To address the high safety risks and frequent accidents associated with oil tanker loading and unloading operations, this paper proposes a data-driven risk assessment method based on Bayesian networks. Guided by systems engineering theory, a three-layer Bayesian network evaluation model comprising 34 nodes is constructed. Using the inference principle of the expectation-maximization algorithm, the conditional probabilities of the network nodes are computed to quantify risk levels within the model. The rationality and reliability of the model are verified through sensitivity and effectiveness analyses. Validation using data from 20 actual tankers demonstrates that the model's output aligns with risk levels assessed by port security personnel and can accurately evaluate the risks during oil tanker loading and unloading operations. The proposed model and method are applicable for assessing safety risk levels in oil tanker operations and can serve as a reference for safety evaluations of loading and unloading operations for other types of dangerous goods carriers.

In accordance with the MARPOL Annex Ⅵ Nitrogen Oxide (NO_x) Technical Code issued by the International Maritime Organization (IMO), this paper proposes a simplified on-board test scheme for measuring NO_x emissions from marine diesel engines. The scheme employs the carbon balance method to calculate exhaust flow, requiring only the measurement of NO_x and CO₂ concentrations in diesel exhaust at selected load points, and adopts the default fuel element content values provided in the IMO Technical Code. The results show that the weighted average NO_x deviation of the simplified method is approximately 2% compared with the standard cycle, while also reducing the complexity of emission testing equipment and shortening the testing process. The findings provide a theoretical basis for streamlining NO_x emission on-board tests.

Complex intersection waters, characterized by heavy ship traffic and frequent route crossings, exhibit complex traffic dynamics and a high risk of collisions, making route optimization in such areas highly important. Traditional route optimization methods tend to focus more on the formulation of traffic rules and traffic control measures. While effective, these approaches often rely on the subjective experience of maritime managers and lack an objective basis. To address these limitations, this paper proposes a route optimization method based on historical ship trajectory data. The ship traffic network is extracted through trajectory clustering and image processing techniques. A node similarity model is constructed, and a clustering algorithm is applied to partition the overall network into multiple local traffic networks. Route optimization is then achieved by merging nodes and reconstructing the network within each community. Experimental results demonstrate that the proposed method reduces the complexity of ship traffic by 50% and the risk of ship collisions by 30% compared to pre-optimized conditions. These improvements significantly enhance navigation safety, alleviate the regulatory burden on maritime authorities, and provide valuable insights for the planning of ship routing systems.

To accurately predict the thermal fire hazard of these containers under different stowage methods on a ship, this paper establishes typical working conditions, such as the position of the fire container, stowage height and wind speed. It then simulates the fire scene of lithium battery containers on the deck using FLACS 10.9 and establishes a prediction model for the temperature and heat flow density of the fire flow field using the CatBoost algorithm. The results demonstrate that the air volume within the upper and lower spaces of the lithium battery container is directly proportional to the change in fire temperature. The maximum temperature occurs when the fire layer is in the 7th layer, and the range of damage caused by high temperatures and heat flux is minimised when the fire layer is between the 7th and 8th layers. Increasing the stowage height decreases airflow, resulting in higher maximum fire temperatures, a larger temperature influence range and a longer vertical diffusion distance of heat flux. When the wind speed is in the range of 1-4 m/s, it helps to dissipate heat and reduce the maximum temperature. However, when the wind speed reaches 5 m/s, the oxygen uptake rate of the flame increases, resulting in a higher maximum temperature. When the wind speed reaches 6 m/s, the heat dissipation effect dominates and the maximum fire temperature decreases again. The higher the wind speed, the smaller the area of damage caused by high temperatures and heat flux. Comparing the temperature and heat flux density values predicted by the CatBoost algorithm with the measured samples shows that the model is highly accurate and can identify overheating spots. These research results can inform the determination of lithium battery container accumulation modes and the corresponding fire monitoring.

Reasonable maritime route planning contributes to enhancing both the safety and economic efficiency of ship navigation. To address the challenges associated with route planning in complex waters, this paper proposes a method for extracting maritime traffic routes based on ship behavior patterns. Using Automatic Identification System (AIS) data, characteristic points of ship behavior patterns are identified through threshold judgment and a sliding window approach. A clustering algorithm is then applied to determine centroid points within each cluster, which represent the distribution of point sets. Finally, connection rules are established to sequentially link these centroids, thereby generating maritime traffic routes. Experimental analysis was conducted using AIS data from the Beibu Gulf waters. The results indicate that the routes extracted by this method align closely with the recommended routes published in the Navigation Guide for Ships in the Guangxi Area of the Beibu Gulf.

Ship trajectory prediction has become increasingly important in marine traffic management, shipping safety, and related fields. Current methods for ship time-series prediction exhibit certain limitations when handling multi-feature data inputs in water traffic scenarios, as they fail to adequately capture the correlations among features or focus on the critical information within time-series data. To address these shortcomings and further improve the accuracy of ship trajectory prediction, this study proposes a method named GCAU, which integrates an improved Graph Convolutional Network (GCN) with a Recurrent Attention Unit (RAU). First, Graph Convolutional Networks are employed to capture interdependencies between features, thereby enhancing the model's capability to extract feature correlations. Second, an attention gate is incorporated into the Recurrent Attention Unit (RAU), enabling selective emphasis on time-level features. Finally, the study evaluates four different ship time-series prediction methods across three distinct scenarios. The results demonstrate that GCAU outperforms the other methods in all tested scenarios, achieving lower values in Mean Squared Error (MSE), Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE), and Mean Absolute Error (MAE). The proposed method can effectively enhance the accuracy and stability of ship trajectory prediction, thereby providing more reliable decision support for maritime traffic management and other related applications.

Ship trajectory prediction and behavior recognition can help effectively assess navigational risks and provide an important basis for decision-making in collision avoidance and traffic management. To improve the accuracy of ship trajectory prediction and behavior recognition, this paper studies a multi-task Informer model for simultaneous trajectory prediction and behavior recognition. Based on the Informer framework, the model incorporates a multi-task learning approach. It addresses the issue that inaccurate ship behavior records in AIS data cannot be directly used as model inputs by designing a multi-task loss function that jointly trains behavior recognition and trajectory prediction in parallel. During training, an adaptive updating strategy for the loss function-based on homoscedastic uncertainty-is designed to automatically allocate weights to the losses of the two tasks. Evaluated using real AIS data from the Taicang sector waters, the multi-task Informer model reduces trajectory prediction loss by 40.2% and 14.7% compared to LSTM and Informer models, respectively. In behavior recognition, the multi-task model improves accuracy by 11.7% and 5.95% compared to LSTM and Informer models, respectively. The results demonstrate that the multi-task model effectively enhances the performance of ship trajectory prediction while achieving accurate recognition of ship behavior.

This study aims to apply deep reinforcement learning to address the challenges of trajectory planning and control for unmanned surface vehicles. In trajectory planning, the Q-learning algorithm is employed to generate trajectories in real-world aquatic environments. For the design of the reward function, factors such as shallow water areas are taken into account, with an emphasis on minimizing the number of turning points along the path. For trajectory tracking control, we integrate the Soft Actor-Critic (SAC) algorithm with the Proportional-Integral-Derivative (PID) control method to alleviate the difficulties of manual parameter tuning associated with conventional PID controllers. This hybrid approach also mitigates the interpretability limitations often found in pure deep reinforcement learning methods. Comparative experiments involving the traditional PID algorithm, Genetic Algorithm (GA), and Deep Deterministic Policy Gradient (DDPG) algorithm demonstrate the superiority of the proposed SAC-PID method. Simulation results show that the planned trajectories effectively incorporate multiple factors, including travel distance, shallow water regions, and number of turning point, the SAC-PID method achieves outstanding performance in trajectory tracking.

In order to reduce the emission of polluting gases from auxiliary generators during ship berthing, and in response to the growing adoption of shore power infrastructure, this paper incorporates the distribution of shore power into the berth scheduling plan of container terminals. Building upon the traditional berth scheduling model, relevant constraints for shore power allocation and carbon emission reduction targets are introduced, establishing a mathematical model that integrates ship-in-port activities, shore power usage, and carbon emissions. To solve the model effectively, an Improved Bat Algorithm (IBA) incorporating a stagnation mutation strategy is proposed. The inertia weight method is employed to update individual optimization speeds, preventing the algorithm from converging to local optima. Case studies show that considering shore power distribution increases the complexity of the berth scheduling problem. When the number of ships does not exceed 25, the mathematical model can be solved accurately with optimal solution quality; however, when the number increases to 30, the model cannot be solved within a reasonable time frame. In comparison, the IBA achieves efficient solutions for all test cases with significantly shorter computation times. The maximum deviation between IBA results and the exact model solutions is only 2.37%. Furthermore, compared to traditional Genetic Algorithms and the basic Bat Algorithm, IBA demonstrates superior performance in terms of solution quality and computational efficiency, with an average increase in computation time of only about 10 seconds compared to the basic bat algorithm. A matching analysis between the shore power retrofit ratio of berths and ships revealed that under a fixed dock berth retrofit ratio, terminal costs decrease as the ship retrofit ratio increases. However, once the two ratios reach equilibrium, the rate of cost reduction levels off and remains largely stable. These results indicate that optimal cost savings are achieved when the shore power retrofit ratios of berths and ships are appropriately matched. Ensuring a balance between supply and demand can effectively prevent resource waste and enhance the efficiency of shore power utilization.

In view of the slow response speed of fuel cells, which limits their ability to promptly respond to dynamic power loads, a composite energy storage power supply is employed to address this issue. Using wavelet transform technology, the steady component of the load is allocated to the fuel cell, while the fluctuating portion is assigned to the composite power supply. Based on Pontryagin's minimum principle, an energy management strategy is formulated with the supercapacitor's energy as the state variable, the output power of the lithium battery as the control variable, and the root mean square current of the lithium battery as the cost function. A simulation model of the ship power system is built in Matlab/Simulink to validate the proposed energy management strategy. The results demonstrate that the proposed control strategy enables stable output power from the fuel cell and achieves rational power distribution according to the charge-discharge characteristics, capacity, and current state of charge of both the supercapacitor and the lithium battery. Compared to hybrid ships without supercapacitors and traditional fixed filter strategies, the proposed approach reduces the rate of current change in the lithium battery and extends the service life of the fuel cell and lithium battery.