Bearing capacity mechanism of vibro-replacement stone columns in coral sand with particle breakage Authors: Xiangji Ye, Zongyuan Pan, Feng Liu, Hongxiang Tang, Zezhou Ji, Wei Zhao, Xin Zhao Published: 2026 ABSTRACT AbstractCoral sand presents significant challenges to ground improvement design due to its high void ratio, high compressibility, and particle breakage. Traditional bearing capacity theories for vibro-replacement stone columns are typically based on the assumption of soil shear dilation or constant volume. However, these assumptions fail to account for the particle breakage-induced energy dissipation and volume contraction characteristic of coral sand under high stress, leading to significant discrepancies in calculation results. Based on an analysis of the physical and mechanical properties and the breakage mechanism of coral sand, this paper proposes a novel three-step method for in-situ porosity measurement. Furthermore, a dual-mechanism coupling model, incorporating both energy-equivalent additional confining pressure and breakage-friction coupling, is proposed. By introducing the particle breakage energy dissipation coefficient and the energy-confining pressure conversion efficiency , a modified limit equilibrium model for bulging failure is established to dynamically couple the void ratio evolution and the mobilized internal friction angle. Validation through an airport runway project in coral sand geological conditions demonstrates that the proposed method accurately captures the lateral restraint enhancement of coral sand particle breakage. The calculated lateral ultimate stress (166.3 kPa) is consistent with the field measured range of 158.0–175.0 kPa from nine parallel plate load tests. Considering the full range of measured data, the relative error of the prediction ranges from − 5.3% to + 5.0%. This approach yields significantly higher accuracy compared to traditional standard methods, providing a reliable theoretical basis for engineering design under the conditions of coral sand geology.Keywords: Coral sand, Particle breakage, Energy dissipation, Bearing capacity, In-situ porositySubject terms: Engineering, Natural hazards, Solid Earth sciences Introduction Coral sand constitutes the primary seabed sediment in extensive offshore and oceanic regions. Globally, these deposits cover an estimated area of 284,300 km² across tropical and subtropical seas1, as illustrated in Fig. 1. Driven by the rapid advancement of marine engineering, construction activities within coral sand environments have become increasingly prevalent. Coral sand originates from the weathering and degradation of coral reef limestone; its porous biogenic skeleton makes the particles highly crushable during high-energy vibro-replacement, which directly challenges conventional stone column bearing capacity theories2. Critical infrastructure projects in these domains include pile foundations for conventional and renewable energy systems, land reclamation initiatives, and the construction of offshore airports. In the stabilization of these highly compressible foundations, vibro-replacement stone column technology is widely employed due to its high construction efficiency, strong resistance to liquefaction, and significant deep reinforcement effects. Global distribution map of coral sand3. Extensive geomechanical studies have characterized the unique biogenic microstructure of coral sand. Unlike terrigenous quartz sand, coral sand particles possess highly irregular shapes and retain numerous intra-particle voids inherited from their original biological skeletons4. While this distinct microstructure contributes to a high initial void ratio and a substantial internal friction angle, it renders the material highly susceptible to breakage. Wang et al.5 confirmed that this porous structure induces skeleton reorganization during early loading stages, resulting in a compression index significantly higher than that of quartz sand. Within the framework of critical state soil mechanics, Xiao et al.6 demonstrated that coral sand undergoes significant particle breakage during shearing, causing a continuous evolution of the particle size distribution (PSD). This breakage behavior involves significant energy dissipation, which fundamentally modifies the soil’s dilatancy characteristics. Consequently, the critical state line (CSL) is non-unique, manifesting instead as a family of rotating curves dependent on the degree of breakage in the plane. The volumetric contraction induced by particle breakage has profound implications for foundation engineering. Traditional soil mechanics principles posit that pile installation induces shear dilation, creating a positive compaction effect and increasing lateral stress. However, empirical derived from studies on coral sand contradicts this assertion. Pioneering work by Poulos7 and Nauroy et al.8 highlighted that the energy input during pile driving causes severe particle breakage and stress relaxation, resulting in shaft resistance reaching only 20%–30% of that in quartz sand. Further quantification analyses conducted by Shan9 and Jiang10 have elucidated that with the augmentation of relative density, there is a corresponding increase in the proportion of energy dissipation attributable to particle breakage. This phenomenon, in turn, undermines the efficacy of conventional design approaches predicated solely on relative density. Despite the understanding of driven piles, the design theory for vibro-replacement stone columns in coral sand remains underdeveloped. Currently, bearing capacity calculations primarily follow the cavity expansion theory (CET) proposed by Aleksandar11 or the bulging failure theory by Hughes and Withers12. These classical theories, predicated on the Mohr-Coulomb criterion for terrestrial soils, assume that the soil volume remains constant or undergoes shear dilation during expansion. They fail to account for the energy dissipation mechanism and volumetric contraction caused by the high-frequency vibration and intense compaction of vibro-replacement construction in breakage soils13. Consequently, the direct application of these theories often leads to significant calculation errors, resulting in either excessive conservatism or safety risks due to the overestimation of lateral restraint forces14. The distinct reaming response between terrigenous quartz sand and coral sand directly causes calculation deviation of lateral restraint, as schematically shown in Fig. 2. Diagram of the reaming mechanism of terrigenous sand and coral sand. To address these deficiencies, this study aims to transcend the framework of traditional terrestrial soil mechanics by establishing a revised calculation theory that explicitly incorporates the energy dissipation of particle breakage. First, the physical and mechanical properties of coral sand are analyzed, and a three-step method is proposed for accurate in-situ porosity measurement. Second, based on the principle of energy conservation, a particle breakage energy dissipation coefficient is introduced to quantify the energy consumed by breakage during vibro-replacement construction. Third, a modified bulging limit equilibrium model is constructed. This model couples the void ratio evolution with stress levels, theoretically filling the gap in the applicability of classical CET to breakage soils. Finally, the proposed method is validated against in-situ plate load tests from a high-precision coral sand geological airport project, demonstrating its accuracy and engineering value. Physical and mechanical properties of coral sand Basic physical indicators and parameter determination of coral sand As a specialized biogenic geomaterial, coral sand exhibits unique engineering mechanical behaviors governed directly by its material composition and microstructure. Distinct from terrestrial siliceous sand, coral sand is predominantly composed of calcium carbonate, with a content typically exceeding 90%15. X-ray diffraction analysis identifies the primary mineral phases as aragonite and high-magnesium calcite. This mineralogical composition results in low particle hardness, rendering the material highly susceptible to cleavage and breakage under high stress. Scanning electron microscopy (SEM) observations reveal that coral sand retains the skeletal microstructure of original biological organisms, such as reef-building corals, foraminifera, and mollusks. Consequently, the particles exhibit extremely irregular morphologies, characterized primarily by dendritic, rod-like, blocky, and flaky shapes. A defining feature of coral sand is the abundance of intra-particle voids. Research indicates that these intra-particle voids can constitute approximately 10% of the total porosity. This porous structure not only yields a large specific surface area and high water absorption capacity but also serves as the primary structural factor facilitating particle breakage during the vibro-replacement construction compaction process. Driven by its unique biogenic origin, the physical indices of coral sand differ significantly from those of conventional quartz sand. Attributable to the high mineral density of aragonite, the particle density of coral sand typically ranges from 2.70 to 2.85, which is notably higher than that of standard quartz sand (approximately 2.65). For the accurate determination of specific gravity, the pycnometer method utilizing kerosene as the medium is recommended to ensure complete de-airing of the intra-particle pores. Furthermore, coral sand exhibits an exceptionally high initial void ratio, varying widely between 0.54 and 2.97. This elevated void ratio stems from a dual-porosity system comprising both inter-particle packing voids and intra-particle biological voids, thereby endowing the soil with significant compressibility and sensitivity to collapse upon saturation. Accurate determination of the in-situ porosity (a core parameter in the bearing capacity calculation model) is a prerequisite for ensuring the design precision of vibro-replacement stone columns. However, due to the high fragility, irregular particle shape, abundant intra-particle voids of coral sand, conventional drilling and sampling methods are prone to cause particle disturbance and pore structure damage, leading to distorted test data. To address this, this study proposes a reliable three-step measurement technique for undisturbed coral sand: Undisturbed sampling and vacuum saturation for lower bound determination The thin-walled soil sampler is used to obtain the undisturbed coral sand sample with the size of by static pressure method on site, to minimize the disturbance to the granular skeleton. The sample is dried at and weighed to obtain the dry mass . Subsequently, vacuum saturation treatment (vacuum degree of 8 kPa for 24 h) is performed, and the saturated mass is measured. The connected porosity is calculated by the buoyancy method: This method is sensitive to connected inter-particle and intra-particle pores, but may overlook enclosed isolated pores, thus serving as the lower bound of porosity. Mercury Intrusion Porosimetry (MIP) for upper bound verification Parallel samples from the same sampling position are used for laboratory MIP testing. Because high-pressure mercury intrusion can penetrate micron-scale intra-particle micropores, the resulting porosity values are typically 8% to 15% higher than those obtained via water saturation. Consequently, MIP results serve as the upper bound reference of the total porosity. CT scanning and digital reconstruction for precise calibration High-resolution industrial CT (resolution: 18 μm) is employed to perform tomographic scanning and 3D digital reconstruction of the entire core. Local threshold segmentation is executed using Avizo software to extract the digital porosity . To mitigate bias caused by partial volume effects in micropores, a linear coupling formula is applied for final calibration: In the formula, K represents the weight coefficient, which is recommended to take a value of based on regression analysis of a large amount of experimental data from our research group. Through the aforementioned three-step method, the measurement error of in-situ porosity of coral sand can be controlled within ± 2%, which effectively solves the problems of sampling disturbance and intra-particle pore omission in conventional test methods, and provides a solid basis for the selection of initial void ratio and post-treatment void ratio in the subsequent bearing capacity calculation model. In the subsequent calculation, the calibrated porosity is first converted into void ratio by e=n/(1-n). The pre-treatment value is used as the initial void ratio , whereas the post-treatment value is used as . These two parameters are then substituted into Eq. (5) to calculate the breakage-induced friction angle increment , which further determines , , and finally the lateral ultimate stress in Eq. (7). The proposed three-step method achieves high accuracy, but has higher requirements for industrial CT and MIP equipment. A detailed comparison with conventional methods is shown in Table 1. Table 1Comparison between the proposed three-step method and conventional methods.Test methodAccuracyInstrument requirementCost per sampleTest cycleApplication scenarioThe three-step method± 2%Thin-walled sampler, CT, MIPHigh7 daysHigh-precision design of major projectsConventional ring knife method± 8%–±15%Basic samplerLow1 dayConventional surveyCPTU in-situ method± 5%–±10%CPTU probeMedium1 dayLarge-area rapid survey Comparison between the proposed three-step method and conventional methods. For large-scale engineering, this method can be combined with CPTU to balance accuracy and efficiency. The overall workflow of the proposed three-step in-situ porosity measurement method is illustrated in Fig. 3. Schematic diagram of the three-step method for in-situ porosity measurement. Coral sand particle breakage characteristics and its influence mechanism on stone column bearing capacity Compared to terrigenous siliceous sand, coral sand has highly irregular particle morphology and abundant intra-particle pores, which makes it significantly more susceptible to particle breakage under stress16,17. The degree of particle breakage in coral sand is jointly affected by consolidation confining pressure, relative density, particle shape, gradation and loading rate, and this breakage behavior will fundamentally change the core mechanical parameters of the soil.Existing studies have confirmed that particle breakage during shearing will directly alter the dilatancy, internal friction angle and consolidation characteristics of coral sand18–21. To quantitatively describe the degree of fragmentation, Hardin22 proposed the classic concept of relative fragmentation potential , which quantifies the degree of fragmentation through the area ratio of fragmentation potential to total fragmentation rate. Unlike terrigenous quartz sand, which exhibits shear dilation under high stress, coral sand mainly shows volumetric contraction due to particle breakage, which means the classical CET based on shear dilation assumption is no longer applicable. However, current studies on coral sand particle breakage mostly focus on the mechanical response in laboratory element tests, and there are still obvious research gaps: The positive “breakage-filling-densification” effect of coral sand particle breakage during vibro-replacement construction has not been clearly revealed; The positive “breakage-filling-densification” effect of coral sand particle breakage during vibro-replacement construction has not been clearly revealed; There is no quantitative theoretical model to characterize the influence of particle breakage on the lateral restraint capacity of soil around vibro-replacement stone columns, which leads to the large deviation of traditional bearing capacity calculation methods in coral sand geology. There is no quantitative theoretical model to characterize the influence of particle breakage on the lateral restraint capacity of soil around vibro-replacement stone columns, which leads to the large deviation of traditional bearing capacity calculation methods in coral sand geology. To fill the above gaps, this study first clarifies the dual effect of coral sand particle breakage in vibro-replacement construction, and then establishes a bearing capacity calculation model that couples the energy dissipation of particle breakage. The classic Hardin relative breakage potential22 is used to quantify the degree of particle breakage in this study, which is the most widely used index in coral sand breakage research with clear physical meaning. Theoretical derivation of bearing capacity of vibro-replacement stone column considering energy consumption due to breakage Analysis and theoretical correction approach of particle breakage effect The estimation of bearing capacity for vibro-replacement stone columns is predominantly based on the CET proposed by Vesic11. Predicated on the assumption of an elastic-plastic medium, this theory postulates that the plastic zone undergoes either constant volume deformation or shear dilation during expansion, thereby inducing a positive compaction effect and enhancing lateral stress. However, field observations from coral sand foundations indicate that the direct application of this theory often yields substantial discrepancies between calculated predictions and measured values. This divergence is primarily attributed to the theory’s failure to account for particle breakage and the associated volumetric contraction characteristic of biogenic carbonate sands under high stress. Before introducing the modified model, it is necessary to first analyze the distinctly different mechanisms of action of driven piles and vibro-replacement stone columns in coral sand. Shear-relaxation mechanism of driven piles Shear-relaxation mechanism of driven piles For solid driven piles or static pressure piles, the intrusion of the pile body primarily relies on forcibly shearing and displacing the soil. In fragile coral sand, the immense shear force leads to the breakage of particle edges and corners around the pile. Due to the absence of high-frequency vibration assistance, the fine particles after breakage cannot effectively fill the pores, but instead cause local volume contraction due to the destruction of the skeleton. This contraction manifests macroscopically as the release of effective stress around the pile, which is the stress relaxation phenomenon described by Poulos7, leading to a significant reduction in lateral friction resistance. (2)The breakage-filling mechanism of vibro-replacement stone columns during construction The breakage-filling mechanism of vibro-replacement stone columns during construction Unlike the forced shear displacement of driven piles, vibro-replacement construction is a high-energy, high-frequency vibration compaction process. The high-frequency horizontal vibration of the vibroprobe instantly liquefies the saturated coral sand around the pile, eliminating the inter-particle interlocking resistance temporarily. In this specific construction state, the fine debris generated by particle breakage will not cause skeleton collapse and stress relaxation. On the contrary, under the action of the vibrational force field, the crushed fine particles are rapidly filled into the original inter-particle and intra-particle large pores, which promotes the forced skeleton reorganization of the soil around the pile, and the void ratio decreases sharply from the natural deposition stat to the dense state close to the minimum void ratio . It should be emphasized that particle breakage of coral sand has a significant dual effect on its mechanical properties: during the vibro-replacement construction stage, the positive effect of particle breakage, namely breakage-filling-densification, is absolutely dominant. The breakage is no longer a destructive factor leading to stress relaxation, but a necessary way to realize the transition of coral sand from loose state to dense state, and finally provides the stone column with lateral restraint stress much higher than that of natural foundation. However, under the action of long-term static load or overlying overload after construction, the continuous particle breakage of coral sand will lead to the wear of particle edges and corners, the weakening of inter-particle interlocking, and then cause the attenuation of internal friction angle, the increase of compressibility and the decrease of long-term bearing capacity, which is an adverse effect that cannot be ignored in engineering design. Based on the above analysis, the energy term used to overcome shear dilation in classical theory does not exist in coral sand, and is replaced by the energy consumed in the breakage-filling process. According to the principle of energy conservation, the mechanical work output by the vibro-replacement device does not disappear, but is transformed into an increase in soil compactness through energy consumption in breakage. This paper proposes a modified bulging limit equilibrium model that considers the energy consumption due to particle breakage. The core idea of this model is to map the energy density consumed in particle breakage during the vibro-replacement and reaming process to the increment of effective lateral confining stress through a dimensionless energy consumption coefficient . This not only corrects the defect of traditional theory that ignores volume contraction, but also quantifies the special engineering gain characteristic of coral sand, which is more prone to breakage and compaction. The schematic framework of the circular hole expansion correction model considering particle breakage is shown in Fig. 4. Schematic diagram of the circular hole expansion correction model. This study focuses on the calculation of the ultimate bearing capacity of vibro-replacement stone columns in coral sand after construction completion. The proposed modified model is based on the core premise that the particle breakage of coral sand has been fully completed during the vibro-compaction construction stage, and the soil around the pile has reached a stable dense state. Within the range of design ultimate load, the secondary particle breakage of the compacted coral sand is not significant, and the strength enhancement effect caused by construction-stage breakage is dominant. For the long-term bearing capacity attenuation caused by continuous particle breakage under long-term load, it needs to be further corrected by combining with the creep characteristics of coral sand, which is not included in the scope of this study. Definition and physical meaning of core parameters Three key parameters are introduced in the modified model, including a material intrinsic parameter, a system response parameter and a coupling coefficient. Their physical definitions, dimensions, calibration methods, and adopted values are summarized in Table 2 to eliminate ambiguity. Table 2Summary of key parameters in the proposed model.ParameterPhysical meaningUnitCalibration methodAdopted value Fraction of input energy dissipated by coral sand particle breakage/High-pressure triaxial calibration0.25 Fraction of breakage energy converted into effective lateral confining pressure/Back-calculation from field plate load–pressuremeter tests0.35 Increment of internal friction angle induced by unit breakage energy dissipation°Triaxial-gradation combined calibration40° Combined energy–confining pressure conversion coefficient/Product of and 0.0875 Summary of key parameters in the proposed model. Product of and Particle breakage energy dissipation coefficient Particle breakage energy dissipation coefficient is an intrinsic material property of coral sand, defined as the ratio of energy consumed by particle breakage to the total input energy per unit volume. It reflects the crushability of coral sand particles and is only determined by mineral composition, particle morphology and pore structure. Different from the Hardin breakage potential 22 (which quantifies the result of particle breakage), quantifies the energy distribution during the breakage process. Typical range for South China Sea coral sand: , calibrated value in this study: . is calibrated by high-pressure triaxial tests, which quantifies the energy proportion of particle breakage under different confining pressures. (2)Energy-confining pressure conversion efficiency Energy-confining pressure conversion efficiency is a system response parameter, representing the fraction of breakage energy that is effectively converted into effective lateral confining pressure. It is affected by vibro-replacement technology, drainage conditions and stress path, and must be calibrated by field tests. Typical range for South China Sea coral sand: ; calibrated value in this study: . (3)Breakage-friction coupling coefficient Breakage-friction coupling coefficient characterizes the increment of internal friction angle induced by unit breakage energy dissipation, which reflects the strengthening effect of particle interlocking after breakage and filling. Calibrated value in this study: . Energy flow relationship and non-circular calibration logic To avoid the logical risk of circular reasoning in parameter back-calculation, a fixed-input iterative calibration framework is established, and the energy flow chain is defined as: where is the energy density per unit volume, and is the effective energy consumed by particle breakage. and are pre-calibrated as fixed constants by laboratory high-pressure triaxial tests, the mobilized internal friction angle is calculated with fixed and , is back-calculated by field measured through one-way iteration, with no mutual unknowns between parameters. The comprehensive conversion coefficient is defined as , which is used for rapid engineering calculation. Revised bulging limit equilibrium formula To construct a vibro-replacement gravel pile reaming model suitable for fragile coral sand, the following basic assumptions are formulated: Assumption of infinite soil: The coral sand surrounding the pile is considered as an infinitely large, isotropic, and homogeneous half-space. Assumption of infinite soil: The coral sand surrounding the pile is considered as an infinitely large, isotropic, and homogeneous half-space. Column hole expansion mode: The process of hole-making and compaction by the vibro-replacement device is simplified to the radial expansion of a cylindrical hole. Column hole expansion mode: The process of hole-making and compaction by the vibro-replacement device is simplified to the radial expansion of a cylindrical hole. Superposition principle: The ultimate lateral confinement stress of the soil surrounding the pile consists of two parts: one is the static equilibrium resistance provided by the shear strength of the soil itself; the other is the additional stress increment converted from the volume contraction caused by particle breakage during the vibro-replacement process. Superposition principle: The ultimate lateral confinement stress of the soil surrounding the pile consists of two parts: one is the static equilibrium resistance provided by the shear strength of the soil itself; the other is the additional stress increment converted from the volume contraction caused by particle breakage during the vibro-replacement process. According to the principle of energy conservation, the total mechanical work imparted by the vibroprobe per unit volume of soil during the cavity expansion process is primarily dissipated through two mechanisms: the plastic shear work required to overcome interparticle resistance, and the volumetric work of breakage responsible for particle breakage and skeleton reorganization. In stark contrast to the shear dilation observed in terrigenous quartz sand, the substantial energy dissipation in coral sand under high-frequency vibration primarily drives particle breakage. This study posits that this breakage-induced compaction process is not merely a simple volumetric contraction; rather, it triggers a dual nonlinear mechanical reinforcement mechanism: an energy-equivalent additional confining pressure and a breakage-friction coupling effect. Calculation of energy-equivalent additional confining pressure Calculation of energy-equivalent additional confining pressure Driven by the kinetic energy of vibro-replacement construction, the soil void ratio within the effective compacted volume decreases precipitously from its initial state to a compacted state. Macroscopically, this forced densification translates into an increment in the effective lateral constraint exerted by the surrounding soil on the stone column, denoted as the equivalent additional confining pressure . Tracing the energy flow, the conversion of input energy density into confining pressure can be expressed as: To ensure the physical consistency of the theoretical framework, two key dimensional parameters are introduced in this equation. The first parameter, the particle breakage energy dissipation coefficient characterizes the inherent susceptibility of the coral sand to breakage. It is an intrinsic material property that can be calibrated through laboratory high-pressure triaxial testing. The second parameter, the energy–confining pressure conversion efficiency , represents the fraction of breakage energy dissipation that is effectively converted into confining pressure. As a system response parameter, is influenced by factors such as the vibro-replacement construction method, drainage conditions, and stress paths, and thus requires calibration through in-situ plate load tests or pressuremeter tests. The combined energy–confining pressure conversion factor, defined as , directly governs the magnitude of the additional stress. (2)Breakage-friction coupling reinforcement Breakage-friction coupling reinforcement Traditional cavity expansion theories typically assume that the internal friction angle within the plastic zone remains constant at its initial value . However, micromechanical analysis reveals that the fines generated by the breakage of coral sand particles not only fill the interstitial voids but also significantly enhance the interlocking of the reorganized granular skeleton. This physical reconfiguration of the soil skeleton directly manifests as a macroscopic increase in the internal friction angle. Consequently, a mobilized internal friction angle must be introduced to capture this evolutionary process: Based on critical state soil mechanics and energy dissipation principle, the increment of internal friction angle caused by particle breakage and filling is expressed as: where is the breakage-friction coupling coefficient, defined physically as the increment in the internal friction angle per unit of breakage energy dissipation; and are the initial and post-treatment void ratios, respectively; represents the vibro-replacement construction energy per point, and is the reference calibration energy. Particle breakage generates fine particles to fill pores, enhances particle interlocking, and thus increases the macroscopic internal friction angle. The logarithmic relationship between energy input and strength increment has been widely verified by triaxial tests on coral sand18–21. The breakage-friction coupling coefficient is a core parameter to characterize the increment of internal friction angle caused by particle breakage, which is defined as the increment of internal friction angle per unit breakage index, with the dimensionality of degree. A standardized triaxial-gradation combined calibration method is adopted for , which can be completed within 7 days through laboratory tests. 4 samples of coral sand specimens with the same initial void ratio are prepared, and 3 parallel samples are set for each sample to conduct consolidated drained triaxial shear tests. Different vibro-replacement energy levels are simulated by controlling the target post-treatment void ratio of each sample, and the mobilized internal friction angle of each sample is measured through tests. The test scheme is shown in Table 3. The corresponding data fitting curves to derive are illustrated in Fig. 5. Table 3Test scheme for triaxial-gradation combined calibration of .Sample numberTarget Simulated energy ratio Measured Calculated 10.951.042.0°0°20.852.543.2°1.2°30.755.044.5°2.5°40.707.545.8°3.8° Test scheme for triaxial-gradation combined calibration of . Calibration of the breakage-friction coupling coefficient based on triaxial-gradation combined tests: (a) relationship between friction angle increment and simulated energy ratio ; (b) linear fitting for , with shown in the figure. Taking the test parameters of this project as an example: set , =500 kJ, the corresponding of each sample is 500 kJ, 1250 kJ, 2500 kJ and 3750 kJ respectively. Substitute the test data into the theoretical definition formula of for regression calculation: For sample 2: For sample 3: For sample 4: The average value of the regression results is taken as the design value of , that is , with a coefficient of variation of 10%, which is consistent with the calibration results for the coral sand geological conditions project in this study. Field calibration methods for energy-confining pressure conversion efficiency The energy-confining pressure conversion efficiency is a dimensionless system parameter, which characterizes the proportion of particle breakage energy dissipation effectively converted into effective confining pressure. It is affected by vibro-replacement technology, drainage conditions and stress path, and must be calibrated through field tests. The core idea is to back-calculate by comparing the measured lateral ultimate stress in the field with the theoretical calculated value. Three standardized calibration methods are proposed for different engineering scenarios, and the calibration can be completed within 3–14 days, which meets the requirements of rapid construction of projects on islands and reefs underlain by coral sand geology. Field plate load-pressuremeter combined test Field plate load-pressuremeter combined test This method is applicable to the detailed design stage of key projects, with the highest calibration accuracy. A 2 m×2 m plate load test is set in the reinforced area, and pressuremeter tests are carried out at the depths of 0.5 m, 1.0 m and 1.5 m directly below the plate to measure the distribution of lateral ultimate stress with depth. The lateral ultimate stress is extracted from the pressuremeter curve in strict accordance with GB 50,021 standards. The limit pressure is adopted as the characteristic value, which corresponds to the pressure when the volume expansion reaches 10% of the initial volume of the probe cavity or the plateau value of the pressuremeter curve. Complete vibro-replacement construction, record the single-point energy , measure the post-reinforcement void ratio , and determine the single-point reinforcement volume ; carry out graded loading on the plate until failure, record the ultimate load and load-settlement curve; obtain the measured lateral ultimate stress through pressuremeter tests; Calculate the mobilized internal friction angle through the calibrated in the laboratory, and back-calculate by the following formula: where is the coefficient of passive earth pressure. (2)Field cone penetration test with pore pressure measurement (CPTU) Field cone penetration test with pore pressure measurement (CPTU) This method is applicable to large-area reinforcement projects that require rapid statistical analysis of distribution. Record the and of each vibro-replacement point during construction, and take soil samples to measure ; conduct CPTU test 3–7 days after reinforcement to obtain the cone tip resistance ; establish the relationship between and through the adjacent plate load test: (for coral sand, the recommended value of is 0.15 ~ 0.20); substitute the converted into the back-calculation formula in Method one to solve . (3)Field settlement-inverse analysis method Field settlement-inverse analysis method This method is applicable to the completed and operated runways with long-term settlement monitoring data. The equivalent lateral ultimate stress is back-calculated through the measured settlement , and then is solved iteratively. Establish the settlement-stress relationship by the layer-wise summation method:; back-calculate the equivalent from the measured long-term settlement ; substitute into the back-calculation formula in Method one, and iteratively solve . General modified formula for lateral ultimate stress considering dual-mechanism coupling Synthesizing the physical evolution processes of the two aforementioned reinforcement mechanisms, the ultimate lateral stress of vibro-replacement stone columns in coral sand foundations is no longer the passive earth pressure under a singular assumption. By adopting the dual mechanisms of energy-equivalent additional confining pressure and breakage-friction coupling, the isolated static equilibrium term and the additional deformation term in traditional models can achieve profound physical integration. Under the limit equilibrium state, the total effective confining pressure at the calculation depth is jointly composed of the natural soil self-weight stress and the additional confining pressure converted from breakage energy dissipation. Simultaneously, due to the enhanced interlocking of the reorganized skeleton, the corresponding coefficient of passive earth pressure must be calculated using the mobilized internal friction angle . Accordingly, this paper proposes a novel general modified formula for lateral ultimate stress: This general modified formula organically couples vibro-replacement construction energy input, soil particle breakage, density evolution, and macroscopic strength enhancement within a unified mechanical framework, with each term matching in magnitude and presenting clear physical significance. The first term realizes the enhancement of total confining pressure: the additional term in the parentheses abandons the traditional empirical estimation based on relative density, directly quantifies the conversion law of construction mechanical energy into effective constraining lateral pressure, and reflects the contribution of passive densification under high-energy vibro-replacement construction. The second term achieves the nonlinear amplification of shear resistance: driven by fines filling and skeleton reorganization, the soil’s shear resistance is strengthened, which is macroscopically manifested as an increase in the internal friction angle and the nonlinear amplification of the passive earth pressure multiplier. Moreover, the product of these two terms forms a superposition effect of the dual mechanisms, which transcends the limitations of single-factor modification and embodies the physical essence of the multiplicative, synergistic growth of lateral ultimate stress under dual-mechanism coupling. More importantly, this theoretical model has rigorous boundary degeneration characteristics: for uncrushable conventional terrigenous quartz sand, the particle breakage energy dissipation coefficient approaches zero due to no significant particle breakage, at which point the derived internal friction angle increment and energy-equivalent additional confining pressure both vanish, and the formula naturally degenerates into the standard passive earth pressure solution without energy input under the classical Terzaghi theory. This mathematical degeneration not only bridges the theoretical gap of the classical CET in the application to crushable carbonate soils but also fundamentally substantiates the universality and rigor of the proposed modified model within the framework of classical geotechnical mechanics. Calculation of bearing capacity of single pile and composite foundation Based on the modified lateral restraining stress , the ultimate bearing capacity of a single pile can be expressed: where is the internal friction angle of the gravel pile body. Further introducing the area replacement ratio m, which is determined based on the pile arrangement form, for the common equilateral triangular arrangement: where d represents the pile diameter and s denotes the pile spacing. In addition, the pile-soil load sharing ratio n is introduced to reflect the stiffness concentration effect of the pile relative to the soil between piles. Typically, n is taken as 3 to 5. The characteristic value of the composite foundation bearing capacity, denoted as , can be expressed based on the weight sharing between piles and soil: In the design of vibro-replacement stone columns, simplified empirical formulas are often used to directly reduce based on the ultimate strength of a single pile: In the formula, K represents the safety factor, typically set at K=2.0. Engineering case verification and comparative analysis Engineering background and parameter determination To verify the accuracy and engineering applicability of the proposed general modified formula considering the breakage-friction coupling dual mechanism, this study conducts an empirical calculation based on a vibro-replacement stone column ground improvement project for an airport runway test section in coral sand geological conditions. The site is primarily composed of hydraulically filled biogenic coral sand, exhibiting typical characteristics of high porosity, high compressibility and high particle crushability. The runway structure adopts a strip foundation (B = 2 m), and the foundation is treated with high-power vibro-replacement stone columns, with crushed stone as the filling material for the pile body. The stone columns are arranged in an equilateral triangular pattern, which is the most common layout form in coral sand engineering. To obtain accurate calculation parameters, laboratory high-pressure triaxial tests and in-situ tests were conducted in advance at the test section. Through a combined triaxial-gradation calibration method, the particle breakage energy dissipation coefficient () and the breakage-friction coupling coefficient () of the coral sand at the site were obtained. The energy-confining pressure conversion efficiency back-calculated from the combined in-situ plate load and pressuremeter tests (Recommended Method 1) is . The key calculation parameters used for the ground improvement are detailed in Table 4. Table 4Calculation parameters for ground treatment in the coral sand geological conditions project.ParameterSymbolValueSourceInitial void ratio 1.10In-situ cutting ringMinimum void ratio 0.65Dry vibration methodMaximum void ratio 1.35Funnel methodPost-treatment void ratio 0.70Measured after vibro-replacement constructionEffective unit weight of coral sand 10In-situ measurementInitial internal friction angle 34Static triaxial testParticle breakage energy dissipation coefficient 0.25High-pressure triaxial calibrationBreakage-friction coupling coefficient 40Calibration as aboveVibro-replacement construction energy per point 2000Construction recordsReference calibration energy 500Calibration baselineCompacted volume per point 8Influence diameter 2 m × depth 2 mEnergy-confining pressure conversion efficiency 0.35In-situ regressionCalculation depth of bulging 2Theoretical assumptionInternal friction angle of gravel pile body 42Direct shear test Calculation parameters for ground treatment in the coral sand geological conditions project. The initial void ratio was obtained using thin-walled tube sampling and the ring-knife method before vibro-replacement, while the post-treatment void ratio et was measured at the center of the soil between columns 7 days after construction using the same procedure. The minimum and maximum void ratios were determined by the dry vibration method and funnel deposition method, respectively. The initial friction angle was obtained from consolidated drained triaxial tests under three confining pressures, and the friction angle of the crushed-stone column material was obtained from large-scale direct shear tests. The construction energy was calculated from the effective vibration duration and the rated power of the vibroprobe, and was estimated from the measured influence diameter and effective improvement depth. The parameter was calibrated from three parallel plate load–pressuremeter tests, yielding values of 0.33, 0.36, and 0.35; therefore, = 0.35 was adopted. Step-by-step calculation based on dual-mechanism coupling Substituting the site and process parameters from Table 4 into the dual-mechanism coupling general modified formula derived in this study, the step-by-step theoretical calculation procedure is as follows: Calculate the breakage-driven internal friction angle increment. Calculate the breakage-driven internal friction angle increment. (2)Determine the mobilized internal friction angle. Determine the mobilized internal friction angle. (3)Calculate the energy-equivalent additional confining pressure. Calculate the energy-equivalent additional confining pressure. (4)Calculate the equivalent total confining pressure and the coefficient of passive earth pressure. Calculate the equivalent total confining pressure and the coefficient of passive earth pressure. (5)Obtain the modified lateral ultimate stress. Obtain the modified lateral ultimate stress. (6)Calculate the ultimate bearing capacity of a single stone column. Calculate the ultimate bearing capacity of a single stone column. Substitute the modified lateral ultimate stress into formula (8), the ultimate bearing capacity of single pile is calculated as: (7)Calculate the characteristic value of composite foundation bearing capacity. Calculate the characteristic value of composite foundation bearing capacity. Taking the common area replacement ratio for coral sand engineering, pile-soil load sharing ratio n=2, safety factor K=2.0, substitute into Eq. (11): In-situ measurement comparison and error analysis To further demonstrate the reliability and dispersion of the field validation data, nine parallel 2 m × 2 m plate load tests were conducted in the reinforced area (Table 5). The ultimate load obtained from each test was converted into the corresponding lateral ultimate stress using the inverse calculation procedure described above. The back-calculated lateral ultimate stress ranges from 158 to 175 kPa, with an average value of 166.5 kPa, indicating a moderate but acceptable field dispersion for coral sand foundations. The ultimate bearing capacity in each plate load test was determined according to the load–settlement curve criterion. Specifically, the load corresponding to the evident turning point of the p-s curve was adopted as the ultimate load; when no distinct turning point was observed, the load corresponding to a settlement ratio of was used as the ultimate bearing capacity. Table 5Back-calculated lateral ultimate stress from nine parallel plate load tests.Test no.Ultimate bearing capacity determination value (kN)Back-calculated lateral ultimate Stress (kPa)RemarksPLT-1648162-2.7%PLT-2672168+ 0.9%PLT-3632158-5.1%PLT-4680170+ 2.1%PLT-5700175+ 5.1%PLT-6656164-1.5%PLT-7676169+ 1.5%PLT-8640160-3.9%PLT-9688172+ 3.3%StatisticsMean: 665.8 kN; Range: 632–700 kN; COV: 3.45%Mean: 166.4 kPa; Range: 158–175 kPa; COV: 3.45% Back-calculated lateral ultimate stress from nine parallel plate load tests. The load–settlement responses of the nine parallel tests are summarized in Fig. 6. The solid line represents the mean settlement response under each loading level, while the shaded region denotes the mean ± standard deviation envelope. It can be observed that the dispersion of settlement increases gradually with increasing plate pressure, forming a typical widening envelope at high load levels. This trend is consistent with the crushable and heterogeneous nature of coral sand, where particle breakage and local skeleton rearrangement become more significant under higher stress levels. Load–settlement response of nine parallel plate load tests. The solid line denotes the mean response, and the shaded area denotes the mean ± one standard deviation envelope. The comparative analysis results of obtained using different theoretical methods are shown in Table 6. Table 6Comparative verification of lateral ultimate stress using different calculation methods.Calculation method RemarksTraditional Terzaghi theory70.8, no energy inputConfining pressure modification only73.6Only considers partial energy effects, with a contribution rate of merely 3.6%Modified dual-mechanism coupling formula166.3Additional confining pressure increased by 109%, friction angle increased by 7.7%Back-calculated value from in-situ plate load test158–175Actual response to site ultimate load Comparative verification of lateral ultimate stress using different calculation methods. The results indicate that if the traditional Terzaghi theory without breakage modification or the empirical formula considering only a single additional confining pressure term is continuously used, the design estimates will be heavily conservative. This is because they completely neglect the stress loss caused by the extremely high compressibility of coral sand, as well as the bidirectional amplification triggered by particle breakage and skeleton reorganization. This leads to an error approaching or exceeding 50% compared to the in-situ measurements. Conversely, the calculated using the dual-mechanism coupling formula proposed in this study falls perfectly within the envelope of in-situ measurements, and the relative error of the prediction ranges from − 5.3% to + 5.0%. From the quantitative analysis of the contribution proportions, the energy-equivalent additional confining pressure increased the equivalent confining pressure term by 109%, while the increase in the internal friction angle caused by enhanced skeleton interlocking amplified the resistance coefficient by 7.7%. The nonlinear coupling and superposition effect of these two major mechanisms reaches up to 135%, accurately and successfully capturing the true physical essence of the macroscopic lateral pressure response in breakage coral sand foundations. This theory significantly reduces the traditional calculation error from 50% to < 5%, greatly improving the accuracy of rapid fill design and settlement control for coral sands. Applicability and limitations of the proposed model The proposed model is mainly applicable to saturated coral sand foundations improved by vibro-replacement stone columns under construction energy levels, drainage conditions, and column spacing comparable to those in the present engineering case. Its accuracy depends on the representativeness of the measured void ratio, the calibration of , , and , and the assumption that construction-stage particle breakage dominates the post-treatment mechanical response. For sites with strong soil heterogeneity, poor drainage, significantly different column spacing, or continued particle breakage and creep under long-term loading, additional calibration or correction may be required before direct application. Conclusion To tackle the pivotal engineering challenges marked by substantial calculation discrepancies and inadequate design foundations pertaining to the bearing capacity of vibro-replacement stone columns embedded in coral sand, this study establishes a modified bulging limit equilibrium theory. This theory explicitly incorporates the energy dissipation and dual coupling mechanisms of particle breakage and has been rigorously verified through in-situ testing and engineering measurements. The primary conclusions drawn from this study are outlined as follows: Elucidation of the dual effect and breakage-friction coupling mechanism of coral sand particle breakage. During the high-frequency vibration compaction stage of vibro-replacement construction, coral sand exhibits significant particle breakage and volumetric contraction. Research demonstrates that this breakage has a dual effect: the dominant positive effect is that the crushed fine particles fill the inter-particle voids, enhance the interlocking action of the reorganized skeleton, and improve the lateral restraint and shear strength of the soil; the adverse effect is that particle edge wear will weaken the inter-particle interlocking, which is not significant under the design load after construction. Failure to account for the dual enhancements (lateral constraint increment and internal friction angle elevation) induced by this breakage-compaction process is the root cause of distortions in traditional methods. Elucidation of the dual effect and breakage-friction coupling mechanism of coral sand particle breakage. During the high-frequency vibration compaction stage of vibro-replacement construction, coral sand exhibits significant particle breakage and volumetric contraction. Research demonstrates that this breakage has a dual effect: the dominant positive effect is that the crushed fine particles fill the inter-particle voids, enhance the interlocking action of the reorganized skeleton, and improve the lateral restraint and shear strength of the soil; the adverse effect is that particle edge wear will weaken the inter-particle interlocking, which is not significant under the design load after construction. Failure to account for the dual enhancements (lateral constraint increment and internal friction angle elevation) induced by this breakage-compaction process is the root cause of distortions in traditional methods. Establishment of a dual-mechanism modified bearing capacity model. A novel bulging limit equilibrium formula is developed based on the principle of energy conservation. By introducing the particle breakage energy dissipation coefficient and the energy-confining pressure conversion efficiency, the model theoretically quantifies the energy-equivalent additional confining pressure. Moreover, the breakage-friction coupling coefficient is defined to characterize the mobilized internal friction angle. This fully coupled model bridges the gap in the applicability of classical CETto highly breakage and compressible soils. Establishment of a dual-mechanism modified bearing capacity model. A novel bulging limit equilibrium formula is developed based on the principle of energy conservation. By introducing the particle breakage energy dissipation coefficient and the energy-confining pressure conversion efficiency, the model theoretically quantifies the energy-equivalent additional confining pressure. Moreover, the breakage-friction coupling coefficient is defined to characterize the mobilized internal friction angle. This fully coupled model bridges the gap in the applicability of classical CETto highly breakage and compressible soils. Proposal of a standardized porosity measurement technique. To overcome the difficulties associated with large sampling disturbances and porosity determination in coral sand limestone, a standardized three-step measurement process is proposed. This comprehensive method effectively addresses the omission of closed pores and microcracks common in traditional techniques, thereby ensuring the accuracy and reliability of the critical model input parameters. Proposal of a standardized porosity measurement technique. To overcome the difficulties associated with large sampling disturbances and porosity determination in coral sand limestone, a standardized three-step measurement process is proposed. This comprehensive method effectively addresses the omission of closed pores and microcracks common in traditional techniques, thereby ensuring the accuracy and reliability of the critical model input parameters. Verification of engineering applicability with high precision. The proposed method has been validated through the case studies of coral sand geology airport projects. Analysis indicates that the lateral ultimate stress calculated by the modified dual-mechanism method is in good agreement with the measured values from on-site plate load tests, yielding close agreement with the mean field response; In contrast to traditional methods, this approach accurately captures the true bearing potential of coral sand, satisfying the rigorous demands of engineering design under the conditions of coral sand geology island construction. Verification of engineering applicability with high precision. The proposed method has been validated through the case studies of coral sand geology airport projects. Analysis indicates that the lateral ultimate stress calculated by the modified dual-mechanism method is in good agreement with the measured values from on-site plate load tests, yielding close agreement with the mean field response; In contrast to traditional methods, this approach accurately captures the true bearing potential of coral sand, satisfying the rigorous demands of engineering design under the conditions of coral sand geology island construction. Acknowledgements This research is supported by the Large R&D project of China Communications Construction Group (Grant No. 2025-ZJKJ-17). Author contributions Xiangji Ye, Zongyuan Pan, Feng Liu and Wei Zhao wrote the main manuscript text. Hongxiang Tang, Zezhou Ji and Xin Zhao reviewed the manuscript. Data availability Data will be made available on request. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement Data will be made available on request.