Structural Damage Pattern Research Articles

Characteristics of strong ground motions and structural damage patterns from the February 6th, 2023 Kahramanmaraş earthquakes, Türkiye

Characteristics of strong ground motions and structural damage patterns from the February 6th, 2023 Kahramanmaraş earthquakes, Türkiye

Vascular Risk Factors and Brain Health in Aging: Insights from a Community-Based Cohort Study.

The aging population and high rates of Alzheimer's disease (AD) create significant medical burdens, prompting a need for early prevention. Targeting modifiable risk factors like vascular risk factors (VRFs), closely linked to AD, may provide a promising strategy for intervention. This study investigates how VRFs influence cognitive performance and brain structures in a community-based cohort. In this cross-sectional study, 4,667 participants over 50 years old, drawn from the Beijing Ageing Brain Rejuvenation Initiative project, were meticulously examined. Cognitive function and VRFs (diabetes mellitus, hypertension, hyperlipidemia, obesity, and smoking), were comprehensively assessed through one-to-one interviews. Additionally, a subset of participants (n = 719) underwent MRI, encompassing T1-weighted and diffusion-weighted scans, to elucidate gray matter volume and white matter structural network organization. The findings unveil diabetes as a potent detriment to memory, manifesting in atrophy within the right supramarginal gyrus and diminished nodal efficiency and degree centrality in the right inferior parietal lobe. Hypertension solely impaired memory without significant structural changes. Intriguingly, individuals with comorbid diabetes and hypertension exhibited the most pronounced deficits in both brain structure and cognitive performance. Remarkably, hyperlipidemia emerged as a factor associated with enhanced cognition, and preservation of brain structure. This study illuminates the intricate associations between VRFs and the varied patterns of cognitive and brain structural damage. Notably, the synergistic effect of diabetes and hypertension emerges as particularly deleterious. These findings underscore the imperative to tailor interventions for patients with distinct VRF comorbidities, especially when addressing cognitive decline and structural brain changes.

Multiaxial Fatigue Damage Analysis of Steel–Concrete Composite Beam Based on the Smith–Watson–Topper Parameter

The fatigue performance of steel–concrete composite beams is crucial for ensuring structural safety. To account for the member’s multiaxial stress state, this study employed the critical surface method, using fatigue damage parameters as an evaluation index for assessing fatigue performance. Static and fatigue performance tests on steel–concrete beams were conducted to identify failure characteristics, which informed the development of a finite element model that incorporates concrete damage. Using the SWT model, the most unfavorable loading parameters were determined by analyzing critical paths on the test beams, providing a basis for predicting how initial defects impact fatigue performance. The impact of initial defects on the fatigue performance of the composite beam is assessed using this criterion. The results indicate that the discrepancy between the actual and predicted load capacities of the test beam is within 5%, and cyclic loading significantly affects the test beam’s mechanical properties, resulting in a 27% reduction in load capacity and a 48% increase in deflection after 2 million cycles. Finite element modeling reveals that components experience multiaxial stress, with test beam mechanical property changes aligning with predicted fatigue damage parameters, confirming the reliability of using these parameters as a criterion. As the strength of the composite beams diminished due to pore defects, the fatigue damage parameter escalated, increasing the likelihood of crack formation. However, once the concrete’s strength fell to a level where the pegs were insufficiently constrained, the structural damage pattern shifted, and the fatigue damage parameter subsequently decreased.

Credal identification of damage patterns in ultra-thin-ply composite bonded/bolted interference-fit joints

Investigating the mechanical behavior and damage patterns of ultra-thin-ply composite joints is essential for ensuring their widespread application in the aerospace field. It is currently an advanced method of exploring the damage patterns of composite joint structures via a combination of acoustic emission and machine learning algorithms. However, the existing research lacks the evaluation indexes for the determination of the number of damage categories, and on the other hand, the overlapping area and noise of the monitored damage data cannot be accurately identified. In this paper, load-bearing performance of ultra-thin-ply composite hybrid bonded/bolted interference-fit joints is studied, then the SDI(Structural Damage Index) evaluation index is proposed to quantitatively analyze the number of damage categories of the complex joints, and finally the adaptive optimization algorithm based on the credal partition for the complex joints of composite is studied. The experimental results demonstrate the accuracy and generalization of the SDI formulae, and the introduction of the credal partition that submits data where overlapping regions cannot be accurately identified to ensembles of special single classes (meta-clusters) obtains deeper insights into the damage data and improves identification and isolation of noise.

Study on the failure effect of aramid reinforced concrete slab under localized blast loading

As a new material with low density, high modulus and high strength, aramid fibers and composite laminates are widely used in protective structures. Aramid reinforced concrete panels as a new type of protective combination of structures, especially in the concrete-aramid laminate composite blast-resistant structure, the optimal thickness combination of reinforced concrete and aramid laminate has not been targeted research. In this paper, three types of scaled down reinforced concrete (RC) slabs, aramid fiber reinforced concrete (AFRC) slabs and aramid fiber reinforced plastic reinforced concrete (AFRPRC) slabs were designed and fabricated to carry out experimental studies under local explosive loading and to compare the blast resistance of different slabs through damage analysis. Based on LS-DYNA software, the damage characteristics of AFRPRC plates were investigated and compared and analysed with the test results to verify the accuracy and applicability of the FEA model. Typical damage patterns of structures under different TNT charges were obtained, and empirical equations between the damage characteristics of reinforced concrete slabs (maximum mid-span displacement, crater diameter, crater depth) and TNT charges were fitted by means of magnitude analysis. The effect of the aramid layer on the structural blast resistance performance was analysed from three perspectives: maximum central displacement, stress wave attenuation rate and internal energy absorption. The results show that AFRC improves the blast resistance of reinforced concrete not significantly relative to ordinary reinforced concrete slabs, while AFRPRC slabs maintain good integrity and show better blast resistance. The best blast resistance performance of AFRPRC panels is achieved when the ratio of reinforced concrete thickness to aramid layer thickness is 50:3. The results of the study can provide theoretical basis and reference for efficient protection of composite structures.

Structural health monitoring on offshore jacket platforms using a novel ensemble deep learning model

Monitoring health condition of offshore jacket platforms is crucial to prevent unexpected structural damages, where a prevailing challenge involves translating available feature information into structural damage patterns. Although the artificial neural network (ANN) models are popular in addressing this challenge, they often fail to capture the temporal correlations between the feature information and the damage patterns, which reduce their capability for discovering the laws governing the structural damage detection. To bridge this research gap, this study proposes a novel ensemble deep learning model to enhance the temporal feature extraction to improve the damage pattern identification. In this approach, a one-dimensional Convolutional Neural Network (CNN) extracts the spatiotemporal features from the structural vibration measurements. Simultaneously, a SENet attention mechanism is introduced to select the most informatic features. Subsequently, a bidirectional long short-term memory network (BiLSTM) is employed to learn the mapping between the extracted features and the structural damage patterns. Furthermore, the particle swarm optimization (PSO) algorithm is used to optimize the BiLSTM hyperparameters to enhance its stability and reliability. Both simulations and experiments are carried out to collect the vibration responses of the offshore jacket structure in different damage scenarios. The analysis results demonstrate that the proposed method produces remarkable improvement with respect to the accuracy and robustness in identifying the structural damages when compared with the ANNs. The overall detection accuracy of the proposed CNN-BiLSTM-Attention ensemble model is beyond 95%, which provides strong applicability to practical structural health monitoring of offshore platforms.

A comparative analysis of damage identification methods based on multi-channel data fusion and convolutional neural network for frame structures

Vibration-based damage identification methods are crucial for structural health monitoring. In recent years, deep learning methods, such as convolutional neural networks (CNNs), have been widely used for damage identification due to their superior performance. To enhance the accuracy of damage identification for complex frame structures and effectively utilize data measured from multiple sensors, this study proposes and compares four different damage identification methods based on CNN and four different kinds of multi-channel data fusion approaches, i.e., data splicing, matrix reconstruction, data dimension elevation, and sub-model integration. Firstly, the original one-dimensional vibration signals of the frame structures are measured. Then, multi-channel data fusion is performed on the signals to construct datasets used as input to train the CNN model. Finally, the sensitive damage features are automatically extracted using the trained CNN models after parameter tuning, and the structural damage patterns are identified. The effectiveness of these damage identification methods is evaluated using the numerical model of IASC-ASCE Benchmark structure, the Qatar University Grandstand Simulator test, and a three-layer stainless steel frame structure test. The efficiency of damage identification methods using multi-channel data is compared to those using only single-channel data. Additionally, the impact of noise on damage identification accuracy is also analyzed. The results demonstrate that the proposed four different damage identification methods based on CNN and multi-channel data fusion, outperform the methods that solely rely on single-channel data. Furthermore, these methods significantly enhance the accuracy of damage identification.

Mapping progressive damage epicenters in epilepsy with generalized tonic-clonic seizures by causal structural covariance network density (CaSCNd)

AimsMapping progressive patterns of structural damage in epilepsies with idiopathic and secondarily generalized tonic-clonic seizures with causal structural covariance networks and multiple analysis strategies. MethodsPatients with idiopathic generalized tonic-clonic seizures (IGTCS) (n = 114) and secondarily generalized tonic-clonic seizures (SGTCS) (n = 125) were recruited. Morphometric parameter of gray matter volume was analyzed on structural MRI. Structural covariance network based on granger causality analysis (CaSCN) was performed on the cross-sectional morphometric data sorted by disease durations of patients. Seed-based CaSCN analysis was firstly carried out to map the progressive and influential patterns of damage to thalamus-related structures. A novel technique for voxel-based CaSCN density (CaSCNd) analysis was further proposed, enabling for identifying the epicenter of structural brain damage during the disease process. ResultsThe thalamus-associated CaSCNs demonstrated different patterns of progressive damage in two types of generalized tonic-clonic seizures. In IGTCS, the structural damage was predominantly driven from the thalamus, and expanded to the cortex, while in SGTCS, the damage was predominantly driven from the cortex, and expanded to the thalamus through the basal ganglia. CaSCNd analysis revealed that the IGTCS had an out-effect epicenter in the thalamus, whereas the SGTCS had equipotent in- and out-effects in the thalamus, cortex, and basal ganglia. ConclusionCaSCN revealed distinct damage patterns in the two types of GTCS, featuring with measurement of structural brain damage from the accumulating effect over a relatively long time period. Our work provided evidence for understanding network impairment mechanism underlying different GTCSs.

Magnetization transfer ratio of the sciatic nerve differs between patients in type 1 and type 2 diabetes

BackgroundPrevious studies on magnetic resonance neurography (MRN) found different patterns of structural nerve damage in type 1 diabetes (T1D) and type 2 diabetes (T2D). Magnetization transfer ratio (MTR) is a quantitative technique to analyze the macromolecular tissue composition. We compared MTR values of the sciatic nerve in patients with T1D, T2D, and healthy controls (HC).Methods3-T MRN of the right sciatic nerve at thigh level was performed in 14 HC, 10 patients with T1D (3 with diabetic neuropathy), and 28 patients with T2D (10 with diabetic neuropathy). Results were subsequently correlated with clinical and electrophysiological data.ResultsThe sciatic nerve’s MTR was lower in patients with T2D (0.211 ± 0.07, mean ± standard deviation) compared to patients with T1D (T1D 0.285 ± 0.03; p = 0.015) and HC (0.269 ± 0.05; p = 0.039). In patients with T1D, sciatic MTR correlated positively with tibial nerve conduction velocity (NCV; r = 0.71; p = 0.021) and negatively with hemoglobin A1c (r = − 0.63; p < 0.050). In patients with T2D, we found negative correlations of sciatic nerve’s MTR peroneal NCV (r = − 0.44; p = 0.031) which remained significant after partial correlation analysis controlled for age and body mass index (r = 0.51; p = 0.016).ConclusionsLower MTR values of the sciatic nerve in T2D compared to T1D and HC and diametrical correlations of MTR values with NCV in T1D and T2D indicate that there are different macromolecular changes and pathophysiological pathways underlying the development of neuropathic nerve damage in T1D and T2D.Trial registrationhttps://classic.clinicaltrials.gov/ct2/show/NCT03022721. 16 January 2017.Relevance statementMagnetization transfer ratio imaging may serve as a non-invasive imaging method to monitor the diseases progress and to encode the pathophysiology of nerve damage in patients with type 1 and type 2 diabetes.Key points• Magnetization transfer imaging detects distinct macromolecular nerve lesion patterns in diabetes patients.• Magnetization transfer ratio was lower in type 2 diabetes compared to type 1 diabetes.• Different pathophysiological mechanisms drive nerve damage in type 1 and 2 diabetes.Graphical

Dynamic Validation and Assessment of a Historical Bell Tower

The work investigates the dynamic validation and seismic assessment of a bell tower located in Caserta Vecchia, emphasizing its relevance in structural engineering. Given the tower historical significance and possible vulnerability, understanding its structural behaviour under seismic loading is vital for preservation efforts. To capture the bell tower structural intricacies, advanced photogrammetric surveying techniques were utilized. This produced a detailed 3D model of the bell tower, which was subsequently integrated into the Abaqus software for finite element analysis. This approach facilitated a comprehensive examination of the tower structural dynamics and vibrational characteristics. The robustness of the study has been further enhanced through a validation process. Frequency analysis outcomes from Abaqus were cross-referenced with real-world seismic data collected using structural health monitoring (SHM) systems. This seismic data was analyzed using the Frequency Domain Decomposition (FDD) technique. This method ensured that the analytical model results were consistent with observed seismic responses, thereby validating the model’s accuracy and reliability. Further dynamic analyses were conducted to evaluate the bell tower potential vulnerabilities during high-intensity seismic events. These analyses are essential for predicting the structural response and potential damage patterns during earthquakes, facilitating better preparedness and mitigation strategies. The applicability of the methodologies used in similar contexts offers opportunities for advancing heritage preservation and optimizing the seismic safety of such complex structures.

Divergent brain regional atrophy and associated fiber disruption in amnestic and non-amnestic MCI

BackgroundUnderstanding the pathological characteristics of various mild cognitive impairment (MCI) subtypes is crucial for the differential diagnosis of dementia. The purpose of this study was to feature divergent symptom-deficit profiles in amnestic MCI (aMCI) and non-amnestic MCI (naMCI).MethodsT1 and DTI MRI data from a total of 158 older adults with 50 normal controls, 56 aMCI, and 52 naMCI were included. The voxel-wise gray matter volumes and the number of seed-based white matter fiber bundles were compared among these three groups. Furthermore, correlation and mediation analyses between the neuroimaging indices and cognitive measures were performed.ResultsThe aMCI with specific memory abnormalities was characterized by volumetric atrophy of the left hippocampus but not by damage in the linked white matter fiber bundles. Conversely, naMCI was characterized by both the altered volume of the right inferior frontal gyrus and the significant damage to fiber bundles traversing the region in all three directions, not only affecting fibers around the atrophied area but also distant fibers. Mediation analyses of gray matter-white matter-cognition showed that gray matter atrophy affects the number of fiber bundles and further affects attention and executive function. Meanwhile, fiber bundle damage also affects gray matter volume, which further affects visual processing and language.ConclusionsThe divergent structural damage patterns of the MCI subtypes and cognitive dysfunctions highlight the importance of detailed differential diagnoses in the early stages of pathological neurodegenerative diseases to deepen the understanding of dementia subtypes and inform targeted early clinical interventions.

A rate-dependent peridynamic model of reinforced concrete subjected to explosive loading

Explosive loads caused by various factors seriously threaten the security of reinforced concrete (RC) structures. To explore the dynamic damage and failure patterns of RC structures under explosive loads, new rate-dependent peridynamic (PD) models for concrete and reinforcement are proposed. For concrete, the behavior of progressive damage and the compressive deformation limit are also taken into account. The ideal microplastic PD model is adopted for reinforcement. Then, this paper investigated the failure process of RC slabs under close-in explosive loads and obtained crack propagation paths and central bending deflections consistent with the test results. The results manifest the established PD model can correctly calculate the damage process of RC structures under explosive conditions.

Earthquake response and damage patterns assessment of two historical masonry churches with bell tower

The earthquake response of two historical masonry churches is investigated through a finite element macro-modelling approach and non-linear dynamic analyses. The main aim of the study is to evaluate the church walls effects on the seismic performance and damage patterns of slender bell towers, which are analyzed also in the isolated configuration to perform a comprehensive comparison with the confined counterpart. The results of the numerical simulations point out that both the churches, which present different sizes, structural features and dynamic properties, are highly vulnerable to low-to-moderate seismic actions and the bell tower is one of the most critical elements, suffering widespread damage due to also the interactions effects with the neighboring church walls. The presence of shorter adjacent structural elements significantly modifies the main structural dynamic characteristics, seismic response and damage patterns of the bell tower. For both the confined towers, the first cracks originate in the connection regions with the adjacent church walls: on the other hand, for both the isolated towers, the first cracks originate mainly at the base. An advance of damage onset is also detected in the case of confined towers with respect to isolated towers. At the end of the numerical analyses, for both the case studies, the confined configurations show more remarkable damage patterns and larger values of energy density dissipated by tensile damage than the isolated ones, presenting relevant cracks concentrations mostly near the connection regions with the church walls and in the middle-upper part of the bell tower. The results of this study emphasize the paramount importance of taking into account the interactions effects with the surrounding church walls for a proper damage and failure mechanisms assessment of bell towers within the church context: simplified analyses performed on isolated configurations without considering such interaction effects may lead to significant underestimations of damage for the bell tower inside the church.

Surface features of frost damaged concrete: A study using multi-point micro displacement laser ranging method

Surface scaling caused by cyclic freezing and thawing (F-T) is one of the major damage patterns of concrete pavements or hydraulic structures in cold regions, and the damage will be more serious when deicing salts combined with F-T. The weight method has been employed almost exclusively to evaluate the surface scaling of concrete since the 1940s. A multi-point micro displacement laser ranging (MMD-LR) method was proposed to characterize the evolution of frost pits and surface roughness of concrete, quantitatively. The vertical displacement of measurement points on the scaling surface of concrete exposed to F-T cycles with 3 wt% NaCl solution was determined after every four F-T cycles using the MMD-LR method. The traditional weight method was also employed as a baseline to measure the mass of scaling substances. Results show that mean vertical displacement (Dn), surface roughness (Rn) and mass of scaling substances per unit area (wn) of concrete increase with increasing F-T cycles and w/c from 0.30 to 0.60. Relations of wn with Dn and Rn are regressed as power function and quadratic function, respectively, showing the removal order of scaling substances from cement pastes to fine aggregates. The frost pit on the test surface of higher w/c concrete tends to develop the depth with increasing F-T cycles, while it tends to expand the edge in lower w/c concrete. It is verified that the proposed MMD-LR method is credible for inspecting the surface damage of frost concrete, and providing a new perspective for studying the mechanism of concrete surface damage induced by F-T cycles.

High pressure and long-term gas diffusion coupled damage of composites through a multi-physical phase field framework

In extreme service environments, composites are often exposed to high pressure and prolonged chemical penetration, leading to irreversible material damage and structural failure. To address this multi-field coupling problem that has long been a concern, we develop a theoretical framework for fully coupled stress distribution, gas diffusion, heat conduction, and damage evolution of composites to reproduce complex failure modes and disclose the damage mechanisms of composites in complicated environments. To verify the theoretical feasibility, the susceptibility of material fracture properties to hydrogen content is well revealed, and a typical problem, i.e., damage of multi-layered composite hydrogen storage vessels in distinct working stages, is investigated and discussed insightfully. We highlight that (i) the fracture phase field theory is rederived under the thermodynamic consistent framework; (ii) the proposed computational model can accurately capture the heat conduction response under the coupled effects of chemical diffusion and crack propagation; (iii) through a detailed analysis of the energy evolution, the model allows precise yield of the dominant mechanisms in the diverse damage stages of the materials. This study can reproduce the different damage patterns of composite structures under the action of multi-physical fields in extreme environments, thus providing a predictive computational approach to prevent disastrous failures.

Electromechanical phase-field fracture modelling of piezoresistive CNT-based composites

We present a novel computational framework to simulate the electromechanical response of self-sensing carbon nanotube (CNT)-based composites experiencing fracture. The computational framework combines electrical-deformation-fracture finite element modelling with a mixed micromechanics formulation. The latter is used to estimate the constitutive properties of CNT-based composites, including the elastic tensor, fracture energy, electrical conductivity, and linear piezoresistive coefficients. These properties are inputted into a coupled electro-structural finite element model, which simulates the evolution of cracks based upon phase-field fracture. The coupled physical problem is solved in a monolithic manner, exploiting the robustness and efficiency of a quasi-Newton algorithm. 2D and 3D boundary value problems are simulated to illustrate the potential of the modelling framework in assessing the influence of defects on the electromechanical response of meso- and macro-scale smart structures. Case studies aim at shedding light into the interplay between fracture and the electromechanical material response and include parametric analyses, validation against experiments and the simulation of complex cracking conditions (multiple defects, crack merging). The presented numerical results showcase the efficiency and robustness of the computational framework, as well as its ability to model a large variety of structural configurations and damage patterns. The deformation-electrical-fracture finite element code developed is made freely available to download.

Structural covariance in subcortical regions in multiple sclerosis and neuromyelitis optica spectrum disorders: An MRI-based study with automated brain volumetry.

This study aimed to investigate the alterations of brain volumetry and associated structural covariance in subcortical regions in multiple sclerosis (MS) and neuromyelitis optica spectrum disorders (NMOSD). Fourty MS patients, 35 NMOSD patients and 34 healthy controls (HC) underwent 3D T1-weighted image and 3D T2 FLAIR of MRI. The volume differences in subcortical regions were compared between the MS, NMOSD, and HC groups by automated brain volumetry. Structural covariance analysis was performed with each pair of these regions to investigate the alterations of anatomical connections in MS and NMOSD compared to HC. Compared with HC, MS patients presented significantly smaller volume in some subcortical and infratentorial regions (P<0.05), while NMOSD patients showed no significant difference of volumetry in any of the brain regions (P>0.05), although they had no significant difference in disease duration (MS 3.95±3.73 ys; NMOSD 3.11±4.61 ys; P>0.05). In addition, the structural covariance analyses revealed synergic volume alteration in subcortical regions both in the MS and NMOSD groups. More extensive additional connections compared with HC were found in MS patients and more extensive missing connections compared with HC were found in NMOSD patients. This study revealed distinct patterns of brain structural damage and reorganization in MS and NMOSD, which could facilitate a better distinction between these two entities.

Synthetic experimental and numerical investigation on the vertical seismic effect on underground structures

Seismic damage patterns of underground structures indicate unignorable effects of vertical seismic actions. To deal with this, this study distinguished the impact of the vertical seismic effect on the seismic safety of underground structures experimentally and theoretically and provides important shaking table test results for this topic. A typical shallow buried subway station-tunnel junction structure was designed by following the Buckingham similarity law, based on which the shaking table test was performed to investigate the vertical seismic effect on underground structures. The vertical wave propagation laws of the free field, the model structure, and the soil–structure interaction (SSI) system were comparatively analyzed. Numerical simulations considering the soil’s nonlinear behaviors under vertical loads were performed using the equivalent linear analysis method and were verified by the test records. By combining both the experimental and numerical results, it was found that the surrounding soils have very weak constraints on the structural vertical movements. By distinguishing lateral soil pressures from horizontal actions, the vertical seismic action is found to be mainly an inertia force that can stimulate high-frequency vibration modes of the system and cause considerable vertical relative deformations within the underground structure. Consistent with the observed structural damage at the Daikai station, the participation of high-frequency vibration modes is a rational approach that can explain the amplification effect of vertical ground motions on structural seismic responses. In addition, the equivalent linear method is generally satisfactory for simulating the soil compressional nonlinearity.

Pattern Matching-Based Structural Damage Identification Using Mode Shape Difference Ratio with Limited Sensors

In order to rapidly identify structural damage, a pattern matching-based structural damage identification using mode shape difference ratio with limited sensors is proposed. The mode shape difference ratio is proposed as the damage indicator for pattern matching. The natural frequencies and mode shapes of the structure under different damage scenarios are obtained by the numerical simulation with finite element software. Using the simulation results, the damage pattern library based on mode shape difference ratio matrix is constructed. Afterwards, for the real structure to be test, the response power spectrum ratio is calculated with the measured acceleration responses. Then the mode shape difference ratio matrix of real structure is constructed using the power spectrum ratio of the measured responses with limited sensors. With the reciprocal of correlation coefficient difference (ROCCD) used as the pattern matching approach, the actual structural damage pattern is matched with the pattern in the damage pattern library. The pattern in the library corresponding to the maximum value of ROCCD is considered to be the damage condition of the structure. Finally, the numerical simulation and experiment of a four-layer interlayer shear model are investigated. The results show that the proposed method well determines the damage location and severity of the structure, and possesses excellent practicability.

Influence of water level on RC caisson subjected to underwater explosions

With the increasing significance of marine safety issues, the studies on the underwater blast-resistant performance of wharves, e.g., reinforced concrete (RC) caissons, has become an urgent demand. At present, the influence of water level on caisson subjected to underwater explosions was studied experimentally and numerically, e.g., underwater explosion loading characteristics and dynamic behaviors of caisson. Firstly, four shots of underwater explosion test were carried out both in free field and on a partially submerged caisson specimen, and the overpressure- and deflection-time histories, as well as the structural damage pattern were recorded. Then, the underwater explosion loading characteristics including the preceding blast wave and the succeeding bubble oscillations, the cut-off effect of the water surface, as well as the dynamic behaviors of caisson are comprehensively discussed. Furthermore, both the 1D and 3D finite element (FE) models were established, and by adopting Coupled Eulerian-Lagrangian method and remapping technology, the reliability of the FE models and analysis approach were validated by comparing with the test results. Finally, the influence of water level on the dynamic behaviors of caisson against underwater explosions were numerically examined. It derives that, the empirical formula derived based on spherical charge for underwater explosion peak overpressure is applicable to group charge, while the formula for impulse provides higher prediction on group charge; bubble oscillation has great influence on the impulse of underwater explosions; the unsubmerged part of hydraulic structures suffers slighter damage during underwater explosions than submerged part; the upper part of caisson is more sensitive to the change of water level; the partially submerged caisson shows better blast-resistant performance than fully submerged caisson. The present work could provide helpful reference for the studies on the underwater explosion loadings, as well as the blast-resistant assessment and design of hydraulic structures.