition of low cement content, and the soil displacement in the lower part of the hard crust was significant at this time. The destruction of the lower grid structure mostly emerged in the working condition of high cement content, and the displacement and deformation of the soil around the bottom of the grid structure were noticeable.

The cement content was the main factor influencing the deformation pattern. First, the cement content influenced the tensile strength of the cement soil and thus determined whether the hard crust was damaged in tension before the ultimate load was reached. The span of the grid structure exerted less influence on the deformation pattern. The thinner thickness of the hard crust was more likely to be damaged by tension when the soil bore less load and the main force was the gridded hard crust. Meanwhile, it had no effect on the change in damage mode.