Computers and Geotechnics最新文献_第6页

Dual-horizon peridynamic modeling of thermally induced fracture in anisotropic materials: A variational energy-based approach 各向异性材料热致裂缝的双层位动力学建模：一种基于变分能量的方法

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

Computers and Geotechnics

Pub Date : 2026-01-20 DOI: 10.1016/j.compgeo.2026.107916

Yongzheng Zhang , Jidong Zhao , Gang Wang , Xuxu Yang , Huilong Ren , Timon Rabczuk , Yunpeng Zhao

Fracturing phenomena driven by temperature variations present substantial challenges across geotechnical engineering applications. Accurate computational representation of such behavior demands robust numerical architectures that can reliably capture discontinuous crack evolution within materials characterized by pronounced directional dependencies. This paper introduces a novel approach that integrates dual-horizon non-ordinary state-based peridynamics (DH-NOSBPD) with a variational damage model for simulating thermo-mechanical fracture in anisotropic media. The proposed framework overcomes the limitations of conventional peridynamic methods in representing anisotropic thermal and mechanical coupling while eliminating numerical instabilities inherent in bond-breaking criteria. A staggered coupling strategy is employed to synchronize thermal and mechanical field updates, incorporating anisotropic constitutive relationships for both heat conduction and stress–strain behavior. The variational damage model introduces a history-dependent scalar damage field derived from strain energy density, thereby circumventing the spurious energy release and mesh dependence associated with abrupt bond deletion. This approach yields physically consistent crack evolution. Numerical examples validate the framework’s accuracy in anisotropic heat transfer, mechanical deformation, and complex fracture patterns under combined thermo-mechanical loading. The model demonstrates superior stability and predictive capability compared to conventional bond-breaking approaches.

温度变化导致的压裂现象给岩土工程应用带来了巨大挑战。这种行为的精确计算表示需要强大的数值架构，可以可靠地捕捉材料中具有明显方向依赖性的不连续裂纹演化。本文提出了一种将双视界非普通状态周动力学（h - nosbpd）与变分损伤模型相结合的方法，用于模拟各向异性介质中的热-机械断裂。提出的框架克服了传统的周动力学方法在表示各向异性热和力学耦合方面的局限性，同时消除了断裂准则固有的数值不稳定性。采用交错耦合策略同步热场和力学场更新，结合热传导和应力-应变行为的各向异性本构关系。变分损伤模型引入了由应变能密度导出的历史相关的标量损伤场，从而避免了与突然键缺失相关的虚假能量释放和网格依赖。这种方法产生物理上一致的裂纹演化。数值算例验证了该框架在热-机复合载荷下的各向异性传热、力学变形和复杂断裂模式的准确性。与传统的断键方法相比，该模型具有更好的稳定性和预测能力。

{"title":"Dual-horizon peridynamic modeling of thermally induced fracture in anisotropic materials: A variational energy-based approach","authors":"Yongzheng Zhang , Jidong Zhao , Gang Wang , Xuxu Yang , Huilong Ren , Timon Rabczuk , Yunpeng Zhao","doi":"10.1016/j.compgeo.2026.107916","DOIUrl":"10.1016/j.compgeo.2026.107916","url":null,"abstract":"<div><div>Fracturing phenomena driven by temperature variations present substantial challenges across geotechnical engineering applications. Accurate computational representation of such behavior demands robust numerical architectures that can reliably capture discontinuous crack evolution within materials characterized by pronounced directional dependencies. This paper introduces a novel approach that integrates dual-horizon non-ordinary state-based peridynamics (DH-NOSBPD) with a variational damage model for simulating thermo-mechanical fracture in anisotropic media. The proposed framework overcomes the limitations of conventional peridynamic methods in representing anisotropic thermal and mechanical coupling while eliminating numerical instabilities inherent in bond-breaking criteria. A staggered coupling strategy is employed to synchronize thermal and mechanical field updates, incorporating anisotropic constitutive relationships for both heat conduction and stress–strain behavior. The variational damage model introduces a history-dependent scalar damage field derived from strain energy density, thereby circumventing the spurious energy release and mesh dependence associated with abrupt bond deletion. This approach yields physically consistent crack evolution. Numerical examples validate the framework’s accuracy in anisotropic heat transfer, mechanical deformation, and complex fracture patterns under combined thermo-mechanical loading. The model demonstrates superior stability and predictive capability compared to conventional bond-breaking approaches.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"192 ","pages":"Article 107916"},"PeriodicalIF":6.2,"publicationDate":"2026-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146022678","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 GPU accelerated LBM-DEM framework for non-spherical convex particles and its application in permeability tests 非球形凸颗粒的GPU加速LBM-DEM框架及其在渗透率测试中的应用

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

Computers and Geotechnics

Pub Date : 2026-01-20 DOI: 10.1016/j.compgeo.2026.107935

Yecheng Qian, Lu Jing

Particle shape plays an important role in particle–fluid systems, but accurate numerical tools for simulating the interactions between non-spherical particles and the fluid are still limited. One of the challenges is the efficient and accurate evaluation of the solid volume fraction in fluid nodes for non-spherical particles. Here, we develop a GPU accelerated computational framework that couples the lattice Boltzmann method (LBM) for the fluid phase with the spheropolyhedra-based discrete element method (DEM) for particles. Using the immersed moving boundary method to resolve fluid particle interactions, we introduce a novel sub-grid and surface-normal approach to precisely calculate the solid volume fraction for each fluid node. Compared to the traditional approximate polyhedron method in LBM-DEM, this approach achieves higher accuracy while maintaining computational efficiency. The improved method is validated by volume calculation of a given geometry and drag coefficient of individual Platonic polyhedral particles under varying Reynolds numbers. Furthermore, the fully coupled LBM–IMB–spheropolyhedra DEM framework is validated through a single free settling non-spherical particle case in a viscous fluid. This framework is then applied to predict the permeability of random packings of various non-spherical particles, and the predicted values are found to be consistent with the classic Kozeny–Carman and Ergun equations. Finally, to provide a more general prediction model accounting for the particle shape, the Ergun equation is modified with the sphericity of non-spherical particles, which shows good agreement with our results and the data from the literature.

颗粒形状在颗粒-流体系统中起着重要的作用，但用于模拟非球形颗粒与流体之间相互作用的精确数值工具仍然有限。其中一个挑战是有效和准确地评估非球形颗粒在流体节点中的固体体积分数。在这里，我们开发了一个GPU加速计算框架，将晶格玻尔兹曼方法（LBM）与基于球多面体的离散元方法（DEM）相结合。利用浸入式移动边界法求解流体-颗粒相互作用，引入了一种新的子网格和表面法向方法来精确计算每个流体节点的固体体积分数。与传统的LBM-DEM近似多面体方法相比，该方法在保持计算效率的前提下实现了更高的精度。通过对给定几何形状和不同雷诺数下单个柏拉图多面体粒子阻力系数的体积计算，验证了改进方法的有效性。此外，通过粘性流体中单个自由沉降的非球形颗粒情况，验证了完全耦合的lbm - imb -球多面体DEM框架。然后将该框架应用于预测各种非球形颗粒随机填料的渗透率，发现预测值与经典的Kozeny-Carman和Ergun方程一致。最后，为了提供一个考虑粒子形状的更通用的预测模型，我们将Ergun方程用非球形粒子的球度进行了修正，结果与我们的结果和文献数据吻合得很好。

{"title":"A GPU accelerated LBM-DEM framework for non-spherical convex particles and its application in permeability tests","authors":"Yecheng Qian, Lu Jing","doi":"10.1016/j.compgeo.2026.107935","DOIUrl":"10.1016/j.compgeo.2026.107935","url":null,"abstract":"<div><div>Particle shape plays an important role in particle–fluid systems, but accurate numerical tools for simulating the interactions between non-spherical particles and the fluid are still limited. One of the challenges is the efficient and accurate evaluation of the solid volume fraction in fluid nodes for non-spherical particles. Here, we develop a GPU accelerated computational framework that couples the lattice Boltzmann method (LBM) for the fluid phase with the spheropolyhedra-based discrete element method (DEM) for particles. Using the immersed moving boundary method to resolve fluid particle interactions, we introduce a novel sub-grid and surface-normal approach to precisely calculate the solid volume fraction for each fluid node. Compared to the traditional approximate polyhedron method in LBM-DEM, this approach achieves higher accuracy while maintaining computational efficiency. The improved method is validated by volume calculation of a given geometry and drag coefficient of individual Platonic polyhedral particles under varying Reynolds numbers. Furthermore, the fully coupled LBM–IMB–spheropolyhedra DEM framework is validated through a single free settling non-spherical particle case in a viscous fluid. This framework is then applied to predict the permeability of random packings of various non-spherical particles, and the predicted values are found to be consistent with the classic Kozeny–Carman and Ergun equations. Finally, to provide a more general prediction model accounting for the particle shape, the Ergun equation is modified with the sphericity of non-spherical particles, which shows good agreement with our results and the data from the literature.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"192 ","pages":"Article 107935"},"PeriodicalIF":6.2,"publicationDate":"2026-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146022679","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

Quantifying the scale effects on shallow foundation bearing capacity induced by shear band on sensitive marine clays 敏感海相粘土剪切带对浅基础承载力的尺度效应量化

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

Computers and Geotechnics

Pub Date : 2026-01-20 DOI: 10.1016/j.compgeo.2026.107932

Chunlei Li , Long Yu , Yu Xin , Yunrui Han

Mapping from small-scale model tests to prototype design remains challenging in geotechnics, primarily due to scale effects induced by progressive failure in sensitive marine clays. While progressive failure is often negligible in small-scale models due to uniform mobilized strain within the failure soil, it becomes critical at the prototype scale. This study investigates these scale effects using a nonlocal Remeshing and Interpolation Technique with Small Strain (RITSS) method, incorporating a strain-softening constitutive model based on the simple modified elastic-perfectly plastic Tresca criterion. Simulations of plan-strain biaxial tests and shallow foundation penetration successfully capture shear band development. The results reveal that the extent of the strain-softening zone depends on the ratio of the model size to the intrinsic internal length of the clay, approximately the shear band thickness. A clear transition from localized to diffuse softening is observed as the model size decreases. Based on this finding, novel rescaling coefficients are proposed, incorporating the ratio of internal length to footing size and a clay ductility parameter. These coefficients enable more reliable prediction of the ultimate bearing capacity when mapping from reduced-scale models to prototypes in strain-softening clays.

在岩土工程中，从小规模模型试验到原型设计的映射仍然具有挑战性，主要是由于敏感的海相粘土中逐渐破坏引起的尺度效应。在小尺度模型中，渐进式破坏通常可以忽略不计，这是由于破坏土内部的均匀动员应变造成的，而在原型尺度上，渐进式破坏变得至关重要。本研究采用小应变非局部Remeshing和插值技术（RITSS）方法，结合基于简单修正弹塑性Tresca准则的应变软化本构模型，研究了这些尺度效应。平面-应变双轴试验和浅基础贯入模拟成功地捕捉到了剪切带的发展过程。结果表明，应变软化区的范围取决于模型尺寸与粘土固有内部长度（近似于剪切带厚度）的比值。随着模型尺寸的减小，可以观察到从局部软化到漫射软化的明显转变。基于这一发现，提出了新的重标系数，包括内部长度与基础尺寸的比例和粘土延性参数。这些系数使得在应变软化粘土中从缩尺模型映射到原型时能够更可靠地预测极限承载力。

{"title":"Quantifying the scale effects on shallow foundation bearing capacity induced by shear band on sensitive marine clays","authors":"Chunlei Li , Long Yu , Yu Xin , Yunrui Han","doi":"10.1016/j.compgeo.2026.107932","DOIUrl":"10.1016/j.compgeo.2026.107932","url":null,"abstract":"<div><div>Mapping from small-scale model tests to prototype design remains challenging in geotechnics, primarily due to scale effects induced by progressive failure in sensitive marine clays. While progressive failure is often negligible in small-scale models due to uniform mobilized strain within the failure soil, it becomes critical at the prototype scale. This study investigates these scale effects using a nonlocal Remeshing and Interpolation Technique with Small Strain (RITSS) method, incorporating a strain-softening constitutive model based on the simple modified elastic-perfectly plastic Tresca criterion. Simulations of plan-strain biaxial tests and shallow foundation penetration successfully capture shear band development. The results reveal that the extent of the strain-softening zone depends on the ratio of the model size to the intrinsic internal length of the clay, approximately the shear band thickness. A clear transition from localized to diffuse softening is observed as the model size decreases. Based on this finding, novel rescaling coefficients are proposed, incorporating the ratio of internal length to footing size and a clay ductility parameter. These coefficients enable more reliable prediction of the ultimate bearing capacity when mapping from reduced-scale models to prototypes in strain-softening clays.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"192 ","pages":"Article 107932"},"PeriodicalIF":6.2,"publicationDate":"2026-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146022680","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

Artificial neural network-based surrogate models for predicting multiple seismic slope engineering demand parameters derived from material-point analyses 基于人工神经网络的多地震边坡工程需求参数预测模型

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

Computers and Geotechnics

Pub Date : 2026-01-19 DOI: 10.1016/j.compgeo.2026.107928

Haixia Huang , Duruo Huang , Gang Wang

The approaches for developing surrogate models to predict seismic slope permanent displacement have progressed from sliding block analyses to stress-deformation analyses. Stress-deformation analyses in existing models are mostly mesh-based approaches incapable of modeling post-failure large deformation in slopes. This constraint restricts the derivation of typical engineering demand parameters (EDPs), such as run-out distance, influence distance, sliding depth and sliding volume. These EDPs are essential for quantitatively assessing post-failure behavior of slopes. The material point method (MPM) overcomes mesh distortion issues inherent in mesh-based approaches, making it possible to simulate large deformation of geomaterials. In this study, surrogate models for multiple seismic slope EDPs are developed based on material-point analyses. The Mohr-Coulomb model with strain softening is incorporated to simulate the strength loss of loose fills. The effect of strength loss on post-failure behavior of slopes is investigated. Artificial neural networks (ANNs) are then trained using the simulated dataset to establish relationships between slope EDPs and intensity measures. The developed ANN-based surrogate models significantly reduce computational time compared with MPM simulations while satisfying sufficiency and proficiency criteria. They also maintain physical consistency, as evidenced by the predicted shear band depth remaining within the slope height. Therefore, coupling material-point analyses for physics-based data generation with ANNs for surrogate model development provides an efficient and physically consistent framework for evaluating the seismic performance and post-failure behavior of slopes.

开发替代模型预测地震边坡永久位移的方法已经从滑块分析发展到应力变形分析。现有的应力变形分析模型大多是基于网格的方法，无法模拟边坡破坏后的大变形。这一约束限制了典型工程需求参数（edp）的推导，如跳动距离、影响距离、滑动深度和滑动体积。这些edp对于定量评估边坡破坏后的行为至关重要。材料点法（MPM）克服了基于网格的方法固有的网格畸变问题，使模拟岩土材料的大变形成为可能。在本研究中，基于物质点分析，建立了多个地震斜坡edp的代理模型。采用考虑应变软化的Mohr-Coulomb模型模拟松散充填体的强度损失。研究了强度损失对边坡破坏后行为的影响。然后使用模拟数据集训练人工神经网络（ann），以建立斜坡edp与强度测量之间的关系。所开发的基于人工神经网络的代理模型与MPM模拟相比显著减少了计算时间，同时满足充分性和熟练度标准。预测剪切带深度保持在坡高范围内，证明了两者在物理上的一致性。因此，将基于物理的数据生成的物质点分析与用于替代模型开发的人工神经网络相结合，为评估边坡的地震性能和破坏后行为提供了一个有效且物理一致的框架。

{"title":"Artificial neural network-based surrogate models for predicting multiple seismic slope engineering demand parameters derived from material-point analyses","authors":"Haixia Huang , Duruo Huang , Gang Wang","doi":"10.1016/j.compgeo.2026.107928","DOIUrl":"10.1016/j.compgeo.2026.107928","url":null,"abstract":"<div><div>The approaches for developing surrogate models to predict seismic slope permanent displacement have progressed from sliding block analyses to stress-deformation analyses. Stress-deformation analyses in existing models are mostly mesh-based approaches incapable of modeling post-failure large deformation in slopes. This constraint restricts the derivation of typical engineering demand parameters (<em>EDPs</em>), such as run-out distance, influence distance, sliding depth and sliding volume. These <em>EDPs</em> are essential for quantitatively assessing post-failure behavior of slopes. The material point method (MPM) overcomes mesh distortion issues inherent in mesh-based approaches, making it possible to simulate large deformation of geomaterials. In this study, surrogate models for multiple seismic slope <em>EDPs</em> are developed based on material-point analyses. The Mohr-Coulomb model with strain softening is incorporated to simulate the strength loss of loose fills. The effect of strength loss on post-failure behavior of slopes is investigated. Artificial neural networks (ANNs) are then trained using the simulated dataset to establish relationships between slope <em>EDPs</em> and intensity measures. The developed ANN-based surrogate models significantly reduce computational time compared with MPM simulations while satisfying sufficiency and proficiency criteria. They also maintain physical consistency, as evidenced by the predicted shear band depth remaining within the slope height. Therefore, coupling material-point analyses for physics-based data generation with ANNs for surrogate model development provides an efficient and physically consistent framework for evaluating the seismic performance and post-failure behavior of slopes.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"192 ","pages":"Article 107928"},"PeriodicalIF":6.2,"publicationDate":"2026-01-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146022772","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

Unified drag-force temporal discretization in the u-U-p fractional step method u-U-p分步法中统一阻力时间离散化

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

Computers and Geotechnics

Pub Date : 2026-01-19 DOI: 10.1016/j.compgeo.2026.107910

Zhaonan Wang , Louis Ngai Yuen Wong

The u-U-p formulation, which utilizes solid displacement, fluid displacement, and pressure as fundamental unknowns, is a preferred approach for computing the dynamic response of fluid-saturated geomaterials. In the past, the implicit monolithic solution was the conventional approach for finite element analysis. However monolithic solution encounters significant computational challenges due to their large-scale linear equation system and pronounced material nonlinearity involved. As a viable alternative, the semi-explicit–implicit fractional step method (FSM) demonstrates strong adaptability for rapid dynamic problems and can substantially reduce computational complexity compared to the monolithic solution. Nevertheless, the temporal discretization of the drag force within the FSM directly impacts the selection of time step sizes and computational accuracy. This study proposes a unified temporal discretization framework for the drag force in the u-U-p formulation. Depending on the choice of time factors, this method can be categorized into one explicit and three implicit schemes. Notably, a new implicit scheme significant improvements in computational accuracy over other schemes as verified by numerical examples. This unified drag-force fractional step method (UDF-FSM) strictly adheres to the process of explicit velocity prediction, implicit pressure solution, and explicit velocity correction, applicable to both high and low permeability problems.

u-U-p公式利用固体位移、流体位移和压力作为基本未知量，是计算流体饱和岩土材料动态响应的首选方法。在过去，隐式整体解是有限元分析的传统方法。然而，单片解由于其大规模的线性方程组和明显的材料非线性，在计算上面临着巨大的挑战。作为一种可行的替代方案，半显式-隐式分步法（FSM）对快速动态问题具有较强的适应性，与单片解决方案相比，可以大大降低计算复杂度。然而，FSM内阻力的时间离散化直接影响时间步长的选择和计算精度。本研究提出了u-U-p公式中阻力的统一时间离散化框架。根据时间因素的选择，该方法可分为一种显式方案和三种隐式方案。通过数值算例验证了该隐式格式在计算精度上的显著提高。统一阻力分步法（UDF-FSM）严格遵循显式速度预测、隐式压力求解、显式速度修正的过程，适用于高、低渗透问题。

{"title":"Unified drag-force temporal discretization in the u-U-p fractional step method","authors":"Zhaonan Wang , Louis Ngai Yuen Wong","doi":"10.1016/j.compgeo.2026.107910","DOIUrl":"10.1016/j.compgeo.2026.107910","url":null,"abstract":"<div><div>The <strong>u</strong>-<strong>U</strong>-<em>p</em> formulation, which utilizes solid displacement, fluid displacement, and pressure as fundamental unknowns, is a preferred approach for computing the dynamic response of fluid-saturated geomaterials. In the past, the implicit monolithic solution was the conventional approach for finite element analysis. However monolithic solution encounters significant computational challenges due to their large-scale linear equation system and pronounced material nonlinearity involved. As a viable alternative, the semi-explicit–implicit fractional step method (FSM) demonstrates strong adaptability for rapid dynamic problems and can substantially reduce computational complexity compared to the monolithic solution. Nevertheless, the temporal discretization of the drag force within the FSM directly impacts the selection of time step sizes and computational accuracy. This study proposes a unified temporal discretization framework for the drag force in the <strong>u</strong>-<strong>U</strong>-<em>p</em> formulation. Depending on the choice of time factors, this method can be categorized into one explicit and three implicit schemes. Notably, a new implicit scheme significant improvements in computational accuracy over other schemes as verified by numerical examples. This unified drag-force fractional step method (UDF-FSM) strictly adheres to the process of explicit velocity prediction, implicit pressure solution, and explicit velocity correction, applicable to both high and low permeability problems.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"192 ","pages":"Article 107910"},"PeriodicalIF":6.2,"publicationDate":"2026-01-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146022683","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

Rolling resistance as a surrogate for angularity in DEM: Macromechanical potentials and micromechanical limitations 滚动阻力作为DEM中角度的替代：宏观力学潜力和微观力学限制

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

Computers and Geotechnics

Pub Date : 2026-01-19 DOI: 10.1016/j.compgeo.2026.107919

Usman Ali , Mamoru Kikumoto

In this study, a calibrated DEM model of biaxial shearing tests was used to simulate polygon-shaped particles with varying corner numbers, representing different degrees of roundness (R). Simultaneously, rolling resistance (μ_r) was applied to conventional circular particles as a surrogate for shape angularity. The μ_r values were systematically calibrated to reproduce the stress–strain behaviour of the polygonal particles. Comparative analyses were then conducted at both macro- and micro-scales to evaluate the validity of the μ_r-based approximation. At the macroscopic level, the rolling resistance approach successfully captures the critical state stress ratio for all polygonal particles. At the microscopic scale, the μ_r approach reasonably reproduced key features such as particle rotations, coordination numbers, and contact force distribution for particles with moderate roundness (R > 0.5). However, for more angular particles (R ≤ 0.5), the μ_r approach failed to replicate the micromechanical behaviours inherent to angular shapes, despite aligning well in macroscopic stress–strain response. This limitation is primarily attributed to the absence of microstructural mechanisms unique to angular geometries, such as interlocking and the eccentric contact force paths, that inherently both induce and resist particle rotation. In contrast, rolling resistance solely resists rotation. This study highlights the importance of recognizing the limitations of rolling resistance in representing the micromechanics of angular particles. Careful interpretation is necessary to avoid misrepresenting underlying behaviours when modelling particle shape effects using this simplified approach.

本研究采用校正后的双轴剪切试验DEM模型，模拟具有不同角数（代表不同圆度R）的多边形颗粒。同时，将滚动阻力（μr）应用于常规圆形粒子，作为形状角度的替代。系统地校准了μr值，以再现多边形颗粒的应力-应变行为。然后在宏观和微观尺度上进行了比较分析，以评估基于μr的近似的有效性。在宏观层面上，滚动阻力方法成功地捕获了所有多边形粒子的临界状态应力比。在微观尺度上，μr方法合理再现了圆度适中（R > 0.5）的颗粒的旋转、配位数和接触力分布等关键特征。然而，对于更多的角状颗粒（R≤0.5），μr方法无法复制角状颗粒固有的微观力学行为，尽管在宏观应力-应变响应中表现得很好。这种限制主要归因于缺乏独特的角度几何结构的微观结构机制，如联锁和偏心接触力路径，固有地诱导和抵抗粒子旋转。与此相反，滚动阻力只抵抗旋转。这项研究强调了认识到滚动阻力在代表角粒子微观力学中的局限性的重要性。仔细的解释是必要的，以避免误读潜在的行为时，建模粒子形状效果使用这种简化的方法。

{"title":"Rolling resistance as a surrogate for angularity in DEM: Macromechanical potentials and micromechanical limitations","authors":"Usman Ali , Mamoru Kikumoto","doi":"10.1016/j.compgeo.2026.107919","DOIUrl":"10.1016/j.compgeo.2026.107919","url":null,"abstract":"<div><div>In this study, a calibrated DEM model of biaxial shearing tests was used to simulate polygon-shaped particles with varying corner numbers, representing different degrees of roundness (R). Simultaneously, rolling resistance (μ<sub>r</sub>) was applied to conventional circular particles as a surrogate for shape angularity. The μ<sub>r</sub> values were systematically calibrated to reproduce the stress–strain behaviour of the polygonal particles. Comparative analyses were then conducted at both macro- and micro-scales to evaluate the validity of the μ<sub>r</sub>-based approximation. At the macroscopic level, the rolling resistance approach successfully captures the critical state stress ratio for all polygonal particles. At the microscopic scale, the μ<sub>r</sub> approach reasonably reproduced key features such as particle rotations, coordination numbers, and contact force distribution for particles with moderate roundness (R > 0.5). However, for more angular particles (R ≤ 0.5), the μ<sub>r</sub> approach failed to replicate the micromechanical behaviours inherent to angular shapes, despite aligning well in macroscopic stress–strain response. This limitation is primarily attributed to the absence of microstructural mechanisms unique to angular geometries, such as interlocking and the eccentric contact force paths, that inherently both induce and resist particle rotation. In contrast, rolling resistance solely resists rotation. This study highlights the importance of recognizing the limitations of rolling resistance in representing the micromechanics of angular particles. Careful interpretation is necessary to avoid misrepresenting underlying behaviours when modelling particle shape effects using this simplified approach.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"192 ","pages":"Article 107919"},"PeriodicalIF":6.2,"publicationDate":"2026-01-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146022682","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

Uplift resistance mechanism of pipes in lightweight backfill material of ceramsite 陶粒轻质回填材料管道抗拔机理研究

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

Computers and Geotechnics

Pub Date : 2026-01-19 DOI: 10.1016/j.compgeo.2026.107927

Le Wang , Bowei Zhang , Yu Li , Bo Liu , Chenxi Tong , Hao Xiong , Guangsi Chen , Zhaohui Hong , Chunhui Zhang , Yinghui Tian

Buried pipelines crossing active fault zones undergo significant seismic displacement, inducing substantial soil resistance that can damage the pipe structure. Employing lightweight backfill materials offers a potential mitigation strategy. This study investigates the suitability of ceramsite—a lightweight material characterized by its smooth surface and low density—as a novel backfill, specifically examining why it generates lower resistance than conventional silica sand. Comprehensive evaluation combined element tests comparing ceramsite and silica sand properties with model experiments and discrete element method (DEM) simulations of pipe uplift. Results demonstrate significantly lower uplift resistance in ceramsite compared to silica sand, indicating its potential to enhance the seismic performance of pipelines in fault zones. DEM simulations attribute this reduced resistance to ceramsite’s lower self-weight, more convergent slip surfaces, and lower mobilized shear stresses on the slip surface. The convergent slip surfaces stem from ceramsite’s lower internal friction angle, enabling surrounding particles to slide into the void beneath the uplifted pipe, causing the slip surfaces to converge centrally. Stress monitoring confirmed significantly lower mobilized shear stresses in ceramsite. Furthermore, the stress path followed by both materials on the slip surface was found to nearly follow drained paths up to peak resistance. This observation challenges the constant-mean stress trajectory assumption prevalent in prior studies, which tends to underestimate uplift resistance.

埋地管道在穿越活动断裂带时，会产生较大的地震位移，从而产生较大的土体阻力，破坏管道结构。采用轻质回填材料提供了一种潜在的缓解策略。本研究调查了陶粒的适用性，陶粒是一种轻质材料，其表面光滑，密度低，是一种新型的回填材料，特别研究了为什么它比传统的硅砂产生更低的阻力。综合评价组合单元试验比较陶粒和硅砂的性能与模型试验和离散单元法（DEM）模拟管道提升。结果表明，陶粒比硅砂具有更低的隆升阻力，表明陶粒具有增强断裂带管道抗震性能的潜力。DEM模拟将这种降低的阻力归因于陶粒较低的自重、更收敛的滑移面以及滑移面上较低的动员剪切应力。趋同的滑移面源于陶粒较低的内摩擦角，使周围的颗粒滑入提升管下方的空隙中，导致滑移面集中聚集。应力监测证实陶粒的动员剪应力明显降低。此外，发现两种材料在滑移面上的应力路径几乎遵循排水路径直到峰值阻力。这一观察结果挑战了先前研究中普遍存在的恒定平均应力轨迹假设，该假设往往低估了隆升阻力。

{"title":"Uplift resistance mechanism of pipes in lightweight backfill material of ceramsite","authors":"Le Wang , Bowei Zhang , Yu Li , Bo Liu , Chenxi Tong , Hao Xiong , Guangsi Chen , Zhaohui Hong , Chunhui Zhang , Yinghui Tian","doi":"10.1016/j.compgeo.2026.107927","DOIUrl":"10.1016/j.compgeo.2026.107927","url":null,"abstract":"<div><div>Buried pipelines crossing active fault zones undergo significant seismic displacement, inducing substantial soil resistance that can damage the pipe structure. Employing lightweight backfill materials offers a potential mitigation strategy. This study investigates the suitability of ceramsite—a lightweight material characterized by its smooth surface and low density—as a novel backfill, specifically examining why it generates lower resistance than conventional silica sand. Comprehensive evaluation combined element tests comparing ceramsite and silica sand properties with model experiments and discrete element method (DEM) simulations of pipe uplift. Results demonstrate significantly lower uplift resistance in ceramsite compared to silica sand, indicating its potential to enhance the seismic performance of pipelines in fault zones. DEM simulations attribute this reduced resistance to ceramsite’s lower self-weight, more convergent slip surfaces, and lower mobilized shear stresses on the slip surface. The convergent slip surfaces stem from ceramsite’s lower internal friction angle, enabling surrounding particles to slide into the void beneath the uplifted pipe, causing the slip surfaces to converge centrally. Stress monitoring confirmed significantly lower mobilized shear stresses in ceramsite. Furthermore, the stress path followed by both materials on the slip surface was found to nearly follow drained paths up to peak resistance. This observation challenges the constant-mean stress trajectory assumption prevalent in prior studies, which tends to underestimate uplift resistance.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"192 ","pages":"Article 107927"},"PeriodicalIF":6.2,"publicationDate":"2026-01-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146022773","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

Stochastic collocation enhanced radial-circumferential importance sampling method for efficiently estimating rare failure probability with high-dimensional inputs and its application in slope structures 随机配置增强径向-周向重要抽样方法高效估计高维输入罕见破坏概率及其在边坡结构中的应用

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

Computers and Geotechnics

Pub Date : 2026-01-17 DOI: 10.1016/j.compgeo.2026.107920

Yuhua Yan , Zhenzhou Lu , Jiangji Qu , Xingcheng Wang

To efficiently address complex reliability problems involving high-dimensional random inputs and rare failure events, this paper proposes a stochastic collocation enhanced radial-circumferential importance sampling method. The proposed method designs an explicit and regular radial importance sampling density to efficiently explore the rare failure domain along the radial direction. By this importance sampling density, stochastic collocation points and corresponding weights of the radial variable are generated. Then the rare failure probability can be estimated by the weighted summation of the conditional failure probabilities at the given radial stochastic collocation points. To further improve the efficiency of estimating the conditional failure probabilities at different radial stochastic collocation points, this paper introduces a circumferential importance sampling strategy. The proposed radial-circumferential importance sampling in the proposed method does not require searching the most probable failure point to determine the importance sampling density, and it gradually explores the rare failure domain from the global to the local and then constructs the importance sampling density, which can cover the domains that contribute significantly to the failure probability with low computational cost, thereby reducing the variance of the failure probability estimation. The feasibility and superiority of the proposed method are verified by numerical and slope structure examples.

为了有效地解决包含高维随机输入和罕见失效事件的复杂可靠性问题，本文提出了一种随机配置增强径向-周向重要抽样方法。该方法设计了一个明确的、规则的径向重要采样密度，以有效地沿径向探索稀有破坏域。通过该重要抽样密度，生成随机配点和相应的径向变量权值。然后通过对给定径向随机配点处的条件失效概率的加权求和来估计罕见失效概率。为了进一步提高在不同径向随机配点处条件失效概率的估计效率，本文引入了一种环向重要抽样策略。提出的径向-周向重要采样方法不需要搜索最可能的失效点来确定重要采样密度，而是从全局到局部逐步探索罕见的失效域，然后构建重要采样密度，以较低的计算成本覆盖对失效概率有重要贡献的域，从而减小失效概率估计的方差。通过数值算例和边坡结构算例验证了该方法的可行性和优越性。

{"title":"Stochastic collocation enhanced radial-circumferential importance sampling method for efficiently estimating rare failure probability with high-dimensional inputs and its application in slope structures","authors":"Yuhua Yan , Zhenzhou Lu , Jiangji Qu , Xingcheng Wang","doi":"10.1016/j.compgeo.2026.107920","DOIUrl":"10.1016/j.compgeo.2026.107920","url":null,"abstract":"<div><div>To efficiently address complex reliability problems involving high-dimensional random inputs and rare failure events, this paper proposes a stochastic collocation enhanced radial-circumferential importance sampling method. The proposed method designs an explicit and regular radial importance sampling density to efficiently explore the rare failure domain along the radial direction. By this importance sampling density, stochastic collocation points and corresponding weights of the radial variable are generated. Then the rare failure probability can be estimated by the weighted summation of the conditional failure probabilities at the given radial stochastic collocation points. To further improve the efficiency of estimating the conditional failure probabilities at different radial stochastic collocation points, this paper introduces a circumferential importance sampling strategy. The proposed radial-circumferential importance sampling in the proposed method does not require searching the most probable failure point to determine the importance sampling density, and it gradually explores the rare failure domain from the global to the local and then constructs the importance sampling density, which can cover the domains that contribute significantly to the failure probability with low computational cost, thereby reducing the variance of the failure probability estimation. The feasibility and superiority of the proposed method are verified by numerical and slope structure examples.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"192 ","pages":"Article 107920"},"PeriodicalIF":6.2,"publicationDate":"2026-01-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146022474","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-element material-point approach for anchored tunnels in large deformation 大变形锚固隧道的材料-点混合有限元方法

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

Computers and Geotechnics

Pub Date : 2026-01-17 DOI: 10.1016/j.compgeo.2026.107921

Jianbo Fei , Jiayan Wei , Chenglong Wu , Xianshun Zhou , Yuxin Jie , Xiangsheng Chen

The stability of deep-buried tunnels undergoing large deformation critically depends on the interaction between rock masses and bolt reinforcement systems. To address the limitations of traditional numerical methods in simulating such scenarios—particularly mesh distortion and poor coupling accuracy—this study proposes a Hybrid Finite-Element Material-Point (HFEMP) method. In this framework, the surrounding rock is modeled using the Material Point Method (MPM), while the rock bolts are represented as axial-force-bearing bar elements within the Finite Element Method (FEM), allowing for efficient and robust simulation of bolt–rock interaction under large strain. The HFEMP method is validated against scaled physical model tests of anchored tunnels subjected to increasing vertical loads. The simulation results highlight the critical role of bolt density and geometry in mitigating tunnel deformation. A subsequent parametric analysis systematically examines the influence of circumferential spacing, longitudinal spacing, length, and diameter of the bolts on the stability of the surrounding rock. Results indicate that decreasing both the circumferential and longitudinal spacing of the bolts, and increasing their diameter and length, markedly suppresses the plastic deformation of the surrounding rock. However, increases in bolt length beyond a threshold yield diminishing returns. Based on these findings, the study develops optimized bolt design charts that balance reinforcement performance with material efficiency, offering both theoretical insight and practical guidance for tunnel reinforcement design. The proposed HFEMP framework offers a powerful and scalable tool for the analysis and design of tunnel reinforcement strategies in complex geomechanical environments, with potential applications in large deforming tunnels.

大变形深埋隧道的稳定性关键取决于岩体与锚杆加固系统之间的相互作用。为了解决传统数值方法在模拟此类场景时的局限性，特别是网格畸变和耦合精度差，本研究提出了一种混合有限元材料点（HFEMP）方法。在该框架中，围岩采用材料点法（MPM）建模，而锚杆采用有限元法（FEM）表示为轴向受力杆单元，从而可以高效、稳健地模拟大应变下的锚杆-岩石相互作用。通过锚固隧道竖向荷载增大的比例物理模型试验，验证了该方法的有效性。模拟结果突出了锚杆密度和几何形状对减轻巷道变形的关键作用。随后的参数分析系统地考察了螺栓的周向间距、纵向间距、长度和直径对围岩稳定性的影响。结果表明，减小锚杆的周向间距和纵向间距，增大锚杆的直径和长度，均能显著抑制围岩的塑性变形。然而，增加螺栓长度超过一个阈值产生递减收益。基于这些发现，本研究开发了优化的锚杆设计图表，以平衡加固性能和材料效率，为隧道加固设计提供理论见解和实践指导。提出的HFEMP框架为复杂地质力学环境下隧道加固策略的分析和设计提供了一个强大且可扩展的工具，在大变形隧道中具有潜在的应用前景。

{"title":"A hybrid finite-element material-point approach for anchored tunnels in large deformation","authors":"Jianbo Fei , Jiayan Wei , Chenglong Wu , Xianshun Zhou , Yuxin Jie , Xiangsheng Chen","doi":"10.1016/j.compgeo.2026.107921","DOIUrl":"10.1016/j.compgeo.2026.107921","url":null,"abstract":"<div><div>The stability of deep-buried tunnels undergoing large deformation critically depends on the interaction between rock masses and bolt reinforcement systems. To address the limitations of traditional numerical methods in simulating such scenarios—particularly mesh distortion and poor coupling accuracy—this study proposes a Hybrid Finite-Element Material-Point (HFEMP) method. In this framework, the surrounding rock is modeled using the Material Point Method (MPM), while the rock bolts are represented as axial-force-bearing bar elements within the Finite Element Method (FEM), allowing for efficient and robust simulation of bolt–rock interaction under large strain. The HFEMP method is validated against scaled physical model tests of anchored tunnels subjected to increasing vertical loads. The simulation results highlight the critical role of bolt density and geometry in mitigating tunnel deformation. A subsequent parametric analysis systematically examines the influence of circumferential spacing, longitudinal spacing, length, and diameter of the bolts on the stability of the surrounding rock. Results indicate that decreasing both the circumferential and longitudinal spacing of the bolts, and increasing their diameter and length, markedly suppresses the plastic deformation of the surrounding rock. However, increases in bolt length beyond a threshold yield diminishing returns. Based on these findings, the study develops optimized bolt design charts that balance reinforcement performance with material efficiency, offering both theoretical insight and practical guidance for tunnel reinforcement design. The proposed HFEMP framework offers a powerful and scalable tool for the analysis and design of tunnel reinforcement strategies in complex geomechanical environments, with potential applications in large deforming tunnels.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"192 ","pages":"Article 107921"},"PeriodicalIF":6.2,"publicationDate":"2026-01-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146022584","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 site demarcation using tailored lego clustering – size of site 数据驱动的站点划分使用量身定制的乐高聚类-站点的大小

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

Computers and Geotechnics

Pub Date : 2026-01-17 DOI: 10.1016/j.compgeo.2026.107924

Yongmin Cai , Kok-Kwang Phoon

Conventional methods for constructing quasi-local transformation models can introduce significant uncertainty in soil property inference, particularly for large-scale geotechnical projects. To address this, a novel data-driven site demarcation (DDSD) framework is proposed. The process begins by defining a target site—a local area within the large-scale project containing the soil property of interest. The large-scale project (excluding the target site) is then demarcated into uniformly sized elemental sites, analogous to “Lego bricks”. These elemental sites are assembled into database sites to form a Big Indirect Database (BID). A tailored Lego clustering algorithm is then developed to retrieve database sites from the BID that are highly similar to the target site. It is termed tailored Lego clustering because the cluster is tailored to the target site and the assembled database sites can vary in size, shape, and/or depth segments depending on how the “Lego bricks” are combined. This study focuses on a specific implementation where the assembled sites are restricted to squares or rectangles, and we optimize their sizes. A quasi-local transformation model is then constructed using these similar database sites that are of different sizes, enabling more accurate inference of soil properties at the target site. The core of this paper is the optimization of the sizes of these square or rectangular database sites (i.e., the assembly configuration of elemental sites) to minimize inference uncertainty, guided by a criterion based on the root mean squared error. The framework is validated through synthetic examples and a real-world case study, demonstrating that our data-driven size optimization for regular-shaped sites significantly reduces inference uncertainty.

传统的构建准局部变换模型的方法会给土壤性质推断带来很大的不确定性，特别是在大型岩土工程中。为了解决这个问题，提出了一种新的数据驱动的站点划分（DDSD）框架。这个过程首先定义一个目标地点——在大型项目中包含感兴趣的土壤属性的局部区域。大型项目（不包括目标场地）然后划分为统一大小的元素场地，类似于“乐高积木”。这些基本站点被组装成数据库站点，形成一个大间接数据库（BID）。然后开发定制的Lego聚类算法，从BID中检索与目标站点高度相似的数据库站点。它被称为定制乐高集群，因为集群是针对目标站点进行定制的，而组装的数据库站点可以根据“乐高积木”的组合方式在大小、形状和/或深度分段上有所不同。本研究的重点是一个特定的实现，其中组装站点被限制为正方形或矩形，我们优化它们的大小。然后使用这些相似的不同大小的数据库站点构建准局部转换模型，从而能够更准确地推断目标站点的土壤特性。本文的核心是在基于均方根误差的准则指导下，优化这些正方形或矩形数据库站点的大小（即元素站点的装配配置）以最小化推理不确定性。该框架通过综合示例和现实世界的案例研究进行了验证，表明我们的数据驱动的规则形状站点的大小优化显着降低了推理的不确定性。

{"title":"Data-driven site demarcation using tailored lego clustering – size of site","authors":"Yongmin Cai , Kok-Kwang Phoon","doi":"10.1016/j.compgeo.2026.107924","DOIUrl":"10.1016/j.compgeo.2026.107924","url":null,"abstract":"<div><div>Conventional methods for constructing quasi-local transformation models can introduce significant uncertainty in soil property inference, particularly for large-scale geotechnical projects. To address this, a novel data-driven site demarcation (DDSD) framework is proposed. The process begins by defining a target site—a local area within the large-scale project containing the soil property of interest. The large-scale project (excluding the target site) is then demarcated into uniformly sized elemental sites, analogous to “Lego bricks”. These elemental sites are assembled into database sites to form a Big Indirect Database (BID). A tailored Lego clustering algorithm is then developed to retrieve database sites from the BID that are highly similar to the target site. It is termed tailored Lego clustering because the cluster is tailored to the target site and the assembled database sites can vary in size, shape, and/or depth segments depending on how the “Lego bricks” are combined. This study focuses on a specific implementation where the assembled sites are restricted to squares or rectangles, and we optimize their sizes. A quasi-local transformation model is then constructed using these similar database sites that are of different sizes, enabling more accurate inference of soil properties at the target site. The core of this paper is the optimization of the sizes of these square or rectangular database sites (i.e., the assembly configuration of elemental sites) to minimize inference uncertainty, guided by a criterion based on the root mean squared error. The framework is validated through synthetic examples and a real-world case study, demonstrating that our data-driven size optimization for regular-shaped sites significantly reduces inference uncertainty.</div></div>","PeriodicalId":55217,"journal":{"name":"Computers and Geotechnics","volume":"192 ","pages":"Article 107924"},"PeriodicalIF":6.2,"publicationDate":"2026-01-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146022771","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