To address the insufficient real-time capability and long-horizon accuracy degradation of ship maneuvering motion prediction under environmental disturbances such as waves, an online prediction method based on an improved Long Short-Term Memory (LSTM) neural network is proposed. A multi-layer LSTM is adopted as the core predictor, and an embedded sliding-window structure is introduced to compute the error metrics within the window in real time. When the window-averaged error exceeds a preset threshold, model retraining and updating are triggered, thereby achieving timely online prediction. The results indicate that, compared with offline prediction, the proposed online method maintains stable prediction accuracy under long-horizon conditions with continuously switching wave states. With the same window length, the online method with a stricter threshold achieves a maximum RMSE improvement of 56.85%, while the cumulative update time is only 3.82 s. The proposed online prediction method delivers satisfactory long-horizon prediction performance for ship maneuvering motion and shows practical value for accurate long-horizon prediction under complex sea conditions. Key words:navigation safety; online prediction; long short-term memory neural network; ship maneuvering; wave influence; sliding time window

In response to recent adjustments in the fluvial shoal~channel pattern of the Tongzhou Shoal Reach in the lower Yangtze River caused by upstream reservoir operations and natural evolution, which threaten the stability of the 12.5-m deep-draft channel, this study investigates the characteristics of recent river regime evolution and the corresponding channel response mechanisms based on measured hydrological, sediment, and topographic data from 2018 to 2024. Spatiotemporal comparison, cross-section analysis, and erosion-deposition calculation were employed. The shoal-channel adjustments in the reach are pronounced and exhibit systematic spatial differences. The annual swing amplitude of the thalweg in Nantong Waterway reaches 0.4 km, and its navigation-obstructing shoal undergoes a three-stage dynamic evolution of "downstream incision-disconnection-aggregation," characterized by channel erosion and bar deposition, together with seasonal patterns of flood-season deposition and dry-season erosion. In contrast, the Tongzhou Shoal Waterway is mainly characterized by continuous retreat along the right margin of Xinkaisha and the development of chutes, which drive the entire Kuzigangsha to migrate southeastward and squeeze the navigation channel. The study further quantifies the key regulatory role of hydrodynamic forcing. During high-flow years, enhanced hydrodynamics induce approximately 30% reduction in the shoal area in the Nantong Waterway, improving channel conditions, but simultaneously intensify chute development and sandbody migration in the Tongzhou Shoal Waterway. During low-flow years, shoal deposition intrude into the navigation channel, deteriorating channel conditions, while the Tongzhou Shoal Waterway exhibits localized adjustments. These findings provide critical scientific basis for predicting the evolution of deep-draft channels and for optimizing the design of dredging and regulation projects, thereby establishing an important theoretical foundation for the long-term stability and sustainable management of the channel.

In order to solve the problems of low positioning accuracy, poor anti-interference performance and actuator wear due to the uncertainty of model parameters and unknown time-varying environmental interference in the control of Dynamic Positioning Vessel (DPV), an adaptive backstepping sliding mode control method based on a dynamic event triggering mechanism was proposed. Firstly, sliding mode control is combined with backstepping technology to ensure the robustness of the system to uncertainty and interference. At the same time, the adaptive law is used to estimate the unknown uncertainty term. Based on this, an adaptive backstepping sliding mode control controller is designed. Finally, a dynamic event-triggered mechanism is designed to reduce the frequency of actuator update, thereby reducing unnecessary wear and tear. In order to verify the effectiveness and stability of the proposed control method, the Lyapunov stability theory method is used to prove that all signals in the system are uniformly and ultimately bounded, and Zeno phenomenon can be effectively avoided. The simulation results further verify the superior performance and wide application prospect of the proposed control method in DPV control system.

Under the Carbon Intensity Indicator (CⅡ) rules of the International Maritime Organization (IMO), most theoretical studies manage ship carbon intensity primarily by reducing carbon emissions. However, reducing carbon emissions at the expense of ship transport work no longer aligns with the goal of carbon peaking intensity. Therefore, considering sulfur emission limits, a model was developed to determine whether fuel switching or scrubber retrofitting should be adopted. Combining with carbon intensity management, a decision model for the ship deployment and scheduling problem is proposed, subject to the constraints on sailing speed, fleet deployment, and carbon intensity compliance. To solve the proposed mixed-integer nonlinear programming model, a hybrid algorithm combining linearization and CPLEX is designed. The model is validated using five routes operated by COSCO Shipping. The results show that, compared with the genetic algorithm, the proposed hybrid algorithm increases the solution time slightly by 7.6%, while reducing the operating cost significantly by 33.4%, and all solutions satisfy the engineering constraints. Without carbon intensity management, the carbon intensity of some routes deteriorates to a non-compliant level, which confirms that carbon intensity management can effectively reduce the risk of ship downgrade and service suspension. Based on the above results, two managerial insights are obtained. First, to reduce fleet fuel consumption, liner companies should reduce ship deadweight while still meeting cargo demand, and lower sailing speed within the allowable range. To reduce fleet carbon intensity, besides lowering speed within the allowable range, liner companies should also increase cargo demand to increase ship deadweight. Second, a higher reduction factor imposes stricter carbon intensity requirement. Limited by the minimum and maximum sailing speeds, carbon intensity management requires the deployment of ships with larger deadweight. To avoid carbon intensity non-compliance and excessively low ship loading rate, liner companies should focus on improving transport work by increasing cargo demand.

With the advancement of global ports' green and low-carbon transformation, port microgrids, as key carriers for integrating high-penetration renewable energy, face the challenge of balancing heterogeneous optimization objectives in practical operation. Existing optimal scheduling methods based on the traditional Multi-Objective Particle Swarm Optimization (MOPSO) algorithm often rely on empirically determined conversion coefficients when coordinating economic and energy-consumption objectives. This approach suffers from strong subjectivity and lacks sufficient criteria for screening the Pareto solution set, making it difficult to consistently obtain globally optimal scheduling schemes. To address these issues, this paper proposes a method that introduces Grey Relational Analysis (GRA) into the traditional MOPSO algorithm to evaluate the Pareto solution set and thereby derive the optimal scheduling scheme. First, considering the high penetration of renewable energy and the source-load characteristics of port microgrids, a multi-objective optimization scheduling model is established, aiming to minimize comprehensive operational costs and maximize the local consumption rate of wind and solar power. Second, within the MOPSO framework, GRA is introduced as a decision-making tool to objectively evaluate the Pareto-optimal solution set generated during iterations, thereby accurately selecting the scheduling scheme with the best overall performance. The effectiveness of the proposed algorithm is verified using typical daily measured data from the Chuanshan Port microgrid demonstration project at Ningbo-Zhoushan Port. The results show that, compared to the scheduling algorithm based on traditional MOPSO, the proposed method significantly improves the consumption of renewable energy while maintaining system economic efficiency, achieving a 5.82% increase in the local consumption rate of wind and solar power and an approximately 9% reduction in the system's comprehensive operational costs, providing a feasible technical pathway for the effective utilization of high-density new energy in ports.

Against the backdrop of increasing global supply chain uncertainties, how to enhance the ability of ports to cope with external shocks has become a hot topic in both academia and industry. To this end, this study is based on panel data of 16 listed Chinese port companies from 2004 to 2023. A web crawling technique was used to obtain the text of corporate annual reports. The term frequency-inverse document frequency method was applied to extract the frequency of digitalization-related keywords, so as to quantify the degree of digital technology application. Meanwhile, the sensitivity index method was used to measure the level of port resilience. On this basis, a fixed-effects model was further constructed to empirically examine the empowering effect of digital technology on port resilience and its underlying mechanism. The results show that the application of digital technology significantly improves the resilience of major Chinese ports. For each standard deviation increase in the digital technology level, port resilience increases by about 0. 2 standard deviations. This finding remains valid after a series of robustness tests. Digital technology exerts its effect by strengthening absorptive capacity and adaptive capacity, among which the enhancing effect on adaptive capacity is particularly prominent. However, the path of improving resilience through innovation capacity has not yet emerged. Under the impact of the 2020 global public health event, the empowering effect of digital technology on port resilience was significantly enhanced. In contrast, under the impact of climate change and the 2008 financial crisis, the empowering effect of digital technology on port resilience did not change significantly. These conclusions provide a new perspective for seeking to improve port resilience in the current context of sharply increasing global uncertainties.

With inland waterways transitioning from linear to networked operation, accurately identifying critical segments is essential for optimizing resource allocation and enhancing system resilience. Existing methods have limitations in effectively identifying segments that play a decisive role in maintaining global connectivity. To address this issue, a community bridge-based method is proposed. Firstly, a weighted topological network is constructed using waterway class and length. Then, the Louvain algorithm is applied to divide the inland waterway network into multiple communities with strong internal connectivity, and edges connecting different communities are identified as critical segments. Finally, attack simulation experiments are conducted to evaluate the effectiveness of the proposed method. Taking the Jiangsu inland waterway network as a case study, the results show a maximum modularity of 0. 901, indicating a pronounced community structure characteristics, and the network can be divided into 18 communities. Currently, 46 critical segments are identified in the network. If all critical segments fail simultaneously, both relative network efficiency and the relative size of the largest connected component decrease by nearly 80%, validating the effectiveness of the identification method. After implementing the 2017—2035 and 2023—2035 waterway network upgrades, the community structure becomes more compact, and the number of identified critical segments decreases while the results remain consistent. The identified critical segments provide theoretical support for routine maintenance and safety supervision of inland waterways, strengthening navigational assurance to enhance network resilience.

To enhance the safety and efficiency of maritime route planning, this study proposes a hybrid clustering method integrating multiple algorithms to evaluate the impact of tropical cyclones in the Western Pacific. Firstly, using tropical cyclone data from the National Oceanic and Atmospheric Administration (NOAA), we characterize genesis patterns, intensity variations, and track features of tropical cyclones from 1924 to 2023. Secondly, the K-means clustering algorithm is employed to analyze spatiotemporal distribution, movement speed and direction, while a Gaussian Mixture Model（GMM）is used to identify hotspot regions and high-frequency activity belts. Finally, to overcome the limitations of two-dimensional clustering, we propose a hybrid clustering approach combining 2D and 3D clustering analyses. By overlaying tropical cyclone tracks with major shipping routes, we reveal their potential impacts on maritime safety. Experimental results demonstrate that tropical cyclones primarily affect latitudes between 10°N and 25°N, with seasonal variations:cyclones are generally weaker and more localized in winter-spring, whereas typhoons are stronger and more frequent in summer-autumn. In terms of movement direction, low-latitude cyclones initially move westward from east and then recurve northward near 15°N, while high-latitude cyclones move westward before turning northeastward. The hybrid clustering method effectively identifies tropical cyclone risk zones, providing critical references for shipping route planning and risk management.

With the continued advancement of China "dual-carbon" goals and the increasingly stringent emission reduction regulations of the International Maritime Organization (IMO), the shipping industry is facing more severe emission reduction challenges and urgently needs to clarify the green transition pathways. Most of the existing research focuses on the selection of alternative fuels, while there is relatively little research on emission reduction strategies from the perspective of the fleet. To address the shortcomings of the existing research, this study identifies the key factors influencing fleet green transition decisions. Based on this, a bi-objective linear programming model jointly considering economic and environmental objectives is established, and a genetic algorithm combined with the ε-constraint method is constructed to solve the model. Finally, taking the 10, 000-11, 000 TEU container fleet of COSCO Shipping Group as a case study, this research derives the optimal green transition strategy for the fleet during the planning horizon, encompassing fuel choices and operational configurations for individual vessels. Sensitivity analysis reveals that fluctuations in fuel prices significantly affect the selection of engine types during vessel retrofitting and renewal decisions, while the stringency of emission reduction targets directly influences fleet transition costs, thereby affecting corporate proactiveness in pursuing decarbonization initiatives. Consequently, policymakers should establish appropriately calibrated emission reduction targets and incentive mechanisms to accelerate the advancement of low-carbon technologies, reduce fleet transition costs, and expedite the achievement of decarbonization objectives in the shipping industry.

During the reconstruction of the ship lock, excavation of the rock masses on both sides is often required. In current practice, slope instability is commonly judged using indicators such as displacement, the extent of the plastic zone, and the number of numerical iterations. These criteria typically rely on manual interpretation, which can be subjective and may not clearly determine whether slope failure has occurred. Thus, this paper proposes a slope stability analysis method based on a kinetic-energy evolution criterion, and applies it to the reconstructed ship-lock slope at Baishi Hydropower Station. A representative excavation cross-section at the ship lock is selected and the strength reduction method is adopted to analyze the kinetic energy evolution of the slope soil under different reduction coefficients, thereby identifying the occurrence of instability and failure. The proposed criterion is further compared with the conventional static assessment method based on displacement. The results show that the dynamic analysis method based on the kinetic energy evolution can obtain an accurate slope safety factor. Compared with the traditional displacement judgment method, the method proposed in this paper provides a more objective and explicit identification of slope instability which has greater advantages. The findings of this paper provide practical value and guidance for the high rock slope engineering, which can serve as useful reference for professionals in related fields.