Cyclic-Induced Soil Disturbance in Structured Soft Clay: Experimental Evidence from Undisturbed and Reconstituted Specimens

Featured Application

This study provides experimental evidence for the mechanism underlying the soil disturbance concept established by the numerical simulations of previous studies, which they defined as the simultaneous degradation of the specimen’s stiffness and strength. The undisturbed (UDS) and reconstituted (REC) soft clays used in the experiments were derived from the same material, thereby ensuring consistent physical properties across specimens and increasing the reliability of the results. The results of this study synthesize numerical model findings and experimental observations to enhance understanding of soil disturbance.

Abstract

Seismic damage has been observed not only in liquefiable sandy soil layers but also in thick deposits of soft clayey soils, which are characterized by the destruction of the soil structure, leading to strain softening. Previous studies conducted numerical simulations and defined this phenomenon as soil disturbance, which refers to the simultaneous reduction in stiffness and peak shear strength. To fill the research gap, this study systematically compares the post-cyclic degradation behavior of stiffness and peak shear strength of UDS and REC specimens derived from the same material. Based on the experimental results, the peak shear strength and rigidity of the UDS specimens simultaneously decrease, as the number of cycles increases. In contrast, the peak shear strength degradation effect is absent in the REC samples; both specimens exhibited loss in stiffness. The reduction in stiffness of UDS specimens was slower than that of REC specimens due to aging effects. Nevertheless, both effects on UDS and REC specimens are due to soil disturbance, which is defined in the numerical simulations of previous studies. Hence, the effects of soil disturbance can be summarized as (1) a reduction in the initial stiffness of soft clay and (2) a reduction in the mean effective stress during cyclic loading.

1. Introduction

The frequent occurrence of earthquakes in many regions of the world has led to significant global economic losses, averaging USD 39 billion per year from 1970 to 2023. This amount represents 25% of the total global economic losses resulting from geophysical disasters [1]. While liquefaction in saturated sandy ground has been the primary cause of seismic damage, significant structural failures on soft clay foundations have also been reported, including the loss of bearing capacity and settlement during and after earthquakes. Although clay soils are generally considered resistant to seismic damage, they exhibit seismic vulnerability under soft soil conditions. Such clay soils are characterized by their high-water content, high void ratio, and high sensitivity ratio. During the 1985 Mexico City earthquake, foundation failures due to overturning were observed. This was due to a significant reduction in the bearing capacity of the soft clay foundation [2,3]. Furthermore, continuous settlements at multiple foundation points of a specific building in Mexico City, on a soft-clay foundation, were recorded in 1973 following the 1957 Guerrero earthquake [4]. In Japan, accelerated ground-settlement damage was reported shortly after the 2007 Chuetsu Offshore Earthquake [5] and the 2011 earthquake off the Pacific coast of Tohoku [6,7]. Similarly, the affected areas were underlain by soft clay layers. The authors of [8] found that clay soils still accumulate excess pore water pressure under cyclic shear loading, which causes consolidation settlement after it dissipates.

Numerous studies have been conducted to understand the dynamic behavior and properties of soft clay soil, addressing its seismic vulnerability. Undrained cyclic triaxial tests were performed using REC soft clay soils to investigate the effects of the overconsolidation ratio, confining pressure, cyclic loading frequency, and stress ratio on their cyclic shear behavior [9,10,11,12,13]. Based on the results of their experiments, the undrained shear behavior is independent from the loading frequency and duration and is influenced by variations in the confining pressure, consolidation pressure, overconsolidation ratio, displacement rate, and inherent anisotropy. Several studies [14,15,16] conducted monotonic shear tests on REC specimens subjected to undrained cyclic loading. Based on their results, the undrained peak shear strength of soft clay soils decreases temporarily with an increasing number of loading cycles due to excess pore water pressure buildup in the specimens. In contrast, the post-cyclic undrained shear strength of clay soil increased after consolidation. Furthermore, cyclic triaxial tests on UDS clayey soil were also previously conducted, focusing on the cyclic shear behavior and axial strain accumulation [17,18,19]. These studies found that the cyclic shear resistance of remolded clay is significantly lower than that of UDS clay because of rapid axial strain accumulation. This is due to the natural soil structure formed by aging and natural cementation between soil particles during deposition. The authors of [20] investigated the cyclic shear strength of both UDS and remolded marine clays, which focused on strength degradation and axial strain accumulation as the basis. Based on their results, the UDS specimens showed a higher cyclic shear strength and larger axial strain resistance. This is due to the natural soil structure and bonding present in UDS specimens. Furthermore, the authors of [21,22] performed numerical simulations of road embankments and river levees underlain by soft clay soil, demonstrating that cyclic loading induces a simultaneous reduction in the stiffness and peak shear strength, a behavior referred to as the soil disturbance phenomenon. The authors of [22] also revealed that seismic damage occurs in normal weak clays beneath embankments (Figure 1a); however, the extent of damage is more pronounced in soft clay deposits under heavy overburden pressure (Figure 1b). However, these studies rely heavily on soil–water coupled finite-deformation analysis, which oversimplifies the complex behavior and properties of soft clays, including the spatial variability, heterogeneity, and aging effects. Additionally, the effect of soil disturbance was defined solely in terms of the constitutive response of soft soils, without the necessary laboratory validation. These studies also lacked discussion of the differences in the effects of soil disturbance on UDS and REC soft clays.

Considering all this, the study aims to bridge the gap between the experimental evidence and numerical simulations of soil disturbance. This paper provides direct experimental data on the soil disturbance concept, which has been extensively examined through numerical modeling in previous studies. To achieve the main objective, this study conducted post-cyclic undrained shear tests on both UDS and REC soft clay specimens to determine their peak shear strength and stiffness degradation behavior. Cyclic shear tests on UDS and REC specimens until failure were also conducted to determine the axial strain propagation behavior of both specimen conditions. Post-cyclic monotonic shear tests after different amounts of consolidation were also conducted to determine the changes in the peak shear strength of soft clay soil during, immediately after, and after the earthquake to confirm the highly overconsolidated behavior of clay soil after post-cyclic consolidation.

The significance of this study lies in the systematic comparison of the post-cyclic degradation behavior of strength and stiffness between UDS and REC soft clays derived from the same original material. This study experimentally visualizes and quantitatively evaluates soil disturbance as a form of structural degradation characterized by reductions in peak shear strength and initial stiffness. Furthermore, the REC specimens were prepared from the trimmed soil of the UDS samples, which effectively controls the mineralogical composition, particle size distribution, and consistency indices (LL and PI). Consequently, this method enables a more rigorous comparison than in most previous studies, where different materials and depositional histories were often used.

2. Materials and Methods

2.1. Soil Specimen Information

The soil specimen used in this study was collected from Kaizu City, Gifu Prefecture, Japan. Figure 2 shows the exact location of the site where the specimens were extracted. Subsurface investigation was conducted through two boreholes, R5-1 and R5-2, whose stratigraphic profiles are shown in Figure 3. Borehole R5-1 has an approximately 3 m embankment, whereas borehole R5-2 does not. Furthermore, Figure 3 also shows a thick deposit of soft clay soil at depths ranging from 8.5 m to 36.5 m, labeled as the Amc group. Soft clay specimens extracted from depths of 17.5 m to 30 m were selected for laboratory testing, as they exhibited uniform specific gravity (G_s = 2.69), liquid limit (LL = 80% to 85%), and compression index (Cc = 0.21 to 0.23).

2.2. Index Properties of Soft Clay Soil

Grain size analyses and Atterberg Limits tests were performed on the extracted specimens to determine their index properties and corresponding soil classification. The results are summarized in

Table 1. Summary of index properties of Kaizu soft clay.

Index Properties	Values
Specific Gravity, Gs	2.69
% Passing No. 200 Sieve (%)	99.40
Liquid Limit, LL (%)	83
Plastic Limit, PL	33
Plasticity Index, PI	50
USCS Classification	Clay, High Plasticity (CH)

. Moreover, Figure 4 shows the grain-size distribution curve for the soil specimen. Soil classification based on the Japanese Geotechnical Society (JGS) standards and the Unified Soil Classification System (USCS) identified the selected material as high-plasticity clay (CH).

2.3. Consolidation Behavior of Highly Structured Soft Clay

Undisturbed and disturbed soil specimens were used in the Standard Consolidation Test to assess differences in their compression behavior. Figure 5 shows that the UDS specimen exhibits bulking behavior, which is beyond the normal consolidation line (impossible region), unlike the disturbed specimen. This consolidation behavior is innate to UDS soft clay [23]. According to previous studies, the soil’s deviation from the normal consolidation line in the e-log—p′ space is referred to as “structure” [24,25].

2.4. Consolidation Behaviors of UDS and REC Specimens

The consolidation behaviors of the UDS and REC specimens were also determined in this study. The researchers conducted four trials of isotropic consolidation using UDS and REC specimens. Based on the test results, the REC specimens consistently reached convergence in the volumetric strain vs. time graphs significantly faster than the UDS specimens, as shown in one of the trials in Figure 6. This is due to the presence of structure and aging effect in the UDS specimens, which caused the slower development of volumetric and vertical strain.

2.5. Undrained Monotonic Shear Behavior of Normally Consolidated and Overconsolidated UDS and REC Specimens

Undrained monotonic shear tests were conducted under isotropic conditions on both UDS and REC specimens, prepared in normally consolidated (NC) and overconsolidated (OC) states. This is to differentiate the stress–strain and effective stress behavior of UDS and REC soft clay. The OC specimens were tested at a mean effective stress of 45 kPa, while the NC specimens were tested at 180 kPa. Figure 7a shows a decreasing shear capacity, as the axial strain develops. This behavior is due to the degradation of the microstructure of undisturbed soft clay, which develops through long-term deposition and various geological processes, resulting in strain softening. On the other hand, Figure 7b presents the “rewinding and bowing” behavior on the q-p′ graph of the UDS soil specimen against the Critical State Line (CSL), a behavior inherent to soft clay soils [26,27]. Figure 8a shows a continuous increase in shear capacity with increasing deformation, reflecting a strain-hardening response of REC specimens. Figure 8b presents the stress paths of the NC and OC REC specimens, where the absence of the characteristic “rewinding and bowing” behavior can be clearly observed. Moreover, the peak shear strengths of both states have minimal to no reduction, indicating that the REC specimens do not exhibit the strain-softening behavior shown in the UDS specimens. The CSL for these graphs was defined using the relationship, $q=M{p}^{\prime }$, where M is the critical state slope of the soil specimen. The same CSL was used for both UDS and REC specimens since they originated from the same parent material.

Furthermore, Figure 9 shows that the graphs of the NC and OC UDS specimens remain above the Normal Consolidation Line (NCL), whereas those of REC specimens do not. The stress paths of the UDS specimens also did not converge with the CSL after the undrained shear test. This response is associated with the soil structure in UDS specimens, which developed during long-term deposition and geological processes. The NCL and CSL used in the figures were obtained from the one-dimensional consolidation graph of the tested specimens.

2.6. Experimental Procedures and Methodology

This chapter outlines the specimen extraction, preparation procedures, and experimental procedures used in the study. Kaizu soft clay soils were meticulously extracted from thin-walled Shelby tube samplers and cut to the required dimensions (Figure 10a). The specimens were sealed with plastic wrap and coated with wax to minimize air exposure and preserve the samples’ in situ moisture content until laboratory testing (Figure 10b). Undrained cyclic triaxial shear tests, followed by monotonic shear tests, were conducted on both UDS and REC soft clay to define the soil disturbance phenomenon brought by cyclic loading history. The detailed experimental procedures are summarized below.

1.

Specimens are placed in the trimming apparatus per JGS 0520 [28] and carefully cut according to the required dimensions ($\phi =35\mathrm{mm},h=80\mathrm{mm})$, Figure 11a,b.

2.

The soil specimen was then placed in the triaxial apparatus for testing.

3.

The saturation process was then initiated using the stepwise method. After reaching the final interval, the soil specimen was saturated for 30 min.

4.

A B-value of 95% and above is desirable before proceeding to the consolidation phase. This is to ensure that the specimen is fully saturated.

5.

Consolidation process is conducted in isotropic stress conditions with 180 kPa confining pressure, which is based on the actual overburden pressure applied on the specimen at the site.

6.

A minimum of 24 h of consolidation was employed to ensure that the consolidation graph of the specimen had already converged. Cyclic triaxial shearing was then applied at a stress-controlled setup with a constant amplitude of 72 kPa and a period of 20 s.

7.

After cyclic loading, the specimens were left for a sufficient time for the excess pore water pressure in the specimen to become uniform. Monotonic shear was then applied at a constant loading rate of 0.007 mm/min.

In preparing REC specimens, soil trimmings from the UDS specimen were blended with water at 1.5 times the liquid limit using a drum mixer to achieve a uniform slurry. A sieve was used to filter out the remaining broken shells from the slurry. The slurry was then consolidated under a pre-consolidation pressure of 150 kPa in a consolidation chamber until the specimen stopped consolidating or for a minimum of 96 h. The consolidated REC specimen was extracted by removing the chamber top cap and carefully applying sufficient pressure to lift it out. The soil specimen was divided into six blocks by using the same wire trimmer and stored similarly to the UDS specimens. Then, sample preparation and triaxial shear tests were conducted using the same procedures as for the UDS specimens in accordance with JGS 0520 [28].

2.7. Reproducibility of Experiments

One limitation of this study was that only Kaizu soft clay was used in the experiments, due to the limited availability of homogeneous UDS specimens. With that, the researchers conducted reproducibility experiments on UDS specimens from the same borehole at a deeper depth, with constant physical properties, to confirm the reliability of the results. Figure 12 shows the undrained monotonic shear tests on UDS specimens under isotropic stress conditions. It can be observed that all the trials have similar q-${\epsilon }_a$ and q-p′ relationships. Therefore, it can be concluded that the methodology used across all experiments in this study yields consistent results.

Table 2. Physical properties of the UDS specimen after subsequent isotropic monotonic shear test.

Parameters	Trial 1	Trial 2	Trial 3
Initial specific volume, v0	3.031	3.175	3.224
Specific volume, v	2.646	2.701	2.648
Water content (cans), wc	66.57%	75.52%	76.34%
Water content (spec.), ws	75.51%	80.06%	81.46%
Shear strength, qmax (kPa)	151.929	143.407	148.264
qmax reduction	0	−5.6	−2.4

summarizes the essential parameters of the soil specimen used before and after the experiments. Figure 13 presents the histogram of the initial specific volumes of the specimens used in this study, which shows a normal distribution. The calculated standard deviation is 0.096, indicating high precision, and the values are closely clustered around the mean of 3.087.

3. Results

This section presents the experimental results, analyses, and interpretations of the laboratory tests conducted in the study. The results of the experiments on UDS and REC specimens are summarized and compared to understand the soil disturbance phenomenon. Undrained shear strengths of soft clay at different amounts of post-cyclic consolidation are also determined.

3.1. Cyclic Shear Test on UDS Soft Clay

The cyclic shear tests in this study, per JGS 0541, were conducted with a constant stress amplitude of 72 kPa and a period of 20 s, following a stress-controlled setup. Figure 14b presents the stress–strain relationship of the UDS specimen after the cyclic shear test. The results show that axial strain continuously developed on both the compression and extension sides of the UDS clay during cyclic loading. After 1295 cycles, the specimen eventually failed on the extension side, resulting in necking (Figure 14a). This behavior is typically associated with naturally deposited soils due to their weaker extension side. Moreover, the mean effective stress, p′, gradually decreased, as the cyclic loading progressed, reaching approximately 20 kPa at failure (Figure 14c). These findings demonstrate that natural clay experiences a progressive reduction in p′ with the continuous accumulation of axial strain under cyclic loading. The parametric values of the soil specimens before and after are summarized in

Table 3. Physical properties of specimen after cyclic shear test on UDS until failure.

Parameters	Values
Initial specific volume, v0	3.190
Specific volume after consolidation, v	3.092

.

3.2. Subsequent Monotonic Shear Test on UDS Soft Clay

Based on the observed number of cycles to failure, subsequent monotonic shear tests were conducted at two cyclic loading stages, representing low and high cyclic histories. Consolidated-undrained triaxial shear tests were conducted in accordance with JGS 0523 [29]. Initially, the monotonic shear behavior and mechanical properties of the UDS specimen were identified without any prior cyclic loading history. Then, subsequent monotonic shear tests were conducted after 48 and 960 cycles.

Based on the experimental results, there is a noticeable reduction in the peak shear strength of the UDS specimen after 48 and 960 cycles. Figure 15a shows the degradation of the specimen’s structure and stiffness as cyclic loading progressed, with a more pronounced reduction at a higher cyclic loading history. Furthermore, Figure 15b presents the peak shear strength reduction in the specimens under a higher number of cycles.

The UDS specimen without cyclic loading histories recorded a peak shear strength of 162.448 kPa. After 48 cycles, the peak shear strength reduced to 148.501 kPa, corresponding to an 8.59% reduction. After 960 cycles, the peak shear strength further declined to 141.586%, representing a 12.84% reduction relative to the specimen without a cyclic loading history. The parametric values of the soil specimens before and after are summarized in

Table 4. Physical properties of UDS specimen after subsequent monotonic shear test.

Parameters	0 Cycles	48 Cycles	960 Cycles
Initial Specific Volume, v0	2.965	3.152	3.041
Specific Volume, v	2.757	2.783	2.854
Shear Strength, qmax (kPa)	162.448	148.501	141.586
Secant Young’s Modulus, E50	212.628	167.271	57.417

.

3.3. Cyclic Shear Test on REC Soft Clay

Reconstituted specimens were prepared following the procedure described in Section 2.5 and subjected to cyclic triaxial shear testing [30], using the same amplitude (72 kPa) and loading period (20 s) applied to the UDS specimens. Figure 16b shows the stress–strain behavior of the REC specimen during cyclic loading. The results show that the axial strain initially propagates toward both the compression and extension sides. However, the axial strain abruptly propagates towards the extension side after several cycles. Moreover, there was a constant reduction in the mean effective stress, p′, of the REC specimen, similar to the UDS samples (Figure 16c). The specimen eventually failed after 39 cycles, and p′ was reduced to approximately 60 kPa, resulting in visible necking, as shown in Figure 16a. The parametric values of the soil specimens before and after are summarized in

Table 5. Physical properties of specimen after cyclic shear test on REC until failure.

Parameters	Values
Initial specific volume, v0	2.965
Specific volume after consolidation, v	2.611

.

3.4. Subsequent Monotonic Shear Test on REC Soft Clay

Consolidated-undrained triaxial shear tests were conducted, in accordance with JGS 0523 [29], to determine the shear behavior and mechanical properties of the REC specimen without prior cyclic loading history. Then, subsequent monotonic shear tests were performed on REC specimens at two cyclic loading stages, representing low (7 cycles) and high (25 cycles) cyclic histories. The number of cycles for cyclic history consideration is based on the number of cycles to failure from the cyclic shear test.

The peak shear of the REC specimen without cyclic loading histories is 138.582 kPa. After seven cycles, the peak shear strength is 138.902 kPa, corresponding to negligible changes in the value. After 25 cycles, the peak shear strength slightly decreased to 135.971 kPa, corresponding to a 1.88% reduction relative to the specimen without a cyclic loading history.

Table 6. Physical properties of the REC specimen after the subsequent monotonic shear test.

Parameters	0 Cycles	7 Cycles	25 Cycles
Initial Specific Volume, v0	2.994	3.056	2.984
Specific Volume, v	2.627	2.553	2.601
Shear Strength, qmax (kPa)	138.582	138.902	135.971
Secant Young’s Modulus, E50	225.294	62.366	64.934

summarizes the parametric values before and after the experiments.

Based on the experimental results, the peak shear strength reductions observed in the REC specimens with cyclic loading histories are negligible (Figure 17b). This indicates the absence of strain softening behavior in REC specimens under cyclic loading. Figure 17a shows the stiffness degradation of the REC specimens as cyclic loading progressed. Based on these findings, it can be concluded that the effect of soil disturbance in REC specimens is the reduction in stiffness without the peak shear strength.

Cyclic and subsequent monotonic shear tests on UDS and REC specimens revealed that there is also a reduction in the mean effective stress, p′, of REC specimens during cyclic loading, indicating that stiffness reduction is a common effect of soil disturbance regardless of the type of specimen. Moreover, stiffness degradation in the REC specimens is more pronounced. In contrast, the reduction in peak shear strength only occurs in UDS specimens, whereas little to no reduction is observed in REC specimens. The REC specimens also failed at significantly fewer loading cycles than the UDS specimens.

3.5. Peak Shear Strength and Stiffness Degradation Behavior of UDS and REC Specimens

Based on the experimental results, the peak shear strength of UDS specimens has a more significant reduction than REC specimens. This is due to soil disturbance, which disrupts the soil structure of UDS specimens, resulting in a greater reduction in peak shear strength, as shown in Figure 18a and Figure 19a. The reduction in stiffness of the specimens was also analyzed by determining the Secant Young’s Modulus, E₅₀, values of all tests and graphing them against the number of cycles, log N, and the mean effective stress reduction ratio, Δp′/p₀′. Figure 18b and Figure 19b show that UDS specimens exhibit a significantly slower reduction in stiffness than REC specimens, which show a sharp decline.

These behaviors can be attributed to the absence of aging effects in the REC soft clay. Aging effects refer to the gradual changes in the engineering properties of natural clay resulting from physicochemical processes and microstructural developments, leading to increased undrained shear strength, enhanced stiffness, and reduced compressibility [31]. The UDS clay specimens possess an inherent soil structure developed through long-term deposition and aging, which gradually degrades under cyclic loading due to disturbance of particle arrangement and natural cementation. As a result, the peak shear strength and stiffness of the UDS specimens continuously reduce until these mechanical properties become similar to those of the REC specimens.

3.6. Soil Structure Degradation During Cyclic Loading

The presence of soil structure in soft clay specimens was previously described in the early sections of this paper by plotting the specific volume vs. effective stress for NC and OC UDS specimens. In this section, the degradation of the structure of UDS specimens after cyclic loading is discussed using a specific volume–effective stress graph. Figure 20a shows the soil structure degradation of the UDS specimen after 48 cycles, indicating that it has already crossed the NCL but has not yet converged with the CSL. On the other hand, Figure 20b presents the soil structure degradation of the UDS specimen after 960 cycles, which closely reaches the CSL. In contrast, the REC specimens in Figure 21a,b exhibited the absence of structure and reached the CSL after cyclic loading and an undrained shear test.

3.7. Post-Cyclic Undrained Monotonic Shear Test at Different Amounts of Consolidation

Post-cyclic undrained monotonic shear tests were conducted on UDS specimens under isotropic conditions after different amounts of consolidation. The consolidation percentages (0%, 30%, 50%, and 100%) were calculated based on the consolidation behavior of specimens from previous experiments. Figure 22a illustrates the stress–strain behavior of the UDS specimens, showing their stiffness degradation and recovery. On the other hand, Figure 22b presents the effect stress path of the UDS specimen, showing the reduction and increase in peak shear strength. This increase in peak shear strength after post-cyclic consolidation is due to the densification of the specimen. Based on the experimental results, strain softening behavior is observed only under monotonic loading (Blue) and at 0% consolidation (Green). At 30% consolidation after cyclic loading, there is a subtle increase in peak shear strength and a drastic recovery of soil stiffness in the specimen compared to 0% consolidation. At 50% consolidation, the specimen’s peak shear strength exceeds its original capacity, and the soil stiffness continues to recover. At 100% consolidation, the soil stiffness fully recovered, and the peak shear strength of the specimen increased significantly.

Table 7. Physical properties of subsequent monotonic shear test on UDS specimens with different post-cyclic consolidation progress.

Parameters	Mono	0%	30%	50%	100%
Initial Specific Volume, v0	3.108	3.088	3.176	3.069	3.258
Specific Volume (after isotropic consolidation), v1	2.795	2.787	2.757	2.711	2.920
Specific Volume (after post-cyclic consolidation), v2	no data	no data	2.714	2.663	2.822
Shear Strength, qmax (kPa)	146.779	136.656	139.895	155.324	166.494
Secant Young’s Modulus, E50	312.974	45.998	126.947	101.011	268.539

summarizes the values of important parameters of the soil specimens before and after the experiments. Figure 23a visualizes the peak shear strength reduction and increase a, while Figure 23b shows stiffness degradation and recovery of the UDS specimens.

4. Discussion

This study enhanced the understanding of the dynamic behavior and properties of soft clay soils within the concept of soil disturbance established by Nakai-Noda’s constitutive framework. This study directly compared the properties and behavior of UDS and REC soft clay from the same material, preserving its physical properties. As a result, the rigor of comparison is substantially higher than in most previous studies [7,8,9,10,11,12,13,14,15,16,17]. Furthermore, this study provided the necessary experimental evidence for the mechanism of soil disturbance established by [19,20], thereby bridging the gap between numerical simulations and experimental observations. However, the experiments were conducted only on Kaizu soft clay, as the focus of this study is to demonstrate the mechanism underlying the soil disturbance concept rather than to generalize statistically. The number of experiments was also limited in this study due to inconsistencies in the physical properties of the specimens, which are caused by uncertainties in natural soil.

The physical properties of Kaizu soft clay indicate that the specimens used in the experiments have a high moisture content, a high sensitivity ratio, and a high void ratio. Undrained monotonic shear tests on UDS specimens under NC and OC also exhibited the “rewinding and bowing” behavior, which is inherent to structured clay soils. These properties and behavior confirm that Kaizu clay is considered highly structured soft clay.

Based on the analysis of the experimental results, the peak shear strength of UDS specimens declined more than that of REC specimens, supporting the findings of [17]. However, this study lacks a discussion of the stiffness degradation in UDS and REC specimens, also referred to as the soil disturbance effect in [19,20]. Therefore, the current study synthesizes the findings of these papers to refine the knowledge about the soil disturbance concept. Figure 16 and Figure 17 show the stiffness degradation graphs of both UDS and REC specimens. It is noticeable that there is a moderate decline in the stiffness of UDS specimens during the early stages of cyclic loading. In contrast, REC specimens exhibited a drastic decrease in stiffness at the start. This behavior can be attributed to the presence of a natural structure in the UDS specimens and the lack of aging effect in the REC specimens.

Additionally, the specific volume vs. effective stress graphs from undrained triaxial and cyclic shear tests on UDS specimens show that the specimens still retain some of their soil structure even after cyclic loading. On the other hand, the REC specimens exhibited no soil structure even before cyclic loading was applied. Such observations of UDS specimens are important both for defining the mechanism of soil disturbance in naturally deposited soft clay and for designing structures on soft clay foundations.

Based on the results of the subsequent monotonic shear tests on UDS specimens at different post-cyclic consolidations, the peak shear strength of the specimen fully recovered, and the soil stiffness continuously recovered after 50% consolidation. At 100% consolidation, the peak shear strength is significantly higher than the original value, and the soil stiffness fully recovered. This is due to the densification effect after post-cyclic consolidation. This confirms the findings of previous studies [12,13] that even soft clay soils perform like highly overconsolidated clay soil after post-cyclic consolidation.

The catastrophic damage caused by the 1985 Mexico City earthquake was due to the amplification of seismic waves within the highly structured soft clay deposits. The resulting reduction in bearing capacity, which caused the overturning failure of buildings, was attributed to the degradation of the soil structure in these highly structured soft clay deposits [2,3]. This behavior was also observed in Kaizu soft clay during the experiments, which is associated with the effects of soil disturbance on undisturbed soft clay. Therefore, engineers need to consider the possible effects of soil disturbance on the bearing capacity and rigidity of highly structured soft clay deposits when designing structures above them.

5. Conclusions

In conclusion, this paper successfully achieved its main objective by providing experimental evidence that supports the numerical simulation results reported in previous studies on the definition of soil disturbance. Specifically, this study found that peak shear strength degradation is observed only in UDS specimens. On the other hand, both UDS and REC specimens experienced stiffness degradation, but the reduction was more pronounced in REC specimens. Furthermore, both specimens experienced a reduction in mean effective stress under cyclic loading. Therefore, soil disturbance can be defined as the drastic reduction in the stiffness of the REC soft clay and as the simultaneous degradation in the peak shear strength and stiffness of the UDS soft clay. Moreover, the effect of soil disturbance is more pronounced in UDS specimens due to the reduced peak shear strength, which is attributed to the presence of structure and other aging effects. From an engineering design perspective, engineers should be more careful when designing structures on UDS soft clay foundations, due to the reduction in shear capacity caused by soil disturbance. In general, soil disturbance can be defined as the reduction in the mean effective stress and stiffness under cyclic loading.

The effect of soil disturbance on the stiffness and peak shear strength has important implications for seismic response analyses and the reinforcement design of structures. It is essential to consider the effects of soil disturbance to avoid overestimating the bearing capacity and ground stability of soft-clay foundations. Furthermore, appropriate soil improvement techniques, such as deep cement mixing and geosynthetic reinforcement, can be adopted to enhance stiffness and reduce excess pore water generation in a soft clay foundation.

The experimental program in this study is limited to Kaizu soft clay, as the primary objective is to clarify the mechanisms of soil disturbance through direct experimental evidence rather than to achieve statistical generalization. Additionally, this study was also limited by the availability of UDS specimens with similar physical properties, as inconsistencies are inevitable for natural soils. Nevertheless, the researchers believe that statistically generalizing the properties and behavior of soft clay soils is very important. Therefore, the researchers aim to conduct experimental programs using various types of soft clay soil at different amplitudes in future work. The researchers are also planning to consider the effects of anisotropy on soil disturbance, as this is an essential property of UDS soils. Moreover, differences in the microstructures of UDS and REC before and after soil disturbance, as revealed by SEM-EDS analysis, are also being considered in the researchers’ future work.