Improved Mandrel System for Prefabricated Vertical Drain Installation: A Macro to Micro Analysis

: Increasing development of infrastructure in Indonesia has driven the need for effective ground improvement methods to accelerate the consolidation of soft soil, which is estimated to occupy around 10% of the country’s land area. A prefabricated vertical drain combined with vacuum preloading is among the most effective methods for this purpose. However, the prefabricated vertical drain creates a smear zone in the surrounding soil area during installation. This study examines the effectiveness of a newly developed mandrel system in reducing the smear zone during prefabricated vertical drain installation. Large-scale consolidation tests at a macro level and microstructure analysis using scanning electron microscopy at a micro level were employed to investigate the effect of soil water content and shear strength. The results show that the water content and shear strength of the soft soil gradually increased in the inner smear zone and transition zone, while both decreased in the radial distance. Furthermore, the soil structure underwent a transformation in which the particle area and pore area became a closed ﬂake structure, and apparent agglomeration occurred. The test results indicate that the newly developed mandrel system can effectively reduce the smear zone. The macro to micro test results demonstrated that the mandrel system is successful in reducing the smear zone effect.


Introduction
In archipelagic countries such as Indonesia, airport infrastructure is becoming more important to increase connectivity and economic growth. Air transportation is one of the priority sectors in the ASEAN Economic Community (AEC) [1]. In particular, more airport facilities have been rebuilt and expanded in Indonesia [2][3][4][5]. However, most of the airports are located on soft soil because the limited areas of stable soil are designated for urban use [6]. Therefore, finding an effective method for engineering design in soft soil areas has become an important issue.
Soft soil has low strength, high compressibility, high water content, and it requires a far longer time to be completely consolidated. Rapid and effective treatment of soft soil is needed to increase shear strength, reduce the post-construction differential settlement, and save construction costs [7]. Therefore, ground improvement methods using a prefabricated vertical drain (PVD) combined with vacuum preloading to accelerate the consolidation of soft soil by improving the mechanical properties have been used since the early 1970s [8,9].
However, the PVD performance is influenced by several factors. Firstly, the effect of PVD installation creates a smear zone in which the surrounding soil area is disturbed due to the mandrel's influence. Moreover, various types of PVD cores are related to the influence zone. The difference in undrained shear strength is primarily due to different excess pore water pressure development associated with the change in soil structure. The particle migration induces a binding effect which significantly weakens clay particles in Table 1. The kaolinite properties.

PVD Characteristics
The PVD consists of a core, drainage channel, and filter. A schematic illustration of the PVD is provided in Figure 1. The material of the PVD was made by polypropylene core. Physical modeling was performed to study particular aspects of the behavior of prototypes.
A scaled-down model conducted for large-scale consolidation testing is recommended by assuming a related governing equation [21]. Therefore, the coefficient of PVD depth and permeability are the significant factors influencing PVD performance. The similarity theory for the scaling factor of PVD can be determined using Equations (1) and (2).
k p k m = n k (2) where z p is the embedded length of the PVD prototype, z m is the embedded length of the PVD model, n z is the scale factor of embedded length, k p is the permeability of the prototype, and k m is the change permeability of the model. The scale factor implies a necessary initial specific volume or void ratio, a scale factor of F = 20 was determined, as expressed in Equation (3). ∆v = ∆e = n k ln(F) (3) where ∆v is the change in initial specific volume, and ∆e is the change in void ratio. The scale factor parameter for three types of PVD is shown in Table 2. A scaled-down model conducted for large-scale consolidation testing is recommended by assuming a related governing equation [21]. Therefore, the coefficient of PVD depth and permeability are the significant factors influencing PVD performance. The similarity theory for the scaling factor of PVD can be determined using Equations (1) and (2).
where is the embedded length of the PVD prototype, is the embedded length of the PVD model, is the scale factor of embedded length, is the permeability of the prototype, and is the change permeability of the model. The scale factor implies a necessary initial specific volume or void ratio, a scale factor of F = 20 was determined, as expressed in Equation (3).
where ∆ is the change in initial specific volume, and ∆ is the change in void ratio. The scale factor parameter for three types of PVD is shown in Table 2.   The main apparatus of the large-scale consolidation test is illustrated in Figure 2, and functions and remarks are provided in Table 3. The chamber apparatus material was made of a thick glass cell with a 460 mm internal diameter and 700 mm height. Furthermore, the axial loading system was an air jack-compressor system with a maximum axial loading capacity of 1000 kPa. A vacuum pump that could generate up to 100 kPa suction pressure was employed at the top of PVD. Pore water pressure (PWP) was placed at the bottom of the chamber and connected directly to the data logger. The LVDT performed the settlement measurement of the axial displacement placed at the top of the acrylic plate. The water tank with a balance was used to measure the volume of water, and it was connected to the center of the PVD connector. In addition, the designed rectangular mandrel was made Appl. Sci. 2023, 13, 6673 4 of 24 from iron steel, with a length of 750 mm and a width of 90 mm; it was used to install the PVD in the ground. of a thick glass cell with a 460 mm internal diameter and 700 mm height. Furthermore, the axial loading system was an air jack-compressor system with a maximum axial loading capacity of 1000 kPa. A vacuum pump that could generate up to 100 kPa suction pressure was employed at the top of PVD. Pore water pressure (PWP) was placed at the bottom of the chamber and connected directly to the data logger. The LVDT performed the settlement measurement of the axial displacement placed at the top of the acrylic plate. The water tank with a balance was used to measure the volume of water, and it was connected to the center of the PVD connector. In addition, the designed rectangular mandrel was made from iron steel, with a length of 750 mm and a width of 90 mm; it was used to install the PVD in the ground.     The mandrel has an influence on the resulting smear zone for the equivalent drain radius of PVD, as proposed in the literature [22][23][24][25]. Furthermore, Equations (4)-(6) were used to describe the drain radius of the smear zone's influence on PVD with mandrel in this study, shown in Figure 3. The mandrel has an influence on the resulting smear zone for the equivalent drain radius of PVD, as proposed in the literature [22][23][24][25]. Furthermore, Equations (4)-(6) were used to describe the drain radius of the smear zone's influence on PVD with mandrel in this study, shown in Figure 3.   The model in a general form is expressed in Equation (7).
where r w.eq is the radius of an equivalent drain system, a is the mandrel width, x i represents different types of PVD thickness variants, and y i is the factor variant of PVD. In this study, x 1 is the thickness of PVD type A, x 2 is the thickness of PVD type B, x 3 is the thickness of PVD type C, and r w.eq was determined as 30 mm on the basis of the designed rectangular mandrel.

The Experiment Procedures
A two-step preloading test was conducted. The first step was pre-consolidation with 20 kPa using a loading system. The second step was applying vacuum suction at 100 kPa during the consolidation process. The setup apparatus is shown in Figure 4.
where . is the radius of an equivalent drain system, is the mandrel width, represents different types of PVD thickness variants, and is the factor variant of PVD. In this study, is the thickness of PVD type A, is the thickness of PVD type B, is the thickness of PVD type C, and . was determined as 30 mm on the basis of the designed rectangular mandrel.

The Experiment Procedures
A two-step preloading test was conducted. The first step was pre-consolidation with 20 kPa using a loading system. The second step was applying vacuum suction at 100 kPa during the consolidation process. The setup apparatus is shown in Figure 4.  All tests were carried out in three main stages: (1) preparation of soil samples, (2) drain installation, and (3) specimen sample collection after the process of consolidation. Firstly, a big mixer thoroughly mixed the soil sample with a water level slightly greater than the liquid limit; then, the sample was set up into the chamber. Controlling the water content and controlling saturation are common problems. Secondly, in order to optimize the method [26] using the mandrel system, the PVD was installed with the driving motor system. The specimen samples were collected after the consolidation.
The distribution of vacuum pressure along the drain boundary varied linearly from (−p0) at the top of the drain to (k p0 ) at the bottom of the drain, where vacuum pressure was applied through the PVD. The suction head could decrease with depth and move laterally, resulting in a constant permeability smear zone (k s ) and horizontal permeability (k h ); hence, a trapezoidal vacuum pressure was assumed [27]. The detailed distribution patterns of vacuum pressure are presented in Figure 5a. than the liquid limit; then, the sample was set up into the chamber. Controlling the water content and controlling saturation are common problems. Secondly, in order to optimize the method [26] using the mandrel system, the PVD was installed with the driving motor system. The specimen samples were collected after the consolidation.
The distribution of vacuum pressure along the drain boundary varied linearly from (− 0) at the top of the drain to ( ) at the bottom of the drain, where vacuum pressure was applied through the PVD. The suction head could decrease with depth and move laterally, resulting in a constant permeability smear zone ( ) and horizontal permeability ( ); hence, a trapezoidal vacuum pressure was assumed [27]. The detailed distribution patterns of vacuum pressure are presented in Figure 5a. The three-zone hypothesis was introduced in a previous study [28]. The locations of specimen samples for microstructure analysis were determined on the basis of the water content reduction in the top, middle, and bottom areas. The total number of samples (including their average values) collected before PVD installation for tests on water content,

Sampling Method
The locations of specimen samples illustrated in the sampling area are shown in Figure 5b. A total of 180 specimens of each sample were used in sections A-A and B-B after consolidation. The diameter of each sample was 30 mm. The water content of the samples was measured; during the test, pore water pressure and drainage were monitored.
The three-zone hypothesis was introduced in a previous study [28]. The locations of specimen samples for microstructure analysis were determined on the basis of the water content reduction in the top, middle, and bottom areas. The total number of samples (including their average values) collected before PVD installation for tests on water content, vane shear strength, and SEM is shown in Table 4, while the number of samples collected after PVD installation is shown in Table 5.   PVD  3  3  3  3   Soil structure  Inner smear zone  3  3  3  3  Transition zone  3  3  3  3  Undisturbed zone  3  3  3  3 Appl. Sci. 2023, 13, 6673 8 of 24

Vane Shear Apparatus
The vane shear equipment was used to measure the undrained shear strength before and after consolidation tests based on ASTM 4648 [29]. The vane shear blade was made of stainless steel; it was 20 mm in diameter and 40 mm in height. The adjustable stainless-steel rod diameter was 5 mm, capable of measuring the shear strengths at different locations; its height was 500 mm from the top of the chamber, 300 mm from the middle, and 100 mm from the bottom area. In the radial distance based on the three-zone hypothesis, the inner smear zone was 9.25 mm, transition zone was 46.25 mm, and undisturbed zone was 145.5 mm.
The test procedure of the vane shear test is described as follows: firstly, the apparatus was cleaned, and grease was applied to the lead screw for smoother handle movement. Subsequently, the soil surface was leveled, and the container was mounted on the base of the vane shear test apparatus using the provided screws. Gradually, the vane was lowered into the soil specimen until it was 10 to 20 mm below the surface, and the initial reading of the pointer was recorded on the circular graduated scale. The vane was rotated inside the soil specimen using the torque-applying handle at a rate of 0.1 • per second. When the specimen fails, the strain indicator pointer moves backward on the circular graduated scale. At this point, the test must be stopped, and the final reading of pointer is noted. Lastly, the procedure must be repeated at different specified points. The details of the vane shear testing apparatus are shown in Figure 6, and the shear strength could be obtained by calculation using Equations (8) and (9).
where T is the torque, S is the shear strength of soil, D is the diameter of the vane (mm), and H is the height of the vane (mm).
where is the torque, is the shear strength of soil, is the diameter of the vane (mm), and is the height of the vane (mm).

SEM Method
The SEM equipment used for observing the microstructure of PVD and soil due to installation and vacuum preloading was the HITACHI SU3500, as shown in Figure 7. After consolidation testing, SEM analysis was performed to investigate the structure of soil and PVD interface side in three types of PVD conditions.

SEM Method
The SEM equipment used for observing the microstructure of PVD and soil due to installation and vacuum preloading was the HITACHI SU3500, as shown in Figure 7. After consolidation testing, SEM analysis was performed to investigate the structure of soil and PVD interface side in three types of PVD conditions. The first step was to cut the PVD filter to size 10 mm × 10 mm before and after the consolidations at specified positions. Soil samples before and after the test were also prepared in the three-zone area with different depths. The second step involved a PVD filter brought to the Central of Advanced Instrumental Analysis (CAIA) laboratory to conduct SEM tests. After these two steps, the specimens were ready for testing. Subsequently, the samples were added to a plate with a diameter of 5 cm to capture pictures for observation. Furthermore, the main step was to put the samples into the SEM machine and start with a low-vacuum setting. The last step was to analyze structure of the PVD and soil from the screen (units = µm), captured with ×300, ×500, and ×3000 magnification and a voltage of 15 kV. The most appropriate image for interpreting the overall morphological characteristics of the PVD, soil and pore microstructure, and their respective ratios was selected on the basis of the magnification results.

Procedures of Microstructure Analysis
In this study, ImageJ was utilized for image processing, as shown in Figure 8, and the microstructure was characterized using quantitative analysis. The objective of image processing was to extract information from the image and the output data of pores and particles. Observations of pore diameters are a basic feature in ImageJ. The ratio of the measured area of the pores to the total area of the image can be expressed using Equation (10).
where is the total area of soil pores as a percentage, is the total area of soil pores, ′ is total area of PVD and pores as a percentage, is the total area of PVD and pores, and is the total area of the SEM image. The first step was to cut the PVD filter to size 10 mm × 10 mm before and after the consolidations at specified positions. Soil samples before and after the test were also prepared in the three-zone area with different depths. The second step involved a PVD filter brought to the Central of Advanced Instrumental Analysis (CAIA) laboratory to conduct SEM tests. After these two steps, the specimens were ready for testing. Subsequently, the samples were added to a plate with a diameter of 5 cm to capture pictures for observation. Furthermore, the main step was to put the samples into the SEM machine and start with a low-vacuum setting. The last step was to analyze structure of the PVD and soil from the screen (units = µm), captured with ×300, ×500, and ×3000 magnification and a voltage of 15 kV. The most appropriate image for interpreting the overall morphological characteristics of the PVD, soil and pore microstructure, and their respective ratios was selected on the basis of the magnification results.

Procedures of Microstructure Analysis
In this study, ImageJ was utilized for image processing, as shown in Figure 8, and the microstructure was characterized using quantitative analysis. The objective of image processing was to extract information from the image and the output data of pores and particles. Observations of pore diameters are a basic feature in ImageJ. The ratio of the measured area of the pores to the total area of the image can be expressed using Equation (10).
n (%) = n 0 n total area × 100 (11) where n is the total area of soil pores as a percentage, n 0 is the total area of soil pores, n is total area of PVD and pores as a percentage, n 0 is the total area of PVD and pores, and n total is the total area of the SEM image.

Water Content Reduction Effect of Preloading
This study examined the extent of the smear zone from the water content reduction 24 h after the initial water content of 135%. For this purpose, samples were collected vertically and horizontally from the chamber apparatus depth. The results from the three types of PVD are shown in Figure 9.

Water Content Reduction Effect of Preloading
This study examined the extent of the smear zone from the water content reduction 24 h after the initial water content of 135%. For this purpose, samples were collected vertically and horizontally from the chamber apparatus depth. The results from the three types of PVD are shown in Figure 9.

Water Content Reduction Effect of Preloading
This study examined the extent of the smear zone from the water content reduction 24 h after the initial water content of 135%. For this purpose, samples were collected vertically and horizontally from the chamber apparatus depth. The results from the three types of PVD are shown in Figure 9.  As a result, the water content in Type A at the 18.5 mm and 55.5 mm radial distances decreased vertically and horizontally by 21-37%.
It is indicated that the radial distance of the PVD has a direct relationship with the water content. For Type B, a more uniform pattern was observed than Type A, with water content decreasing by 25-36% at 18.5 mm and 55.5 mm radial distances. On the other hand, the water content in Type C decreased by 30-47% at radial distances of 18.5 mm and 55.5 mm. The difference in Type C is that the water content significantly decreased around the middle of the chamber at 300 mm, with the lowest water content at 45% at the radial distance of 55 mm, resulting in a local buckling. However, the uniformity of the result shows that Type B was the least disturbed PVD among all three.
From this study, it was found that the smallest smear zone occurred during Type B installation because of the effect of decreasing water content. Moreover, when the compressive load was generated and the sudden deflection changed the PVD structure, the effects from buckling and migrating soil particles were small. Therefore, the differences in PVD shape influenced the soil disturbance during installation with the mandrel, consequently decreasing the water content.
For a PVD that works ideally, the water content should decrease more in the area smear and transition zone, or near the PVD, compared with the undisturbed zone. However, there are times when water content does not significantly change at a given depth in all radial locations as expected, implying reduced performance of the PVD. The average water content in the transition area of the extended smear zone as a function of the radial distance due to the effect of the mandrel is shown in Figure 10. As a result, the water content in Type A at the 18.5 mm and 55.5 mm radial distances decreased vertically and horizontally by 21-37%.
It is indicated that the radial distance of the PVD has a direct relationship with the water content. For Type B, a more uniform pattern was observed than Type A, with water content decreasing by 25-36% at 18.5 mm and 55.5 mm radial distances. On the other hand, the water content in Type C decreased by 30-47% at radial distances of 18.5 mm and 55.5 mm. The difference in Type C is that the water content significantly decreased around the middle of the chamber at 300 mm, with the lowest water content at 45% at the radial distance of 55 mm, resulting in a local buckling. However, the uniformity of the result shows that Type B was the least disturbed PVD among all three.
From this study, it was found that the smallest smear zone occurred during Type B installation because of the effect of decreasing water content. Moreover, when the compressive load was generated and the sudden deflection changed the PVD structure, the effects from buckling and migrating soil particles were small. Therefore, the differences in PVD shape influenced the soil disturbance during installation with the mandrel, consequently decreasing the water content.
For a PVD that works ideally, the water content should decrease more in the area smear and transition zone, or near the PVD, compared with the undisturbed zone. However, there are times when water content does not significantly change at a given depth in all radial locations as expected, implying reduced performance of the PVD. The average water content in the transition area of the extended smear zone as a function of the radial distance due to the effect of the mandrel is shown in Figure 10. The average water content in the area close to the PVD was increased. The indication of reduced water content ( ) at 55.5 mm radial distance near the PVD clarifies that the inner smear zone ( ) at 18.55 mm had nonuniform discharging water content. Therefore, the results were in accordance with the analysis of the PVD depth. However, analyzing the water content due to vacuum preloading proved that the effect on soil surrounding the PVD differed at each measured height and radial distance.

Shear Strength of the Soil
The PVD combined with vacuum preloading improved the shear strength of soft soil [30]. The increased shear strength of soft clay deposits was subjected to a constant increase in effective vertical stress with depth, resulting from vacuum preloading in the field measurements [31,32]. In this study, the shear strength was measured in a specified area before and after vacuum preloading in the inner smear zone, transition zone, and undisturbed zone. The average water content in the area close to the PVD was increased. The indication of reduced water content (r s ) at 55.5 mm radial distance near the PVD clarifies that the inner smear zone (s) at 18.55 mm had nonuniform discharging water content. Therefore, the results were in accordance with the analysis of the PVD depth. However, analyzing the water content due to vacuum preloading proved that the effect on soil surrounding the PVD differed at each measured height and radial distance.

Shear Strength of the Soil
The PVD combined with vacuum preloading improved the shear strength of soft soil [30]. The increased shear strength of soft clay deposits was subjected to a constant increase in effective vertical stress with depth, resulting from vacuum preloading in the field measurements [31,32]. In this study, the shear strength was measured in a specified area before and after vacuum preloading in the inner smear zone, transition zone, and undisturbed zone.
The initial condition of the soft ground was measured using the vane shear test, and the results indicated a shear strength of 2 kPa. Furthermore, after consolidation, an increase in shear strength in the area close to the PVD was determined for the three types of PVD. It can be seen from Figure 11 that the shear strength was affected by radial distance at each location and depth. The initial condition of the soft ground was measured using the vane shear test, and the results indicated a shear strength of 2 kPa. Furthermore, after consolidation, an increase in shear strength in the area close to the PVD was determined for the three types of PVD. It can be seen from Figure 11 that the shear strength was affected by radial distance at each location and depth. After vacuum preloading, the variation in shear strength for the three types of PVD increased from 2 kPa to 14 kPa at different depths and radial distances. PVD type B led to an increase in shear strength from 10 kPa to 14 kPa in the smear zone at all depths compared with type A, where shear strength increased from 6.8 kPa to 14 kPa, and type C, where shear strength increased from 7 kPa to 14 kPa.
The soil shear strength decreased with increasing radial distance from the PVD, whereas it was highest in the soil close to the PVD, in line with the water content distribution properties. Furthermore, the study found that the PVD worked more effectively in the inner smear zone and transition zone than in the undisturbed zone with decreasing radial distance. The following general relationship was determined for the three types of PVD: where is the slope of the shear strength of soil, is the radial distance of the actual distance, and is the initial shear strength. After vacuum preloading, the variation in shear strength for the three types of PVD increased from 2 kPa to 14 kPa at different depths and radial distances. PVD type B led to an increase in shear strength from 10 kPa to 14 kPa in the smear zone at all depths compared with type A, where shear strength increased from 6.8 kPa to 14 kPa, and type C, where shear strength increased from 7 kPa to 14 kPa.
The soil shear strength decreased with increasing radial distance from the PVD, whereas it was highest in the soil close to the PVD, in line with the water content distribution properties. Furthermore, the study found that the PVD worked more effectively in the inner smear zone and transition zone than in the undisturbed zone with decreasing radial distance. The following general relationship was determined for the three types of PVD: s u.cal = β ln r d0 + s u0 (12) where β is the slope of the shear strength of soil, r d0 is the radial distance of the actual distance, and s u0 is the initial shear strength.
The soil strength decreased significantly after being disturbed, which indicated that the structure of the soft soil was improved considerably by the reagent after vacuum preloading. The trend observed for the vane shear strength was comparable with the reduction in water content, indicating better consolidation in the inner smear zone and transition zone compared to the undisturbed zone. PVD filter and pore interaction before vacuum preloading A PVD is typically composed of a geotextile filter-wrapped plastic band with molded channels designed to aid in the consolidation of soils. These serve as drainage pathways for pore water in soft compressible soils that consolidate more quickly when subjected to continuous vacuum preloading [33]. The PVD filter should meet the filter design criteria as they are important for enhancing the PVD performance. A commonly used criterion was given in previous studies [34][35][36][37].
The SEM pictures from the three types of PVD before vacuum preloading showed the arrangement of the straight filter and pore area. At 500× magnification, microscopic images in representative sections of the three types of PVD were captured, as shown in Figure 12. The pore size or the apparent opening size (AOS) must be small enough to prevent the fine particles of the soil from entering the filter and the drain. The standards [34,35] commonly used can be expressed as O 50 ≤ (10 − 12)D 50 (15) where O 95 is the AOS of the filter, D 85 is the size larger than 50% of the fabric pores, and D 85 and D 50 refer to the sizes for 85% and 50% of particles in the soil by weight. O 95 ≤ 0.075 m (or 75 mm) is often specified for a prefabricated vertical drain. As a result, to achieve the goals of the vacuum preloading effect, it is critical to select a suitable pore size of the filter. The soil strength decreased significantly after being disturbed, which indicated that the structure of the soft soil was improved considerably by the reagent after vacuum preloading. The trend observed for the vane shear strength was comparable with the reduction in water content, indicating better consolidation in the inner smear zone and transition zone compared to the undisturbed zone.

Microstructure Analysis
4.2.1. PVD Structure 1. PVD filter and pore interaction before vacuum preloading A PVD is typically composed of a geotextile filter-wrapped plastic band with molded channels designed to aid in the consolidation of soils. These serve as drainage pathways for pore water in soft compressible soils that consolidate more quickly when subjected to continuous vacuum preloading [33]. The PVD filter should meet the filter design criteria as they are important for enhancing the PVD performance. A commonly used criterion was given in previous studies [34][35][36][37].
The SEM pictures from the three types of PVD before vacuum preloading showed the arrangement of the straight filter and pore area. At 500× magnification, microscopic images in representative sections of the three types of PVD were captured, as shown in Figure 12. The pore size or the apparent opening size (AOS) must be small enough to prevent the fine particles of the soil from entering the filter and the drain. The standards [34,35] commonly used can be expressed as (10 − 12) where is the AOS of the filter, is the size larger than 50% of the fabric pores, and and refer to the sizes for 85% and 50% of particles in the soil by weight. ≤ 0.075 m (or 75 mm) is often specified for a prefabricated vertical drain. As a result, to achieve the goals of the vacuum preloading effect, it is critical to select a suitable pore size of the filter. In addition, to interpret the SEM images, the pores and PVD filter size in the area were analyzed as shown in Figure 13. In addition, to interpret the SEM images, the pores and PVD filter size in the area were analyzed as shown in Figure 13.
The results show that the pore size area of the filter for PVD types A, B, and C varied in the range of 0.1-12.136 µm. The average O 95 values for PVD filters types A, B, and C of 83, 116, and 265 µm were employed in this study according to the criterion proposed in [35]. The results show that the pore size area of the filter for PVD types A, B, and C varied in the range of 0.1-12.136 µm. The average values for PVD filters types A, B, and C of 83, 116, and 265 µm were employed in this study according to the criterion proposed in [35].
In addition, quantitative analysis of the soil microstructure area showed that PVD type B had a large filter area of 55% and a pore area of 45%, whereas types A and C had a filter area of 43% and a higher pore area of 57%. Therefore, the SEM images could characterize PVD filters with different pore sizes. It is essential to observe the PVD type, and a large PVD filter ensures a higher drainage rate.
A large PVD filter area in percentage means that more filter material is present per unit area of soil, thus increasing the filtering capacity of the system and helping prevent soil particles from clogging the PVD [38]. However, if the pore area is smaller in percentage, there is less space available for water to flow through the soil and into the PVDs; this can reduce the overall drainage capacity of the system and slow down the consolidation process.
The balance between the PVD filter and pore areas is important in designing an effective PVD system. The filter material must be sufficient to prevent clogging and maintain long-term performance. At the same time, the available pore space must be adequate to allow for efficient drainage and consolidation. Therefore, selecting the PVD type is essential to ensure its successful implementation in the field. According to this study, PVD type B over types A and C according to the PVD filter structure.

The PVD filter and pore interaction after vacuum preloading
The interaction between the PVD filter and the pores, also known as clogging, which is caused by vacuum preloading, depends on several factors, including the properties of the soil, the filter material, and the drainage capacity of the system [38][39][40]. On the other hand, if the filter material has a low permeability or the PVD system is not designed correctly, the interaction between the PVD filter and pores may not be optimal [41]. This can lead to clogging of the filter and a reduction in the drainage capacity of the system. If this occurs, the consolidation process may be slowed down or ineffective, leading to potential long-term stability issues.
Therefore, it is crucial to understand the interaction between the PVD filter and the pores caused by vacuum preloading to optimize the design and properties of the system accordingly. Thus, the effectiveness of the PVD system during installation using a mandrel was ensured in this study. Figure 14 shows the structure of the PVD filter and pore changes caused by vacuum preloading with reference to deformation modes [18]. In addition, quantitative analysis of the soil microstructure area showed that PVD type B had a large filter area of 55% and a pore area of 45%, whereas types A and C had a filter area of 43% and a higher pore area of 57%. Therefore, the SEM images could characterize PVD filters with different pore sizes. It is essential to observe the PVD type, and a large PVD filter ensures a higher drainage rate.
A large PVD filter area in percentage means that more filter material is present per unit area of soil, thus increasing the filtering capacity of the system and helping prevent soil particles from clogging the PVD [38]. However, if the pore area is smaller in percentage, there is less space available for water to flow through the soil and into the PVDs; this can reduce the overall drainage capacity of the system and slow down the consolidation process.
The balance between the PVD filter and pore areas is important in designing an effective PVD system. The filter material must be sufficient to prevent clogging and maintain long-term performance. At the same time, the available pore space must be adequate to allow for efficient drainage and consolidation. Therefore, selecting the PVD type is essential to ensure its successful implementation in the field. According to this study, PVD type B over types A and C according to the PVD filter structure.

2.
The PVD filter and pore interaction after vacuum preloading The interaction between the PVD filter and the pores, also known as clogging, which is caused by vacuum preloading, depends on several factors, including the properties of the soil, the filter material, and the drainage capacity of the system [38][39][40]. On the other hand, if the filter material has a low permeability or the PVD system is not designed correctly, the interaction between the PVD filter and pores may not be optimal [41]. This can lead to clogging of the filter and a reduction in the drainage capacity of the system. If this occurs, the consolidation process may be slowed down or ineffective, leading to potential long-term stability issues.
Therefore, it is crucial to understand the interaction between the PVD filter and the pores caused by vacuum preloading to optimize the design and properties of the system accordingly. Thus, the effectiveness of the PVD system during installation using a mandrel was ensured in this study. Figure 14 shows the structure of the PVD filter and pore changes caused by vacuum preloading with reference to deformation modes [18].
After vacuum preloading, SEM images of both the soil-facing and drain channelfacing sides were captured. Additionally, the opening in the filter was filled with fine soil particles on both sides, with notable differences existing. On the side contacting the soil, more fine particles filled the filter interspacing. The filter structure changed due to the effect of vacuum preloading. In this study, the approach used was based on the shape of the deformation effect of the PVD. The PVD filter could present one of three different structures after consolidation, namely, uniform, local, and sinusoidal bending, referring to the deformation modes in a previous study [42]. Appl  After vacuum preloading, SEM images of both the soil-facing and drain channel-facing sides were captured. Additionally, the opening in the filter was filled with fine soil particles on both sides, with notable differences existing. On the side contacting the soil, more fine particles filled the filter interspacing. The filter structure changed due to the effect of vacuum preloading. In this study, the approach used was based on the shape of the deformation effect of the PVD. The PVD filter could present one of three different structures after consolidation, namely, uniform, local, and sinusoidal bending, referring to the deformation modes in a previous study [42].
For more detail, Figure 15a shows the quantitative analysis results in terms of area and percentage to understand the effectiveness of PVD depth. The results show that, in the bottom area of PVDs, the particle area was 4.3 × 10 4 µm 2 and pore area was 8.9 × 10 4 µm 2 for type A, 4.9 × 10 4 µm 2 and 8.4 × 10 4 µm 2 for type B, and 4.7 × 10 4 µm 2 and 8.6 × 10 4 µm 2 for type C. These results show that the bottom area of all PVD types in this study had a higher particle area due to vacuum preloading, and that using the PVD filter decreased the pore size. These results indicate that the fine particles moved to fill the areas with large For more detail, Figure 15a shows the quantitative analysis results in terms of area and percentage to understand the effectiveness of PVD depth. The results show that, in the bottom area of PVDs, the particle area was 4.3 × 10 4 µm 2 and pore area was 8.9 × 10 4 µm 2 for type A, 4.9 × 10 4 µm 2 and 8.4 × 10 4 µm 2 for type B, and 4.7 × 10 4 µm 2 and 8.6 × 10 4 µm 2 for type C. These results show that the bottom area of all PVD types in this study had a higher particle area due to vacuum preloading, and that using the PVD filter decreased the pore size. These results indicate that the fine particles moved to fill the areas with large pores during vacuum preloading. This can be considered a clogging effect [38], as also proven by D 85 being larger than 50% of n in Figure 15b. pores during vacuum preloading. This can be considered a clogging effect [38], as also proven by being larger than 50% of n′ in Figure 15b. The general form of the PVD filter, pore, and particle areas was analyzed using Im-ageJ and expressed as follows: where is the pore area, ′ is the PVD and pore area, α is the slope of the pore area, ∅ is the particle area, and ′ is the initial PVD and pore area before the test. In summary, the clogging effect on PVD filters caused by vacuum preloading was influenced by the filter's geometry, stiffness, and surrounding soil properties. The design and properties of the PVD filter should be optimized to ensure that it can withstand external loads and maintain its stability during installation and use.

Soil structure before vacuum preloading
Kaolinite is a clay mineral that can form a soft soil with high water content. In Figure  16, an SEM image of the soil before the consolidation test is shown, and the corresponding size distribution is described in Table 6. The SEM image of kaolinite specimens with different fabrics is consistent with the conclusion obtained from the sedimentation test [43,44]. In addition, the samples varied in several factors, such as the sample preparation method, imaging conditions, and the nature of the material [45]. The general form of the PVD filter, pore, and particle areas was analyzed using ImageJ and expressed as follows: where n is the pore area, n is the PVD and pore area, α is the slope of the pore area, ∅ is the particle area, and n 0 is the initial PVD and pore area before the test. In summary, the clogging effect on PVD filters caused by vacuum preloading was influenced by the filter's geometry, stiffness, and surrounding soil properties. The design and properties of the PVD filter should be optimized to ensure that it can withstand external loads and maintain its stability during installation and use.

1.
Soil structure before vacuum preloading Kaolinite is a clay mineral that can form a soft soil with high water content. In Figure 16, an SEM image of the soil before the consolidation test is shown, and the corresponding size distribution is described in Table 6. The SEM image of kaolinite specimens with different fabrics is consistent with the conclusion obtained from the sedimentation test [43,44]. In addition, the samples varied in several factors, such as the sample preparation method, imaging conditions, and the nature of the material [45].    In this study, SEM images showed a relatively smooth and uniform surface at lower magnifications, with some voids and visible pores. The edges of the particles appeared slightly rounded or irregular due to the presence of water. At higher magnifications, the SEM image revealed the platy structure of the kaolinite particles, which were stacked or arranged in a random orientation. The particles were tightly packed together, with little space between them, or more loosely packed, with visible voids and gaps.
In some areas, the SEM image showed small aggregates of particles, indicating the presence of localized bonding between the particles. The aggregates were irregular in shape, and the surfaces showed some textural variation, including cracks, fissures, or other features. Overall, the structure of kaolinite soft soil with high water content appeared relatively homogeneous and uniform, with some variability in the particle arrangement and texture visible under higher magnification. The SEM image provides valuable information about the microstructure and properties of the soil, which can help to inform engineering and construction decisions.

2.
Soil structure after vacuum preloading SEM images were used to study the changes in soil structure due to vacuum preloading. SEM can provide high-resolution images of soil particles and their arrangements, allowing us to observe the changes in the soil microstructure due to consolidation [46][47][48].
Taking into consideration the PVD, pore, and particle results, the bottom area of the PVD was determined as a clogging area. Further analysis was necessary to understand the effect of PVD in the three areas. Figure 17 shows the SEM images from the three areas, highlighting the effect of PVD after vacuum preloading on the soil structure. The SEM images show that the soil particles became more closely packed after vacuum preloading.
The soil particles became more tightly interlocked, resulting in a denser soil structure with face-to-face contact between particles and domains [49]. Furthermore, the morphology of soil particles was affected by vacuum preloading [50,51]. For example, clay particles became more angular or irregularly shaped, indicating that vacuum preloading altered the soil particle properties.
To explain these results more clearly, an analytical approach was carried out in this study, and the results are shown in Figure 18. It can be seen that the pore and particle areas for PVD type B were 0.5 × 10 3 µm 2 and 0.7 × 10 3 µm 2 in the inner smear zone, 0.3 × 10 3 µm 2 and 1.1 × 10 3 µm 2 in the transition zone, and 0.7 × 10 3 µm and 0.7 × 10 3 µm in the undisturbed zone. The corresponding values for type A were 1.2 × 10 3 µm 2 and 2.4 × 10 3 µm 2 , 0.9 × 10 3 µm 2 and 2.7 × 10 3 µm 2 , and 0.8 × 10 3 µm 2 and 2.7 × 10 3 µm 2 , respectively. The corresponding values for type C were 1.2 × 10 3 µm 2 and 2.4 × 10 3 µm 2 , 0.9 × 10 3 µm 2 and 2.7 × 10 3 µm 2 , and 1.9 × 10 3 µm 2 and 1.7 × 10 3 µm 2 , respectively. The results show that the particle area increased whereas the pore area decreased for all types of PVD. However, the particle size used in PVD type B was smaller than that used in PVD types A and C. The results indicate that PVD type B worked optimally during consolidation due to the clogging effect. Furthermore, when using type B, the particle area increased by 40-80%. The soil particles became more tightly interlocked, resulting in a denser soil structure with face-to-face contact between particles and domains [49]. Furthermore, the morphology of soil particles was affected by vacuum preloading [50,51]. For example, clay particles became more angular or irregularly shaped, indicating that vacuum preloading altered the soil particle properties.
To explain these results more clearly, an analytical approach was carried out in this study, and the results are shown in Figure 18. It can be seen that the pore and particle areas for PVD type B were 0.5 × 10 3 µm 2 and 0.7 × 10 3 µm 2 in the inner smear zone, 0.3 × 10 3 µm 2 and 1.1 × 10 3 µm 2 in the transition zone, and 0.7 × 10 3 µm and 0.7 × 10 3 µm in the undisturbed zone. The corresponding values for type A were 1.2 × 10 3 µm 2 and 2.4 × 10 3 µm 2 , 0.9 × 10 3 µm 2 and 2.7 × 10 3 µm 2 , and 0.8 × 10 3 µm 2 and 2.7 × 10 3 µm 2 , respectively. The corresponding values for type C were 1.2 × 10 3 µm 2 and 2.4 × 10 3 µm 2 , 0.9 × 10 3 µm 2 and 2.7 × 10 3 µm 2 , and 1.9 × 10 3 µm 2 and 1.7 × 10 3 µm 2 , respectively. The results show that the particle area increased whereas the pore area decreased for all types of PVD. However, the particle size used in PVD type B was smaller than that used in PVD types A and C. The results indicate that PVD type B worked optimally during consolidation due to the clogging effect. Furthermore, when using type B, the particle area increased by 40-80%. This study indicated that a reduction in pores and particle size impacted the soil structure. SEM images revealed the changes in the soil fabric and provided insights into how these changes influenced the soil properties. Overall, SEM can provide valuable insights into the changes in soil microstructure caused by vacuum preloading. This information can be used to understand the mechanisms of vacuum preloading and to optimize the technique for specific soil conditions. This study indicated that a reduction in pores and particle size impacted the soil structure. SEM images revealed the changes in the soil fabric and provided insights into how these changes influenced the soil properties. Overall, SEM can provide valuable insights into the changes in soil microstructure caused by vacuum preloading. This information can be used to understand the mechanisms of vacuum preloading and to optimize the technique for specific soil conditions.

Correlations of Shear Strength and Pores of the Soil Structure
The relationship between the shear strength and pore structure of a soil structure can be influenced by several factors, such as soil composition, particle size, soil structure, and moisture content [52]. In general, the shear strength of soil is related to its depth and distance [53], and it is influenced by the interlocking of soil particles and the bonding between them, which can be affected by the pore structure. Therefore, a reduction in pores can improve soil stability and increase the shear strength of soil [54]. To determine the shear strength of the soil in the three zones around the PVD installation, the number of pores in the soil structure was measured and correlated with the data in Figure 19. This study indicated that a reduction in pores and particle size impacted the soil structure. SEM images revealed the changes in the soil fabric and provided insights into how these changes influenced the soil properties. Overall, SEM can provide valuable insights into the changes in soil microstructure caused by vacuum preloading. This information can be used to understand the mechanisms of vacuum preloading and to optimize the technique for specific soil conditions.

Correlations of Shear Strength and Pores of the Soil Structure
The relationship between the shear strength and pore structure of a soil structure can be influenced by several factors, such as soil composition, particle size, soil structure, and moisture content [52]. In general, the shear strength of soil is related to its depth and distance [53], and it is influenced by the interlocking of soil particles and the bonding between them, which can be affected by the pore structure. Therefore, a reduction in pores can improve soil stability and increase the shear strength of soil [54]. To determine the shear strength of the soil in the three zones around the PVD installation, the number of pores in the soil structure was measured and correlated with the data in Figure 19. The results in this study indicate that PVD type B in the inner smear zone had more influence on the shear strength of soil compared with the transition zone and undisturbed zone. Type C exhibited a large shear strength of soil in the inner smear zone, transition zone, and undisturbed zone, with the former two having a larger impact than the latter. The results in this study indicate that PVD type B in the inner smear zone had more influence on the shear strength of soil compared with the transition zone and undisturbed zone. Type C exhibited a large shear strength of soil in the inner smear zone, transition zone, and undisturbed zone, with the former two having a larger impact than the latter. The results of this experiment, similar to the findings in [55], revealed an increasing edgeto-face orientation of the microstructures after consolidation with increasing levels of PVD installation.
Smaller pores can increase the interlocking between soil particles and improve the shear strength. On the other hand, the soil pore structure can also affect its hydraulic properties, such as permeability and water retention characteristics. Large pores can increase the soil's permeability, allowing water to flow more easily. However, this can also lead to reduced water retention, which can affect the stability of the soil structure. The presence of smaller pores can improve water retention, which can improve the stability of the soil structure.
The results of this study showed that the relationship between the shear strength and pore structure of the soil is complex and can vary depending on the specific properties of the soil. The relationship between the pore structure and shear strength is a critical consideration in soil engineering and geotechnical analysis, as it can affect the stability of the soil structure and its suitability for various applications.

Conclusions
This study investigated the micromechanical properties and microstructure of soil by examining the effect of a mandrel system using three types of PVD combined with vacuum preloading in soft soil. The inner smear zone, transition zone, and undisturbed zone surrounding the PVD were evaluated. On the basis of the test results, the main findings are summarized as follows: 1.
The large-scale consolidation apparatus efficiently investigated the effect on the inner smear zone, transition zone, and undisturbed zone of the installation and clogging effect by evaluating the water content and shear strength of soil.

2.
The inner smear zone was evaluated by measuring the change in water content surrounding the PVD. In this study, it was found that the water content decreased both vertically and horizontally by 21% to 37% at all radial distance measurements of the inner smear zone and the transition zone. 3.
The highest increase in shear strength (s u ) was achieved in the smear zone and gradually decreased with the radial distances according to the type of PVD.

4.
A microstructure analysis allowed examining the clogging effect by estimating the pore area (n) and the particle area (Ø) of the structure after consolidation.

5.
The correlation between shear strength (s u ) and pore area (n) revealed an inverse relationship, which can characterize the effect of the mandrel system on consolidation.
The practical implication of this study is that optimal mandrel system selection is essential because it can affect the surrounding soil and influence PVD performance. Therefore, finding a suitable mandrel system to reduce the smear zone due to PVD installation is essential for ground improvement in geotechnical engineering design.

Data Availability Statement:
The data that support the findings of this study are available from the corresponding author upon reasonable request.

Acknowledgments:
The authors are grateful to M. Nakashima and S. Nagamatsu for providing technical support. The authors also thank E.H. Gronlund, M. Gobin, R.E.A. Rifa'i, and I.W.A. Basoka for their helpful review of the manuscript.

Conflicts of Interest:
The authors declare no conflict of interest. The embedded length of the PVD prototype (m) z m

Abbreviations
The embedded length of the PVD model (m) n z The scale factor of embedded length k p The permeability of the prototype (m/s) k m The change permeability of the model (m/s) ∆v The change in initial specific volume (m 3 /kg) ∆e The change in void ratio k v Permeability in vertical direction (m/s) k h Permeability in horizontal direction (m/s) k s Permeability in smear zone area (m/s) E Young's modulus (kPa) r w(Type A) The radius of PVD type A (mm) r w (Type B) The radius of PVD type B (mm) r w (Type C) The radius of PVD type C (mm) r w.eq The radius of an equivalent drain system (mm) x 1 The thickness of PVD type A (mm) The thickness of PVD type B (mm) x 3 The thickness of PVD type C (mm) x i Different types of PVD thickness variants (mm) y i The factor variant of PVD −p0 The vacuum pressure (kPa) S Torque D Diameter of vane (mm) H Height of vane (mm) n Total area of soil pores as a percentage (%) n Pore area (µm 2 ) n 0 Total area of soil pores (µm 2 ) n Total area of PVD and pores as a percentage (%) n PVD and pore area (µm 2 ) n 0 Total area of PVD and pores (µm 2 ) n total Total area of SEM image (µm 2 ) s u.cal Initial shear strength (kPa) β Slope of the strength ( • ) r d0 Radial distance (mm) ∅ Particle area (µm) α Slope of the pore area ( • ) s u Vane shear strength (kPa)