Study on the interaction mechanism between the contra-rotational shear deep soil mixing drilling tool and the soil

Deep soil mixing piles are established for soft soil stabilization, yet the internal mechanism by which drilling tools achieve complete soil-cement homogenization remains unclear, hindering precise quality control in column formation. In this paper, a new type of contra-rotational shear deep soil mixing (CS-DSM) drilling tool was proposed and the detailed design information was thoroughly introduced. The discrete element modeling method was applied to establish a coupling model of CS-DSM drilling tool and soil verified by a laboratory model experiment. Soil particle moving path analysis quantified mixing dynamics during drilling-down and drilling-up phases. The results explained the internal mechanism that the soil particles could be fully sheared, stirred, and mixed in the horizontal and vertical directions to form a uniform soil cement column. The average VC of particle X1-X3 was 1.9 times larger than that of particles Y1-Y3, which indicated that soil particles along the x-axis and y-axis exhibited distinct moving patterns. The mixing efficiency during drilling-down and drilling-up was quite different, and the VC during the drilling-down process could be 146.7 times larger than that during the drilling-up process. In practice, reducing drilling speed or increasing rotational velocity during drilling-down enhances CS-DSM pile quality under identical conditions.

1 Introduction

The DSM pile technology was first proposed and applied to the treatment of soft soil foundations by the American company Intrusion Prepakt. Subsequently, it had been widely used in engineering fields, such as foundation reinforcement and foundation pit support. During this period, the mixing pile technology had undergone multiple updates and iterations, resulting in three technical systems: DSM piles, double deep mixing (DDM) piles, and CS-DSM piles. Among them, DSM piles (Wang et al., 2025; Zeng et al., 2024) mainly referred to traditional single-direction mixing piles. DDM piles mainly referred to mixing piles constructed using the double-direction mixing drilling tools developed by Raito, Inc. in Japan and Liu et al. (Cai et al., 2012; Yi et al., 2019) in China. This drilling tool could achieve single-layer shearing at the inner and outer blades adjacent to the inner and outer drilling rods. The CS-DSM piles mainly referred to mixing piles constructed using the multi-layer mutual shearing mixing drilling tools and the CS-DSM construction method proposed by Aoya-ma Kiko (Kizuki et al., 2018) in Japan and Zhejiang Kunde Innovative Geotechnical Engineering Co., Ltd. (Ge et al., 2024a; Ge et al., 2024b) in China. The drilling tool adopted a coaxial double-layer drilling pipe structure, which allowed the inner and outer drilling rods to rotate in opposite directions.

Research on DSM piles mainly focuses on areas such as drilling tools, solidifying materials, pile structure form, and construction techniques. In terms of drilling tools, Cai et al. (2012), Yi et al. (2019) proposed a bidirectional mixing drilling tool and the DDM construction technique. They conducted on-site and laboratory tests and compared the results with those of traditional DSM piles. The results showed that the engineering properties of DDM piles were significantly superior to the DSM piles, such as the bearing capacity. Ge et al. (2024a), Ge et al. (2024b) carried out a series of model tests on CS-DSM piles and compared the apparent uniformity and pile strength between traditional DSM piles and CS-DSM piles. The results indicated that, when the number of soil shearing times per unit height was the same, the strength of the CS-DSM pile was 1.4–6.4 times that of the DSM pile. The CS-DSM pile strength was relatively high as the rotational speed ratio of the inner and outer drilling rod was between 1.8 and 2.2. Mosadegh et al. (2017) conducted laboratory tests on cement-soil specimens utilizing two distinct types of mixing equipment and three curing conditions (humid environment, water immersion, and sealed containers with varying salt concentrations), concluding that both mixing methods and curing conditions significantly influenced the strength of cement-soil. Masaki (Kitazume, 2021) studied the phenomenon of blade clogging during mixing and founded that spraying cement slurry and adding dispersants could improve the fluidity of cement soil, thereby reducing the blade clogging issue. For solidifying materials, Ekmen et al. (2020) studied the strength and stiffness variations of DSM pile materials treated with different amounts of cement (5%, 10%, 15%, and 20% by dry soil weight) and F-class fly ash (10%, 20%, and 30% by cement weight) as binders. Using unconfined compression strength (UCS) tests after curing of 28 and 90 days, they concluded that for target UCS values of 1.0 MPa and 2.1 MPa, the optimum secant elastic moduli range was 425–435 MPa and 846–863 MPa, respectively. Yin et al. (2024) conducted experimental research on the load-bearing performance of carbonated seawater cement soil mixing piles, finding that after carbonation, the 28-day strength of the cement soil increased by 6.8%. Claver et al. (2024) developed an alternative alkali-activated cement (AAC) made from ladle slag precursor mixed with sodium hydroxide and sodium silicate to enhance the load-bearing capacity of estuarine soil through DSM piles. The results showed that the AAC-stabilized piles had their flexural and compressive strengths nearly double after 28 days, and the AAC mixtures exhibited a higher stiffness increase rate. Arul et al. (2018) used fly ash (FA) and slag (S) as alternative green binders and compared their performance with traditional cement and lime control binders. The results showed that using the new FA + S binders significantly improved the strength and stiffness of soft clay, confirming that it could serve as alternatives to traditional cement or lime binders.

For pile structure form, Shen et al. (2024) employed the finite element method to simulate the behavior of single pile foundations in reinforced soft soil through the DSM pile technique. The factors influencing the reinforcement efficiency of this method were systematically investigated, including reinforcement width, reinforcement depth, and layout configuration. Based on the simulation results, n design approach termed the equivalent pile diameter method was proposed specifically for DSM piles. Zhang et al. (2021) investigated the seismic response of stiffened deep cement mixing (SDCM) piles through shaking table tests and finite element simulations. Zhao et al. (2025) utilized Abaqus to perform full-scale modeling, revealing the influence of bidirectional mixing section parameters on the load response and damage patterns of the bidirectional helix-stiffened soil mixing piles. Zhuang et al. (2022) developed empirical formulas to calculate the uplift bearing capacity of helix-stiffened soil mixing piles under different damage types. For construction techniques, Nikbakhtan and Osanloo (2009) investigated the effects of grouting pressure and grout flow on the UCS of piles using jet grouting. Based on the numerical results from the experiments, increasing the grout pressure and flow rate leads to a logarithmic increase in the UCS of the soil. Additionally, jet grouting significantly enhances properties such as cohesion and friction angle. Nikbakhtan and Ahangari (2010) proposed a new relationship between the uniaxial compressive strength of soil-cement piles and their diameter, along with lift speed, rotation speed, ratio between the cement and water, and grouting pressure. Han et al. (2022) performed two series of laboratory model tests on the compaction and expansion effects induced by the penetration of the core pile during the construction of the SDCM pile. The results revealed that during the core-pile penetration, the excess pore-water pressure of the surrounding soil rises until the end of the core pile reached the corresponding depth. The influence radius was less than 7.90 times the diameter of the core pile.

It should be noted that current research mainly focused on the influence of solidifying agent materials, construction techniques, and pile structural configurations on the deformation and bearing capacity of mixing piles. However, theoretical investigations into the intrinsic mechanisms of enhanced bearing capacity in mixing piles, the interaction mechanisms between mixing tools and soil, as well as studies specifically targeting CS-DSM piles, remain notably scarce.

In this study, a new CS-DSM drilling tool was proposed and introduced. The discrete element software was employed to establish a coupling model of the CS-DSM drilling tool and soil particles. The numerical model was verified through indoor model tests. Subsequently, the variation rules of the moving paths of soil particles at different positions relative to the mixing drilling tool were analyzed during the mixing process. The similarities and differences in the moving paths of soil particles between the drilling-down and drilling-up processes were compared. Finally, the interaction mechanism between the CS-DSM drilling tools and soil particles was revealed.

2 Methods

2.1 The CS-DSM drilling tool

To resolve persistent challenges associated with DSM piles—specifically tool clogging and material adhesion—a newly designed CS-DSM implement was introduced. This system integrates dual mixing blade sets operating in opposing rotational directions. Figure 1 displays schematic comparisons of conventional DSM, DDM, and the proposed CS-DSM apparatus. Traditional DSM tools (Figure 1A) restrict rotation to a single direction during stratum penetration and soil blending processes. While DDM equipment (Figure 1B) enables bidirectional rotation in drilling/lifting sequences, it produces merely one shear surface. Conversely, the CS-DSM mechanism (Figure 1C) achieves multidirectional shearing through 4-6 intersecting failure planes spanning its operational height, enabling complete soil homogenization and substantially enhancing composite pile quality and structural consistency.

Figure 1

Figure 1. The diagram of DSM, DDM, and CS-DSM drilling tools: (a) DSM tool; (b) DDM tool; (c) CS-DSM tool.

Figure 1 demonstrated that the drilling tool mainly consisted of an inner rod, an outer rod, and mixing wing plates fixed on the rods. The drilling tool employed in this study had a total of 6 layers wing plates. The drilling and mixing wing plates were equipped with 2-3 cutting blades. The total length of the mixing drilling tool was approximately 2.0 m, the effective length was about 1.6 m, and the total weight of the drilling tool was around 0.6 t. Note that the given design parameters were of one type of the CS-DSM drilling tool that was manufactured and applied in the engineering. Through interchangeable modular components within the dual-pipe architecture, operators could modify the angular disposition, interlamellar spacing, and geometric configuration of the drilling apparatus.

For the structural design, each single piece of the outer rod had a figure-eight structure to enhance its stiffness. The radial and axial supports were added to the lower guiding structure of the outer rod to ensure the stiffness of the tool. The framed bidirectional CS-DSM drilling tool could achieve multi-level mutual shearing and uniform mixing of the in-situ soil and the solidification material slurry during the process of sinking, drilling, shearing, and mixing. Consequently, it ensured the uniform mixing of the pile body and improved the continuity of the pile body.

2.2 The coupling discrete element model between soil and the drilling tool

The Discrete Element Method is a numerical simulation technique for non-continuous media. The core of this method is to discretize the complex material structure into particle bodies with independent physical properties, and characterize the macroscopic mechanical behavior through the dynamic interaction between the particles. This method breaks through the limitations of traditional continuous medium assumptions and is particularly suitable for the numerical analysis of composite discrete materials such as soil (Zhang et al., 2025; Elabd et al., 2025). Compared with the continuous medium method, such as finite element method, it can visually simulate the interaction of non-continuous particles and exhibit the evolution processes such as soil mixing from the micro perspective. Hence, the discrete element software was employed to construct a coupling model of the CS-DSM drilling tool and soil particles, as showed in Figure 2. In the model, soil particles were simulated using the “Ball” element, while the mixing drilling tool was represented by “Wall” elements. A multi-phase computational strategy was proposed as the shape of drilling tool was too complicated to model in the discrete element software. The implemented methodology progressed through sequential operational stages: (1) Precision geometric modeling of the drilling apparatus was first accomplished using the finite element analysis platform MIDAS GTS, generating three-dimensional structural data exported in standardized STL file format. (2) This digitally reconstructed tool model was then imported into the discrete element software environment. (3) Suitable contact model was implemented to precisely describe the dynamic interfacial interactions between the drilling tool surfaces and soil particles.

Figure 2

Figure 2. The discrete element model of CS-DSM drilling tool and the soil: (A) The soil foundation; (B) The numerical drilling tool; (C) The coupling model of drilling tool and soil particles.

Figure 2A stated that the foundation exhibited a cylindrical geometry with a height of 0.75 m and a diameter of 0.2 m. To balance the computational accuracy and efficiency of the discrete element simulation, parametric optimization analysis was conducted, ultimately selecting a soil particle diameter of 4 mm. The numerical model comprised 56,105 discrete soil particles in total. For enhanced visualization of soil particle mixing patterns after mixing, the soil particles were partitioned into four distinct groups within the horizontal plane, each demarcated by a unique color.

As showed in Figure 2B, the mixing drilling tool was composed of the inner drilling rod and outer drilling rod. These two components were coaxial but remain unconnected with a 2 mm gap maintained between them to enable independent rotational speeds and bidirectional rotation. The dimensions of each blade were 28 × 20 × 5 mm with a 20° inclination angle. Specifically, on the inner drilling rod, the blade spacing varied. It was 60 mm at the upper section and increased to 110 mm at the lower section. In contrast, the outer drilling rod had a consistent blade spacing of 50 mm. The outer rod of the assembly had a diameter of 100 mm. The above dimensions of the drilling tool were coincident with that used in the indoor model test described in Section 2.3.1.

According to reference (Xu et al., 2025), the linear contact model was effective to simulate the interaction between soil particles. Hence, it was used to simulate the contact behavior between soil particle and soil particle, soil particles and the stirred drilling tool in the discrete element model. As showed in Figure 3, the linear model provided an infinitely small behavior of the interface and did not resist relative rotation with the contact torque Mc equal to zero. The contact load was decomposed into the linear force and damping force (F_c = F^l + F^d). The linear force provided linear elastic (no tension) and frictional properties, while the damping force declared the viscous properties. The linear force was produced by a linear spring with constant normal k_n and shear stiffness k_s and the damping force was produced by a damper with the normal and shear critical damping ratios (β_n, β_s). The linear spring acted in parallel with the damper. The surface clearance (g_s) was defined as the difference between the contact clearance (g_c) and the reference clearance (g_r). When the reference clearance was zero, the conceptual surface coincided with the workpiece surface. The contact was an active contact only if the surface clearance was less than or equal to zero.

Figure 3

Figure 3. Schematic diagram of the linear model.

It should be noted that the primary objective of this study was to investigate the moving paths of soil particles during the CS-DSM mixing process, thereby elucidating the interaction mechanisms between the mixing drilling tool and soil. Previous studies (Ge et al., 2024a; Nakao et al., 2021) had demonstrated that employing uniform particle sizes for soil simulation could effectively achieve the aforementioned research objectives.

2.3 The verification of the discrete element model

2.3.1 Indoor model test

To validate the correctness of the discrete element model, indoor colored sand model test was conducted. The colored sands used in the tests consisted of red, green, blue, and white quartz sand. The white sand was natural quartz sand, while the red, green, and blue sands were prepared by dyeing quartz sand with acrylic pigments. The size of the sands was 0.8–1.25 mm with density of 2,630 kg/m³. Before test, the sands were dry in the sun to remove the water. The four-colored sands were hierarchically compacted into a foundation within a cylindrical container with diameter of 36 cm. The colored foundation comprised four equal vertical sectors with each occupying 1/4 of the cylinder volume, and the division boards were used to divide the space and removed after the sands were compacted. The prototype drilling tool employed in experimental investigations maintained geometric congruity with the schematic representation in Figure 2. Under the test conditions, the rotation velocity of the mixing blade was 7 rpm, and the drilling-down/up speed of the drilling tool was 0.3 m/min. As indicated by Figure 4, the model test was carried out in the following steps.

1. The layered method was used to prepare the simulated foundation with good uniformity.

2. Start the model CS-DSM drilling tool system for mixing.

3. After mixing, use a vacuum cleaner to absorb the upper colored sand with thickness of 5 cm.

4. Take photos and analyze the horizontal plane of sand.

Figure 4

Figure 4. Test steps.

2.3.2 The results of numerical simulation and model tests

Experimental validation of the discrete element computational framework developed in Section 3 was performed with the test results. Figure 5 showed the relative positions of the drilling tool and soil particles at different moments of drilling-down and drilling-up. Comparative analysis of soil particle distribution characteristics at depth of 5 cm was conducted between numerical simulations and experimental models under the same mixing conditions, as illustrated in Figure 6. The soil particles distribution of the simulation before and after mixing was showed in Figure 6A and (b) respectively, while Figure 6C demonstrated that mixing result of test. Chromatic demarcation in Figure 6A employed yellow, green, blue, and red circular boundaries corresponding to radial extensions of 1.2, 1.3, 1.4, and 1.5 times the CS-DSM tool diameter D. Quantitative evaluation revealed minimal soil particles displacement beyond the yellow (1.2D) circumferential boundary post-mixing, establishing the principal influence zone of CS-DSM apparatus within 1.2 times its operational diameter. Consequently, comparative validation between numerical and experimental results was constrained to soil particles within this critical radial domain.

Figure 5

Figure 5. The relative positions of the drilling tool and soil particles at different moments: (A) Drilling-down 20s; (B) Drilling-down 60s; (C) Drilling-down 120s; (D) Drilling-up 20s; (E) Drilling-up 60s; (F) Drilling-up 120s.

Figure 6

Figure 6. Distribution of soil particles at depth of 5 cm: (A) distribution of numerical model before mixing; (B) Distribution of numerical model after mixing; (C) Distribution in the model test after mixing.

As illustrated in Figure 6C, the mixing range of the drilling tool approximates 1.2 times the outer drilling rod diameter, which exhibited a high degree of consistency with the discrete element simulation results showed in Figure 6B. The particle counts of colored sand with distinct hues within the range in both Figures 6B,C were statistically analyzed. The detailed quantification methodology was described as follows.

1. Numerical simulation: Prior to the drilling tool’s mixing operation, the “ball group” command in discrete element software was utilized to categorize particles within four sector-shaped regions (as showed in Figure 6A) into four distinct groups (Group 1 to Group 4) based on their color attributes, and each color represented a unique particle group. Upon completion of the mixing process, the “count” command was employed to statistically analyze the distribution characteristics of soil particles. This involved quantifying the number of particles belonging to Groups 1–4 within each sector-shaped region through coordinate-based particle identification, thereby obtaining the spatial distribution pattern of soil particles after mixing.

2. Model test: Firstly, the images of Figure 6C were magnified to manually count the pixel occupancy of individual sand particles. Subsequently, machine vision techniques were implemented to quantify the colored pixels within each fan-shaped region in Figure 6C. The particle count was determined through pixel-based calculations. For instance, with an average particle occupying 9 pixels, the blue particle quantity in Sector 1 was calculated as 2,567/9 ≈ 285.2 particles, where 2,567 represented the total blue pixels identified in that sector.

Using the above method, the distribution of four-color particles in each sector was showed in Table 1. It should be noted that due to the limitation of computer computing capacity and operation time, the diameter of soil particles in numerical simulation was about 4 times that in model test. Therefore, when comparing the distribution of soil particles in a certain buried depth between numerical simulation and model test, the number of particles should be multiplied by the area ratio, that was, 4² = 16. As could be seen from Table 1, the error of particles of different colors in different regions was less than or equal to 10%, and the numerical simulation results were very close to the model test results, thus verifying the applicability of the coupling discrete element model between the CS-DSM drilling tool and soil particles.

Table 1

Table 1. The comparison between results of numerical simulation and model test.

In addition, through calibration with the model test results, the parameters of the contact model in the discrete element model were determined. The normal and tangential stiffness of the contact between Ball and Ball were both 1E4 N/m, and that between Ball and Wall were both 1E6 N/m.

2.3.3 Numerical simulation conditions

To systematically investigate the influence mechanisms of operational parameters, on the movement paths of soil particles at different positions, 5 distinct numerical simulation cases were established with operational parameters as detailed in Table 2. The drilling rod rotation velocity was set as 5, 7, 10 rpm, the drilling/lifting speed was set as 0.21, 0.30, 0.42 m/min.

Table 2

Table 2. Numerical simulation conditions.

3 Moving paths analysis of soil particles

3.1 The selection of the soil particles

To comprehensively analyze the moving paths of soil particles at different positions, 3 particles were selected along the x-axis, y-axis, and z-axis respectively. As showed in Table 3, there were a total of 8 soil particles with the coordinates of soil particles X1 and Z1 coinciding. The central axis of the drilling tool was defined as the z-axis. Particles with x and y coordinate of 0.03 were located between the inner drilling rod and the outer drilling rod; those with x and y coordinate of 0.05 were positioned at the outer rod; and particles with x and y coordinate of 0.06 were situated at a distance of 1.2 times the diameter of the drilling tool, which marked the edge of the mixing influence range of the drilling tool. Figure 7 presented the relative positions between the drilling tool and the aforementioned 8 particles at three critical moments: the start of drilling-down, the end of drilling-down, and the end of drilling-up. Note that the soil particles with other coordinates and that under different condition parameters had similar moving paths. Due to space constraints, this section took the above 8 soil particles as example to investigate the interaction mechanism between the CS-DSM drilling tools and soil particles.

Table 3

Table 3. The coordinates of the selected soil particles.

Figure 7

Figure 7. The relative position between the soil particles and drilling tool: (A) At start moment of the drilling-down; (B) At end moment of drilling-down and start moment of the drilling-up; (C) At end moment of drilling-up.

In order to quantitatively evaluate the path richness of particles, a Volume Coefficient (VC) index of soil particle was presented, and its value was the ratio of the product of the x-direction, y-direction and z-direction moving ranges of moving path to the volume of the foundation as described by Equation 1.

$k=\frac{L_{\Delta x}\times L_{\Delta y}\times L_{\Delta z}}{V}(1)$

where k was the VC of particle moving path; L_x, L_y, L_z was the moving range of the particle path in the x direction, y direction and z direction, respectively; V was volume of the soil foundation.

A higher VC corresponded to an expanded particle moving range and enhanced path complexity, whereas a reduced coefficient was indicative of constrained particle displacement and diminished path diversity. This inverse relationship quantitatively demonstrated that the volumetric parameter serves as a reliable indicator for evaluating the path multiplicity of soil particle.

3.2 Moving paths of soil particles along x-axis

Figure 8 showed the three-dimensional moving paths of the three soil particles along x-axis under different rotation velocities and different drilling/lifting speeds. Figures 8A–D,F represented the three-dimensional moving paths of the particles under the various conditions listed in Table 2. As could be seen from the figure, the moving paths of particle X2 was the most complex, followed by that of particle X1, while the moving paths of particle X3 was relatively less abundant. In fact, particle X3 did not even form a closed circle in the xoy plane.

Figure 8

Figure 8. Three-dimensional moving paths of particles along the x-axis: (A) Condition 1; (B) Condition 2; (C) Condition 3; (D) Condition 4; (E) Condition 5.

Using condition 1 as the standard condition, Figure 9 showed the projections of the particle’s three-dimensional moving path on the xoy, xoz, and yoz planes under condition 1. From the figure, it could be seen that the moving ranges of particle X1 in the x, y, and z directions were −0.03–0.035 m, −0.03–0.05 m, and −0.15 to −0.35 m respectively. For particle X2, the corresponding moving ranges were −0.03–0.05 m, −0.025–0.04 m, and −0.15 to −0.50 m respectively. The moving ranges of particle X3 were −0.05–0.06 m, −0.05 to 0.01 m, and −0.1 to −0.15 m respectively. Particle X1 had the largest moving range in the y direction, particle X2 had the largest moving range in the z direction, and particle X3 had the largest moving range in the x direction. The fact that soil particle X3 had no downward moving paths indicated that it did not move downward with the drilling tool during the drilling-down process. However, it moved upward by 0.05 m with the drilling tool during the drilling-up process. The VC of particle X1, X2, X3 was 0.110, 0.143, 0.014. It was observed that particle X2 has more complicated moving path and the moving path richness of particle X1 and X2 was much larger than particle X3.

Figure 9

Figure 9. The moving paths of particles along the x-axis: (A) The moving paths in xoy plane; (B) The moving paths in xoz plane; (C) The moving paths in yoz plane.

Analysis of the aforementioned computational results revealed the following: (1) although particle X1 initially resided closer to the drilling rod compared to particle X2, it exhibited a greater moving range in the y direction. This suggested that soil particles in proximity to the drilling rod might experience more extensive moving ranges than those at greater distances during mixing processes. (2) As illustrated in Figure 9A, all particles demonstrated reciprocating paths in the xoy plane, involving both outward moving away from the drilling rod and inward approach toward it. This observation implied potential positional interchange between soil particles located at varying distances from the rod during mixing. (3) Figure 9B demonstrated that three particles initially aligned along the x-axis at z = −0.15 m exhibited final vertical coordinates of z = −0.1 m, z = −0.3 m, and z = −0.46 m after mixing. These vertical displacements confirmed bidirectional soil movement during mixing, encompassing both upward and downward moving components.

3.3 Moving paths of soil particles along y-axis

Figure 10 showed the three-dimensional moving paths of the three soil particles along y-axis under different rotation velocities and different drilling/lifting speeds. Figures 10A–D,F represented the three-dimensional moving paths of the particles under the various conditions listed in Table 2. As depicted in the figure, particle Y1 exhibited the most intricate moving paths, followed by particle Y2, while particle Y3 displayed relatively limited movement. Notably, both particles Y2 and Y3 failed to form closed circular paths within the xoy plane.

Figure 10

Figure 10. Three-dimensional moving paths of particles along the y-axis: (A) Condition 1; (B) Condition 2; (C) Condition 3; (D) Condition 4; (E) Condition 5.

Using condition 1 as the standard condition, Figure 11 showed the projections of the particle’s three-dimensional moving path on the xoy, xoz, and yoz planes under condition 1. From the figure, it could be seen that the moving ranges of particle Y1 in the x, y, and z directions were −0.035–0.040 m, −0. 040 to 0.040 m, and −0.150 to −0.550 m, respectively. For particle Y2, the corresponding ranges were −0.040 to 0.040 m, −0.040–0.050 m, and −0.150 to −0.200 m, respectively. Particle Y3 exhibited moving ranges of −0.020–0.040 m, −0.060 to 0.040 m, and −0.100 to −0.200 m, respectively. Comparative analysis revealed that particle Y1 demonstrated the largest moving ranges in both the x and z directions, whereas particle Y3 showed the most extensive moving range in the y direction. The VC of particle Y1, Y2, Y3 was 0.102, 0.015, 0.025. It was observed that the moving path richness of particle Y1 was much larger than particle Y2 and Y3.

Figure 11

Figure 11. The moving paths of particles along the y-axis: (A) The moving paths in xoy plane; (B) The moving paths in xoz plane; (C) The moving paths in yoz plane.

Analysis of the above calculation results revealed the following: (1) the initial position of particle Y1 was closer to the drilling rod than that of particle Y2. However, its moving range in the x-direction was larger than that of particle Y2. This indicated that during the mixing process, the moving range of soil particles closer to the drilling rod might exceed that of those farther away from the drilling rod. (2) all particles exhibited reciprocating moving within the xoy plane, involving both moving away from and towards the drilling rod. This suggested that during mixing, the positions of soil particles at different distances from the drilling rod might be interchanged. (3) after mixing, the vertical coordinates of the three particles on the y-axis at z = -0.15 m became −0.17 m, −0.20 m, and −0.53 m, respectively. This demonstrated that soil particles had a relatively large moving range in the vertical direction during mixing.

Comparative analysis of VC of X1-X3 and Y1-Y3 revealed that soil particles along the x-axis and y-axis exhibited distinct moving patterns. This discrepancy arose from the differential relative positions between drilling tool and soil particles on each axis. As showed in Figure 7A, during drilling-down, particles along x-axis were positioned directly beneath a specific blade set, while particles along y-axis resided in the inter-blade region between two adjacent blade groups. The phenomena demonstrated that particles at distinct spatial positions exhibited a certain degree of stochasticity in their moving paths. It was precisely such stochastic characteristics that enabled the soil to achieve a homogenized state through sufficient operational cycles of auger blade mixing.

3.4 Moving paths of soil particles along z-axis

Figure 12 showed the three-dimensional moving paths of the three soil particles along z-axis under different rotation velocities and different drilling/lifting speeds. Figures 12A–D,F represented the three-dimensional moving paths of the particles under the various conditions listed in Table 2. As could be observed from the figure, the moving paths of particles Z1 and Z2 were the most diverse, while those of particle Z3 were relatively scarce, and it failed to form a closed circle in the xoy plane.

Figure 12

Figure 12. Three-dimensional moving paths of particles along the z-axis: (A) Condition 1; (B) Condition 2; (C) Condition 3; (D) Condition 4; (E) Condition 5.

Using condition 1 as the standard condition, Figure 13 showed the projections of the particle’s three-dimensional moving path on the xoy, xoz, and yoz planes under condition 1. From the figure, it could be seen that the moving ranges of particle Z1 in the x, y, and z directions were −0.030–0.040 m, −0.030–0.050 m, and −0.150 to −0.350 m, respectively. For particle Z2, the corresponding moving ranges in the x, y, and z directions were −0.020–0.030 m, −0.020 to 0.020 m, and −0.350 to −0.600 m, respectively. Regarding particle Z3, the moving ranges in the x, y, and z directions were 0.030–0.050 m, 0–0.030 m, and −0.550 to −0.600 m, respectively. Particle Z1 exhibited the largest moving ranges in the x and y directions, and both particles Z1 and Z2 had relatively large moving ranges in the z direction. The VC of particle Z1, Z2, Z3 was 0.048, 0.021, 0.001. It was observed that the moving path richness of particle Z1 and Z2 was larger than particle Z3.

Figure 13

Figure 13. The moving paths of particles along the z-axis: (A) The moving paths in xoy plane; (B) The moving paths in xoz plane; (C) The moving paths in yoz plane.

A more in-depth analysis revealed that particle Z1 first moved outward to the vicinity of the outer rod in the xoy plane and then moved inward to the vicinity of the inner drilling rod, indicating that some soil particles undergo reciprocating moving.

3.5 Moving paths of soil particles during drilling-up and drilling-down

Figure 14 compared the drilling-down and drilling-up moving paths of three soil particles along the x-axis at burial depths of 0.15 m and 0.45 m below the ground surface. The coordinates of the three particles at the 0.45 m depth were specifically defined as X4 (0.03, 0, −0.45), X5 (0.05, 0, −0.45), and X6 (0.06, 0, −0.45). It should be noted that the drilling tool employed in this study had a vertical dimension of 190 mm as showed in Figure 2A, with both drilling-down and drilling-up operations achieving a maximum depth of 0.7 m. Due to the physical constraints of the tool configuration, soil particles within the bottom 190 mm stratum of the foundation cannot achieved complete mixing during the actual mixing process. Therefore, this section focused on analyzing the moving paths of particles at the 0.15 m and 0.45 m burial depth to investigate their displacement characteristics during tool drilling-down and drilling-up.

Figure 14

Figure 14. Comparison of particle moving paths during drilling-down and drilling-up. (A) Three-dimensional moving paths of particles at depth of 0.15 m during drilling-down; (B) Three-dimensional moving paths of particles at depth of 0.15 m during drilling-up; (C) Three-dimensional moving paths of particles at depth of 0.45 m during drilling-down; (D) Three-dimensional moving paths of particles at depth of 0.45 m during drilling-up.

As showed in Figure 14, during the drilling-down process, the moving paths of particles were significantly more diverse than those during the drilling-up process. The moving ranges of soil particles at depths of 0.15 m and 0.45 m during the drilling-down process were several times larger than those during the drilling-up process. For instance, during the drilling-down process, at depth of 0.15 m, the moving ranges of particles with x coordinate of 0.03 in the x, y, and z directions were −0.030–0.035 m, −0.030–0.050 m, and −0.150 m to −0.350 m, respectively, with a VC of 0.044. During the drilling-up process, those ranges were −0.030 to 0.018 m, −0.020–0.030 m, and −0.029 to −0.032 m, respectively, with a VC of 0.0003. The moving ranges in the x, y, and z directions during the drilling-down process were 1.4, 1.6, and 66.7 times larger than those during the drilling-up process, respectively, and the VC during the drilling-down process was 146.7 times larger than that during the drilling-up process. This was because during the drilling-up process, the upward moving of particles was affected by the gravitational field of the upper soil particles, which restricted these soil particles to move upward in a relatively small range along with the drilling tool.

The following conclusions provided practical implications for engineering construction practices. When the ground conditions, dosage of solidification materials, and mixing frequency per unit height of the drilling tool remained consistent, two approaches could be adopted to enhance mixing efficiency: (1): decreasing the drilling-down speed while increasing the drilling-up speed; (2) increasing the rotational velocity during drilling-down and reducing it during drilling-up.

4 Conclusions and limitations

4.1 Conclusions

In this paper, a combined method of discrete element software and MIDAS GTS software was adopted to establish a mesoscopic model of CS-DSM drilling tools and soil particles. The variation laws of the moving paths of soil particles under the mixing action of the drilling tools were systematically studied, and the interaction mechanism between the CS-DSM drilling tools and soil particles was revealed. The main conclusions were as follows.

1. Horizontal moving patterns: Under drilling tool rotation, soil particles reciprocate horizontally. Proximal particles exhibit broader displacement ranges than distal ones, enabling spatial exchange and facilitating horizontal homogenization through dynamic redistribution.

2. Vertical moving patterns: Particles at identical depths show extensive vertical displacement, with some migrating upward and others downward relative to initial positions, promoting effective vertical soil homogenization.

3. Movement path variability: Not all particles undergo large-scale complex paths, and displacement depends on spatial relation to the tool. Particles directly beneath blades displace more significantly than those in inter-blade regions, and this stochasticity ensures homogenization via sufficient mixing cycles.

4. Drilling-down vs. drilling-up comparison: During drilling-down, particles exhibit larger horizontal and vertical displacement ranges, enabling full 3D movement and position exchange for uniform pile mixing.

5. Construction suggestion: The CS-DSM drilling tool can effectively mix the soil particles in the horizontal and vertical direction that replaces the DSM drilling tool. In practice, reducing drilling speed or increasing rotational velocity during drilling-down enhances CS-DSM pile quality under identical conditions.

4.2 Limitations

1. The use of circular particle for soil neglects the angularity effects on interlocking and fracture. Future modeling will establish a fine discrete element model including the soil particle angularity.

2. The discrete element model for soil and drilling tool does not account for soil cohesion or groundwater effects, and assumes uniformly sized soil particles without considering gradation influences. Subsequent research should develop more sophisticated models incorporating these critical factors.

3. To enhance computational efficiency, the developed model employs a scaled-down configuration without accounting for scale effects. Subsequent research will establish full-scale prototypes replicating actual engineering conditions to conduct comprehensive analyses.

Data availability statement

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

Author contributions

ZX: Writing – review and editing, Conceptualization, Resources, Supervision. ZL: Writing – review and editing, Data curation, Validation. SF: Data curation, Investigation, Visualization, Writing – original draft. LX: Investigation, Writing – original draft, Formal Analysis. KW: Funding acquisition, Methodology, Project administration, Validation, Writing – review and editing.

Funding

The authors declare that financial support was received for the research and/or publication of this article. This research was funded by the Tianjin Municipal Education Commission (2021KJ055).

Conflict of interest

Authors ZX and SF were employed by China Jingye Engineering Co., Ltd., Central Research Institute of Building and Construction Co., Ltd.

Author ZL was employed by ZheJiang Kunde Innovate Geotechnical Engineering Co., Ltd.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The authors declare that no Generative AI was used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fbuil.2025.1729162/full#supplementary-material

References

Arul, A., Mohammadjavad, Y., Mahdi, M. D., Suksun, H., Myint, W. B., and Melvyn, L. (2018). Evaluation of fly ash- and slag-based geopolymers for the improvement of a soft marine clay by deep soil mixing. Soils Found. 58 (6), 1358–1370. doi:10.1016/j.sandf.2018.07.005

CrossRef Full Text | Google Scholar

Cai, G. J., Liu, S. Y., and Puppala, A. J. (2012). Assessment of soft clay ground improvement from SCPTU results. Proc. Institution Civ. Engineers-Geotechnical Eng. 165 (2), 83–95. doi:10.1680/geng.9.00081

CrossRef Full Text | Google Scholar

Claver, P., Sara, R., António, V. F., and Nuno, C. (2024). Stabilisation of estuarine sediments with an alkali-activated cement for deep soil mixing applications. J. Rock Mech. Geotechnical Eng. 16 (4), 1398–1410. doi:10.1016/j.jrmge.2023.08.020

CrossRef Full Text | Google Scholar

Ekmen, A. B., Algin, H. M., and Özen, M. (2020). Strength and stiffness optimisation of fly ash-admixed DCM columns constructed in clayey silty sand. Transp. Geotech. 24, 100364. doi:10.1016/j.trgeo.2020.100364

CrossRef Full Text | Google Scholar

Elabd, M., Rui, R., Yi, K., He, S. K., Chen, C., and Ye, Y. Q. (2025). Investigation of soil arching evolution in unreinforced piled embankments with three-dimensional discrete element method-simulated trapdoor tests. Comput. Geotechnics 188, 107561. doi:10.1016/j.compgeo.2025.107561

CrossRef Full Text | Google Scholar

Ge, C. W., Liu, Z., Lan, W., Yang, N. H., Wen, L., and Zhou, J. (2024a). Model tests on performances of DSM and CS-DSM piles. Chin. J. Geotechnical Eng. 46 (11), 2420–2428.

Google Scholar

Ge, C. W., Liu, Z., Yu, T. X., Lan, W., Yang, N. H., and Zhao, M. Y. (2024b). Model test study of key factors of deep soil mixing mechanism using contra-rotational shear method. Rock Soil Mech. 45 (1), 68–76.

Google Scholar

Han, Y., Cheng, J., Zhou, M., Ni, P. P., and Wang, Y. L. (2022). Experimental study of compaction and expansion effects caused by penetration of core pile during construction of SDCM pile. Int. J. Geomechanics 22 (5), 04022041. doi:10.1061/(asce)gm.1943-5622.0002353

CrossRef Full Text | Google Scholar

Kitazume, M. (2021). Recent development and future perspectives of quality control and assurance for the deep mixing method. Appl. Sci. 11 (19), 9155. doi:10.3390/app11199155

CrossRef Full Text | Google Scholar

Kizuki, T., Sawaguchi, H., Imai, T., Takaue, T., Tsuchiya, J., and Inazumi, S. (2018). An introduction of real-time management system for applying deep-mixing method with large diameter and improvement depth (DCS method). J. Soc. Mater. Sci. Jpn. 67 (1), 93–98. doi:10.2472/jsms.67.93

CrossRef Full Text | Google Scholar

Mosadegh, A., Szymkiewicz, F., and Nikraz, H. (2017). An experimental investigation of the impact of specimen preparation and curing conditions on cement-treated material strength (deep mixing method). Aust. J. Civ. Eng. 15 (1), 49–60. doi:10.1080/14488353.2017.1372685

CrossRef Full Text | Google Scholar

Nakao, K., Inazumi, S., Takaue, T., Tanaka, S., and Takayuki, S. (2021). Evaluation of discharging surplus soils for relative stirred deep mixing methods by MPS-CAE analysis. Sustainability 14 (1), 58. doi:10.3390/su14010058

CrossRef Full Text | Google Scholar

Nikbakhtan, B., and Ahangari, K. (2010). Field study of the influence of various jet grouting parameters on soilcrete unconfined compressive strength and its diameter. Int. J. Rock Mech. Min. Sci. 47 (4), 685–689. doi:10.1016/j.ijrmms.2010.03.004

CrossRef Full Text | Google Scholar

Nikbakhtan, B., and Osanloo, M. (2009). Effect of grout pressure and grout flow on soil physical and mechanical properties in jet grouting operations. Int. J. Rock Mech. Min. Sci. 46 (3), 498–505. doi:10.1016/j.ijrmms.2008.10.005

CrossRef Full Text | Google Scholar

Shen, K., Zhang, H., Liu, J. K., and Zhang, Y. H. (2024). Study of cement-soil mixed piles reinforcement method for offshore wind turbine pile foundation. Ocean. Eng. 313 (2), 119423. doi:10.1016/j.oceaneng.2024.119423

CrossRef Full Text | Google Scholar

Wang, Z. J., Peng, Z. K., Yan, Y. W., Zeng, J. C., and Wu, Z. S. (2025). Computed tomography of layered defects in large-diameter pile foundation using transmission of hybrid ultrasonic modes. Structures 74, 108545. doi:10.1016/j.istruc.2025.108545

CrossRef Full Text | Google Scholar

Xu, M. Q., Pan, K., Duan, B., Wu, Q. X., and Yang, Z. X. (2025). Investigating the influence of particle shape on discrete element modeling of granular soil under multidirectional cyclic shearing. Soil Dyn. Earthq. Eng. 189, 109097. doi:10.1016/j.soildyn.2024.109097

CrossRef Full Text | Google Scholar

Yi, Y. L., Liu, S. Y., Puppala, A. J., and Jing, F. (2019). Variable-diameter deep mixing column for multi-layered soft ground improvement: laboratory modeling and field application. Soils Found. 59 (3), 633–643. doi:10.1016/j.sandf.2019.01.009

CrossRef Full Text | Google Scholar

Yin, K. S., Shen, P. L., Zhang, L. M., Cai, Y. M., Xuan, D. X., and Poon, C. S. (2024). Carbonized seawater cement slurries for offshore deep cement mixing: carbonation mechanism, strength enhancement and microstructure evolution. Cem. Concr. Compos. 154, 105788. doi:10.1016/j.cemconcomp.2024.105788

CrossRef Full Text | Google Scholar

Zeng, B., Zhang, D. W., Hou, J., Zhang, A. J., Cheng, C. H., He, L., et al. (2024). Performance enhancement of large diameter bored pile reinforced by cement-soil in clay under static and cyclic loadings. Structures 60, 105923. doi:10.1016/j.istruc.2024.105923

CrossRef Full Text | Google Scholar

Zhang, D. W., Wang, A. H., and Ding, X. M. (2021). Seismic response of pile groups improved with deep cement mixing columns in liquefiable sand: shaking table tests. Can. Geotechnical J. 59 (6), 994–1006. doi:10.1139/cgj-2020-0505

CrossRef Full Text | Google Scholar

Zhang, J. H., Liu, K., and Yang, H. (2025). Influence mechanisms of gravel shape on the triaxial shear characteristics of soil-rock mixtures based on discrete element numerical simulation. Adv. Powder Technol. 36 (10), 105025. doi:10.1016/j.apt.2025.105025

CrossRef Full Text | Google Scholar

Zhao, J. N., Zong, Z. L., Huang, Y. H., Lu, M. M., Huang, D. L., and Guo, C. Y. (2025). Vertical and lateral bearing performance of bidirectional helix-stiffened cement mixing piles on marine soft soil sites: field tests and numerical simulations. Ocean. Eng. 323, 120678. doi:10.1016/j.oceaneng.2025.120678

CrossRef Full Text | Google Scholar

Zhuang, X. X., Zong, Z. L., Huang, Y. H., Wang, C. S., and Lin, X. J. (2022). Parametric study on analyzing the effect of soil-cement strength on the uplifting behavior of HSCM piles installed in marine soft clay. Appl. Sci. 13 (1), 330. doi:10.3390/app13010330

CrossRef Full Text | Google Scholar