The reliability of dam-break scour prediction remains sensitive to the calibration of evolving bed resistance and to uncertainty in the associated model parameters. This study develops a probabilistic inversion framework that integrates a two-phase Smoothed Particle Hydrodynamics (SPH) model with Bayesian inference implemented through Markov chain Monte Carlo (MCMC) sampling for uncertainty-aware parameter inference in transient water-soil interaction. The framework captures the progressive loss of bed resistance through a time-dependent degradation of the internal friction angle. It is evaluated against the classical Louvain dam-break experiment using two spatial parameterization strategies: a vertically layered representation and a longitudinally partitioned representation. Bayesian updating improves the reconstruction of the observed transient soil–water interface relative to the deterministic baseline. Posterior inference and multi-chain convergence diagnostics reveal parameter-dependent practical identifiability, with the initial internal friction angle more clearly constrained than the degradation coefficients over the short observation window. A within-experiment temporal holdout and complexity-adjusted model comparison indicate that the longitudinally partitioned model better represents the spatially non-uniform scour response in this benchmark, although its improved morphological fidelity does not ensure reliable estimation of every regional parameter. Overall, the proposed framework provides a tractable route from deterministic calibration toward uncertainty-aware inference in SPH-based scour simulations and a physically interpretable basis for evaluating evolving bed response.

Crack identification under ambient excitation is crucial for long-term monitoring of in-service concrete dams, yet damage-induced changes are often weak and masked by environmental and operational variability. This paper proposes a novel vibration-based crack identification method based on dynamic kernel principal component analysis (DKPCA) and multivariate statistical process control (MSPC). The method operates directly on multi-channel vibration responses without relying on conventional modal parameter tracking. First, the DKPCA with time-delay embedding is used to construct an intact-state baseline feature space, from which Hotelling's T2 statistic is computed for multivariate monitoring. Specifically, two quantitative damage sensitive indicators are derived from the T2 statistic, namely the anomaly rate Ca and the maximum anomaly rate Cr, to characterize the frequency and severity of control limit exceedance. Next, a screening step based on Mahalanobis distance is introduced to remove occasional abnormal indicator points and enhance robustness to transient interference. Finally, crack existence is determined using a binary 3σ elliptical confidence region constructed from the intact-state indicator distribution, enabling objective and repeatable decisions. The proposed method is verified through numerical simulations on a two-dimensional gravity dam and a three-dimensional arch dam. Sensitivity studies examine the effects of sensor spacing, signal-to-noise ratio, crack depth, and crack location, while an array-type study on the arch dam demonstrates crack localization and propagation assessment. Results indicate reliable crack identification under ambient excitation with improved interpretability and applicability when conventional modal-feature methods lack sensitivity.

The Concrete Damaged Plasticity (CDP) model has been widely used for analyzing the seismic response of concrete dams. In the case of extreme earthquakes, the seismic response of gravity dams may be quite sensitive to the CDP model parameters due to substantial shear stress at its base. This paper focuses on the influence of the CDP model parameters, including the eccentricity , the ratio of biaxial compressive strength to uniaxial compressive strength , the ratio of the second stress invariant on the tensile meridian to that on the compressive meridian , the dilation angle , and the compressive stiffness recovery factor , on the response of concrete dams under extreme earthquakes. Taking a 71-m-high concrete dam as example, the damage distribution under varied CDP model parameters is analyzed considering design and hypothetical extreme earthquakes, respectively. The results show that the parameters , , and have little effect on the seismic response of the dam. However, the effects of the parameters and on the seismic response of the dam are intensity-dependent. Under the extreme earthquake, a larger dilation angle overestimates shear resistance because it amplifies the volume expansion effect (dilation effect) caused by shear damage, and a larger compressive stiffness recovery factor also underestimates the seismic response. When approaches 0, a vibration-isolation layer will form at the dam base, which relieves damage at the upper dam body. These results provide important references for the safety assessment of concrete dams under seismic action and the rational selection of CDP model parameters.

Gate hoist structures are safety-critical appurtenant systems in hydropower projects, yet their seismic behavior has received far less attention than that of dam bodies. This study investigates the seismic performance of a gate hoist structure in a roller-compacted concrete gravity dam using a three-dimensional finite element model of the foundation-dam-gate hoist structure system. Linear-elastic time-history analyses are performed for both unreinforced and brace-reinforced configurations under the design earthquake. Seismic performance is evaluated using principal stress, the point safety factor, the section safety factor, and demand-to-capacity ratio (DCR). The results show that tensile stress is concentrated mainly at beam-column joints in the unreinforced structure, where the maximum principal stress exceeds the concrete tensile strength and local cracking may occur. Brace reinforcement reduces the peak tensile stress and improves the safety level at critical locations, while vulnerable regions may shift to the beam-brace joints rather than being completely eliminated. These findings indicate that the seismic safety of gate hoist structures is better assessed within an integrated dam-structure system, while multi-index evaluation provides a more comprehensive basis for identifying vulnerable regions and assessing reinforcement effectiveness.

As critical infrastructure, the safe operation of hydropower dams is of paramount importance. A reliable structural health monitoring (SHM) system is therefore essential to ensure this safety. This paper presents LLM4SHM, a large language model (LLM)-based multi-agent system for dam SHM, which employs an LLM as the central orchestrator to coordinate four professional agents responsible for behavior prediction, anomaly detection and early warning, monitoring data management, and engineering knowledge retrieval. LLM4SHM establishes a closed-loop, end-to-end intelligent SHM framework by integrating the model context protocol (MCP) for tool invocation and retrieval-augmented generation (RAG) for domain-knowledge fusion. Demonstrated at the LYX and XW dams, LLM4SHM autonomously executes SHM tasks through natural language interaction, improving monitoring timeliness, diagnosis accuracy, and decision-making efficiency. This paper integrates LLMs with a multi-agent architecture for dam SHM, advancing automation and intelligent decision support for large-scale hydraulic infrastructure and providing a scalable foundation for next-generation SHM systems.

Traditionally, the optimization of tuned liquid column dampers (TLCDs) has favored larger length ratios (the ratio of the horizontal liquid column length to the total liquid column length), assuming they always yield better performance. However, this conventional approach often overlooks the vulnerability of TLCDs with large length ratios to strong vibrations, which can lead to air being drawn into the horizontal pipe and reducing damping efficiency. To address this, a novel two-phase design framework that treats the excitation intensity as a critical design parameter is proposed. This framework optimizes three key parameters: length ratio, area blocking ratio of the orifice (related to flow resistance), and frequency tuning ratio (the liquid-to-structure frequency ratio), while explicitly considering the negative impact of air entering the horizontal pipe on TLCD performance. The process begins with the use of closed-form solutions and numerical search methods, based on the classical TLCD model, to determine the optimized frequency tuning ratio in the frequency domain. Subsequently, a newly developed generalized TLCD model, which accounts for the effects of air entering the horizontal pipe, is employed to perform time-domain analyses across a wide range of excitation amplitudes and parameter combinations. The results identify the optimized length and blocking ratios that balance efficiency and robustness, revealing specific conditions under which conventional designs underperform. These findings provide a more realistic design philosophy for TLCDs, ensuring reliable vibration control performance under design excitation amplitudes, especially during high-intensity events.

Strong earthquakes or heavy rainfall may trigger landslides, generating surge waves that pose serious threats to dam safety. This study develops and validates a coupled SPH-FEM framework to simulate multi-hazard interactions involving landslide-generated surges, reservoir dynamics, and their subsequent impact on dam structures. The methodology includes two key stages. First, a unified GPU-accelerated SPH framework simulates the landslide-induced surge dynamics and bidirectional fluid-structure interactions, with both fluid dynamics and auxiliary structural responses described using SPH particles. Second, spatiotemporally varying surge pressure fields, derived from the unified SPH framework, are mapped onto the high-resolution FEM substructure model for high-fidelity structural analysis. This hybrid approach effectively balances computational efficiency and physical fidelity, enabling comprehensive safety assessments of concrete dams subjected to landslide-induced surges. Parametric analyses reveal that surge dynamics and hydrodynamic pressure on the dam surface are predominantly governed by landslide characteristics, including volume, fragmentation degree, and proximity to the dam. Among these factors, the landslide fragmentation degree exhibits a non-monotonic influence. Moderate fragmentation leads to amplified secondary and tertiary wave amplitudes due to dynamic arrangement and reduced energy dissipation of the landslide mass. This finding underscores the significance of incorporating landslide fragmentation effects and site-specific topography into hazard assessments.

We propose a dynamic hypoplastic model to evaluate the influence of grain breakage on soil-structure interface behavior under cyclic loading conditions. A variable critical state line (CSL) is introduced based on the relative breakage ratio, with the latter formulated as a function of plastic strain within the hypoplastic framework. The model incorporates both the relative breakage ratio and intergranular strain to simulate the deformation response of gravelly soils under cyclic loading, accounting for grain breakage. Validation is conducted through simulations of a series of laboratory tests, including monotonic and cyclic loading experiments on gravel-concrete, gravel-steel, and sand-geosynthetic interfaces.

In large-scale three-dimensional fluid-structure interaction (FSI) problems, complex interfaces between heterogeneous media often require locally highly refined meshes, especially near fluid-structure interfaces where coupling behavior occurs. This fine-scale discretization imposes severe restrictions on global time steps, making standard explicit schemes with lumped mass matrices inefficient or inapplicable without manual adjustments. Although fully implicit methods allow for the choice of larger time steps, they involve solving large monolithically coupled systems, leading to high computational and memory costs. Moreover, traditional finite element methods suffer from poor mesh transition near material interfaces, further reducing efficiency in multi-material domains. To address these issues, this study presents an implicit-explicit (IMEX) scaled boundary finite element method (SBFEM) combined with automatic octree-based mesh generation. An octree-based discretization facilitates localized mesh refinement near interfaces, allowing efficient representation of complex domains, while SBFEM provides a natural framework due to its compatibility with polyhedral elements. Then, an IMEX time integration scheme is developed, in which implicit methods are applied locally to regions with fine meshes and coupling domains, such as fluid- structure interfaces, while explicit methods are used elsewhere to maintain overall computational efficiency. This targeted treatment alleviates time step restrictions and robustly handles interface coupling, significantly enhancing computational efficiency. The approach is validated by benchmark FSI problems and applied to the dynamic analysis of a dam-reservoir-foundation system, demonstrating its accuracy and practicality for large-scale engineering simulations.

Tuned liquid column dampers (TLCDs) embedded in rotating wind turbine blades are effective in reducing edgewise vibrations. However, their dynamic behavior under the combined influence of centrifugal and gravitational fields has not been rigorously verified, and existing designs may rely on untested simplifications. This study systematically revisits these assumptions and develops improved control strategies based on the findings. First, intrinsic dynamic analyses incorporate both centrifugal and gravitational effects to evaluate the validity of the conventional “centrifugal-dominant” frequency assumption. Second, comparative load analyses quantify the relative contributions of aerodynamic and gravitational excitations, examining the “aerodynamic-dominant” load assumption. Third, the design error from neglecting gravitational restoring forces is evaluated by comparing a complete model including gravity with a simplified centrifugal-only model under identical excitations. The verification results show that these assumptions may break down under some operating conditions, particularly at low rotor speeds. Building upon these insights, several semi-active TLCD strategies, including adjustable water level control, latching control, air spring-based control, and variable cross-section control, are proposed and evaluated. The proposed designs offer guidance for optimizing TLCD performance under varying rotational conditions.

In order to simulate the whole process of concrete dams from small deformation damage to large deformation failure under seismic load, this study proposed a tension-compression-shear coupled plastic-damage thin-layer element model. Based on continuum mechanics and cohesive zone theory, this model condenses the plasticdamage behavior of concrete into the thin-layer element. The constitutive model of the thin layer element is constructed using the yield surface based on the Coulomb's law and a non-associated flow rule. Subsequently, the tensile-shear test of concrete with double notches, the direct shear test of concrete, and the cyclic loading test are simulated to verify the proposed model. Finally, the failure process of a concrete dam is simulated, the results show that the proposed tension-compression-shear coupled plastic-damage thin-layer element model aligns well with existing continuous nonlinear model in small deformation stage and effectively captures failure processes in large deformation stage.

The seismic response of concrete-faced rockfill dam (CFRD) is intrinsically linked to the complex interaction between the dam body, foundation rock and reservoir water. This paper introduces a comprehensive finite element model accounting for all physical factors in the rockfill dam-reservoir-foundation system. The model employs a generalized plasticity model for the rockfill and a plastic damage model for the concrete face slab. Reservoir water is simulated using acoustic elements. The interaction between concrete face slab and rockfill and the contraction joints is modeled using the contact boundary method. The radiation damping due to infinite foundation is considered with the Viscoelastic artificial boundary. The model is applied to the real-world case of the Zipingpu Dam during the 2008 Wenchuan earthquake. Compared to conventional approaches like the Westergaard's added mass for modeling hydrodynamic pressure or simplified constitutive models for rockfill and concrete face slab, the proposed model shows superior agreement with the observed seismic behavior, highlighting its enhanced accuracy in simulating the dynamic response of rockfill dam-reservoir-foundation system.

Objective High-arch dams require reliable long-term monitoring to ensure safety in complex operating environments and under extreme loads. Vibration-based operational modal analysis under ambient excitation is well-suited for continuous deployment due to its passive and minimally intrusive nature. However, the vibration response under these conditions is often weak and susceptible to noise and non-stationary excitation. When covariance-driven stochastic subspace identification (SSI-COV) is applied, the stabilization diagram frequently becomes cluttered with a mixture of physical poles and spurious poles, complicating manual pole selection and diminishing the effectiveness of automated pole clustering, particularly for densely spaced modes and weakly excited higher-order modes. This study aims to enhance automated modal identification for high-arch dams by refining the clustering distance metric to better separate physical poles from spurious ones. Methods A classical workflow that combines SSI-COV with density-based spatial clustering of applications with noise (DBSCAN) is adopted and enhanced by redesigning the clustering distance metric. Conventional metrics typically use a weighted summation of frequency difference and mode-shape similarity, which may not fully capture the relationships between the two features and may falter when identifying densely spaced modes. In this study, a coupled distance formulation is introduced that directly integrates the modal assurance criterion (MAC) with the absolute frequency deviation and is placed in the denominator. When the mode-shape correlation between two poles is weak and MAC approaches 0, the distance increases significantly. By contrast, when the correlation is strong and MAC approaches 1, the distance reduces to the absolute frequency deviation. Consequently, pole pairs with simultaneously exhibit small frequency differences and highly consistent mode shapes are assigned minimal clustering distances, whereas those with large frequency differences or inconsistent mode shapes are pushed apart. This leads to a clearer separation of physical and spurious modes in the stabilization diagram, thus meeting requirements for automated clustering-based interpretation. A statistical analysis of clustering distances is then performed using the stabilization diagram from a high-arch dam dataset. Finally, the method is validated through two case studies. The first involves a five-degree-of-freedom numerical system excited by broadband white noise with added measurement noise; the responses are segmented into consecutive windows to test both single-window identification and continuous modal tracking. The second case utilizes multisensor field vibration data from an actual high-arch dam, including a representative short-duration record and a multiday dataset for continuous monitoring. For each case, the proposed formulation computes clustering distances, DBSCAN clusters the poles, and modal frequencies and damping ratios are extracted to evaluate clustering accuracy and the performance of automated identification. Results The distance-based statistical analysis reveals that the proposed metric enhances separability. Pole pairs that meet both feature-consistency conditions are clustered within a compact distance interval, whereas partially consistent or inconsistent pairs shift toward larger distances. In the numerical example, the proposed method produces physical clusters that are less prone to absorbing noise points compared to the baseline metric, leading to an approximately 31% increase in identified modal poles for weakly excited higher-order modes. In the real dam case, the baseline metric generates excessive clusters that are closely packed, making it difficult to form effective clusters with clear and interpretable boundaries. By contrast, the proposed method clearly identifies three clusters for the high-arch dam and achieves a 34% increase in recognized poles for the relatively higher-order mode during continuous identification. This suggests that the improvement is most significant for relatively higher-order modes, where the number of identified modal poles increases by approximately one-third compared to the baseline approach. Conclusions By integrating frequency and MAC in a division-based formulation, the proposed metric enhances the compactness of the identified clusters and enables stable distinction between physical and spurious poles, while also improving the identification of weakly excited higher-order vibration modes. This directly enhances the robustness of DBSCAN-based automated modal identification and continuous modal tracking for high-arch dams under ambient excitation. The method can be easily incorporated into existing SSI-COV workflows, as it mainly updates the distance-computation step, providing a practical solution for reliable long-term vibration-based dam monitoring.

Historical earthquakes have demonstrated that radial gates on spillway sections, despite being critical appurtenant structures, are highly seismically vulnerable. Their failure under strong seismic excitations can precipitate secondary disasters. Nevertheless, seismic evaluations of high dams predominantly focus on the dam body, largely overlooking these appurtenant systems and thereby compromising comprehensive project safety. To bridge this research gap, this study develops a refined finite element model of the radial gate-spillway system to systematically evaluate its seismic performance. A 200 m-high roller-compacted concrete gravity dam spillway section is selected as a case study, and nonlinear time-history analyses are conducted under six seismic intensity levels ranging from 1.0 to 2.0 times the design peak ground acceleration (PGA). A set of seismic performance indices related to acceleration, relative displacement, damage, and buckling instability is introduced to characterize the coupled response of the gate–spillway system. The results show that the gate exhibits more pronounced acceleration amplification than the pier and is more sensitive to seismic input intensity, whereas the pier response remains relatively stable. The maximum gate–pier relative displacement is dominated by the stream-direction component, while the residual relative displacement remains much smaller and shows limited variation with increasing seismic intensity, indicating that the gate–pier relative motion is governed mainly by recoverable elastic deformation. The damage and stability analyses indicate that the dam heel, the downstream pier transition, and the upper gate arms are vulnerable regions under strong earthquakes. In particular, the upper gate arms may become critical components because plastic deformation can reduce their effective stiffness and increase the risk of buckling instability. These findings highlight the necessity of evaluating the seismic safety of spillway radial gates within the integrated dam–gate system, with particular attention to coupled dynamic response, local damage, and potential arm instability.

Three-point bending and splitting tensile tests were conducted on four-graded concrete containing coarse aggregates up to 150 mm to investigate tensile strain rate effects over 10_6 to 10_2 s_ 1. Splitting specimens were core-drilled from the tested beams to ensure material consistency between the two configurations. Both methods exhibited strain rate dependence, whereas the core-drilled splitting results showed markedly greater scatter because the reduced specimen size amplified material heterogeneity. To better interpret this variability, splitting specimens were classified according to aggregate-splitting characteristics. Using an aggregate-splitting area threshold of 18%, the influence of aggregate-induced randomness on the extracted rate sensitivity was reduced. After normalization by the corresponding quasi-static strengths, the strain rate sensitivity obtained from the grouped splitting results became consistent with that from the large-scale bending tests. These results indicate that, when aggregate-related heterogeneity is properly accounted for, core samples can provide a calibrated experimental basis for evaluating the rate-dependent tensile behavior of existing dam concrete within the tested strain-rate range, helping bridge laboratory constraints and engineering-scale assessment.

Scouring can significantly impact the natural frequencies and dynamic responses of monopile-supported offshore wind turbines (OWTs) under the combined effects of wind, waves, and seismic loads. As scouring progresses, the passive vibration reduction devices commonly used in OWTs gradually lose their optimal tuning effects. In this study, we present an innovative numerical analysis of a semi-active toroidal tuned liquid column damper (S-TTLCD), specifically designed to effectively suppress multi-directional vibrations in monopile-supported OWTs. This is particularly targeted at enhancing the structural stability under the combined effects of wind, wave, and seismic forces in areas prone to scouring. The soil-pile interaction is modeled utilizing a P-y curve relationship, while the model of OWTs comprehensively integrates the impacts of multi-hazard loading conditions and diverse scouring depths to enable a rigorous systematic analysis. A sophisticated semi-active control strategy based on short-time Fourier transform (STFT) and displacement based groundhook (DBG) is devised, in which the natural frequency and damping of the S-TTLCD are precisely adjusted to match the responses of OWTs affected by scouring. The findings demonstrate that the S-TTLCD mitigates the dynamic responses of OWTs, outperforming the passive system, especially when the structural dynamic properties is altered due to scouring.

Continuous identification of modal parameters of large civil structures under ambient excitation is essential for structural health monitoring (SHM). In practice, however, automated operational modal analysis (OMA) rarely delivers clear and directly applicable mode-evolution curves in real-time online. In this paper, a two-stage clustering method is proposed for online OMA. The method first applies covariance-driven stochastic subspace identification (SSI-COV) in short time windows to extract modal points. On this basis, density-based spatial clustering of applications with noise (DBSCAN) is performed at two different scales (single and cross windows), defined as two stages in this study. The mode symmetry coefficient (MSC) is further used as a geometric constraint for symmetrical structures. A one-year dataset of vibration measurements from a concrete gravity arch dam is employed to systematically investigate the effects of clustering control parameters on the performance of the proposed method. Moreover, a quality evaluation metric is developed to jointly evaluate the bandwidth and density of the identified mode-evolution curves, leading to the determination of practical parameter ranges. The transferability of both the method and the parameter recommendations is then demonstrated on a concrete arch dam with different excitation levels and closely spaced modes. The results show that the proposed method can automatically produce clear mode-evolution curves in real-time online, significantly suppress spurious modes, and reliably separate closely spaced modes, providing a robust and reproducible solution for online modal parameter tracking in long-term structural monitoring.