Pub Date : 2026-05-01Epub Date: 2026-02-16DOI: 10.1016/j.compgeo.2026.107995
Wendu Xie , Zheng Han , Gonghui Wang , Guangqi Chen , Bin Su , Yange Li , Haohui Ding , Bangjie Fu , Changli Li
Soil-rock mixture fluidized landslide (SRMFL) consists of fine-grained soil and high-strength rock and often causes serious damage to infrastructures and human lives. Its computation and analysis involve the accurate characterization of irregular rock blocks and complex soil-rock coupling effects, where traditional numerical methods face limitations and deficiencies in correct and efficient contact analysis between phases. In this paper, an improved 3D-SPH-DEM coupling method is proposed, incorporating a reorientation algorithm to ensure correct normal vector orientations for rock feature surfaces and a spherical cap cutting algorithm to simplify the computation of mutual contact volume between soil particles and rock blocks. A high-resolution 3D digital rock model library based on field investigations and high-precision 3D scanning is constructed to achieve high-fidelity modeling results. Additionally, the accuracy and applicability of the proposed method are validated by reproducing a gravity-driven fresh concrete flow experiment and the 2018 Yabakei landslide event in Japan. A series of simulations is then performed considering different rock contents and shapes. Results show that the proposed method performs well in reproducing the landslide event. The kinetic energy in SRMFL follows a rise-and-fall pattern, with peak impact energy not necessarily occurring before impact. Higher rock content can enhance dual-phase behavior, localized rock disturbances, and energy dissipation through collisions and friction. Sharper rock shapes will strengthen soil-rock-terrain interactions and reduce overall landslide mobility.
{"title":"Simulation of soil-rock mixture fluidized landslide dynamics using the 3D-SPH-DEM method enhanced by reorientation and spherical cap cutting algorithms","authors":"Wendu Xie , Zheng Han , Gonghui Wang , Guangqi Chen , Bin Su , Yange Li , Haohui Ding , Bangjie Fu , Changli Li","doi":"10.1016/j.compgeo.2026.107995","DOIUrl":"10.1016/j.compgeo.2026.107995","url":null,"abstract":"<div><div>Soil-rock mixture fluidized landslide (SRMFL) consists of fine-grained soil and high-strength rock and often causes serious damage to infrastructures and human lives. Its computation and analysis involve the accurate characterization of irregular rock blocks and complex soil-rock coupling effects, where traditional numerical methods face limitations and deficiencies in correct and efficient contact analysis between phases. In this paper, an improved 3D-SPH-DEM coupling method is proposed, incorporating a reorientation algorithm to ensure correct normal vector orientations for rock feature surfaces and a spherical cap cutting algorithm to simplify the computation of mutual contact volume between soil particles and rock blocks. A high-resolution 3D digital rock model library based on field investigations and high-precision 3D scanning is constructed to achieve high-fidelity modeling results. Additionally, the accuracy and applicability of the proposed method are validated by reproducing a gravity-driven fresh concrete flow experiment and the 2018 Yabakei landslide event in Japan. A series of simulations is then performed considering different rock contents and shapes. Results show that the proposed method performs well in reproducing the landslide event. The kinetic energy in SRMFL follows a rise-and-fall pattern, with peak impact energy not necessarily occurring before impact. Higher rock content can enhance dual-phase behavior, localized rock disturbances, and energy dissipation through collisions and friction. Sharper rock shapes will strengthen soil-rock-terrain interactions and reduce overall landslide mobility.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"193 ","pages":"Article 107995"},"PeriodicalIF":6.2,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147385394","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-05-01Epub Date: 2026-02-19DOI: 10.1016/j.compgeo.2026.107991
Yuki Kobayashi, Yota Togashi, Masahiko Osada
The evaluation of face stability during tunnel excavation is an essential technical issue in tunnel construction. Conventional methods of evaluating tunnel face pressure and its stability consist of assuming a failure area bounded by two vertical slip lines directly above the face and considering the silo effect. However, assuming that all slip lines observed in previous experiments are vertical lines, particularly in the case of loose ground, is unreasonable. In addition, in conventional earth pressure theory, the coefficient of the lateral earth pressure is undetermined parameter. This study proposes a novel method for estimating the tunnel face pressure using curved slip lines. This method is based on the lower-bound theory, and the differential equation is derived by modeling slip lines in the failure area as a logarithmic spiral and solving the equilibrium for vertical and horizontal forces and moments on a trapezoidal asymmetric infinitesimal element. This method eliminates the need for a lateral earth pressure coefficient by introducing a stress- state relationship on the slip line. Several calculations were conducted to confirm the behavior of the proposed solution and verify its applicability. As a result, the proper behavior of the proposed solution as a frictional material due to the soil covering and frictional resistance was confirmed. The proposed solution adequately and accurately expresses the face stability indices obtained in a previous centrifuge model test.
{"title":"Evaluation of tunnel face stability using slip curves based on lower bound theory","authors":"Yuki Kobayashi, Yota Togashi, Masahiko Osada","doi":"10.1016/j.compgeo.2026.107991","DOIUrl":"10.1016/j.compgeo.2026.107991","url":null,"abstract":"<div><div>The evaluation of face stability during tunnel excavation is an essential technical issue in tunnel construction. Conventional methods of evaluating tunnel face pressure and its stability consist of assuming a failure area bounded by two vertical slip lines directly above the face and considering the silo effect. However, assuming that all slip lines observed in previous experiments are vertical lines, particularly in the case of loose ground, is unreasonable. In addition, in conventional earth pressure theory, the coefficient of the lateral earth pressure is undetermined parameter. This study proposes a novel method for estimating the tunnel face pressure using curved slip lines. This method is based on the lower-bound theory, and the differential equation is derived by modeling slip lines in the failure area as a logarithmic spiral and solving the equilibrium for vertical and horizontal forces and moments on a trapezoidal asymmetric infinitesimal element. This method eliminates the need for a lateral earth pressure coefficient by introducing a stress- state relationship on the slip line. Several calculations were conducted to confirm the behavior of the proposed solution and verify its applicability. As a result, the proper behavior of the proposed solution as a frictional material due to the soil covering and frictional resistance was confirmed. The proposed solution adequately and accurately expresses the face stability indices obtained in a previous centrifuge model test.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"193 ","pages":"Article 107991"},"PeriodicalIF":6.2,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147385467","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-05-01Epub Date: 2026-02-19DOI: 10.1016/j.compgeo.2026.107996
Carolina Castro-Malaver , Manuel Cárdenas-Barrantes , David Cantor , Mathieu Renouf , Emilien Azéma , Carlos Ovalle , Arcesio Lizcano
Static liquefaction, defined as the loss of strength after strain softening under undrained shear, is often the cause of catastrophic failures in loose earthfills, waste dumps and tailings storage facilities. The number of failures does not cease to increase yearly, revealing operational faults, construction defects, and a lack of knowledge regarding liquefaction triggering. While liquefaction is well understood at the scale of a representative elementary volume of soil, the triggering mechanisms at the particle scale are seldom studied. This study aims to analyze the links between the micro- and macro-mechanisms controlling static liquefaction by means of numerical simulations using the Discrete Element Method. We consider mono to highly size polydisperse 2D samples of discs, prepared at varied densities from loose to dense packings. As expected, loose samples liquefy, while dense ones continue to exhibit resistance under undrained shearing. Some medium-dense cases liquefy only temporarily, recovering their shear strength at larger strains. We reveal a dual mechanism for liquefaction through multi-scale analysis: a collapse of the contact network—marked by the coordination number dropping below 3—and the emergence of low-density regions. Temporary liquefaction involves transient connectivity loss with minor density fluctuations, enabling stress recovery. In contrast, full liquefaction combines both effects, leading to an irreversibly disconnected and heterogeneous microstructure.
{"title":"Static liquefaction: The role of grain size polydispersity from a micro-structural perspective","authors":"Carolina Castro-Malaver , Manuel Cárdenas-Barrantes , David Cantor , Mathieu Renouf , Emilien Azéma , Carlos Ovalle , Arcesio Lizcano","doi":"10.1016/j.compgeo.2026.107996","DOIUrl":"10.1016/j.compgeo.2026.107996","url":null,"abstract":"<div><div>Static liquefaction, defined as the loss of strength after strain softening under undrained shear, is often the cause of catastrophic failures in loose earthfills, waste dumps and tailings storage facilities. The number of failures does not cease to increase yearly, revealing operational faults, construction defects, and a lack of knowledge regarding liquefaction triggering. While liquefaction is well understood at the scale of a representative elementary volume of soil, the triggering mechanisms at the particle scale are seldom studied. This study aims to analyze the links between the micro- and macro-mechanisms controlling static liquefaction by means of numerical simulations using the Discrete Element Method. We consider mono to highly size polydisperse 2D samples of discs, prepared at varied densities from loose to dense packings. As expected, loose samples liquefy, while dense ones continue to exhibit resistance under undrained shearing. Some medium-dense cases liquefy only temporarily, recovering their shear strength at larger strains. We reveal a dual mechanism for liquefaction through multi-scale analysis: a collapse of the contact network—marked by the coordination number dropping below 3—and the emergence of low-density regions. Temporary liquefaction involves transient connectivity loss with minor density fluctuations, enabling stress recovery. In contrast, full liquefaction combines both effects, leading to an irreversibly disconnected and heterogeneous microstructure.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"193 ","pages":"Article 107996"},"PeriodicalIF":6.2,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147385468","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-05-01Epub Date: 2026-02-22DOI: 10.1016/j.compgeo.2026.108014
Hanze Li, Chaofa Zhao, Yi Hong, Aoxi Zhang, Yanni Chen
Natural soils are typically unsaturated, and their mechanical behaviour is significantly influenced by capillary cohesion. By incorporating the attractive forces generated by capillary bridges, the Discrete Element Method (DEM) enables direct simulation of interactions between the soil skeleton and pore fluids, providing microscopic insights into the macroscopic behaviour of unsaturated granular soils. However, most existing DEM simulations adopt capillary bridge models based on the assumption that particles are smooth, which tends to an overestimation of capillary stresses. In this study, an enhanced capillary bridge model that accounts for particle surface roughness is implemented to conduct DEM simulations of triaxial and biaxial tests on unsaturated granular materials. The results demonstrate that incorporating surface roughness into the capillary bridge model allows the Love-Weber type capillary stress to more accurately capture the transition from the dry to the pendular state. During triaxial shearing, capillary cohesion significantly enhances the shear strength and dilatancy of unsaturated granular materials. However, this effect diminishes with increasing particle surface roughness, resulting in mechanical responses that more closely resemble those of dry granular materials. During biaxial shearing, shear bands form in unsaturated granular materials, with capillary effects leading to narrower and more localised deformation zones. Within shear bands, the rupture of capillary bridges occurs more frequently, resulting in lower capillary stress values. This study provides a more realistic approach for future investigations into the hydro-mechanical response of unsaturated granular materials.
{"title":"Effects of particle surface roughness on the mechanical behaviour of unsaturated granular materials: DEM simulations","authors":"Hanze Li, Chaofa Zhao, Yi Hong, Aoxi Zhang, Yanni Chen","doi":"10.1016/j.compgeo.2026.108014","DOIUrl":"10.1016/j.compgeo.2026.108014","url":null,"abstract":"<div><div>Natural soils are typically unsaturated, and their mechanical behaviour is significantly influenced by capillary cohesion. By incorporating the attractive forces generated by capillary bridges, the Discrete Element Method (DEM) enables direct simulation of interactions between the soil skeleton and pore fluids, providing microscopic insights into the macroscopic behaviour of unsaturated granular soils. However, most existing DEM simulations adopt capillary bridge models based on the assumption that particles are smooth, which tends to an overestimation of capillary stresses. In this study, an enhanced capillary bridge model that accounts for particle surface roughness is implemented to conduct DEM simulations of triaxial and biaxial tests on unsaturated granular materials. The results demonstrate that incorporating surface roughness into the capillary bridge model allows the Love-Weber type capillary stress to more accurately capture the transition from the dry to the pendular state. During triaxial shearing, capillary cohesion significantly enhances the shear strength and dilatancy of unsaturated granular materials. However, this effect diminishes with increasing particle surface roughness, resulting in mechanical responses that more closely resemble those of dry granular materials. During biaxial shearing, shear bands form in unsaturated granular materials, with capillary effects leading to narrower and more localised deformation zones. Within shear bands, the rupture of capillary bridges occurs more frequently, resulting in lower capillary stress values. This study provides a more realistic approach for future investigations into the hydro-mechanical response of unsaturated granular materials.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"193 ","pages":"Article 108014"},"PeriodicalIF":6.2,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147385471","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-05-01Epub Date: 2026-02-04DOI: 10.1016/j.compgeo.2026.107940
Fei Song , Antonio Gens , Stefano Collico , Yuan Mi , Huaning Wang , Xiaocong Lyu
In Belgium, Boom Clay is selected as a potentially suitable host rock for the geological nuclear waste disposal. As part of the European Joint Programme on Radioactive Waste Management programme, this study concentrates on the analysis of the couple thermo-hydro-mechanical (THM) processes of the in-situ PRACLAY Heater Test throughout all the construction stages, including gallery excavation, support installation, waiting period, artificial water injection and heating stages. For this purpose, an elasto-viscoplastic model with a hyperbolic Mohr-Coulomb yield surface is selected to characterise the mechanical behaviour of Boom Clay, considering anisotropy of stiffness and strength as well as hardening and softening behaviour. One of the innovative points of this work is the implementation of air gap elements to represent over-excavation in a fully coupled THM framework. The new element is verified against existing numerical results for over-excavation simulations using contact elements and validated against measured data from the in-situ PRACLAY Heater Test. Finally, a comprehensive parametric program of analyses is conducted to investigate the effect of key parameters, such as elastic moduli, Poisson’s ratios, hydro-mechanical coupling properties as well as hydraulic boundary conditions on the THM behaviour of geological repositories for nuclear waste disposal.
{"title":"Numerical investigations of in-situ PRACLAY heater test: over-excavation, strain localisation and THM responses","authors":"Fei Song , Antonio Gens , Stefano Collico , Yuan Mi , Huaning Wang , Xiaocong Lyu","doi":"10.1016/j.compgeo.2026.107940","DOIUrl":"10.1016/j.compgeo.2026.107940","url":null,"abstract":"<div><div>In Belgium, Boom Clay is selected as a potentially suitable host rock for the geological nuclear waste disposal. As part of the European Joint Programme on Radioactive Waste Management programme, this study concentrates on the analysis of the couple thermo-hydro-mechanical (THM) processes of the in-situ PRACLAY Heater Test throughout all the construction stages, including gallery excavation, support installation, waiting period, artificial water injection and heating stages. For this purpose, an elasto-viscoplastic model with a hyperbolic Mohr-Coulomb yield surface is selected to characterise the mechanical behaviour of Boom Clay, considering anisotropy of stiffness and strength as well as hardening and softening behaviour. One of the innovative points of this work is the implementation of air gap elements to represent over-excavation in a fully coupled THM framework. The new element is verified against existing numerical results for over-excavation simulations using contact elements and validated against measured data from the in-situ PRACLAY Heater Test. Finally, a comprehensive parametric program of analyses is conducted to investigate the effect of key parameters, such as elastic moduli, Poisson’s ratios, hydro-mechanical coupling properties as well as hydraulic boundary conditions on the THM behaviour of geological repositories for nuclear waste disposal.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"193 ","pages":"Article 107940"},"PeriodicalIF":6.2,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146174385","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-05-01Epub Date: 2026-02-03DOI: 10.1016/j.compgeo.2026.107949
Tingfang Liu , Gang Wang , Changsheng Wang , Yujing Jiang , Xuezhen Wu , Wen Zheng , Feng Jiang , Weimin Yang , Jinglong Li
The shear behavior and failure mechanisms of non-persistent joints are key to the stability of jointed rock masses, whose shear responses are jointly governed by geometric parameters such as joint aperture and joint persistence. In this study, direct shear tests were performed on specimens containing coplanar non-persistent joints, and the shear-failure process was simulated using the finite element method–cohesive zone model (FEM–CZM) method. The combined effects of joint aperture and joint persistence on shear behavior were investigated from both macroscopic and mesoscopic perspectives, and an improved Jennings shear strength criterion incorporating the weakening effect of joint aperture was derived. The tests revealed two typical post-peak failure patterns: a “sudden drop followed by arcuate recovery” and a “stepwise decline”. Increases in both the joint aperture and joint persistence reduce the peak shear strength, with joint persistence exerting a more pronounced influence. Larger joint apertures increase the degrees of rock bridge fracture surface undulation and specimen surface spalling, whereas higher joint persistence flattens the fracture surface and mitigates surface spalling. Simulations indicate that stress initially concentrates at the rock bridge ends and extends towards the middle during shearing. The number of cracks increases sharply at the peak shear stress, with tensile cracks consistently dominating. Larger joint apertures intensify the stress concentration at the rock bridge ends, leading to earlier crack initiation, a more vigorous crack propagation trend, and more dispersed crack paths, whereas higher joint persistence narrows the stress concentration zone and accelerates crack coalescence across the rock bridge. Finally, based on the test and simulation results, an improved Jennings shear strength criterion is proposed by introducing a cohesion reduction coefficient η(d) that decays exponentially with joint aperture. The validation results demonstrate that the predicted peak shear strengths agree well with the measured values and external data.
{"title":"Effects of the joint aperture and persistence on the shear behavior of coplanar non-persistent jointed rock masses and an improved Jennings shear strength criterion","authors":"Tingfang Liu , Gang Wang , Changsheng Wang , Yujing Jiang , Xuezhen Wu , Wen Zheng , Feng Jiang , Weimin Yang , Jinglong Li","doi":"10.1016/j.compgeo.2026.107949","DOIUrl":"10.1016/j.compgeo.2026.107949","url":null,"abstract":"<div><div>The shear behavior and failure mechanisms of non-persistent joints are key to the stability of jointed rock masses, whose shear responses are jointly governed by geometric parameters such as joint aperture and joint persistence. In this study, direct shear tests were performed on specimens containing coplanar non-persistent joints, and the shear-failure process was simulated using the finite element method–cohesive zone model (FEM–CZM) method. The combined effects of joint aperture and joint persistence on shear behavior were investigated from both macroscopic and mesoscopic perspectives, and an improved Jennings shear strength criterion incorporating the weakening effect of joint aperture was derived. The tests revealed two typical post-peak failure patterns: a “sudden drop followed by arcuate recovery” and a “stepwise decline”. Increases in both the joint aperture and joint persistence reduce the peak shear strength, with joint persistence exerting a more pronounced influence. Larger joint apertures increase the degrees of rock bridge fracture surface undulation and specimen surface spalling, whereas higher joint persistence flattens the fracture surface and mitigates surface spalling. Simulations indicate that stress initially concentrates at the rock bridge ends and extends towards the middle during shearing. The number of cracks increases sharply at the peak shear stress, with tensile cracks consistently dominating. Larger joint apertures intensify the stress concentration at the rock bridge ends, leading to earlier crack initiation, a more vigorous crack propagation trend, and more dispersed crack paths, whereas higher joint persistence narrows the stress concentration zone and accelerates crack coalescence across the rock bridge. Finally, based on the test and simulation results, an improved Jennings shear strength criterion is proposed by introducing a cohesion reduction coefficient <em>η</em>(<em>d</em>) that decays exponentially with joint aperture. The validation results demonstrate that the predicted peak shear strengths agree well with the measured values and external data.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"193 ","pages":"Article 107949"},"PeriodicalIF":6.2,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146174473","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-05-01Epub Date: 2026-01-29DOI: 10.1016/j.compgeo.2026.107939
Wentao Yang, Shaobo Jin, Qiuhua Rao, Wei Yi, Zelin Liu
The seepage-stress coupling effect is critical to the stability of engineering structures, but existing models have certain limitations: the equivalent continuum model neglects seepage differences between pores and fractures; the discrete fracture network model overlooks pore permeability and involves complex calculations; and the dual medium model fails to account for non-Darcy seepage effects and the spatiotemporal evolution of physical parameters. Consequently, solving the seepage-stress coupling problem of spatially heterogeneous pore-fracture dual media remains challenging. In this study, we incorporate two key effects into the modeling framework: the non-Darcy seepage and the nonlinear seepage-stress coupling (including the spatiotemporal evolution of seepage parameters, solid deformation parameters, and effective stress coefficients). The model is developed via self-programming in COMSOL Multiphysics, and its reliability is validated through lab experiments. Results show that both effects suppress the dissipation of fluid pressure, the consolidation rate, and the final settlement of the soil. The degree of such suppression is governed by fracture permeability, pore permeability, and their ratio. Additionally, the practicality of this model is verified through a 3D engineering case. The model can be further extended to a multi-scale fracture network model (explicitly characterizing main fractures and implicitly describing randomly distributed secondary fractures and pores) with consideration of permeability anisotropy.
{"title":"A modified nonlinear seepage-stress coupling model of heterogeneous pore-fracture dual medium for engineering geomaterials","authors":"Wentao Yang, Shaobo Jin, Qiuhua Rao, Wei Yi, Zelin Liu","doi":"10.1016/j.compgeo.2026.107939","DOIUrl":"10.1016/j.compgeo.2026.107939","url":null,"abstract":"<div><div>The seepage-stress coupling effect is critical to the stability of engineering structures, but existing models have certain limitations: the equivalent continuum model neglects seepage differences between pores and fractures; the discrete fracture network model overlooks pore permeability and involves complex calculations; and the dual medium model fails to account for non-Darcy seepage effects and the spatiotemporal evolution of physical parameters. Consequently, solving the seepage-stress coupling problem of spatially heterogeneous pore-fracture dual media remains challenging. In this study, we incorporate two key effects into the modeling framework: the non-Darcy seepage and the nonlinear seepage-stress coupling (including the spatiotemporal evolution of seepage parameters, solid deformation parameters, and effective stress coefficients). The model is developed via self-programming in COMSOL Multiphysics, and its reliability is validated through lab experiments. Results show that both effects suppress the dissipation of fluid pressure, the consolidation rate, and the final settlement of the soil. The degree of such suppression is governed by fracture permeability, pore permeability, and their ratio. Additionally, the practicality of this model is verified through a 3D engineering case. The model can be further extended to a multi-scale fracture network model (explicitly characterizing main fractures and implicitly describing randomly distributed secondary fractures and pores) with consideration of permeability anisotropy.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"193 ","pages":"Article 107939"},"PeriodicalIF":6.2,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146080661","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-05-01Epub Date: 2026-01-30DOI: 10.1016/j.compgeo.2026.107957
Zhenjie Tang, Li He
Characterizing spatially variable hydraulic properties in geotechnical subsurface systems is a high-dimensional and ill-posed inverse problem, particularly in unsaturated flow conditions. We introduce TKLE-BPINN, a novel Bayesian framework that integrates Bayesian physics-informed neural networks (B-PINNs) with truncated Karhunen–Loève expansion (KLE) to address these challenges efficiently. By representing unknown fields with latent KLE coefficients and employing Ensemble Kalman Inversion (EKI) for posterior inference, TKLE-BPINN achieves accurate parameter estimation and robust uncertainty quantification. We validate the framework on three subsurface flow and transport test cases: (i) a linear diffusion–reaction problem with unknown diffusivity, (ii) a nonlinear diffusion–reaction problem with unknown reaction coefficients, and (iii) a multi-parameter inversion of diffusivity and contaminant source locations, where TKLE-BPINN consistently outperforms standard B-PINNs in accuracy and stability. We further apply TKLE-BPINN to a realistic hydrogeotechnical scenario, inverting saturated hydraulic conductivity in nonlinear unsaturated flow for heterogeneous soils, yielding reliable parameter estimates and uncertainty quantification essential for subsurface flow modeling. These results indicate that TKLE-BPINN provides a flexible and robust alternative for geotechnical subsurface characterization.
{"title":"TKLE-BPINN: A Bayesian physics-informed inversion framework for high-dimensional parameter identification in geotechnical subsurface systems","authors":"Zhenjie Tang, Li He","doi":"10.1016/j.compgeo.2026.107957","DOIUrl":"10.1016/j.compgeo.2026.107957","url":null,"abstract":"<div><div>Characterizing spatially variable hydraulic properties in geotechnical subsurface systems is a high-dimensional and ill-posed inverse problem, particularly in unsaturated flow conditions. We introduce TKLE-BPINN, a novel Bayesian framework that integrates Bayesian physics-informed neural networks (B-PINNs) with truncated Karhunen–Loève expansion (KLE) to address these challenges efficiently. By representing unknown fields with latent KLE coefficients and employing Ensemble Kalman Inversion (EKI) for posterior inference, TKLE-BPINN achieves accurate parameter estimation and robust uncertainty quantification. We validate the framework on three subsurface flow and transport test cases: (i) a linear diffusion–reaction problem with unknown diffusivity, (ii) a nonlinear diffusion–reaction problem with unknown reaction coefficients, and (iii) a multi-parameter inversion of diffusivity and contaminant source locations, where TKLE-BPINN consistently outperforms standard B-PINNs in accuracy and stability. We further apply TKLE-BPINN to a realistic hydrogeotechnical scenario, inverting saturated hydraulic conductivity in nonlinear unsaturated flow for heterogeneous soils, yielding reliable parameter estimates and uncertainty quantification essential for subsurface flow modeling. These results indicate that TKLE-BPINN provides a flexible and robust alternative for geotechnical subsurface characterization.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"193 ","pages":"Article 107957"},"PeriodicalIF":6.2,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146080662","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This study presents a multiphysics model and the corresponding algorithm for the numerical simulation of millimetre-wave ablation of geological materials. Simulations of this process are particularly challenging due to the low thermal conductivity and limited melt mobility of rock, as well as the high latent heat of evaporation and melt-to-vapour density ratio (exceeding 2500:1), which induce strong velocity divergence effects at the melt–vapour interface. Specifically, the model introduced in this work addresses the low-intensity, volumetric heating of rocks, taking place over significantly larger spatial and temporal scales (approximately 300 and 100 times greater, respectively) compared to similar processes encountered in laser drilling of metals. A comprehensive description is provided of the mathematical formulation, beam model, and numerical algorithm employed to perform robust, fully three-dimensional simulations lasting up to 25 s. These demanding computations are made tractable through MPI parallelisation and hierarchical Adaptive Mesh Refinement. Following the presentation of the model and solution methodology, an experiment is introduced which was devised specifically to produce data for model calibration and validation purposes. Comparison of the numerical results against the experimental reference test case and a parametric study of varying beam intensity demonstrate that the model can make good predictions of the penetration depth and the volume of material removed. The bounds and limitations of the model are also explored through the parametric study, with desirable future extensions identified to improve the accuracy of the model predictions.
{"title":"Multiphysics modelling of millimetre-wave ablation of geological materials","authors":"Nandan Gokhale , Candace Gilet , Franck Monmont , Nikos Nikiforakis","doi":"10.1016/j.compgeo.2026.107946","DOIUrl":"10.1016/j.compgeo.2026.107946","url":null,"abstract":"<div><div>This study presents a multiphysics model and the corresponding algorithm for the numerical simulation of millimetre-wave ablation of geological materials. Simulations of this process are particularly challenging due to the low thermal conductivity and limited melt mobility of rock, as well as the high latent heat of evaporation and melt-to-vapour density ratio (exceeding 2500:1), which induce strong velocity divergence effects at the melt–vapour interface. Specifically, the model introduced in this work addresses the low-intensity, volumetric heating of rocks, taking place over significantly larger spatial and temporal scales (approximately 300 and 100 times greater, respectively) compared to similar processes encountered in laser drilling of metals. A comprehensive description is provided of the mathematical formulation, beam model, and numerical algorithm employed to perform robust, fully three-dimensional simulations lasting up to 25 s. These demanding computations are made tractable through MPI parallelisation and hierarchical Adaptive Mesh Refinement. Following the presentation of the model and solution methodology, an experiment is introduced which was devised specifically to produce data for model calibration and validation purposes. Comparison of the numerical results against the experimental reference test case and a parametric study of varying beam intensity demonstrate that the model can make good predictions of the penetration depth and the volume of material removed. The bounds and limitations of the model are also explored through the parametric study, with desirable future extensions identified to improve the accuracy of the model predictions.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"193 ","pages":"Article 107946"},"PeriodicalIF":6.2,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146080668","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-05-01Epub Date: 2026-01-27DOI: 10.1016/j.compgeo.2026.107947
Guoliang Ma , Yang Xiao , Jinxuan Zhang , Zhichao Zhang
Microbially induced carbonate precipitation (MICP) has been widely studied for its broad application potential in geotechnical engineering. Exploring the pore-scale CaCO3 precipitation processes is one of the key aspects of fully understanding the mechanism of MICP for enhancing mechanical properties and controlling hydraulic properties of soils and rocks. Additionally, solution tests are typically the first step to understanding the reaction processes of MICP before conducting pore-scale and macro-scale tests. However, most explanations remain qualitative and do not establish a direct link between solution tests, pore-scale processes, and macro-scale performance. In this study, a saturation-dependent kinetic model is developed to describe the chemical reactions involved in MICP, and a phase-field model is employed to simulate the growth of CaCO3 crystals at the pore scale. The two models are coupled through the CaCO3 precipitation rate, enabling direct translation of solution test parameters into pore-scale simulations. Model predictions are validated against experimental data from microdroplet tests, showing good agreement in crystal growth dynamics under varying bacterial densities. The results demonstrate the capability of the proposed approach to quantitatively link solution chemistry with pore-scale mineralization processes. This work provides a robust framework for further investigation of MICP-induced pore sealing and particle bridging, contributing to the optimization and design of bio-mediated ground improvement strategies.
{"title":"Relating solution tests to pore-scale CaCO3 crystal growth: Numerical simulation based on the phase field method","authors":"Guoliang Ma , Yang Xiao , Jinxuan Zhang , Zhichao Zhang","doi":"10.1016/j.compgeo.2026.107947","DOIUrl":"10.1016/j.compgeo.2026.107947","url":null,"abstract":"<div><div>Microbially induced carbonate precipitation (MICP) has been widely studied for its broad application potential in geotechnical engineering. Exploring the pore-scale CaCO<sub>3</sub> precipitation processes is one of the key aspects of fully understanding the mechanism of MICP for enhancing mechanical properties and controlling hydraulic properties of soils and rocks. Additionally, solution tests are typically the first step to understanding the reaction processes of MICP before conducting pore-scale and macro-scale tests. However, most explanations remain qualitative and do not establish a direct link between solution tests, pore-scale processes, and macro-scale performance. In this study, a saturation-dependent kinetic model is developed to describe the chemical reactions involved in MICP, and a phase-field model is employed to simulate the growth of CaCO<sub>3</sub> crystals at the pore scale. The two models are coupled through the CaCO<sub>3</sub> precipitation rate, enabling direct translation of solution test parameters into pore-scale simulations. Model predictions are validated against experimental data from microdroplet tests, showing good agreement in crystal growth dynamics under varying bacterial densities. The results demonstrate the capability of the proposed approach to quantitatively link solution chemistry with pore-scale mineralization processes. This work provides a robust framework for further investigation of MICP-induced pore sealing and particle bridging, contributing to the optimization and design of bio-mediated ground improvement strategies.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"193 ","pages":"Article 107947"},"PeriodicalIF":6.2,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146080673","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}