A new method for rapid preparing high-strength saturated clay samples in large-scale model tests

The preparation of high-strength saturated clay samples for large-scale model tests presents a significant challenge in geotechnical engineering. The slurry consolidation method has been conventionally employed to prepare saturated clay, despite its time-consuming and labor-intensive nature. Therefore, this study proposes a rapid preparation technique for clayey soils utilizing the dynamic compaction method, enabling the facile preparation of saturated clay samples by compacting the soils from an unsaturated state. During compaction, the void ratio decreases, thereby increasing the degree of saturation and enhancing the soil strength. Critical to this method are two variables: the moisture water content and the soil density, which are determined through bench-scale compaction tests using the Proctor compaction test apparatus. These tests establish the relationships between moisture content and density, degree of saturation, and soil strength. The moisture content aligning with the target soil strength is selected as the target moisture content for model-scale soil preparation, whereas the moisture content-density relationship sets the target density value. The laboratory tests validate that the soil strength of the saturated model-size clay samples prepared using the proposed method fulfills the requisite criteria, indicating its effectiveness for rapid preparation of high-strength saturated clay samples in large-scale model tests.


Introduction
In the field of large-scale modeling to simulate marine strata, there exists a pronounced demand for high-strength, saturated clay samples.In marine geotechnical engineering, the pile shoe dimensions of marine platform pile foundations vary significantly from 3 m to 60 m [1], which ensures adequate bearing capacity with soil strengths ranging from a few kPa to several tens of kPa [2].Conventional methods for generating clay samples predominantly rely on soil consolidation from a slurry, a process that is notably protracted, particularly when aiming for higher soil strengths.Although the dynamic compaction technique has demonstrated effectiveness for sandy soils, its applicability to clay samples has been met with skepticism, primarily because of inconsistent achievements in soil strength and saturation levels.
This study introduces an innovative methodology for the rapid preparation of high-strength, saturated clay samples, employing the dynamic compaction technique [3,4].The preparation process involves two stages: initial bench-scale compaction tests conducted at varied water content levels to set precedents for model-scale clay sample preparation, and subsequent model-scale experiments.These preliminary tests aim to ascertain the optimal water content and density parameters for 1330 (2024) 012029 IOP Publishing doi:10.1088/1755-1315/1330/1/012029 2 subsequent model-scale trials.The relationships between water content w, soil density ρ, saturation ratio Sr, and qu (a soil strength metric) were derived to guide the selection of specific parameters that fulfill the desired soil strength and saturation in model-scale samples.

Theoretical foundation
Based on the fundamental principles of soil mechanics, the degree of saturation Sr is determined by its moisture content w and void ratio e.The relationship between the degree of saturation Sr, moisture content w, and void ratio e is described as follows: where s G denotes the soil-specific gravity.The relationship between the void ratio e and the moisture content w is expressed in equation 2, where s  denotes the dry density (g/cm 3 ), and  indicates the mass density (g/cm 3 ).
The research framework of this proposed method is illustrated in figure 1, indicating how repeated impact loads during compaction expel air from the soil samples and facilitate soil particle rearrangement into a denser configuration.This compaction process reduces the void ratio as well as increases the soil density and saturation ratio (Sr), transforming initially loose and unsaturated clay into compact, saturated samples with enhanced soil strength.The strength of the soil is intrinsically related to its void ratio, i.e., a decreased void ratio corresponds to increased soil strength.To determine the optimal moisture content that satisfies the desired soil strength and saturation degree, the relationships between water content w and soil density ρ, saturation ratio Sr, and qu need to be established.In addition to moisture content, compaction is influenced by factors such as the number of drops and layer thickness.

Experimental procedures
The implementation framework for the proposed method is presented in figure 2, highlighting the necessity of preliminary bench-scale compaction tests to determine two key parameters: moisture content and density.These parameters are instrumental in achieving the target strength.
Bench-scale compaction tests are conducted to evaluate the relationships between w and ρ, w and Sr, and w and qu relationships at varying moisture water contents, which enabled swift acquisition of the results because of the reduced volume of soil specimens.The number of drops (N) and layer thickness (Z) during these tests is crucial, which demonstrates the systematic approach of the method to optimize soil sample preparation for achieving desired soil strength and saturation levels.  1 outlines the optimal moisture content ranges for various clay soils, establishing the lower limit as the minimum moisture content for unit-level compaction tests.The compaction process progresses by incrementally increasing the number of compaction passes until no further soil settlement occurs.At this juncture, the compaction passes are designated as the single-layer compaction passes N. The rationale for conducting tests at the minimum moisture content is that, although compaction can enhance soil density under lower moisture conditions, the density increase may stagnate or become negligible [5].Therefore, this number of compaction passes can more accurately reflect the compaction ability of the soil, i.e., the optimal number of single-layer compaction passes for the soil layer.

Determination of layer thickness Z.
Regarding layer thickness, denoted as Z, its value should not exceed the thickness of the influence zone, identified as d, to avoid significant variations in soil layer properties.Utilizing an inappropriate Z can lead to nonuniform distributions of soil strength and moisture content.This section recommends a methodology for determining suitable layer thickness.
Typically, the influence zone thickness d is empirically determined using the Menard formula, described in equation ( 3): where d denotes effective impact depth (m), t W denotes rammer weight (kN), and H denotes rammer drop height (m), which is the correction factor closely linked to soil conditions.Table 2 lists the recommended correction factor values for different clay soil types, typically ranging from 0.4 to 0.8.By adopting the suggested value of nt, the calculated layer thickness using equation ( 3) is labeled as Z1, which serves as the layer thickness for initial bench-scale tests.Upon reaching the height of Z1 in the first layer, its density is measured and documented as the initial density of the sample ρ1.After finishing the compaction test, the soil density is measured and recorded as the overall density of the sample ρ2.A criterion of ρ1 = ρ2 needs to be satisfied after consistently adjusting the correction factor to obtain a suitable layer thickness Z2.

Determination of the moisture content w and density ρ.
Based on the relationships between moisture content (w), soil density (ρ), saturation ratio (Sr), and soil strength (qu) determined from the bench-scale tests, the specific moisture content (w) correlating with the desired strength and saturation level for model-scale tests can be established.Simultaneously, the soil density () corresponding to the selected moisture content is established as the design parameters for both moisture content and density in model-scale soil preparation.
If the achieved soil strength does not satisfy the expected criteria, the modifications in the compaction process (e.g., increasing the hammer weight) become necessary.The bench-scale tests are iteratively conducted until the w-qu relationship aligns with the target design strength, ensuring the attainment of specified soil properties.

Determination of the proportion of soil and water.
After determining the target moisture water content, the proportion of soil and water of the soil-water mixture prepared before compaction can be calculated using equation (4).
where Mw denotes the mass of water and Ms indicates the dry soil mass per layer.The total soil mass used per layer can be obtained, where M denotes the total soil mass per layer (kg),  denotes the design value of the soil density (kg/m 3 ), V denotes soil volume per layer (m 3 ), Z indicates the layer thickness, S denotes the base area of the model box (m 2 ).

Preparation of bench-scale soil samples
To validate the suggested preparation technique, experiments were performed using Shanghai soft clay from Layer 4. The soil properties are enumerated below: liquid limit stands at wL 40.93%, plastic limit wp is recorded at 23.33%, and the specific gravity Gs is 2.73.
The sample preparation apparatus is illustrated in figure 3(a), which is set up with a permeable stone, top cover, tri-part mold, sleeve, and a base.In the bench-scale tests, the number of compactions per layer, N, is designated as 20 times.The effective impact depth correction factor, nt, is assigned a value of 0.5, and the layer thickness Z is determined to be 4 mm using equation (3).
The test group with a layer thickness of 4 mm is labeled Group W1.In adherence to the standard for geotechnical test methods [13], two control test groups were established: one with a layer thickness of 27 mm (Group W2) and another with a thickness of 16 mm (Group W3).

Analysis of soil sample parameters
3.2.1.Relationship between moisture content and density.Figure 4(a) illustrates that soil density attains its maximum at the optimal moisture content (wop = 23.5% ± 1%).The moisture-density relationship curve was deduced from the unit-level test results, revealing that the soil density peaks at an optimal moisture content of 23.5% (  1%).Under identical moisture content conditions, the soil density in Group W1 consistently exceeds that of the control groups (W2 and W3), demonstrating that a 4 mm layer thickness during compaction yields the most effective compaction outcome.Prior to achieving the optimal moisture content, the density gradually increased with the moisture content; however, beyond this optimal point, the density decreased as the moisture content increased further.The underlying cause of this phenomenon is that at lower moisture contents, the predominance of strongly bound water within the soil increases interparticle friction and impedes particle mobility, challenging compaction efforts.As moisture content increases to the optimal level, the water film between the soil particles provides lubrication, thereby facilitating smoother soil particle relocation and enhanced pore filling or compaction, resulting in higher compaction density.

3.2.2.
Relationship between moisture content and degree of saturation.Figure 4(b) indicates that Group W1 attains a higher degree of saturation in comparison to Groups W2 and W3.Experimental findings show that, under constant moisture content conditions, the degree of saturation for Group W1 significantly exceeds that of the control groups W2 and W3.Before reaching the optimal moisture content, an upward trend in saturation degree is observed with increasing moisture content.This is attributed to the water starting to fill the voids between the soil particles at lower moisture contents, thereby elevating the saturation degree.As the moisture content continues to rise, the contact among the soil particles diminishes, yielding a disjointed pore structure.Beyond the optimal moisture content, a further rise in moisture causes the re-distribution of the originally clustered soil particles, thereby promoting a more uniform pore structure, and consequently, a higher degree of saturation.

3.2.3.
Relationship between moisture content and strength.Figure 4(c) depicts the relationship between moisture content and soil strength (w-qu).Below the optimal moisture content, the soil strength in Group W1 surpasses that of the comparison groups W2 and W3 under identical moisture conditions.However, once the moisture content exceeds the optimal threshold, the strength of the soil samples prepared with varying layer compaction thicknesses demonstrates a parallel trend.The shear strength of unsaturated soil is influenced by variables such as density, interparticle friction, and adhesive forces, all of which are closely associated with the moisture content.In completely dry conditions, soil particles lack any binding force.Nonetheless, moisture introduction tightens soil particle contact.At 21% moisture content, moisture serves as a "binder," enhancing soil strength to its peak through the combined effects of adhesion and friction.However, at this stage, minor pores within the soil persist, indicating that the maximal strength of the soil does not coincide with the optimal moisture content.Further moisture increments fill these minuscule pores, and the "lubricating effect" of the water progressively reduces the friction between the soil particles, thereby reducing soil strength.
In summary, when moisture content is below the optimal level, a reduced compacted layer thickness enhances the compaction effectiveness.This finding is supported by the considerable improvements in saturation and density, although with only marginal increment in strength.As the moisture content ascends above the optimal level, the impact of the compacted layer thickness on compaction effectiveness wanes.

Preparation and verification of model-scale soil samples
The preceding sections discussed methods for determining the moisture content and density for target strength soil via bench-scale tests.This section introduces the preparation process for soil in largescale model tests.

Preparation of model-level soil samples
For the model-scale experiments, a transparent acrylic model box with a side length of 80 cm was employed.A compaction machine exerting an impact force of 10 kN and featuring a 0.1 m hammer drop height was utilized for sample preparation.Employing a correction factor nt of 0.5, the layer thickness calculated using equation ( 3) was 16 cm.
Based on the bench-scale tests, the target moisture content for preparing the model-scale soil samples was determined to be 25%, with a target soil density of 1.97 g/cm 3 .The calculation of the dry soil mass (Ms) and water mass (Mw) required for each layer is stated as follows: To achieve uniform water distribution within the soil-water mixture, it was allowed to rest for 12 h post-mixing.Following the designated layer thickness, the compaction process was repeated for three layers of soil, with a total clay sample height of 48 cm.The compaction process extended over a duration of 3 h.

Analysis of model-scale soil sample results
To validate the quality of the prepared model-scale clay samples, measurements of soil density and degree of saturation were recorded at five distinct testing locations (holes 1 to 5) across three depth intervals: 15 cm, 30 cm, and 45 cm.Soil strength assessments were similarly executed at holes 6-9, at the specified depths, as depicted in figure 5. To validate the preparation quality of the model-scale clayey samples, the soil density and the degree of saturation were measured at five different testing locations (holes 1-5) at three depths (15 cm, 30 cm, and 45 cm).The soil strength was measured at holes 6-9 at depths of 15 cm, 30 cm, and 45 cm.
Figure 6 indicates that both soil density and shear strength display a generally uniform distribution across the depth, with a marginal increase in both parameters observed as depth increases.This trend may be attributed to the transmission of the compaction effect, where multiple compaction impacts allow a minor amount of energy to gradually permeate the lower soil layers, thereby rendering these layers denser with a reduced soil volume.This process consequently augments both soil density and strength.Notably, the degree of saturation at each evaluated point consistently exceeds 95%.

Deviation comparison
A comparative analysis focusing on soil characteristics such as density, degree of saturation, and soil strength was conducted between the model-and bench-scale levels.The averaged results from the three depths of the model-scale tests are comparatively presented against the findings of the benchscale experiments in table 4. The observations revealed that deviations in soil density, degree of saturation, and shear strength fall below a 2% margin, remaining well within the acceptable 5% threshold.
These results validate the feasibility of transferring design specifications for density and moisture content, as determined from the bench-scale experiments, to model-scale tests.Overall, the present findings demonstrate the effectiveness and reliability of the proposed method in preparing highstrength, saturated clay samples for comprehensive model evaluations.

Conclusions
This study presents a novel approach for the rapid preparation of high-strength, saturated clayey soils in large-scale model tests.Utilizing the dynamic compaction technique, this method focuses on two crucial properties: soil density and moisture content.Through a comprehensive series of bench-scale compaction tests-each conducted at varying moisture content levels, we established correlations between moisture content, soil density, degree of saturation, and soil strength.These established relationships enabled the determination of design values for both soil density and moisture content, which were then applied to model-scale tests.The model-scale testing validated the feasibility of this rapid preparation technique, demonstrating its capability to consistently fulfill the desired soil strength and saturation criteria.

Figure 2 .
Figure 2. Implementation framework for the preparation method.
Figure 3(b) illustrates the rammer weighing 300 g with a drop height of 20 mm, whereas figure 3(c) showcases a clay specimen extracted from a tri-part mold.

Figure 4 .
Figure 4. Bench-scale test results (a) density-water content relationship, (b) degree of saturationwater content relationship, and (c) strength-water content relationship.

Figure 5 .
Figure 5. Arrangements of the measuring points (top view).

Table 1 .
Optimal moisture content ranges for various types of clayey soils.

Table 2 .
Recommended values for the correction coefficient of effective reinforcement depth.

Table 3
Comparison between the bench-scale tests and the model-scale tests.