Computers and Geotechnics最新文献_第6页

A nonlinear flow model for 3D infilled crossed fractures incorporating the influence of filling particles 考虑充填颗粒影响的三维充填交叉裂缝非线性流动模型

IF 6.2 1区工程技术 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers and Geotechnics

Pub Date : 2025-12-22 DOI: 10.1016/j.compgeo.2025.107854

Xiao Li , Jianmin Zhong , Zhanyuan Zhu , Weimin Xiao

Crossed fractures typically form within fractured rock masses through tectonic processes and constitute essential components of fracture networks. However, most natural rock fractures are filled with different types of fine-grained materials to form infilled rock fractures. Therefore, quantifying the impact of filling particles on flow behavior within crossed fractures is critical for understanding the fluid flow through complex fracture networks. Up to now, the flow behavior within infilled crossed fractures has not yet been comprehensively investigated. To address this gap, this study proposes a method for constructing numerical models of infilled crossed fractures that can replicate natural particle stacking patterns while allowing precise control over porosity. Using this approach, infilled crossed fracture models with different porosities and a corresponding unfilled model were constructed, and a series of seepage simulations were conducted at varying inlet flow rates. The results indicate that filling particles promote channelized flow within crossed fractures, enhancing eddy formation in interstitial regions. This channelization produces a pronounced multi-peak velocity distribution, creating substantial velocity differences between adjacent flow paths. Additionally, the particles alter the flow redistribution and form breakthrough flow through specific local pore structures. Consequently, the nonlinear flow behavior is more pronounced in infilled crossed fractures than in unfilled counterparts. Subsequently, a nonlinear flow model for infilled crossed fractures is proposed by incorporating semi-empirical formulas for hydraulic aperture b_h and nonlinear coefficient β to modify coefficients A and B in the Forchheimer equation. Validation results demonstrate that the model accurately predicts hydraulic gradients, with errors typically below 15%. Furthermore, the model is further extended to predict hydraulic gradients in outlet branches and modified using numerical data. Our study provides valuable insights for predicting and evaluating the fluid flow in complex geological setting.

交叉裂缝通常是在断裂岩体中通过构造作用形成的，是裂缝网络的重要组成部分。然而，大多数天然岩石裂缝被不同类型的细粒材料填充，形成充填岩石裂缝。因此，量化充填颗粒对交叉裂缝内流动特性的影响对于理解流体在复杂裂缝网络中的流动至关重要。迄今为止，对充填交叉裂缝内的流动特性研究还不够全面。为了解决这一问题，本研究提出了一种构建充填交叉裂缝数值模型的方法，该模型可以复制自然颗粒堆积模式，同时可以精确控制孔隙度。利用该方法建立了不同孔隙度的充填交叉裂缝模型和相应的未充填交叉裂缝模型，并进行了不同进口流量下的一系列渗流模拟。结果表明，充填颗粒促进了交叉裂缝内的通道化流动，增强了间隙区涡的形成。这种通道化产生了明显的多峰速度分布，在相邻的流道之间产生了巨大的速度差异。此外，颗粒改变了流动的再分配，并通过特定的局部孔隙结构形成突破流。因此，充填交叉裂缝的非线性流动行为比未充填交叉裂缝更为明显。随后，采用水力孔径bh半经验公式和非线性系数β修正Forchheimer方程中的系数a和系数B，建立了充填交叉裂缝的非线性渗流模型。验证结果表明，该模型准确地预测了水力梯度，误差一般在15%以下。在此基础上，对模型进行了扩展，以预测出水口支路的水力梯度，并利用数值数据对模型进行了修正。本研究为复杂地质环境下流体流动的预测和评价提供了有价值的见解。

{"title":"A nonlinear flow model for 3D infilled crossed fractures incorporating the influence of filling particles","authors":"Xiao Li , Jianmin Zhong , Zhanyuan Zhu , Weimin Xiao","doi":"10.1016/j.compgeo.2025.107854","DOIUrl":"10.1016/j.compgeo.2025.107854","url":null,"abstract":"<div><div>Crossed fractures typically form within fractured rock masses through tectonic processes and constitute essential components of fracture networks. However, most natural rock fractures are filled with different types of fine-grained materials to form infilled rock fractures. Therefore, quantifying the impact of filling particles on flow behavior within crossed fractures is critical for understanding the fluid flow through complex fracture networks. Up to now, the flow behavior within infilled crossed fractures has not yet been comprehensively investigated. To address this gap, this study proposes a method for constructing numerical models of infilled crossed fractures that can replicate natural particle stacking patterns while allowing precise control over porosity. Using this approach, infilled crossed fracture models with different porosities and a corresponding unfilled model were constructed, and a series of seepage simulations were conducted at varying inlet flow rates. The results indicate that filling particles promote channelized flow within crossed fractures, enhancing eddy formation in interstitial regions. This channelization produces a pronounced multi-peak velocity distribution, creating substantial velocity differences between adjacent flow paths. Additionally, the particles alter the flow redistribution and form breakthrough flow through specific local pore structures. Consequently, the nonlinear flow behavior is more pronounced in infilled crossed fractures than in unfilled counterparts. Subsequently, a nonlinear flow model for infilled crossed fractures is proposed by incorporating semi-empirical formulas for hydraulic aperture <em>b<sub>h</sub></em> and nonlinear coefficient <em>β</em> to modify coefficients <em>A</em> and <em>B</em> in the Forchheimer equation. Validation results demonstrate that the model accurately predicts hydraulic gradients, with errors typically below 15%. Furthermore, the model is further extended to predict hydraulic gradients in outlet branches and modified using numerical data. Our study provides valuable insights for predicting and evaluating the fluid flow in complex geological setting.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"192 ","pages":"Article 107854"},"PeriodicalIF":6.2,"publicationDate":"2025-12-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145839293","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}

引用次数: 0

A unified thermo-mechanical constitutive model for sand and clay based on hypoplasticity 基于低塑性的砂土统一热-力学本构模型

IF 6.2 1区工程技术 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers and Geotechnics

Pub Date : 2025-12-22 DOI: 10.1016/j.compgeo.2025.107827

Qianqian Liu, Chao Shi

Existing hypoplastic models treat sands and clays as different materials with distinct mathematical formulations. However, sands and clays share many common features in their mechanical behavior, and there is still no unified hypoplastic framework that can concisely describe the behavior of different soil types under various environmental factors, such as temperature variations. To address these challenges, this study develops a unified hypoplastic model that can describe the thermo-mechanical behavior of various soil types. A key compression hardening parameter is first introduced into the reference hypoplastic framework for clays, enabling a unified description of the strength, stiffness, and volumetric behavior of both sands and clays. Validation against shear test results for Toyoura sand demonstrates that the model effectively captures the mechanical responses of sands under varying initial states and confining pressures. Subsequently, the unified model incorporates a temperature-dependent isotropic compression line and extends the hypoplastic formulation to capture the influence of temperature on compressibility, shear behavior, and volumetric changes. The resulting unified thermo-mechanical constitutive model (Hypo-UT) demonstrates strong capability in reproducing the thermo-mechanical responses of both granular and fine-grained soils under complex thermal and mechanical loading paths, as validated by a series of drained and undrained triaxial shear tests and heating–cooling experiments. The model naturally reduces to the isothermal hypoplastic formulation at a reference temperature and further simplifies to the respective sand and clay models under specific conditions, highlighting its unifying and extensible structure. The Hypo-UT model can be readily calibrated through standard tests, providing a practical and robust tool for simulating and analyzing soil behavior under thermal loading in geotechnical engineering.

现有的发育不良模型将砂和粘土视为不同的材料，并采用不同的数学公式。然而，砂土和粘土在力学行为上有许多共同的特征，目前还没有统一的发育不全框架能够简明地描述不同土壤类型在不同环境因素（如温度变化）下的行为。为了解决这些挑战，本研究开发了一个统一的发育不良模型，可以描述各种土壤类型的热力学行为。一个关键的压缩硬化参数首次被引入到粘土的参考低塑性框架中，从而能够统一描述砂土和粘土的强度、刚度和体积行为。对Toyoura砂的剪切试验结果验证表明，该模型有效地捕捉了砂在不同初始状态和围压下的力学响应。随后，统一模型纳入了与温度相关的各向同性压缩线，并扩展了发育不良公式，以捕捉温度对可压缩性、剪切行为和体积变化的影响。通过一系列排水和不排水三轴剪切试验和加热-冷却试验验证了所建立的统一热-力学本构模型（sub - ut）能够较好地再现颗粒土和细粒土在复杂热-力学加载路径下的热-力学响应。该模型在参考温度下自然简化为等温欠塑性公式，在特定条件下进一步简化为各自的砂和粘土模型，突出了其统一性和可扩展性结构。该模型可以很容易地通过标准测试进行校准，为模拟和分析岩土工程中热载荷作用下的土壤行为提供了实用和可靠的工具。

{"title":"A unified thermo-mechanical constitutive model for sand and clay based on hypoplasticity","authors":"Qianqian Liu, Chao Shi","doi":"10.1016/j.compgeo.2025.107827","DOIUrl":"10.1016/j.compgeo.2025.107827","url":null,"abstract":"<div><div>Existing hypoplastic models treat sands and clays as different materials with distinct mathematical formulations. However, sands and clays share many common features in their mechanical behavior, and there is still no unified hypoplastic framework that can concisely describe the behavior of different soil types under various environmental factors, such as temperature variations. To address these challenges, this study develops a unified hypoplastic model that can describe the thermo-mechanical behavior of various soil types. A key compression hardening parameter is first introduced into the reference hypoplastic framework for clays, enabling a unified description of the strength, stiffness, and volumetric behavior of both sands and clays. Validation against shear test results for Toyoura sand demonstrates that the model effectively captures the mechanical responses of sands under varying initial states and confining pressures. Subsequently, the unified model incorporates a temperature-dependent isotropic compression line and extends the hypoplastic formulation to capture the influence of temperature on compressibility, shear behavior, and volumetric changes. The resulting unified thermo-mechanical constitutive model (Hypo-UT) demonstrates strong capability in reproducing the thermo-mechanical responses of both granular and fine-grained soils under complex thermal and mechanical loading paths, as validated by a series of drained and undrained triaxial shear tests and heating–cooling experiments. The model naturally reduces to the isothermal hypoplastic formulation at a reference temperature and further simplifies to the respective sand and clay models under specific conditions, highlighting its unifying and extensible structure. The Hypo-UT model can be readily calibrated through standard tests, providing a practical and robust tool for simulating and analyzing soil behavior under thermal loading in geotechnical engineering.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"192 ","pages":"Article 107827"},"PeriodicalIF":6.2,"publicationDate":"2025-12-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145839227","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}

引用次数: 0

Erratum to “Application of advanced soil constitutive models to investigate complex soil-structure interaction issues in integral abutment bridges”. [Comput. Geotech. 191 (2026) 107768] “应用先进土本构模型研究整体桥台复杂土-结构相互作用问题”的勘误。(第一版。岩土工程。191 (2026)107768 [j]

IF 6.2 1区工程技术 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers and Geotechnics

Pub Date : 2025-12-22 DOI: 10.1016/j.compgeo.2025.107842

M.S.K. Hassan , D.S. Liyanapathirana , W. Fuentes , J. Machaček

引用次数: 0

A novel physics–data hybrid approach for slope stability assessment considering future rainfall patterns 考虑未来降雨模式的边坡稳定性评估的一种新的物理-数据混合方法

IF 6.2 1区工程技术 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers and Geotechnics

Pub Date : 2025-12-21 DOI: 10.1016/j.compgeo.2025.107853

Zongyu Zhang , Kaiwen Liu , Qian Su , David P. Connolly , Xinzhuang Cui , Tengfei Wang , Guotao Hu

Accurately and dynamically assessing slope stability under changing rainfall patterns is essential for mitigating landslide risks in construction projects. Currently, slope stability analysis commonly relies on data-driven or physics-based methods, neglecting the complementary relationship between both approaches. Thus, this research proposes a slope stability prediction approach that combines numerical analysis techniques and data-driven artificial intelligence techniques. Using slope displacement monitoring data, geotechnical strength parameters are inferred via a random finite element analysis based on a neural network surrogate model. Furthermore, a data-driven time-series prediction model is established to forecast changes in slope geotechnical parameters, before using the strength reduction method to calculate the safety factor. The proposed approach is validated by analyzing two real-world slope cases and comparing it with traditional time-series data mapping methods for safety factor prediction. The results demonstrate that, when combined with meteorological forecast data, the proposed approach effectively captures the evolving trends in slope stability. The stability predictions outperform those based on time-series data mapping methods, providing accurate safety factor estimations and supporting the assessment of future slope failure risks in construction projects.

准确、动态地评估降雨模式变化下的边坡稳定性，对于降低建设工程中的滑坡风险至关重要。目前，边坡稳定性分析通常依赖于数据驱动或基于物理的方法，忽视了两者之间的互补关系。因此，本研究提出了一种结合数值分析技术和数据驱动的人工智能技术的边坡稳定性预测方法。利用边坡位移监测数据，通过基于神经网络代理模型的随机有限元分析，推断岩土强度参数。建立数据驱动的时间序列预测模型，预测边坡岩土参数变化，并采用强度折减法计算安全系数。通过对两个实际边坡实例的分析，并与传统的时序数据映射方法进行安全系数预测的比较，验证了该方法的有效性。结果表明，结合气象预报资料，该方法能有效地反映边坡稳定性的变化趋势。稳定性预测优于基于时间序列数据映射的方法，提供了准确的安全系数估计，并支持对未来建设项目中边坡失稳风险的评估。

{"title":"A novel physics–data hybrid approach for slope stability assessment considering future rainfall patterns","authors":"Zongyu Zhang , Kaiwen Liu , Qian Su , David P. Connolly , Xinzhuang Cui , Tengfei Wang , Guotao Hu","doi":"10.1016/j.compgeo.2025.107853","DOIUrl":"10.1016/j.compgeo.2025.107853","url":null,"abstract":"<div><div>Accurately and dynamically assessing slope stability under changing rainfall patterns is essential for mitigating landslide risks in construction projects. Currently, slope stability analysis commonly relies on data-driven or physics-based methods, neglecting the complementary relationship between both approaches. Thus, this research proposes a slope stability prediction approach that combines numerical analysis techniques and data-driven artificial intelligence techniques. Using slope displacement monitoring data, geotechnical strength parameters are inferred via a random finite element analysis based on a neural network surrogate model. Furthermore, a data-driven time-series prediction model is established to forecast changes in slope geotechnical parameters, before using the strength reduction method to calculate the safety factor. The proposed approach is validated by analyzing two real-world slope cases and comparing it with traditional time-series data mapping methods for safety factor prediction. The results demonstrate that, when combined with meteorological forecast data, the proposed approach effectively captures the evolving trends in slope stability. The stability predictions outperform those based on time-series data mapping methods, providing accurate safety factor estimations and supporting the assessment of future slope failure risks in construction projects.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"192 ","pages":"Article 107853"},"PeriodicalIF":6.2,"publicationDate":"2025-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145839294","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}

引用次数: 0

Data-driven ANN model for estimating unfrozen water content in the thermo-hydraulic simulation of frozen soils 基于数据驱动的人工神经网络模型估算冻土热水力模拟中未冻水含量

IF 6.2 1区工程技术 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers and Geotechnics

Pub Date : 2025-12-21 DOI: 10.1016/j.compgeo.2025.107846

Mingpeng Liu , Peizhi Zhuang , Raul Fuentes

This study integrates a data-driven model for estimating the unfrozen water content into the thermo-hydraulic coupled simulation of frozen soils. An artificial neural network (ANN) was employed to develop this data-driven model using a dataset from the literature. Thereafter, a numerical algorithm was developed to implement the data-driven model into the thermo-hydraulic simulation. In the numerical algorithm, the frozen and unfrozen zones are distinguished first according to the freezing temperature, where the unfrozen water at frozen nodes is updated using the ANN model. Subsequently, discretized hydraulic and thermal equations are solved sequentially and iteratively using the Newton-Raphson method. Horizontal and vertical freezing experiments are used to verify the reliability of the proposed algorithm. The computed variations in temperature, total water, unfrozen water, and ice content achieve good agreement with measured data. Some key features of frozen soils, such as water migration and ice formation, and the increase in total water content, are reproduced by the developed algorithm. Additionally, the comparison between the ANN model and existing empirical equations for determining unfrozen water content demonstrates that the ANN model offers better performance.

本研究将数据驱动的冻土水含量估算模型集成到冻土热-水耦合模拟中。使用文献中的数据集，采用人工神经网络（ANN）来开发该数据驱动模型。在此基础上，提出了一种数值算法，将数据驱动模型应用到热液仿真中。在数值算法中，首先根据冻结温度区分冻结区和未冻结区，其中使用人工神经网络模型更新冻结节点的未冻结水。随后，采用牛顿-拉夫逊法对离散的水力和热工方程进行了顺序迭代求解。通过水平和垂直冻结实验验证了算法的可靠性。计算得到的温度、总水量、未冻水量和冰含量的变化与实测数据吻合良好。该算法可以再现冻土的一些关键特征，如水分迁移和结冰，以及总含水量的增加。此外，将人工神经网络模型与现有的确定未冻水含量的经验方程进行比较，表明人工神经网络模型具有更好的性能。

{"title":"Data-driven ANN model for estimating unfrozen water content in the thermo-hydraulic simulation of frozen soils","authors":"Mingpeng Liu , Peizhi Zhuang , Raul Fuentes","doi":"10.1016/j.compgeo.2025.107846","DOIUrl":"10.1016/j.compgeo.2025.107846","url":null,"abstract":"<div><div>This study integrates a data-driven model for estimating the unfrozen water content into the thermo-hydraulic coupled simulation of frozen soils. An artificial neural network (ANN) was employed to develop this data-driven model using a dataset from the literature. Thereafter, a numerical algorithm was developed to implement the data-driven model into the thermo-hydraulic simulation. In the numerical algorithm, the frozen and unfrozen zones are distinguished first according to the freezing temperature, where the unfrozen water at frozen nodes is updated using the ANN model. Subsequently, discretized hydraulic and thermal equations are solved sequentially and iteratively using the Newton-Raphson method. Horizontal and vertical freezing experiments are used to verify the reliability of the proposed algorithm. The computed variations in temperature, total water, unfrozen water, and ice content achieve good agreement with measured data. Some key features of frozen soils, such as water migration and ice formation, and the increase in total water content, are reproduced by the developed algorithm. Additionally, the comparison between the ANN model and existing empirical equations for determining unfrozen water content demonstrates that the ANN model offers better performance.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"192 ","pages":"Article 107846"},"PeriodicalIF":6.2,"publicationDate":"2025-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145839292","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}

引用次数: 0

A fractal model of the gas diffusion coefficient in unsaturated bentonite accounting for the dual-porosity microstructure 考虑双重孔隙结构的非饱和膨润土气体扩散系数分形模型

IF 6.2 1区工程技术 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers and Geotechnics

Pub Date : 2025-12-21 DOI: 10.1016/j.compgeo.2025.107851

Pu-huai Lu , Wei-min Ye , Yu-heng Ji , Qiong Wang , Yong-gui Chen , Ju Wang

As a buffer/backfill material, bentonite is characterized by a dual-porosity structure that permits gas diffusion into the surrounding environment with possible leakage of nuclides. Accurate identification of gas diffusion mechanisms in multi-scale pore structures and prediction of the diffusion coefficient are crucial for assessing repository sealing performance. In this study, a pore diameter-based criterion was developed to identify gas diffusion patterns, including dissolved gas diffusion, Knudsen diffusion, combined diffusion, and molecular diffusion, and a fractal model was established for predicting the gas diffusion coefficient in unsaturated bentonite. Results reveal that the critical saturated pore diameter is enlarged to approximately 10^-7 m by increases in gas pressure and saturation, which enhances dissolved gas diffusion. Under rigid boundary conditions, a greater volume of desiccated pores is preserved compared to that with flexible boundaries, promoting molecular and combined diffusions. The model was validated using 18 gas diffusion tests on GMZ bentonite. Good agreement between predictions and measurements is observed under rigid conditions, where saturation remains stable. In contrast, under flexible boundaries, the diffusion coefficients were underestimated due to a 10–20 % saturation increase during testing. Model comparison confirms the superiority of the proposed model, with the smallest error (overall RMSE = 6.40 × 10^-11 m²/s) being achieved. By comparison, the gas diffusion coefficient is overestimated by 3–4 orders of magnitude using empirical models that employs an empirical reduction of the binary diffusion coefficient, such as the M−Q model, and is underestimated by the SAT model (total bias = -4.62 × 10^-11 m²/s), which only considers the dissolved gas diffusion.

膨润土作为一种缓冲/回填材料，具有双重孔隙结构，允许气体扩散到周围环境中，可能会有核素泄漏。准确识别气体在多尺度孔隙结构中的扩散机制并预测扩散系数是评价储库密封性能的关键。本文建立了一种基于孔径的气体扩散模式，包括溶解气体扩散、Knudsen扩散、复合扩散和分子扩散，并建立了一个分形模型来预测不饱和膨润土中的气体扩散系数。结果表明：随着气体压力和饱和度的增加，临界饱和孔径增大至10 ~ 7 m左右，溶解气体扩散加快；在刚性边界条件下，与柔性边界条件相比，保留了更大体积的干燥孔隙，促进了分子和组合扩散。通过18次GMZ膨润土气体扩散试验对模型进行了验证。在饱和保持稳定的刚性条件下，预测和测量之间的一致性很好。相反，在柔性边界下，由于测试过程中饱和度增加10 - 20%，扩散系数被低估。模型对比证实了所提模型的优越性，实现了最小误差（总体RMSE = 6.40 × 10-11 m2/s）。相比之下，使用经验模型（如M−Q模型）对二元扩散系数进行经验化简，气体扩散系数被高估了3-4个数量级，而仅考虑溶解气体扩散的SAT模型（总偏差= -4.62 × 10-11 m2/s）低估了气体扩散系数。

{"title":"A fractal model of the gas diffusion coefficient in unsaturated bentonite accounting for the dual-porosity microstructure","authors":"Pu-huai Lu , Wei-min Ye , Yu-heng Ji , Qiong Wang , Yong-gui Chen , Ju Wang","doi":"10.1016/j.compgeo.2025.107851","DOIUrl":"10.1016/j.compgeo.2025.107851","url":null,"abstract":"<div><div>As a buffer/backfill material, bentonite is characterized by a dual-porosity structure that permits gas diffusion into the surrounding environment with possible leakage of nuclides. Accurate identification of gas diffusion mechanisms in multi-scale pore structures and prediction of the diffusion coefficient are crucial for assessing repository sealing performance. In this study, a pore diameter-based criterion was developed to identify gas diffusion patterns, including dissolved gas diffusion, Knudsen diffusion, combined diffusion, and molecular diffusion, and a fractal model was established for predicting the gas diffusion coefficient in unsaturated bentonite. Results reveal that the critical saturated pore diameter is enlarged to approximately 10<sup>-7</sup> m by increases in gas pressure and saturation, which enhances dissolved gas diffusion. Under rigid boundary conditions, a greater volume of desiccated pores is preserved compared to that with flexible boundaries, promoting molecular and combined diffusions. The model was validated using 18 gas diffusion tests on GMZ bentonite. Good agreement between predictions and measurements is observed under rigid conditions, where saturation remains stable. In contrast, under flexible boundaries, the diffusion coefficients were underestimated due to a 10–20 % saturation increase during testing. Model comparison confirms the superiority of the proposed model, with the smallest error (overall RMSE = 6.40 × 10<sup>-11</sup> m<sup>2</sup>/s) being achieved. By comparison, the gas diffusion coefficient is overestimated by 3–4 orders of magnitude using empirical models that employs an empirical reduction of the binary diffusion coefficient, such as the M−Q model, and is underestimated by the SAT model (total bias = -4.62 × 10<sup>-11</sup> m<sup>2</sup>/s), which only considers the dissolved gas diffusion.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"192 ","pages":"Article 107851"},"PeriodicalIF":6.2,"publicationDate":"2025-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145839393","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}

引用次数: 0

Seismic risk assessment of railway embankments on spatially variable loose deposit slopes: A stacking ensemble machine learning-based approach 空间可变松散沉积斜坡上铁路路堤地震风险评估：基于叠加集成机器学习的方法

IF 6.2 1区工程技术 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers and Geotechnics

Pub Date : 2025-12-21 DOI: 10.1016/j.compgeo.2025.107841

Pan Si , Liang Tang , Shuang Tian , Yanfang Liu , Xianzhang Ling

This study presents a stacking ensemble machine learning (SEML) framework to evaluate the seismic risk of railway embankments on spatially variable loose deposits. A deterministic, three-dimensional finite difference model was employed to establish a bilinear probabilistic seismic demand model to quantify the relationship between seismic intensity and structural damage. Results demonstrated that the minor, moderate, and severe limited states (LS) were closely associated with peak permanent settlement (PPS) at the embankment crests. Multicriteria optimization identified characteristic intensity as the optimal intensity measure for seismic risk assessment. Furthermore, extensive seismic analyses were conducted using random fields of the internal friction angle in loose deposits, thereby establishing a comprehensive training database. Using this database, the trained SEML model accurately predicted PPS with improved computational efficiency compared to other ML models. The SEML model was subsequently employed to perform random field–based incremental dynamic analyses (20 ground motion records × 16 intensity levels × 280 random fields) to develop seismic fragility curves. Convolution with type-II extreme-value seismic hazard models yielded risk curves, which indicated elevated annual exceedance probabilities for LSs with increasing coefficient of variation (COV), horizontal correlation distance (δ_h), and vertical correlation distance (δ_v). The sensitivity ranking was as follows: COV > δ_h > δ_v. This study thus provides a computationally efficient probabilistic risk quantification tool for the seismic design of railway embankments on spatially variable loose deposit slopes.

本文提出了一种基于叠加集成机器学习（SEML）的框架来评估空间可变松散沉积物上铁路路堤的地震风险。采用确定性三维有限差分模型，建立双线性概率地震需求模型，量化地震烈度与结构损伤之间的关系。结果表明，轻度、中度和重度极限状态（LS）与路堤波峰永久沉降（PPS）密切相关。多准则优化确定了特征烈度作为地震危险性评价的最优烈度指标。此外，利用松散沉积物内摩擦角随机场进行了广泛的地震分析，从而建立了全面的训练数据库。使用该数据库，与其他ML模型相比，训练后的SEML模型准确地预测了PPS，计算效率提高。随后采用SEML模型进行基于随机场的增量动力分析（20个地震动记录× 16个烈度级别× 280个随机场），得到地震易损性曲线。与ii型极值地震灾害模型进行卷积得到的风险曲线表明，随着变异系数（COV）、水平相关距离（δh）和垂直相关距离（δv）的增加，LSs的年超越概率增加。灵敏度排序为：COV >； δh > δv。因此，该研究为空间可变松散沉积物边坡上的铁路路堤抗震设计提供了一种计算效率高的概率风险量化工具。

{"title":"Seismic risk assessment of railway embankments on spatially variable loose deposit slopes: A stacking ensemble machine learning-based approach","authors":"Pan Si , Liang Tang , Shuang Tian , Yanfang Liu , Xianzhang Ling","doi":"10.1016/j.compgeo.2025.107841","DOIUrl":"10.1016/j.compgeo.2025.107841","url":null,"abstract":"<div><div>This study presents a stacking ensemble machine learning (SEML) framework to evaluate the seismic risk of railway embankments on spatially variable loose deposits. A deterministic, three-dimensional finite difference model was employed to establish a bilinear probabilistic seismic demand model to quantify the relationship between seismic intensity and structural damage. Results demonstrated that the minor, moderate, and severe limited states (LS) were closely associated with peak permanent settlement (PPS) at the embankment crests. Multicriteria optimization identified characteristic intensity as the optimal intensity measure for seismic risk assessment. Furthermore, extensive seismic analyses were conducted using random fields of the internal friction angle in loose deposits, thereby establishing a comprehensive training database. Using this database, the trained SEML model accurately predicted PPS with improved computational efficiency compared to other ML models. The SEML model was subsequently employed to perform random field–based incremental dynamic analyses (20 ground motion records × 16 intensity levels × 280 random fields) to develop seismic fragility curves. Convolution with type-II extreme-value seismic hazard models yielded risk curves, which indicated elevated annual exceedance probabilities for LSs with increasing coefficient of variation (COV), horizontal correlation distance (<em>δ</em><sub>h</sub>), and vertical correlation distance (<em>δ</em><sub>v</sub>). The sensitivity ranking was as follows: COV > <em>δ</em><sub>h</sub> > <em>δ</em><sub>v</sub>. This study thus provides a computationally efficient probabilistic risk quantification tool for the seismic design of railway embankments on spatially variable loose deposit slopes.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"192 ","pages":"Article 107841"},"PeriodicalIF":6.2,"publicationDate":"2025-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145839394","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}

引用次数: 0

Roof stability analysis of deep buried tunnel based on the spatial discretization technique 基于空间离散化技术的深埋隧道顶板稳定性分析

IF 6.2 1区工程技术 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers and Geotechnics

Pub Date : 2025-12-20 DOI: 10.1016/j.compgeo.2025.107847

Fu Huang , Yongtao Wang , Longwu Pan , Qiujing Pan

The roof collapse induced by blast vibration and unfavorable geologies is common in tunnels excavated by the drill-and-blast method. Since many previous collapse mechanisms of rock mass above tunnel roof are constructed by multiple rigid blocks, the real collapse characteristics of the rock mass are not well investigated. In this study, new 2D and 3D collapse mechanisms of deep buried tunnel roofs are generated “point by point” on the basis of the spatial discretization technique for the first time. The tensile strength cut-off criterion is employed to construct the collapse mechanisms, which makes these mechanisms more consistent with the collapse surfaces observed in actual projects. The objective functions of the safety factor and the required supporting pressure are derived by the equation of virtual work. Comparisons between the theoretical results of the proposed method and those of the existing study and numerical simulation have been conducted, which prove the effectiveness of the proposed method. Parametric analysis indicates that the influence of the reduction in tensile strength of rock mass, along with the rock mass strength parameters and tunnel dimensions, cannot be ignored in assessing the stability of deep buried tunnel roof.

在钻爆法开挖的隧道中，爆破振动和不利地质条件引起的顶板垮塌是常见的。由于以往巷道顶板上方岩体的崩塌机制多由多个刚性块体构成，对岩体的真实崩塌特征研究不足。本研究首次基于空间离散化技术“逐点”生成了深埋巷道顶板新的二维和三维坍塌机理。采用抗拉强度截止准则构建了破坏机理，使其与实际工程中观察到的破坏面更加吻合。利用虚功方程推导出安全系数和所需支承压力的目标函数。将所提方法的理论结果与已有研究结果和数值模拟结果进行了比较，验证了所提方法的有效性。参数分析表明，在评价深埋巷道顶板稳定性时，岩体抗拉强度降低以及岩体强度参数和巷道尺寸的影响是不可忽视的。

{"title":"Roof stability analysis of deep buried tunnel based on the spatial discretization technique","authors":"Fu Huang , Yongtao Wang , Longwu Pan , Qiujing Pan","doi":"10.1016/j.compgeo.2025.107847","DOIUrl":"10.1016/j.compgeo.2025.107847","url":null,"abstract":"<div><div>The roof collapse induced by blast vibration and unfavorable geologies is common in tunnels excavated by the drill-and-blast method. Since many previous collapse mechanisms of rock mass above tunnel roof are constructed by multiple rigid blocks, the real collapse characteristics of the rock mass are not well investigated. In this study, new 2D and 3D collapse mechanisms of deep buried tunnel roofs are generated “point by point” on the basis of the spatial discretization technique for the first time. The tensile strength cut-off criterion is employed to construct the collapse mechanisms, which makes these mechanisms more consistent with the collapse surfaces observed in actual projects. The objective functions of the safety factor and the required supporting pressure are derived by the equation of virtual work. Comparisons between the theoretical results of the proposed method and those of the existing study and numerical simulation have been conducted, which prove the effectiveness of the proposed method. Parametric analysis indicates that the influence of the reduction in tensile strength of rock mass, along with the rock mass strength parameters and tunnel dimensions, cannot be ignored in assessing the stability of deep buried tunnel roof.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"191 ","pages":"Article 107847"},"PeriodicalIF":6.2,"publicationDate":"2025-12-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145797499","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}

引用次数: 0

An Interface-enhanced Nonlocal Model for Rocks with Heterogeneous Mineral Properties 非均质岩石的界面增强非局部模型

IF 6.2 1区工程技术 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers and Geotechnics

Pub Date : 2025-12-20 DOI: 10.1016/j.compgeo.2025.107855

Xiaokun Hu , Haitao Yu , Yuqi Sun

Rock consists of various mineral phases that exhibit significant differences in mechanical properties. Consequently, stress concentrations tend to develop at the interfaces between different minerals, promoting fracture initiation and ultimately reducing the overall strength of the rock. To address this issue, an interface-enhanced nonlocal model (IENM) is proposed in this study to capture the interfacial stress concentrations and predict the macroscopic mechanical response of rock based on its mineralogical composition. First, an interfacial layer is introduced at the boundary between different minerals to partition the strain energy according to the material properties within the nonlocal interaction domain. Second, the second-order peridynamic differential operator (PDDO) is employed near the interfacial layer to enhance the accuracy of the deformation gradient at the interface between different minerals. Then, the force state formulation of the interfacial layer is derived from the variational principle. The accuracy of the proposed model is demonstrated through several numerical examples, including a one-dimensional piecewise function analytical solution and a bi-mineral square plate under tension. Numerical results demonstrate that the proposed model accurately captures the strain discontinuities across mineral interfaces. Finally, the model is applied to simulate the mechanical response of a rock specimen reconstructed from mineralogical imaging, which provides an efficient numerical tool for investigating the interfacial compatibility and macro-meso mechanical performance affected by different mineral phases.

岩石由不同的矿物相组成，这些矿物相在力学特性上表现出显著的差异。因此，应力集中倾向于在不同矿物之间的界面处发展，促进破裂的发生，最终降低岩石的整体强度。为了解决这一问题，本研究提出了一种界面增强非局部模型（IENM）来捕捉界面应力集中，并根据岩石的矿物组成预测岩石的宏观力学响应。首先，在不同矿物之间的边界处引入界面层，根据非局部相互作用域内材料的性质划分应变能；其次，在界面层附近采用二阶动态微分算子（PDDO），提高了不同矿物界面处变形梯度的精度；然后，利用变分原理推导了界面层的力态表达式。通过一维分段函数解析解和双矿物方板在张力作用下的数值算例，验证了该模型的准确性。数值计算结果表明，该模型能准确地捕捉矿物界面上的应变不连续。最后，将该模型应用于矿物学成像重建岩石试样的力学响应模拟，为研究不同矿物相对界面相容性和宏观细观力学性能的影响提供了有效的数值工具。

{"title":"An Interface-enhanced Nonlocal Model for Rocks with Heterogeneous Mineral Properties","authors":"Xiaokun Hu , Haitao Yu , Yuqi Sun","doi":"10.1016/j.compgeo.2025.107855","DOIUrl":"10.1016/j.compgeo.2025.107855","url":null,"abstract":"<div><div>Rock consists of various mineral phases that exhibit significant differences in mechanical properties. Consequently, stress concentrations tend to develop at the interfaces between different minerals, promoting fracture initiation and ultimately reducing the overall strength of the rock. To address this issue, an interface-enhanced nonlocal model (IENM) is proposed in this study to capture the interfacial stress concentrations and predict the macroscopic mechanical response of rock based on its mineralogical composition. First, an interfacial layer is introduced at the boundary between different minerals to partition the strain energy according to the material properties within the nonlocal interaction domain. Second, the second-order peridynamic differential operator (PDDO) is employed near the interfacial layer to enhance the accuracy of the deformation gradient at the interface between different minerals. Then, the force state formulation of the interfacial layer is derived from the variational principle. The accuracy of the proposed model is demonstrated through several numerical examples, including a one-dimensional piecewise function analytical solution and a bi-mineral square plate under tension. Numerical results demonstrate that the proposed model accurately captures the strain discontinuities across mineral interfaces. Finally, the model is applied to simulate the mechanical response of a rock specimen reconstructed from mineralogical imaging, which provides an efficient numerical tool for investigating the interfacial compatibility and macro-meso mechanical performance affected by different mineral phases.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"191 ","pages":"Article 107855"},"PeriodicalIF":6.2,"publicationDate":"2025-12-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145797495","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}

引用次数: 0

A hybrid Finite Volume and Material Point Method (FVMPM) for simulating multiphase flow in and around deformable porous media 有限体积与物质点法（FVMPM）用于模拟可变形多孔介质内及周围多相流动

IF 6.2 1区工程技术 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers and Geotechnics

Pub Date : 2025-12-19 DOI: 10.1016/j.compgeo.2025.107831

Jun-Hao Wang, Wen-Jie Xu, Yang-Yang Zhang

Accurately and efficiently capturing the fluid–solid coupling behaviors of granular materials is of great significance. To address the challenge of simulating multiphase flow within and around porous media, a hybrid finite volume and material point method (FVMPM) algorithm is proposed. This approach combines the advantages of the finite volume method (FVM) for simulating complex flow fields and the material point method (MPM) for capturing large deformation behaviors of solids. The Navier-Stokes equations are solved to describe fluid dynamics around the porous media, while the Darcy-Brinkman-Stokes equations are used to model the seepage process within porous media. Additionally, the volume of fluid (VOF) method and capillary model are incorporated to handle the fluid interface and achieve the simulations of saturated and unsaturated flow. Therefore, the proposed FVMPM algorithm can accurately capture the complex interactions and kinematics between multiphase fluids and deformable porous media. In addition, GPU acceleration is employed to enhance the computational efficiency of the algorithm. To validate the accuracy and effectiveness of the proposed method, the one-dimensional consolidation test of a saturated soil column and unsaturated seepage in porous media are simulated, with the numerical results showing good agreement with analytical solutions. To further demonstrate the applicability of the algorithm, the Liakopoulos’ test and embankment seepage are performed, and the numerical results show excellent consistency with both the laboratory results and the results from finite element analysis. As a complex example of multiphase flow in and around deformable porous media, the dam breach process under the seepage failure and overtopping is simulated. The results indicate that the proposed FVMPM algorithm provides a unified numerical method for modeling the deformation and failure processes of porous media under the influence of both surface and subsurface water.

准确、高效地捕捉颗粒材料的流固耦合行为具有重要意义。为了解决多孔介质内部和周围多相流动模拟的难题，提出了一种有限体积-物质点法（FVMPM）混合算法。该方法结合了模拟复杂流场的有限体积法（FVM）和捕捉固体大变形行为的物质点法（MPM）的优点。通过求解Navier-Stokes方程来描述多孔介质周围的流体动力学，采用Darcy-Brinkman-Stokes方程来模拟多孔介质内部的渗流过程。结合流体体积法（VOF）和毛细管模型处理流体界面，实现了饱和和非饱和流动的模拟。因此，所提出的FVMPM算法可以准确地捕捉多相流体与可变形多孔介质之间复杂的相互作用和运动学。此外，采用GPU加速来提高算法的计算效率。为验证该方法的准确性和有效性，对饱和土柱的一维固结试验和多孔介质中非饱和渗流进行了模拟，数值结果与解析解吻合较好。为进一步验证该算法的适用性，进行了Liakopoulos试验和路堤渗流试验，数值计算结果与室内计算结果和有限元分析结果均具有较好的一致性。作为可变形多孔介质内部和周围多相流动的复杂实例，模拟了渗流破坏和漫顶作用下的溃坝过程。结果表明，本文提出的FVMPM算法为模拟地表和地下水作用下多孔介质的变形和破坏过程提供了统一的数值方法。

{"title":"A hybrid Finite Volume and Material Point Method (FVMPM) for simulating multiphase flow in and around deformable porous media","authors":"Jun-Hao Wang, Wen-Jie Xu, Yang-Yang Zhang","doi":"10.1016/j.compgeo.2025.107831","DOIUrl":"10.1016/j.compgeo.2025.107831","url":null,"abstract":"<div><div>Accurately and efficiently capturing the fluid–solid coupling behaviors of granular materials is of great significance. To address the challenge of simulating multiphase flow within and around porous media, a hybrid finite volume and material point method (FVMPM) algorithm is proposed. This approach combines the advantages of the finite volume method (FVM) for simulating complex flow fields and the material point method (MPM) for capturing large deformation behaviors of solids. The Navier-Stokes equations are solved to describe fluid dynamics around the porous media, while the Darcy-Brinkman-Stokes equations are used to model the seepage process within porous media. Additionally, the volume of fluid (VOF) method and capillary model are incorporated to handle the fluid interface and achieve the simulations of saturated and unsaturated flow. Therefore, the proposed FVMPM algorithm can accurately capture the complex interactions and kinematics between multiphase fluids and deformable porous media. In addition, GPU acceleration is employed to enhance the computational efficiency of the algorithm. To validate the accuracy and effectiveness of the proposed method, the one-dimensional consolidation test of a saturated soil column and unsaturated seepage in porous media are simulated, with the numerical results showing good agreement with analytical solutions. To further demonstrate the applicability of the algorithm, the Liakopoulos’ test and embankment seepage are performed, and the numerical results show excellent consistency with both the laboratory results and the results from finite element analysis. As a complex example of multiphase flow in and around deformable porous media, the dam breach process under the seepage failure and overtopping is simulated. The results indicate that the proposed FVMPM algorithm provides a unified numerical method for modeling the deformation and failure processes of porous media under the influence of both surface and subsurface water.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"191 ","pages":"Article 107831"},"PeriodicalIF":6.2,"publicationDate":"2025-12-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145796887","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}

引用次数: 0