The ultimate strength of vessels under combined loads is increasingly highlighted. Physical scale model testing remains an indispensable and critical way for evaluating and studying the ultimate strength of ship hull structures. The similarity method is essentially important in guiding such tests. Large deck openings on the main stream carriers, like container carriers and bulk carriers, affect the ultimate strength and collapse mode greatly. However, studies on the influence of the large deck opening itself on the ultimate strength and failure modes remain insufficient, leaving a lack of similarity criteria to account for the effects of large deck openings. In this paper, the effects of deck openings on the ultimate strength and failure modes are systematically investigated through numerical analysis. Based on the findings, similarity criteria for the effects of openings are proposed, along with the design procedure for the deck openings in scale models. Subsequently, a buckling strength driven similarity method for box girders is proposed, incorporating the proposed criteria for the opening region. To enable experimental validation, an eccentric four-point bending loading scheme is proposed to apply bending moment and torque simultaneously. The proposed method is verified through various numerical examples and physical model tests.

Ship structural failure mode classification still relies heavily on subjective expert judgment, which is time-consuming and may introduce uncertainty in safety assessment. Although deep learning provides a promising avenue for automation, many existing learning approaches rely on 2D image representations and may therefore suffer from geometric occlusion and information loss when projecting complex 3D stiffened structures. To address these challenges, we propose a Physics-Guided Graph Convolutional Network (PGGCN) for failure mode classification. Specifically, our method models finite-element (FE) meshes directly as graphs, preserving the holistic topology and displacement-field fidelity without viewpoint dependency. We further incorporate domain knowledge through a hybrid strategy: a Deep Graph Convolutional Network (DeepGCN) first detects local component buckling states such as plate or web buckling, and a logic matrix derived from classical failure definitions subsequently determines panel-level failure modes. To enable systematic evaluation, we construct a dataset spanning diverse stiffened-panel geometries via Latin Hypercube Sampling. Progressive analysis states from each loading case are organized into task-specific graph samples for supervised learning. Experiments on the test set achieve accuracies of 95.48% and 91.42% for plate- and web-buckling classification, respectively, and 89.56% for panel-level failure mode discrimination. These results demonstrate that the proposed method provides an interpretable framework for automated failure mode classification from FE meshes in ship stiffened panels.

This study explores the fatigue behavior of prestressed CFRP-reinforced concrete (RC) beams in subtropical environments, addressing the gap in understanding their long-term performance under natural exposure. Five specimens were subjected to one year of natural exposure under sustained load conditions before fatigue testing. During the exposure period, the strain on the CFRP laminate was monitored, and a decrease of 5.37% in prestress loss after exposing 300 days was observed. The digital image correlation (DIC) method was employed to capture the crack initiation and growth during the fatigue tests, with detailed analyses of crack growth rates, deflection evolution, and failure modes. Experimental results indicated that natural exposure accelerates fatigue crack growth compared to the results of the unexposed RC beams tested by this group. Moreover, the finite element model (FEM) was developed to account for CFRP-concrete interface degradation due to natural exposure and material nonlinearities, allowing for the determination of the J-integral of the main crack. Fatigue life prediction models, based on experimentally obtained crack growth rates and stress intensity factors (SIF) calculated using the J-integral, were derived. The predicted fatigue lives were within a 20% error margin of the experimental results, demonstrating the reliability of the proposed models.

This study presents a novel fatigue design method to ensure that externally bonded fiber-reinforced polymerstrengthened reinforced concrete (RC) beams achieve infinite fatigue life (>2 million cycles). It utilizes smallsample machine learning (ML) techniques to develop a classification model that predicts whether a sample's fatigue life surpasses this threshold, emphasizing the importance of multiple variables beyond just stress or strain ranges applied to steel reinforcement and considering potential failure modes. A classification probability-stress level (P -S) curve, developed using the ML model, was constructed to estimate fatigue limits, enhancing model reliability with an average relative error of 6.05 %. Based on virtual data analysis, a new fatigue design equation was proposed, offering greater accuracy for infinite life design compared to existing guidelines. This improved accuracy stems from its foundation in strengthening beam-specific data and consideration of potential stress transfer issues between FRP and steel reinforcement due to FRP-concrete bond interface degradation. To enhance design reliability, it is recommended to use the new equation alongside existing ones. This approach provides a more comprehensive and accurate method for designing FRP-strengthened RC beams for infinite fatigue life.

To investigate the impact of frost-layer formation on the heat transfer characteristics of cryogenic valves in LNG vessels, this study derived the temporal variation in the valve’s temperature field based on the thermodynamic characteristic parameters of the low-temperature valve. Additionally, a frost formation model was developed for the drip tray. This considered the physical model characteristics of the tray, and the frost thickness was calculated for different times. The morphology of the calculated frost layer was coupled with the low-temperature valve model for heat transfer calculations to explore the influence of the frost layer on valve heat transfer characteristics. The results show that, in the initial stage of frost formation, the frost layer acts similarly to a finned heat sink, enhancing the thermal exchange efficiency at the surface of the drip tray, which results in a temperature increase in the drip tray and stuffing box compared to the frost-free condition. However, as the frost layer grows on the surface of the drip tray, the surface heat transfer resistance increases, gradually diminishing the enhancing effect of the frost layer on the heat dissipation of the drip tray. The results validate the dual role of the frost layer in the heat transfer process of low-temperature valves, providing important insights for the design and optimization of such valves.

Corrosion fatigue damage significantly affects the long-term service of marine platforms such as propellers. Fatigue testing of pre-corrosion specimens is essential for understanding damage mechanisms and accurately predicting fatigue life. However, traditional seawater-based tests are time-consuming and yield inconsistent results, making them unsuitable for rapid evaluation of newly developed equipment. This study proposes an accelerated corrosion testing method for ZCuAl8Mn13Fe3Ni2 nickel–aluminum bronze, simulating the marine full immersion zone by increasing temperature, adding H2O2, reducing the solution pH, and preparing the special solution. Coupled with the fatigue test of pre-corrosion specimens, the corrosion damage characteristics and their influence on fatigue performance were analyzed. A numerical simulation method was developed to predict the fatigue life of pre-corrosion specimens, showing an average error of 13.82%. The S–N curves under different pre-corrosion cycles were also established. The research results show that using the test solution of 0.6 mol/L NaCl + 0.1 mol/L H3PO4-NaH2PO4 buffer solution + 1.0 mol/L H2O2 + 0.1 mL/500 mL concentrated hydrochloric acid for corrosion acceleration testing shows good corrosion acceleration. Moreover, the test methods ensure accuracy and reliability of the fatigue behavior evaluation of pre-corrosion specimens of the structure under actual service environments, offering a robust foundation for the material selection, corrosion resistance evaluation, and fatigue life prediction of marine structural components.

The inverse Finite Element Method (iFEM) is crucial in structural health monitoring and virtual-real integrated testing, as it allows for the reconstruction of structural deformation from discrete strain data. This study introduces two iFEM approaches specifically designed to handle substantial deformations and nonlinear strains effectively. According to Green-Lagrange strain, four-node inverse shell element is employed to accurately reconstruct large structural deformations. The development of the nonlinear iFEM equilibrium equations is thoroughly derived using both incremental and iterative methods. Furthermore, an analysis is conducted on the results and convergence processes of these two methodologies. Ultimately, the two nonlinear iFEM approaches are applied to the deformation reconstruction of stiffened plates and box girders to evaluate their precision in deformation reconstruction. The findings reveal that for stiffened plates, both the incremental and iterative methods achieve reconstruction variations within 4 % during both linear and nonlinear phases. With respect to the box girder, reconstruction discrepancies between the two nonlinear iFEM approaches remain within 10 % before reaching the ultimate state, and the deformation contours reconstructed by nonlinear iFEM closely align with those computed by nonlinear FEM.

Laser Shock Peening (LSP) is a highly effective surface enhancement method that generates deep compressive residual stresses, which can increase the fatigue life and mechanical strength of metals. One of the challenges in LSP is the presence of tensile stresses in overlapping areas of the laser shock. This study aims to identify laser processing parameters that minimise tensile stresses in the overlapping areas while maximising the desirable compressive residual stresses. A numerical approach is adopted by coupling Finite Element Analysis (FEA) with a data-driven predictive model based on Artificial Neural Networks (ANN). The finite element simulations capture the time-dependent response of Al 6061-T6 alloy to laser shocks and serve as the training database. When compared against validation data, the ANN model yielded high predictive accuracy with performance criteria of R2 = 0.988, MAE = 1.058, and RMSE = 1.786, demonstrating potential for good generalisation and reliability. The ANN model predicted a maximum compressive residual stress of - 224.3 MPa in the shocked area and was able to predict both uniform and non-uniform stress distributions successfully. At a laser frequency of 1 Hz, the maximum process efficiency achieved was 3.24 mm2/s. The results show that using FEA coupled with ANN modelling provides a better cost-effective way to develop and better control residual stress profiles on surfaces after laser shock peening.

For the fatigue life prediction of titanium alloy welded structures used in deep-sea submersibles, a modified model considering the welding characteristics and the load sequence effect is established in this paper. Based on the modified model, the fatigue crack propagation behaviour of the new titanium alloy welded structure under different loading conditions is predicted. The comparison between the prediction results and the test data is analyzed. The prediction ability of the modified model over the conventional model is comparatively analyzed. The results show that the modified model has a strong ability to predict the fatigue crack propagation behaviour of new titanium alloy welded structures under various load spectra.

The sealing performance of subsea connectors plays a vital role in the safe operation of subsea oil and gas production systems, which is directly determined by compression deformations of the gasket of the subsea connector. Therefore, the changing rules of the sealing structural deformations under various axial loads are key problems when designing subsea connector products. In this paper, experimental measurements are performed in order to investigate the structural stresses of the sealing structure of the subsea connector, i.e. the hub and the gasket. As for the real products of the hub and gasket structures, test solutions are designed. Structural stress components of some representative points are observed when the axial preloads are loaded. Test results show the maximum tensile stress of the outer wall is more than that of the inner wall, which indicates the weak point of the hub is on the outer wall surface nearby the joint region between the hub's shoulder and the thick-walled cylinder. The maximum tensile stress of the outer wall is less than that of the inner wall, which indicates the weak point of the hub is on the inner wall surface.

The 980 high-strength steel is widely used in marine structures due to its superior mechanical properties. The damage evolution of 980 high-strength steel caused by fatigue loadings should be quantified to ensure safety of marine equipment. However, fatigue crack growth (FCG) prediction for curved cracks usually shows significant deviations, owing to the incompatible 2D FCG model for the 3D cracks. Here the FCG of the 980 high-strength steel of standard specimens with straight-through cracks are tested, based which a unified 3D FCG model for curved cracks is established by combining the equivalent thickness concept with the out-of-plane constraint theory. Then the FCG for typical curved cracks are predicted using the proposed unified model, showing high agreement with tests in both growth life and crack tracks. The lagging phenomenon near the surface for the curved cracks in tests is predicted in high consistency based on the proposed 3D unified model, which is not captured by traditional 2D crack growth model. This work provides feasible way for FCG prediction of 980 highstrength steel with curved cracks based on standard tests, as well as methodology of damage tolerance design in marine structures.

Laser Shock Peening (LSP) is an advanced technique for enhancing surface properties, drawing significant interest for its ability to induce beneficial residual stresses in materials. Traditional LSP design processes, reliant on manual parameter selection, often result in imprecise control over the stress distribution, necessitating multiple iterations and high costs. This study introduces a machine learning (ML)-based approach, utilizing the Random Forest (RF) algorithm, to automate and optimize the design of LSP parameters for nickel-aluminium bronze surfaces. Our findings demonstrate the RF model’s capability to accurately predict and optimize residual stress distributions, achieving compressive stresses up to 472 MPa with a notable reduction in design iterations. The model forecasts both uniform and non-uniform stress patterns, particularly identifying areas susceptible to Residual Stress Holes (RSH) with improved precision. With an Absolute Percentage Error (APE) of only 6.2 %, our approach significantly outperforms traditional ML algorithms, offering a novel method for efficiently designing complex residual stress fields in LSP applications.

The corrosion-fatigue test is widely used to assess marine equipment. However, due to the low-frequency and high-cycle loading of marine equipment in service, conventional test methods require long cycle times and high costs. This makes it challenging to meet the rapid evaluation demands associated with equipment development effectively. In this study, an accelerated corrosion test method was proposed for steel by adjusting the composition, temperature, pH and H2O2 concentration of environment. Fatigue crack growth tests were conducted under various corrosion environments and load frequencies. The matching relationships between corrosion fatigue acceleration multiplier, corrosion environment parameters and load parameters were established. Additionally, a corrosion fatigue acceleration test method was proposed based on the equivalent principle of fatigue crack growth. The results show that effective coupling between corrosion and fatigue can be achieved by gradually reducing the load frequency as the stress intensity factor amplitude increases throughout the crack growth process. It is consistent with the trend of fatigue crack growth rate in the seawater environment. The 4.85% fatigue life deviation and the 15.68 times corrosion fatigue acceleration can be achieved by our method. This work improves the accuracy and reliability of fatigue performance evaluation for offshore equipment.

In order to study repair effect of carbon fiber composites on the ultimate strength of pitting reinforcement plates, feasibility of the numerical simulation method is firstly verified by comparing the numerical simulation with the test results of repairing cracked steel plates with carbon fiber composite.Considering loading characteristics of the reinforced plate unit, ultimate load carrying capacity of the CFRP-repaired pitting reinforced plate was investigated by uniaxial compression, biaxial compression, lateral load and combined load（biaxial compression, lateral load） respectively. The results show that ultimate strength of the repaired pitting reinforced plate can be effectively improved.

本文针对海洋结构物用的一种新型钛合金材料,开展疲劳裂纹扩展速率和复杂载荷谱下保载-疲劳裂纹扩展速率试验研究,综合探讨上峰值保载时间和下峰值保载时间对该新型钛合金保载-疲劳裂纹扩展行为的影响规律;基于断裂力学理论,建立新型钛合金保载-疲劳裂纹扩展速率预报模型,采用试验方法验证该模型对新型钛合金保载-疲劳裂纹扩展行为的预报能力.研究结果表明:新型钛合金材料疲劳裂纹扩展行为对上、下峰值保载时间较敏感,随着上峰值和下峰值保载时间的增加,裂纹扩展速率增加,但是裂纹扩展速率增加的趋势逐渐减小,即上、下保载时间对新型钛合金材料裂纹扩展速率的影响具有一定的饱和值.提出了新型钛合金保载-疲劳裂纹扩展速率预报模型,对复杂载荷谱下的保载-疲劳裂纹扩展速率进行了预报研究,预报结果与试验结果吻合较好.

Marine flow-passing components are susceptible to cavitation erosion (CE), and researchers have worked to find ways to reduce its effects. Laser Shock Peening (LSP), a material strengthening method, has been widely used in aerospace and other cutting-edge fields. In recent years, LSP has been used in cavitation resistance research. However, the current LSP research does not realize a comprehensive predictive assessment of the material's CE resistance. This paper uses m stresses to develop a comprehensive set of strengthening effect prediction models from LSP to CE using finite element analysis (FEA). Results show that the LSP-1 sample (4 mm spot, 10 J energy) introduced a compressive residual stress value of 37.4 MPa, better than that of 16.6 MPa with the LSP-2 sample (6 mm spot, 10 J energy), which is generally consistent with the experimental findings; the model predicts a 16.35% improvement in the resistance of LSP-1 sample to water jet damage, which is comparable to the experimental result of 14.02%; additionally, interactions between micro-jets do not predominate the cavitation erosion process and the final CE effect of the material is mainly due to the accumulation of jet-material interaction.

Origami offers a novel design possibility for periodic sandwich cores. The curved-crease origami is adopted to enhance the anti-buckling performance and reduce the abrupt change of the carbon fibers in the origami sandwich cores. CFRP curved-crease origami sandwich structures were fabricated using a hot press molding method. Analytical models were developed for predicting the shear stiffness and strength concerning different failure modes including debonding, compressive and shear buckling, tensile, compressive and shear fracture in both transverse and longitudinal directions. The shear deformation histories, failure modes, strength and stiffness of the sandwich structures were investigated through experiments and finite element analysis (FEA). The analytical models, experiments, and simulations agree with each other reasonably. The origami sandwich undergoes buckling failure when the wall thickness of the origami core is sufficiently thin. With the increase of the wall thickness, debonding between the face sheets and the origami cores usually occurs instead of fracture. The exploration of the shear behaviors lays the foundation for the applications of the origami sandwich structures in the lightweight construction fields.

The Cu-Mn based spinel coatings without Cr show good application prospects due to their high conductivity and thermal expansion coefficient matching with metallic interconnectors. Based on the Cu-Mn-O phase diagram, the CuMn alloy layers with different Cu/Mn ratios were designed by electro-deposition and high energy micro-arc alloying (HEMAA) processes. The Mn-33Cu-17Co layer was prepared by HEMAA for thermal growth of the CuMn spinel oxides, and the fine-grained 430 layer was prepared by HEMAA as an interlayer at the interface between the 430 SS substrate and the coating. The microstructure and composition of the composite layers were analyzed. The results show that after oxidation treatment at 750 degrees C for 500 h, the oxide products of the Mn-33Cu17Co layer were mainly cubic spinel Cu1.4Mn1.6O4 and CuO, which were different from those of the Mn-35Cu oxide layer, spinel Cu1.2Mn1.8O4 and Mn2O3. The aggregation of Cu-rich oxides was not found in the crosssection and on the top surface of the Mn-33Cu-17Co oxide layer, and the outer spinel layer was uniform and compact. The designed Mn-33Cu-17Co layers, Mn-35Cu layers and Mn1.5Cu1.5 layers showed good high temperature oxidation resistance and electrical conductivity.