Engineering Geology最新文献_第4页

Field and numerical investigations of canal damage characteristics and mechanisms under coupled drying-wetting and freezing-thawing cycles 干湿-冻融耦合循环条件下渠系损伤特征及机理的现场与数值研究

IF 8.4 1区工程技术 Q1 ENGINEERING, GEOLOGICAL

Engineering Geology

Pub Date : 2026-01-26 DOI: 10.1016/j.enggeo.2026.108587

Haoyuan Jiang , Mingyi Zhang , Zhengzhong Wang , Wansheng Pei , Jianrui Ge , Zhengyi Wang

Canals are essential for water transfer in cold and arid regions. However, they are vulnerable to damage under drying-wetting and freezing-thawing cycles, with wetting during freezing–thawing periods significantly intensifying frost damage. Yet, the underlying mechanisms involving heat-water-deformation interactions remain insufficiently quantified. This study employs field monitoring to investigate temperature, water content, and deformation behaviors in canals. Then, a coupled water-heat numerical model incorporating seepage effects was developed and applied to analyze how leakage increases soil moisture and exacerbates frost damage. Results show that canal temperatures are controlled by water temperature and leakage during water supply, and by air temperature and solar radiation during water cut-off, with freeze–thaw transitions further modulating thermal behavior. Shady slopes freeze more readily, whereas sunny slopes experience greater evaporative loss. Spring irrigation sharply lowers the 0 °C isotherm below water level, triggering peak frost heave at its onset, followed by rapid thawing. Frost heave is lowest on sunny slopes and greatest at mid-shady slopes, with maximum differences exceeding sixfold. After thawing, minor slippage occurs on shady slopes, while sunny slopes remain stable. Canal frost damage is driven by elevated soil water from leakage and winter irrigation, as well as sharp thermal imbalances induced by spring irrigation and solar radiation disparities between slopes. Preserving the integrity of anti-seepage systems proves more effective in mitigating frost damage than early termination of winter irrigation. This study provides the first integrated field-based and modeling evidence on frost damage mechanisms in cold-arid canals, establishing a scientific foundation for mitigation strategies.

在寒冷和干旱地区，运河是输水的必要条件。而冻融循环和干湿循环对冻害的影响较大，冻融期间的润湿作用显著加剧了冻害。然而，涉及热-水-变形相互作用的潜在机制仍然没有充分量化。本研究采用现场监测的方法来调查水渠的温度、含水量和变形行为。在此基础上，建立了考虑渗流效应的水-热耦合数值模型，并应用该模型分析了渗漏对土壤水分增加和霜害加剧的影响。结果表明，供水期间的水温和渗漏以及断水期间的气温和太阳辐射控制着运河的温度，冻融转变进一步调节了运河的热行为。阴坡更容易冻结，而阳坡蒸发损失更大。春灌急剧降低0 °C以下的等温线，在开始时触发霜胀峰值，随后迅速融化。霜胀在阳坡上最小，在中阴坡上最大，最大差异超过6倍。解冻后，背阴坡发生轻微滑移，而阳坡保持稳定。渗漏和冬季灌溉导致的土壤水分升高，以及春灌和坡间太阳辐射差异导致的急剧热不平衡，都是运河冻害的驱动因素。保持防渗系统的完整性比提前终止冬季灌溉更有效地减轻霜冻损害。该研究首次提供了基于现场和模型的寒旱区水渠冻害机制的综合证据，为减灾策略奠定了科学基础。

{"title":"Field and numerical investigations of canal damage characteristics and mechanisms under coupled drying-wetting and freezing-thawing cycles","authors":"Haoyuan Jiang , Mingyi Zhang , Zhengzhong Wang , Wansheng Pei , Jianrui Ge , Zhengyi Wang","doi":"10.1016/j.enggeo.2026.108587","DOIUrl":"10.1016/j.enggeo.2026.108587","url":null,"abstract":"<div><div>Canals are essential for water transfer in cold and arid regions. However, they are vulnerable to damage under drying-wetting and freezing-thawing cycles, with wetting during freezing–thawing periods significantly intensifying frost damage. Yet, the underlying mechanisms involving heat-water-deformation interactions remain insufficiently quantified. This study employs field monitoring to investigate temperature, water content, and deformation behaviors in canals. Then, a coupled water-heat numerical model incorporating seepage effects was developed and applied to analyze how leakage increases soil moisture and exacerbates frost damage. Results show that canal temperatures are controlled by water temperature and leakage during water supply, and by air temperature and solar radiation during water cut-off, with freeze–thaw transitions further modulating thermal behavior. Shady slopes freeze more readily, whereas sunny slopes experience greater evaporative loss. Spring irrigation sharply lowers the 0 °C isotherm below water level, triggering peak frost heave at its onset, followed by rapid thawing. Frost heave is lowest on sunny slopes and greatest at mid-shady slopes, with maximum differences exceeding sixfold. After thawing, minor slippage occurs on shady slopes, while sunny slopes remain stable. Canal frost damage is driven by elevated soil water from leakage and winter irrigation, as well as sharp thermal imbalances induced by spring irrigation and solar radiation disparities between slopes. Preserving the integrity of anti-seepage systems proves more effective in mitigating frost damage than early termination of winter irrigation. This study provides the first integrated field-based and modeling evidence on frost damage mechanisms in cold-arid canals, establishing a scientific foundation for mitigation strategies.</div></div>","PeriodicalId":11567,"journal":{"name":"Engineering Geology","volume":"363 ","pages":"Article 108587"},"PeriodicalIF":8.4,"publicationDate":"2026-01-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146048048","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 innovative transform mapping and visualization of fracture persistence from borehole-group image analysis: MFPbia 基于井眼群图像分析的裂缝持续性创新变换映射与可视化：MFPbia

IF 7.4 1区工程技术 Q1 ENGINEERING, GEOLOGICAL

Engineering Geology

Pub Date : 2026-01-24 DOI: 10.1016/j.enggeo.2026.108583

Zhenhao Xu, Yihui Li, Dongdong Pan, Shengzhe Zhao

引用次数: 0

A simplified approach to analyzing the safety of earthquake-triggered landslides 地震引发滑坡安全性分析的简化方法

IF 8.4 1区工程技术 Q1 ENGINEERING, GEOLOGICAL

Engineering Geology

Pub Date : 2026-01-24 DOI: 10.1016/j.enggeo.2026.108575

Fumin Wang , Alberto Ledesma , Francisco López-Almansa

The sliding of slopes caused by earthquakes is a critical issue, resulting in significant losses and casualties worldwide. This paper presents a simplified analytical formulation for evaluating the safety factor; the approach is essentially coupled, as it incorporates the flexibility of the sliding block. Several issues are considered: soil slope, ground motion direction (horizontal and parallel to the slope), and presence of underground water; however, the transient response and damping are neglected. A Factor of Safety is defined, and closed-form expressions are derived; as well, expressions for the yield accelerations are presented and compared with the Newmark one. Numerical results obtained with the proposed approach provide conclusions that explain some differences between coupled analysis and the traditional uncoupled Newmark method. Finally, validation with experiments, numerical analyses, and actual sliding of slopes demonstrates that the proposed methodology is reliable.

地震引起的斜坡滑动是一个严重的问题，在世界范围内造成巨大的损失和人员伤亡。本文提出了安全系数的简化分析公式；这种方法本质上是耦合的，因为它结合了滑块的灵活性。考虑了几个问题：土壤坡度，地面运动方向（水平和平行于斜坡），以及地下水的存在；然而，暂态响应和阻尼被忽略。定义了安全系数，导出了安全系数的封闭表达式；给出了屈服加速度的表达式，并与Newmark模型进行了比较。用该方法得到的数值结果解释了耦合分析与传统非耦合Newmark方法之间的一些差异。最后，通过实验、数值分析和边坡实际滑动验证了该方法的可靠性。

引用次数: 0

Landslide displacement prediction based on deep displacement state recognition and similarity propagation 基于深度位移状态识别和相似传播的滑坡位移预测

IF 8.4 1区工程技术 Q1 ENGINEERING, GEOLOGICAL

Engineering Geology

Pub Date : 2026-01-23 DOI: 10.1016/j.enggeo.2026.108570

Hanming Zhang, Xiangchen Guo, Kang Liu, Qingsong Liu, Jianhao Xiao, Yu Ke, Yong Liu

Current landslide forecasting is constrained by an over-reliance on surface kinematics, often failing to capture the hidden, non-linear state transitions occurring at depth. To address this, we propose a Trend–State Coupled Framework that shifts the framework from purely externally driven to state-dependent prediction. A trend submodel maps hydrometeorological triggers to displacement baselines, while a state submodel explicitly characterizes internal sensitivity using Non-parametric Bispectral Analysis. Unlike conventional time-frequency methods, this indicates nonlinear phase couplings associated with incipient instability, which are then classified into evolutionary states by a Convolutional Neural Network (CNN). To enhance spatial generalizability, a Slope Cell Matrix is constructed to integrate excitation–state–response information, enabling the cross-site propagation of state variables via a particle-swarm-optimized similarity metric. Validation at the Baishuihe landslide (Three Gorges Reservoir) demonstrates that the coupled model achieves an RMSE of 7.22 mm, significantly outperforming traditional BP (68.06 mm) and SVR (63.81 mm) models. Crucially, the deep-displacement state indicators exhibit a distinct precursory surge ahead of surface acceleration episodes, confirming the framework's capability to capture early failure mechanisms. This study provides a reliable, physically interpretable approach for landslide early warning by linking external drivers with deep-seated evolutionary dynamics.

目前的滑坡预测受到过度依赖地表运动学的限制，往往无法捕捉到发生在深度的隐藏的非线性状态转换。为了解决这个问题，我们提出了一个趋势-状态耦合框架，将框架从纯粹的外部驱动转变为依赖于状态的预测。趋势子模型将水文气象触发器映射到位移基线，而状态子模型则使用非参数双谱分析明确表征内部敏感性。与传统的时频方法不同，这表明与早期不稳定性相关的非线性相位耦合，然后通过卷积神经网络（CNN）将其分类为进化状态。为了增强空间泛化能力，构建了一个斜率细胞矩阵来整合激励-状态-响应信息，通过粒子群优化的相似性度量实现状态变量的跨站点传播。在白水河滑坡（三峡库区）的验证表明，该耦合模型的RMSE为7.22 mm，显著优于传统的BP模型（68.06 mm）和SVR模型（63.81 mm）。至关重要的是，深层位移状态指标在地表加速度发作之前表现出明显的前兆波动，这证实了该框架能够捕捉早期失效机制。该研究通过将外部驱动因素与深层演化动力学联系起来，为滑坡预警提供了一种可靠的物理解释方法。

{"title":"Landslide displacement prediction based on deep displacement state recognition and similarity propagation","authors":"Hanming Zhang, Xiangchen Guo, Kang Liu, Qingsong Liu, Jianhao Xiao, Yu Ke, Yong Liu","doi":"10.1016/j.enggeo.2026.108570","DOIUrl":"10.1016/j.enggeo.2026.108570","url":null,"abstract":"<div><div>Current landslide forecasting is constrained by an over-reliance on surface kinematics, often failing to capture the hidden, non-linear state transitions occurring at depth. To address this, we propose a Trend–State Coupled Framework that shifts the framework from purely externally driven to state-dependent prediction. A trend submodel maps hydrometeorological triggers to displacement baselines, while a state submodel explicitly characterizes internal sensitivity using Non-parametric Bispectral Analysis. Unlike conventional time-frequency methods, this indicates nonlinear phase couplings associated with incipient instability, which are then classified into evolutionary states by a Convolutional Neural Network (CNN). To enhance spatial generalizability, a Slope Cell Matrix is constructed to integrate excitation–state–response information, enabling the cross-site propagation of state variables via a particle-swarm-optimized similarity metric. Validation at the Baishuihe landslide (Three Gorges Reservoir) demonstrates that the coupled model achieves an RMSE of 7.22 mm, significantly outperforming traditional BP (68.06 mm) and SVR (63.81 mm) models. Crucially, the deep-displacement state indicators exhibit a distinct precursory surge ahead of surface acceleration episodes, confirming the framework's capability to capture early failure mechanisms. This study provides a reliable, physically interpretable approach for landslide early warning by linking external drivers with deep-seated evolutionary dynamics.</div></div>","PeriodicalId":11567,"journal":{"name":"Engineering Geology","volume":"363 ","pages":"Article 108570"},"PeriodicalIF":8.4,"publicationDate":"2026-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146033600","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

Lithology dependent pathways of forming rock mass weakness plane shape regional landslide susceptibility 岩体软弱面形成的岩性依赖路径决定了区域滑坡易感性

IF 8.4 1区工程技术 Q1 ENGINEERING, GEOLOGICAL

Engineering Geology

Pub Date : 2026-01-23 DOI: 10.1016/j.enggeo.2026.108586

Yifan Zhang , Zhenni Jiang , Chunhao Wu , Peng Cui , Weilin Kong , Yusheng Li , Shujian Yi

Rock mass weakness planes are mechanically unfavored discontinuities that control slope stability. However, their spatial pattern is poorly understood, limiting advances in regional landslide assessment. In this study, we propose a multi-explainable machine learning framework to predict the distribution of weakness planes and quantify their contribution to landslides. Focusing on southeastern Tibet, the study investigates 194 field outcrops of rock mass weakness planes, integrating geological section comparisons and driving factor analyses to reveal spatial heterogeneity and the evolvement of weakness plane. Rock mass weakness planes are denser in lithologically weak and structurally damaged rock masses, and less developed in gentler, sparsely faulted terrains. Comparative analysis of geological sections indicates that pre-existing discontinuities are essential preconditions for developing weakness planes, whereas precipitation acts as an activator. The controlling factors vary with lithology, indicating that material and structure govern how efficiently exogenic processes transform discontinuities into weakness planes. By considering weakness plane during susceptibility assessment, performances were improved (Recall increased by 4.6–6.8%, AUC increased by 3.1–5.6%). This work constructs the regional-scale model for continuous prediction of rock mass weakness planes. It links lithology-dependent formation mechanisms of rock mass weakness plane to slope instability, providing a process-based framework for interpreting landslide development and improving landslide susceptibility assessment.

岩体软弱面是控制边坡稳定性的机械上不利的不连续面。然而，人们对其空间格局了解甚少，限制了区域滑坡评价的进展。在这项研究中，我们提出了一个多可解释的机器学习框架来预测软弱面的分布并量化它们对滑坡的贡献。以藏东南地区为研究对象，对194个野外露头岩体软弱面进行了调查，结合地质剖面对比和驱动因素分析，揭示了岩体软弱面的空间异质性和演化规律。在岩性薄弱和构造破坏的岩体中，岩体软弱面密度较大，而在较平缓、断裂较少的地形中，岩体软弱面发育较少。地质剖面对比分析表明，已存在的不连续面是弱面发育的必要前提条件，而降水则是弱面发育的激活剂。控制因素随岩性的不同而不同，表明物质和结构控制着外生过程如何有效地将不连续面转化为弱面。在药敏评价中考虑弱点面，提高了评价效果（召回率提高4.6 ~ 6.8%，AUC提高3.1 ~ 5.6%）。本文建立了岩体软弱面连续预测的区域尺度模型。它将岩体软弱面岩性依赖的形成机制与边坡失稳联系起来，为解释滑坡发展和改进滑坡易感性评价提供了一个基于过程的框架。

{"title":"Lithology dependent pathways of forming rock mass weakness plane shape regional landslide susceptibility","authors":"Yifan Zhang , Zhenni Jiang , Chunhao Wu , Peng Cui , Weilin Kong , Yusheng Li , Shujian Yi","doi":"10.1016/j.enggeo.2026.108586","DOIUrl":"10.1016/j.enggeo.2026.108586","url":null,"abstract":"<div><div>Rock mass weakness planes are mechanically unfavored discontinuities that control slope stability. However, their spatial pattern is poorly understood, limiting advances in regional landslide assessment. In this study, we propose a multi-explainable machine learning framework to predict the distribution of weakness planes and quantify their contribution to landslides. Focusing on southeastern Tibet, the study investigates 194 field outcrops of rock mass weakness planes, integrating geological section comparisons and driving factor analyses to reveal spatial heterogeneity and the evolvement of weakness plane. Rock mass weakness planes are denser in lithologically weak and structurally damaged rock masses, and less developed in gentler, sparsely faulted terrains. Comparative analysis of geological sections indicates that pre-existing discontinuities are essential preconditions for developing weakness planes, whereas precipitation acts as an activator. The controlling factors vary with lithology, indicating that material and structure govern how efficiently exogenic processes transform discontinuities into weakness planes. By considering weakness plane during susceptibility assessment, performances were improved (Recall increased by 4.6–6.8%, AUC increased by 3.1–5.6%). This work constructs the regional-scale model for continuous prediction of rock mass weakness planes. It links lithology-dependent formation mechanisms of rock mass weakness plane to slope instability, providing a process-based framework for interpreting landslide development and improving landslide susceptibility assessment.</div></div>","PeriodicalId":11567,"journal":{"name":"Engineering Geology","volume":"363 ","pages":"Article 108586"},"PeriodicalIF":8.4,"publicationDate":"2026-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146033603","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 groundwater level variability and annual slope failure probability using multi-year groundwater level observations 利用多年地下水位观测量化地下水位变化和年度边坡破坏概率

IF 8.4 1区工程技术 Q1 ENGINEERING, GEOLOGICAL

Engineering Geology

Pub Date : 2026-01-23 DOI: 10.1016/j.enggeo.2026.108584

Dianqing Li , Junjie Wang , Yu Wang , Xin Liu

The variation in groundwater level (GL) has been recognized as an important triggering factor of landslides and often exhibits high uncertainty. Most existing slope reliability analysis and landslide risk assessment studies ignored the uncertainty of GL and focused on estimating slope failure probability that is not related to a time period. This study proposes a rigorous method for quantifying the annual failure probability (P_FA) of slopes considering both uncertainties in GL and soil properties. Multi-year groundwater monitoring data are utilized to quantify the annual exceedance probability of GL through a statistical analysis. A series of GL scenarios corresponding to different return periods or exceedance probabilities is generated and used to estimate the corresponding conditional slope failure probabilities. These conditional probabilities are combined using the Total Probability Theorem to estimate P_FA. The results from a real slope indicate that the conditional slope failure probability increases exponentially as GL rises. The variability in GL dominates P_FA when the variability of soil properties is relatively low (e.g., coefficient of variation, COV = 0.1). Conversely, when the variability of soil properties is relatively high (e.g., COV ≥ 0.3), P_FA is dominated by soil uncertainties, and the conditional failure probability becomes insensitive to GL fluctuations. This highlights the importance of monitoring groundwater conditions in quantifying and mitigating landslide risks. Preliminary validation indicates that GL scenarios corresponding to a return period of 10 to 15 years can be used as representative GL conditions for evaluating P_FA, offering a practical guidance for slope design engineering.

地下水位变化是滑坡发生的重要触发因素，具有较高的不确定性。现有的大多数边坡可靠度分析和滑坡风险评估研究都忽略了GL的不确定性，而侧重于估算与时间段无关的边坡破坏概率。本文提出了一种考虑GL和土壤性质不确定性的边坡年破坏概率（PFA）量化方法。利用多年地下水监测资料，通过统计分析，量化了GL年超标概率。生成对应不同回归期或超过概率的一系列GL情景，用于估算相应的条件边坡破坏概率。使用全概率定理将这些条件概率组合起来估计PFA。对一个实际边坡的计算结果表明，随着GL的增大，边坡的条件破坏概率呈指数增长。当土壤性质的变异性较低时（如变异系数，COV = 0.1）， GL的变异性占主导地位。相反，当土壤性质变异性较大时（如COV≥0.3），PFA以土壤不确定性为主，条件失效概率对GL波动不敏感。这突出了监测地下水状况在量化和减轻滑坡风险方面的重要性。初步验证表明，回归期为10 ~ 15年的GL情景可作为评价PFA的代表性GL条件，对边坡设计工程具有实际指导意义。

{"title":"Quantifying groundwater level variability and annual slope failure probability using multi-year groundwater level observations","authors":"Dianqing Li , Junjie Wang , Yu Wang , Xin Liu","doi":"10.1016/j.enggeo.2026.108584","DOIUrl":"10.1016/j.enggeo.2026.108584","url":null,"abstract":"<div><div>The variation in groundwater level (GL) has been recognized as an important triggering factor of landslides and often exhibits high uncertainty. Most existing slope reliability analysis and landslide risk assessment studies ignored the uncertainty of GL and focused on estimating slope failure probability that is not related to a time period. This study proposes a rigorous method for quantifying the annual failure probability (<em>P</em><sub><em>FA</em></sub>) of slopes considering both uncertainties in GL and soil properties. Multi-year groundwater monitoring data are utilized to quantify the annual exceedance probability of GL through a statistical analysis. A series of GL scenarios corresponding to different return periods or exceedance probabilities is generated and used to estimate the corresponding conditional slope failure probabilities. These conditional probabilities are combined using the Total Probability Theorem to estimate <em>P</em><sub><em>FA</em></sub>. The results from a real slope indicate that the conditional slope failure probability increases exponentially as GL rises. The variability in GL dominates <em>P</em><sub><em>FA</em></sub> when the variability of soil properties is relatively low (e.g., coefficient of variation, COV = 0.1). Conversely, when the variability of soil properties is relatively high (e.g., COV ≥ 0.3), <em>P</em><sub><em>FA</em></sub> is dominated by soil uncertainties, and the conditional failure probability becomes insensitive to GL fluctuations. This highlights the importance of monitoring groundwater conditions in quantifying and mitigating landslide risks. Preliminary validation indicates that GL scenarios corresponding to a return period of 10 to 15 years can be used as representative GL conditions for evaluating <em>P</em><sub><em>FA</em></sub>, offering a practical guidance for slope design engineering.</div></div>","PeriodicalId":11567,"journal":{"name":"Engineering Geology","volume":"363 ","pages":"Article 108584"},"PeriodicalIF":8.4,"publicationDate":"2026-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146033604","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

Force distribution due to impact of bidisperse granular avalanches against a retention barrier 由于双分散颗粒雪崩对保留屏障的冲击而产生的力分布

IF 8.4 1区工程技术 Q1 ENGINEERING, GEOLOGICAL

Engineering Geology

Pub Date : 2026-01-23 DOI: 10.1016/j.enggeo.2026.108585

Francisco Martínez , Tomás Trewhela , Simón Guerra , Bastián Araya

Landslides, rock, debris, and snow avalanches are frequent disasters characterized by their destructive power. Retention structures are therefore essential for mitigating their effects, particularly near urban areas. Although widely studied, the role of grain-size polydispersity remains poorly addressed, even for the simplified case of bidisperse mixtures. In this work, we experimentally investigate the force histories generated by dense, dry, bidisperse granular flows impacting a rigid barrier at different relative concentrations

C

of the finer fraction. Experiments were conducted in an inclined flume, where a granular mass collided with a downstream rigid wall. A central strip of load cells measured the local force simultaneously under varying conditions. Time series of the force distribution, peak values, and the net force on the barrier were analyzed, showing a strong dependence on particle concentration, with maximum values often occurring near

C = 20 %

. This behavior is interpreted as the outcome of size segregation and its inherent asymmetry, which can be incorporated into dynamic force predictions using a Bagnold velocity scaling. Deposit profiles were also characterized, showing significant morphological variability with respect to

C

. Finally, we propose empirical relations linking maximum loads to deposit properties, providing scaling laws that may be applied to the design and assessment of real retention structures.

山崩、岩石、碎石和雪崩是经常发生的灾害，其特点是破坏力大。因此，保留结构对于减轻其影响至关重要，特别是在城市地区附近。尽管研究广泛，但即使对于双分散混合物的简化情况，粒径多分散性的作用仍然很少得到解决。在这项工作中，我们通过实验研究了致密、干燥、双分散颗粒流在不同相对浓度C下撞击刚性屏障所产生的力历史。实验是在一个倾斜水槽中进行的，其中颗粒体与下游刚性壁碰撞。测压元件的中心条带在不同条件下同时测量局部力。通过对力分布、峰值和作用在屏障上的净力的时间序列分析，发现其与颗粒浓度有很强的依赖性，最大值往往出现在C=20%附近。这种行为被解释为尺寸偏析及其固有的不对称性的结果，可以使用Bagnold速度标度将其纳入动态力预测中。最后，我们提出了将最大载荷与沉积物性质联系起来的经验关系，提供了可应用于实际保留结构的设计和评估的标度定律。

{"title":"Force distribution due to impact of bidisperse granular avalanches against a retention barrier","authors":"Francisco Martínez , Tomás Trewhela , Simón Guerra , Bastián Araya","doi":"10.1016/j.enggeo.2026.108585","DOIUrl":"10.1016/j.enggeo.2026.108585","url":null,"abstract":"<div><div>Landslides, rock, debris, and snow avalanches are frequent disasters characterized by their destructive power. Retention structures are therefore essential for mitigating their effects, particularly near urban areas. Although widely studied, the role of grain-size polydispersity remains poorly addressed, even for the simplified case of bidisperse mixtures. In this work, we experimentally investigate the force histories generated by dense, dry, bidisperse granular flows impacting a rigid barrier at different relative concentrations <span><math><mi>C</mi></math></span> of the finer fraction. Experiments were conducted in an inclined flume, where a granular mass collided with a downstream rigid wall. A central strip of load cells measured the local force simultaneously under varying conditions. Time series of the force distribution, peak values, and the net force on the barrier were analyzed, showing a strong dependence on particle concentration, with maximum values often occurring near <span><math><mrow><mi>C</mi><mo>=</mo><mn>20</mn><mtext>%</mtext></mrow></math></span>. This behavior is interpreted as the outcome of size segregation and its inherent asymmetry, which can be incorporated into dynamic force predictions using a Bagnold velocity scaling. Deposit profiles were also characterized, showing significant morphological variability with respect to <span><math><mi>C</mi></math></span>. Finally, we propose empirical relations linking maximum loads to deposit properties, providing scaling laws that may be applied to the design and assessment of real retention structures.</div></div>","PeriodicalId":11567,"journal":{"name":"Engineering Geology","volume":"363 ","pages":"Article 108585"},"PeriodicalIF":8.4,"publicationDate":"2026-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146033605","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

Machine learning approaches to predicting the uniaxial compressive strength of granite from image-derived mineralogical features 从图像矿物学特征预测花岗岩单轴抗压强度的机器学习方法

IF 8.4 1区工程技术 Q1 ENGINEERING, GEOLOGICAL

Engineering Geology

Pub Date : 2026-01-19 DOI: 10.1016/j.enggeo.2026.108571

Changdi He , Brijes Mishra , Yuan Li , Jessica Michelle Wempen

This study uses machine learning (ML) models to predict the numerically simulated uniaxial compressive strength (UCS) of granite directly from digital images. Specifically, the images were processed using an in-house Digital Image Processing (DIP) tool to estimate mineralogical features (used as input features for the ML models), including mineral content, grain size, and spatial distribution. Mineral content and distribution were quantified using

m

-harmonic Fourier series equations, whereas mineral grain size was determined using the 4-connectivity method. The target UCS values were derived from the 2D physically informed Subspring Network Breakable Voronoi (SNBV) microstructural models, replicating the mineralogical features observed in the granite images. Extreme Gradient Boosting (XGBoost) models with different input combinations and hyperparameter optimization methods were trained and evaluated on 126 granite images using a single train/test split and repeated 5-fold cross-validation. Results indicate that the input combination of mineral content, grain size, and spatial distribution parameters from

m

-harmonic Fourier series combined with SHapley Additive Explanations (SHAP)-based feature selection, yield the best and robust performance, whereas increasing the harmonic order has a limited effect on accuracy. Among the tested optimization methods, the Optuna–XGBoost model achieved the best performance. In addition, UCS prediction is controlled mainly by the content and grain size of biotite and plagioclase, while the corresponding attributes of quartz/K-feldspar, as well as overall mineral distribution play a comparatively minor role.

本研究使用机器学习（ML）模型直接从数字图像中预测数值模拟的花岗岩单轴抗压强度（UCS）。具体来说，使用内部数字图像处理（DIP）工具对图像进行处理，以估计矿物学特征（用作ML模型的输入特征），包括矿物含量、粒度和空间分布。矿物含量和分布采用m-调和傅立叶级数方程量化，矿物粒度采用4连通性方法确定。目标UCS值来自2D物理信息Subspring Network Breakable Voronoi （SNBV）微观结构模型，复制了花岗岩图像中观察到的矿物学特征。采用单训练/测试分割和重复5倍交叉验证的方法，对126幅花岗岩图像进行了训练和评估，并对不同输入组合和超参数优化方法的极端梯度增强（XGBoost）模型进行了训练和评估。结果表明，矿物含量、粒度和空间分布参数的m-谐波傅立叶级数输入组合与SHapley加性解释（SHAP）特征选择相结合，可以获得最佳的鲁棒性性能，而增加谐波阶数对准确性的影响有限。在测试的优化方法中，Optuna-XGBoost模型的性能最好。此外，UCS预测主要受黑云母和斜长石的含量和粒度控制，石英/钾长石的相应属性以及整体矿物分布的影响相对较小。

{"title":"Machine learning approaches to predicting the uniaxial compressive strength of granite from image-derived mineralogical features","authors":"Changdi He , Brijes Mishra , Yuan Li , Jessica Michelle Wempen","doi":"10.1016/j.enggeo.2026.108571","DOIUrl":"10.1016/j.enggeo.2026.108571","url":null,"abstract":"<div><div>This study uses machine learning (ML) models to predict the numerically simulated uniaxial compressive strength (UCS) of granite directly from digital images. Specifically, the images were processed using an in-house Digital Image Processing (DIP) tool to estimate mineralogical features (used as input features for the ML models), including mineral content, grain size, and spatial distribution. Mineral content and distribution were quantified using <span><math><mi>m</mi></math></span>-harmonic Fourier series equations, whereas mineral grain size was determined using the 4-connectivity method. The target UCS values were derived from the 2D physically informed Subspring Network Breakable Voronoi (SNBV) microstructural models, replicating the mineralogical features observed in the granite images. Extreme Gradient Boosting (XGBoost) models with different input combinations and hyperparameter optimization methods were trained and evaluated on 126 granite images using a single train/test split and repeated 5-fold cross-validation. Results indicate that the input combination of mineral content, grain size, and spatial distribution parameters from <span><math><mi>m</mi></math></span>-harmonic Fourier series combined with SHapley Additive Explanations (SHAP)-based feature selection, yield the best and robust performance, whereas increasing the harmonic order has a limited effect on accuracy. Among the tested optimization methods, the Optuna–XGBoost model achieved the best performance. In addition, UCS prediction is controlled mainly by the content and grain size of biotite and plagioclase, while the corresponding attributes of quartz/K-feldspar, as well as overall mineral distribution play a comparatively minor role.</div></div>","PeriodicalId":11567,"journal":{"name":"Engineering Geology","volume":"363 ","pages":"Article 108571"},"PeriodicalIF":8.4,"publicationDate":"2026-01-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146001414","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

Experimental study on the benefits of nature-based solutions for debris-flow mitigation via synergistic eco-geotechnical measures 通过协同生态岩土工程措施缓解泥石流的基于自然的解决方案效益的实验研究

IF 8.4 1区工程技术 Q1 ENGINEERING, GEOLOGICAL

Engineering Geology

Pub Date : 2026-01-19 DOI: 10.1016/j.enggeo.2026.108547

Songtang He , Songyuan Zou , Yuchao Qi , Yong Li , Zengli Pei , Peng Zhao , Daojie Wang

Research on nature-based solutions (NbSs) for mitigating debris-flow hazards has increased interest in eco-geotechnical systems. Most studies focused on the efficiency of isolated mitigation measures, while the benefits and mechanisms of coordinated approaches remain unclear. Consequently, this study proposed a novel approach to mitigating debris-flow velocity, sediment transport, and energy by utilizing tree-shrub mixed-vegetation filter strips (T-SMVFS) along S-shaped flow paths combined with dams. The optimal design determination involved four steps: 1) optimal T-SMVFS row and stem spacings were determined; 2) S-shaped flow path parameters were set based on width ratios; 3) effects of synergistic and individual measures on debris-flow reduction were compared; and 4) a flow velocity reduction equation was constructed, considering the influence of topographic features, vegetation planting patterns, and debris flow properties. Results revealed that a completely covered T-SMVFS with row and stem spacings of 10 and 6 cm in 1/50th scale, respectively, exhibited the best reduction effects with 50% energy reduction, 55% sediment interception, and 53% flow discharge regulation. As the S-shaped flow path width increased, the flow reduction and sediment interception rates decreased sequentially while the transportation capacity increased. Synergistic measures achieved 60% and 70% in energy and sediment interception reductions, respectively, outperforming pure geotechnical and biological measures. Comparisons between different synergistic approaches indicated that a coupled S-shaped vegetation filter strip with a 45% flow path proportion and a beam dam was more effective in reducing debris flow. These findings provide a reference for subsequent optimal mitigation solutions involving NbSs that integrate synergistic eco-geotechnical measures.

基于自然的解决方案（nbs）减轻泥石流灾害的研究增加了人们对生态岩土工程系统的兴趣。大多数研究侧重于孤立的缓解措施的效率，而协调办法的效益和机制仍不清楚。因此，本研究提出了一种新的方法，即利用树-灌木混合植被过滤带（T-SMVFS）沿s型流道与水坝结合，以减轻泥石流速度，泥沙运输和能量。优化设计的确定分为四个步骤：1)确定最佳T-SMVFS排杆间距；2)基于宽度比设置s型流道参数；3)比较了协同措施和单项措施对减少泥石流的效果；4)考虑地形特征、植被种植方式和泥石流性质的影响，建立了泥石流流速减小方程。结果表明，在1/50尺度下，行间距为10 cm、茎间距为6 cm的全覆盖T-SMVFS，其减能效果最佳，减能50%，截留泥沙55%，调节流量53%。随着s型流道宽度的增大，减流量和截沙率依次减小，而输沙能力则依次增大。协同措施在能量和沉积物截留方面分别减少了60%和70%，优于单纯的岩土和生物措施。不同协同方式的对比表明，流道比例为45%的s型植被过滤带与梁坝耦合的协同方式对泥石流的减少效果更好。这些发现为后续涉及nbs的最佳缓解解决方案提供了参考，这些解决方案结合了协同生态岩土工程措施。

{"title":"Experimental study on the benefits of nature-based solutions for debris-flow mitigation via synergistic eco-geotechnical measures","authors":"Songtang He , Songyuan Zou , Yuchao Qi , Yong Li , Zengli Pei , Peng Zhao , Daojie Wang","doi":"10.1016/j.enggeo.2026.108547","DOIUrl":"10.1016/j.enggeo.2026.108547","url":null,"abstract":"<div><div>Research on nature-based solutions (NbSs) for mitigating debris-flow hazards has increased interest in eco-geotechnical systems. Most studies focused on the efficiency of isolated mitigation measures, while the benefits and mechanisms of coordinated approaches remain unclear. Consequently, this study proposed a novel approach to mitigating debris-flow velocity, sediment transport, and energy by utilizing tree-shrub mixed-vegetation filter strips (T-SMVFS) along S-shaped flow paths combined with dams. The optimal design determination involved four steps: 1) optimal T-SMVFS row and stem spacings were determined; 2) S-shaped flow path parameters were set based on width ratios; 3) effects of synergistic and individual measures on debris-flow reduction were compared; and 4) <strong>a flow velocity reduction equation was constructed, considering the influence of topographic features, vegetation planting patterns, and debris flow properties.</strong> Results revealed that a completely covered T-SMVFS with row and stem spacings of 10 and 6 cm in 1/50th scale, respectively, exhibited the best reduction effects with 50% energy reduction, 55% sediment interception, and 53% flow discharge regulation. As the S-shaped flow path width increased, the flow reduction and sediment interception rates decreased sequentially while the transportation capacity increased. Synergistic measures achieved 60% and 70% in energy and sediment interception reductions, respectively, outperforming pure geotechnical and biological measures. Comparisons between different synergistic approaches indicated that a coupled S-shaped vegetation filter strip with a 45% flow path proportion and a beam dam was more effective in reducing debris flow. These findings provide a reference for subsequent optimal mitigation solutions involving NbSs that integrate synergistic eco-geotechnical measures.</div></div>","PeriodicalId":11567,"journal":{"name":"Engineering Geology","volume":"363 ","pages":"Article 108547"},"PeriodicalIF":8.4,"publicationDate":"2026-01-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146001415","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 implicit feature learning approach for automatic digital twinning of three-dimensional subsurface stratigraphy from limited boreholes 有限钻孔三维地下地层自动数字孪生的隐式特征学习方法

IF 8.4 1区工程技术 Q1 ENGINEERING, GEOLOGICAL

Engineering Geology

Pub Date : 2026-01-19 DOI: 10.1016/j.enggeo.2026.108568

Zehang Qian, Chao Shi

Digital twinning of underground space is a transformative paradigm for life-cycle risk assessment, management, and operation of geotechnical infrastructure. A prerequisite for this transformation is the automated construction of subsurface geological domains from sparse borehole data with quantified stratigraphic uncertainty. However, most three-dimensional (3D) stratigraphic modelling methods are categorical variable (CaV)-based models and primarily rely on site-specific prior geological knowledge or calibration of geological settings to replicate intricate 3D spatial features. This inevitably involves significant computational burdens and frequent human intervention, which impede real-time data transmission and model updating. In this study, a continuous variable (CoV)-based 3D stratigraphic modelling method leveraging multiple Signed Distance Functions (MSDFs) and Bayesian Compressive Sensing (BCS) is proposed to address these challenges. MSDFs and a spatial partitioning technique are introduced to accelerate the transformation of borehole CaVs into CoVs before leveraging BCS for spatial interpolation of CoVs. Subsequently, random field samples of predicted CoVs are transformed back into the most probable geological domain with quantified uncertainty. The proposed implicit method is applied to both synthesized stratigraphy and real Nan-Chang Urban Rail project. Results demonstrate MSDF-BCS-3D excels in reconstructing 3D stratigraphy, comprising multiple interlayers and folded stratification, with uncertainty quantification in a non-parametric and near real-time manner.

地下空间的数字孪生是岩土基础设施生命周期风险评估、管理和运营的变革范式。实现这种转变的先决条件是利用具有量化地层不确定性的稀疏钻孔数据自动构建地下地质域。然而，大多数三维（3D）地层建模方法是基于分类变量（CaV）的模型，主要依赖于特定地点的先验地质知识或地质环境校准来复制复杂的三维空间特征。这不可避免地带来了巨大的计算负担和频繁的人为干预，阻碍了实时数据传输和模型更新。在这项研究中，提出了一种基于连续变量（CoV）的三维地层建模方法，利用多个签名距离函数（msdf）和贝叶斯压缩感知（BCS）来解决这些挑战。在利用BCS对cov进行空间插值之前，引入了msdf和空间分区技术来加速井眼cav向cov的转换。随后，将预测冠状病毒的随机场样本转换回最可能的地质区域，并量化不确定性。本文提出的隐式方法应用于综合地层和实际的南昌城市轨道工程。结果表明，MSDF-BCS-3D能够以非参数和接近实时的方式进行不确定度量化，在重建包括多夹层和褶皱层的三维地层方面表现出色。

{"title":"An implicit feature learning approach for automatic digital twinning of three-dimensional subsurface stratigraphy from limited boreholes","authors":"Zehang Qian, Chao Shi","doi":"10.1016/j.enggeo.2026.108568","DOIUrl":"10.1016/j.enggeo.2026.108568","url":null,"abstract":"<div><div>Digital twinning of underground space is a transformative paradigm for life-cycle risk assessment, management, and operation of geotechnical infrastructure. A prerequisite for this transformation is the automated construction of subsurface geological domains from sparse borehole data with quantified stratigraphic uncertainty. However, most three-dimensional (3D) stratigraphic modelling methods are categorical variable (CaV)-based models and primarily rely on site-specific prior geological knowledge or calibration of geological settings to replicate intricate 3D spatial features. This inevitably involves significant computational burdens and frequent human intervention, which impede real-time data transmission and model updating. In this study, a continuous variable (CoV)-based 3D stratigraphic modelling method leveraging multiple Signed Distance Functions (MSDFs) and Bayesian Compressive Sensing (BCS) is proposed to address these challenges. MSDFs and a spatial partitioning technique are introduced to accelerate the transformation of borehole CaVs into CoVs before leveraging BCS for spatial interpolation of CoVs. Subsequently, random field samples of predicted CoVs are transformed back into the most probable geological domain with quantified uncertainty. The proposed implicit method is applied to both synthesized stratigraphy and real Nan-Chang Urban Rail project. Results demonstrate MSDF-BCS-3D excels in reconstructing 3D stratigraphy, comprising multiple interlayers and folded stratification, with uncertainty quantification in a non-parametric and near real-time manner.</div></div>","PeriodicalId":11567,"journal":{"name":"Engineering Geology","volume":"363 ","pages":"Article 108568"},"PeriodicalIF":8.4,"publicationDate":"2026-01-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146001416","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