In order to enhance the perception ability of ships to the surrounding environment under poor visibility conditions such as foggy weather, a maritime image dehazing method based on the cycle generative adversarial network is introduced. It enhances the stability of dehazing effects, avoiding excessive image deformation and the loss of important features by introducing cycle consistency loss and identity loss. Experimental validation is conducted on both simulated and real foggy maritime images, and then compared with mainstream dehazing methods. The results demonstrates that our method outperforms others in terms ofPSNR (Peak Signal-to-Noise Ratio), structural similarity, and color difference. An average PSNR of 21.92 dB, surpassing traditional dehazing methods by at least 8.79 dB, is achieved. It is seen that the fog is effectively removed from images, and the details and texture information are restored.

Based on the guidelines for compiling air pollutant and greenhouse gas emission inventories, a comprehensive emission inventory for typical coastal container terminals is established. By integrating on-site investigation data, emission factors, and activity level data, emissions during four operational stages (quayside loading/unloading, yard loading/unloading, horizontal transportation, auxiliary and ancillary production) in 2023 are estimated. The Monte Carlo simulation is performed to quantify uncertainties. The research shows that NO_x and CO are the main air pollutants at the terminals, with container trucks and stackers as the primary sources of air pollutants. Indirect emissions from electricity consumption contribute the most to carbon emissions, with container trucks, yard cranes, and shore bridge being the main sources, accounting for 25.57%, 23.78%, and 23.43% respectively. The uncertainty of the inventory ranges from -67% to 139%. This comprehensive inventory reflects the characteristics of air pollutant and carbon dioxide emissions at coastal container terminals, providing scientific support for air pollution control and carbon emission reduction.

A novel tail fin propulsion mechanism is developed for large-scale BCF (Body and Caudal Fin) biomimetic fish. This mechanism employs the crank-rocker principle to convert steady motor rotation into sinusoidal oscillations, enabling stable tail fin undulation and precise dual-joint angular adjustments through a single power source. This mechanism is characterized by compact structure and simplified control. The prototype of the large biomimetic fish adopts a metal skeleton-skin configuration, effectively resolving assembly precision issues in large-scale motion systems while enhancing mechanical reliability. Self-propulsion tests in a towing tank are conducted to measure swimming speeds by tail fin oscillations, validating the propulsion mechanism's efficacy. The achievement provides technical support for applications of large biomimetic fish in marine monitoring and underwater rescue.

An offset pre-swirl energy-saving device to improve ship propeller's propulsion efficiency is developed and evaluated usingCFD (Computational Fluid Dynamics). For illustration, the device particularly for a new bulk carrier is designed. TheRANS (Reynolds-Averaged Navier-Stokes) method is employed to numerically calculate the resistance on the ship and the wake field at the propeller disk plane. Based on the wake field characteristics, the offset pre-swirl energy-saving device is designed, and its key design parameters are optimized for efficiency of the wake flow rotational energy recovery. Numerical calculation shows that this device can lead to an energy-saving of 3.1%, approximately 20% improvement over conventional pre-swirl energy-saving devices.

In order to augment the traffic perception capability of the ferry terminal and improve the overall digitalization level, a Transformer-MLP(Multi Layer Perceptron)-based method for short-term prediction of ferry terminal passenger flow is introduced, The design is capable of predicting the passenger flow in 4 predefined time periods of following day. The prediction method performs ETL (Extract-Transform-Load) and get passenger flow data in the form of time series from ferry business data. A combined time series model PatchTSTiDE is built based on the collected time series data transformed from ferry business data by means of the time series forecasting toolkit NeuralForecast. Tests show that, Compared with TiDE (Time-Series Dense Encoder) and PatchTST, the developed model is about 4.44% and 30.59% better in MAE (Mean Absolute Error) respectively.

The environmental impact assessment project for reconstruction and expansion of expressway in Guangdong Province is presented. The acoustic environment quality of sensitive buildings along the expressway was measured and the noise impact after highway expansion was predicted. The noise reduction effects of four different forms of sound barriers, including vertical sound barriers, bent sound barriers, semi closed sound barriers, and fully closed sound barriers were analyzed. Low noise pavement, and speed limit setting were investigated. Based on the requirements of relevant laws, regulations, and technical policies, a comprehensive noise reduction plan was devised for the entire area affected by the project, including speed limit setting, low noise pavement, and multiple forms of sound barriers. This paper also suggests that big data technology should be introduced to highway noise prevention design to address the problems of updating lag and narrow valid parameter value of the current formula for predicting the intensity of expressway noise sources; Comprehensive information platform should be developed to collect wider range of data, such as road traffic flow, vehicle type ratio, and driving speed of each vehicle type and support the environment projects.

Drawbacks of traditional soil and water conservation monitoring methods, such as low efficiency, terrain limitations, and incomplete data integrity is overcome by introduction of UAV (Unmanned Aerial Vehicle) in the comprehensive soil erosion control project in Qingshan Small Watershed. By processing UAV imagery to generate high-precision DOM (Digital Orthophoto Maps) and DSM ( Digital Surface Models), geographic information such as bare land area, vegetation coverage, slope length, slope width, and slope gradient were automatically extracted. Using the acquired slope length and gradient data, the RUSLE (Revised Universal Soil Loss Equation) was applied to simulate soil loss. A comparison between simulated and field-measured data from six monitoring sample areas was conducted. The experimental results demonstrated minor discrepancies between simulated and measured values, with relative errors ranging from 3.13% to 11.80%, indicating that the simulation model effectively reflects soil loss dynamics. This project has confirmed that UAV technology combined with modeling can provide robust support for soil and water conservation monitoring.

The fuel switching process of a large methanol dual-fuel powered container ship is simulated and analyzed to accurately predict the nitrogen purging effectiveness in the fuel supply system and investigate methanol fuel purging patterns. The numerical simulations of the purging process are performed under the ship's actual operation conditions based on the 3D model of the fuel supply system, methanol fuel's physical properties, and purging process flow. The purging performance under constant inlet pressure is predicted and the results are compared with freshwater flushing simulation results. This analysis clarified the effects of different media and pipeline volumes on purging time and methanol residue rates, providing practical references for safely and effectively executing nitrogen purging operations.

The safety of ammonia fuel transmission is a critical issue for application of the clean energy source. This paper examines fuel transmission performance of a newly developed marine ammonia fuel system through simulation to verify the design. Based on the physical characteristics of ammonia fuel, the mechanism model of the ammonia fuel transmission system is built, and the effectiveness of the model is verified. The real time simulation is performed with the transmission system model to test and verify the performance of the controller of the system. The simulation operation is effective in system performance testing, not only for normal working conditions but also for extreme working conditions.

The scale effect in model tests and numerical simulations of a large bulk carrier with combined energy-saving devices is studied. The energy-saving performance achieved by the devices is tested and calculated in two different model scales respectively. The test data are compared with the calculation results. The comparison shows that the tests with larger model scale is better in predicting performance improvement and the trends of propulsive performance parameter change caused by the energy-saving devices. From the perspective of engineering application, it is recommended that, within the reasonable range of test conditions, large-scale ship models should be used for towing tank tests of ships equipped with energy-saving devices.

To address the congestion issues on highways caused by significant traffic flow variation to improve the road transportation efficiency and reduce the operational costs of traffic management departments, this paper conducts vehicle profiling research based on vehicle travel behavior and trajectory characteristics. The key vehicles that frequently appear on the road network are identified and occasional driving data are ignored; A travel behavior portrait system is constructed. The system operates based on a three-level labeling system, covering six core dimensions including vehicle information, travel frequency, duration, speed, distance, and travel preferences. It produces comprehensive and multidimensional vehicle portraits for peak hours and off-peak hours; An improved Analytic Hierarchy Process (AHP) algorithm and entropy weight method are used to decide the weights of lower level labels and calculate the score of the top level labels; The vehicle portraits are categorized into 6 categories by the K-Means clustering algorithm based on their scores; The PrioritSpan algorithm is used to mine the typical path patterns of the six types of vehicles at different time periods, generating sets of travel patterns for different types of vehicles. The effectiveness of the system is proved by experiments.

To effectively predict congestion at highway toll stations and solve it, a digital twin solution to predicting congestion happening at highway toll stations based on random forest model is introduced. This solution mainly includes the following functional modules: traffic scene construction; real-time data collection; congestion analysis and prediction; toll station intervention; and visualization display. It collects traffic information through integrated sensing devices such as door frames, vehicle inspectors, and millimeter wave radar. It uses the Unity3D platform to create high-precision 3D models of toll stations and uses CART (Classification and Regression Tree) algorithm to construct a random forest model for traffic flow prediction. Research shows that this scheme can help alleviate congestion at toll stations, improve the efficiency of highway operation and management, and improve the feelings of drivers and passengers.

To enhance the safety of maritime operations, an in-depth analysis of cybersecurity risks faced by oceangoing vessels is conducted, along with the exploration of corresponding mitigation strategies. The study systematically categorizes major cybersecurity risks for oceangoing vessels, including external hacker intrusions, malware attacks, internal human errors, and inherent system vulnerabilities. Seven targeted mitigation strategies are proposed, emphasizing the adoption of advanced cybersecurity technologies and solutions-particularly the integration of AI (Artificial Intelligence) and ML (Machine Learning) technologies to strengthen threat detection capabilities. The strategies also highlight the need to intensify cybersecurity training for crew members to improve overall security awareness, implement rigorous data backup and recovery mechanisms, and employ dedicated networks and isolation technologies to reduce risks from external attacks.

To address schedule management challenges in resource-constrained ship outfitting projects, a progress management model is developed based on global project resource scheduling. This model utilizes the AHP (Analytic Hierarchy Process) to decide resource allocation priorities for subtasks, establishing a resource scheduling framework under constrained conditions. Network planning techniques and priority rule-based heuristic algorithms are applied to separately allocate resources for critical and non-critical tasks. Trial results demonstrate that this model enables proactive resource configuration interventions, effectively improves resource utilization efficiency, and resolves issues of resource shortages or wastage.

The electric power system for a multi-purpose heavy lift vessel needs special design effort because it has to ensure the operation of the cranes onboard, which cause great surge loading. This paper explores the impact of cranes on the design of power plants on multi-purpose heavy lift vessels, focusing on the power station capacity configuration, the power management, reverse power handling, and the selection of cable and protection devices. The power requirements of the cranes are determined based on demand coefficients and probability calculation to ensure that the generator capacity meets actual needs. The design includes heavy load request and power limitation functions in the power management system to ensure safe operation. Various solutions for reverse power handling are investigated, including reverse power absorption resistors, power limitation, and battery absorption.

To continually Improve the maintenance timeliness for traffic mechanical and electrical equipment has been one of the major interests of traffic management departments. A Logistic regression model is built based on the related maintenance work order records kept in Shanghai. The cases that did not satisfy the system operators in terms of maintenance time are studied and the causes are analyzed. The analysis indicates that maintenance timeliness is strongly dependent on the following factors: the equipment type, the equipment location and trouble happening season. The camera equipment and the equipment for license plate recognition, information board are found to need more maintenance time. The equipment installed at the outer ring highway needs extra time. Besides, February, June and July are the higher fault rate months. The following improvement measures are suggested: to enhance the maintenance standard for key equipment; to set up operation and maintenance sub-centers at the outer ring highway; to increase the maintenance force in bad weather and holidays.

With the globalization trend of the chemical industry, the operational efficiency and safety resilience of liquid chemical terminals have attracted much attention, and traditional management models are no longer suitable for the needs of modern logistics. This paper, based on real-life application scenarios, designs and develops a digital twin platform for smart liquid chemical terminals. The platform adopts a layered architecture, including data source layer, base engine layer, general capability layer, architecture fusion layer, and scenario empowerment layer. Implementation of integrated, visualized, and intelligent management and control of terminal operations at all levels is realized. Among them, the dock environment monitoring subsystem integrates multiple heterogeneous sensors to achieve real-time and comprehensive monitoring of the dock operation environment; The production safety management subsystem integrates multiple key systems to build a comprehensive safety protection system; The intelligent operation and maintenance subsystem of the dock utilizes digital twin technology to achieve real-time mapping of equipment status and intelligent operation and maintenance management. The system has been developed and integrated in a liquid chemical terminal shore area, achieving succesful results.

The ship mast vibration caused by excitation source coupling is analyzed as an element of ship system. The case of a Panamax bulk carrier is analyzed for illustration. The mast vibration prediction method is derived based on the ship'sglobal-structure-mast-coupled finite element model established in ship design phase. During sea trials, multi-channel vibration measurement technology is applied to conduct full-scale vibration measurement of the mast. The accuracy of the numerical predict method is verified against the full-scale measurement data. The results show that the error of the vibration predicting method can be controlled within 3%.

The wave added resistance is calculated based on full-scale ship navigation data, using two methods: numerical calculation and empirical formula, and the calculation accuracies achieved by the two methods are investigated. A bulk carrier is selected for illustration. The ship type information of the bulk carrier is collected. Based on the ship type information, the wave added resistance response curve is calculated; The navigation monitoring data of bulk carriers are collected, and the corresponding wave added resistance response curves for each sampled data are calculated using Python programming. The corresponding wave added resistance values are obtained according to the wave height and wave period of each sampled data usingITTC (International Towing Tank Conference) dual parameter wave spectrum; Then, based on the ISO 15016:2015 regulations, wind, waves, temperature, and water density are corrected, and the shaft power is corrected to the ideal calm water power through wave added resistance correction, water temperature correction, wind resistance correction, etc. Finally, the correction results are compared with the resistance analysis results of the fast model test for bulk carriers, verifying that the SNNM(SHOPERA-NTUA-NTU-MARIC) method is better corelated to the wave added resistance on the actual ship than the slice method used.

Addressing the potential safety issues of hull structure and docking blocks during the strength test of large container ships in dock, this paper uses FEA (Finite Element Analysis) to simulate the process and verify the safety of hull structure. A suitable simulation method for docking blocks achieving both high FEA calculation efficiency and good reaction force calculation accuracy are introduced. An optimized scheme of docking block arrangement is devised. The fact is that the arrangement of docking blokes has been designed for ensuring the safety of ship construction in dock. Further requirements for in-dock strength test demands arrangement adjustment. The research shows that the knowledge on the distribution of reaction forces from the docking blocks can be the guide for the docking block arrangement adjustment, keeping it partial in range and minimum in quantity.

A study on pollution control performance evaluation indicators for stormwater drainage systems is conducted to support urban stormwater management planning and to safeguard river water quality. Focusing on stormwater pumping systems within separated drainage systems, this research systematically analyzes pollutant sources in stormwater networks, and key factors influencing their discharge, through literature review and practical case studies. The formation mechanism of overflow pollution is studied. Based on drainage engineering characteristics, multiple critical indicators reflecting pollution control effectiveness in stormwater networks are set up. The analysis reveals that pollutant management in stormwater systems can be performed in three aspects: source control, process control, and end-of-pipe control. Specifically refined indicators are developed for each aspect with clear definitions, providing foundational references for establishing a comprehensive evaluation framework for pollution control in stormwater drainage systems.

The impact of errors in measurements of swing amplitude, cycloid length, and arm of force on the calculation results of the metacentric height h_GM of the ship in inclining tests is studied. The 5 000 ton barge “Shipping Barge 1” is analyzed for illustration. The influencing factors for the ship’s inclining test are analyzed. A mathematical model based on error propagation theory is built for analyzing the influence of measurement errors in swing amplitude, cycloid length, and arm of force on h_GM under the test conditions, The sensitivity analysis is conducted to clarify the degree of influence of each factor. The results are verified through calculation of some typical ship cases. The research results indicate that the swing error has the greatest impact on the calculation results of h_GM; The larger the h_GM, the greater the calculation error of h_GM the swing amplitude change causes.

In response to the demands of large-scale Liquefied Natural Gas (LNG) ship development, as a support to major projects of national ministries and commissions in the 14th Five Year Plan, development and application of domestically produced propellers for large-scale LNG ships is carried out. For particular ship under construction, the technical design objective is set that the propeller efficiency and weight indicators should be no lower than those of similar foreign ship designs. A systematic study is conducted on the quantitative influence of parameters on the propulsion efficiency and excitation force of large LNG ships, including the number of propeller blades, diameter, disc ratio, side inclination angle, longitudinal inclination, profile form, and radial load. The basic design scheme is completed based on the researches, and the stability of the back plate air bubbles in the basic design scheme is improved by optimizing the chordwise load. The model test results show that the final propeller design scheme has excellent performance in terms of propulsion efficiency, excitation force, and weight. The propeller manufacturing process complies with ISO 484 class S precision requirements; The weight deviation between the left and right propellers is only 0.06%. The results of the actual ship trial show that the propeller operates stably and the speed index meets the design requirements.

In response to shipyards’ demand for correcting the extensive data management and uneven granularity in the process of group erection assembly and welding, the structured format, integration and control of the assembly and welding data are researched to achieve full sharing of various data and promote the intelligent transformation of shipbuilding. Based on the structured analysis method Integration DEfinition Method (IDEF0), a process model of the group assembly and welding operations is built to achieve effective hierarchical control of the overall process and operation details. The composition of the data is analyzed and, with the Systems Modeling Language (SysML), a data model for the group assembly welding system is constructed from four aspects: process, resources, planning, and execution, to achieve a unified structured description of the diverse, heterogeneous, and multi-source unstructured data; A data collection plan for the ship group assembly and welding process is established to achieve comprehensive data collection; A group erection assembly and welding data control system is developed to achieve transparent and refined integrated control of the system's data. The practical application results show that the system can effectively improve the data management capability of the assembly welding workshop, making data recording and querying clear and fast.

In order to meet the demand for intelligent development in the field of tunnel engineering traffic monitoring in China, a traffic monitoring platform based on microservice architecture is designed to address the problems of low emergency rescue efficiency, insufficient intelligent control, and inadequate data analysis capabilities of traditional monolithic traffic monitoring systems. The traffic monitoring platform adopts a layered architecture design and modular design. By integrating GIS (Geographic Information System), video, Internet of Things, digital twins, and data governance services, sub platforms such as road network monitoring, tunnel monitoring, and emergency response are built to connect rescue parties quickly, improve the emergency rescue efficiency and enhance the intelligent control. The system can predict traffic situations through big data analysis and effectively improve the platform’s intelligence level. The platform has been deployed in Qinghai and Chongqing, and the results show that it can significantly improve the intelligence level of traffic management through full factor data perception and full process business control.

A nanoporous carbon is synthesized taking Metal Organic Framework-5(MOF-5) as templates without any carbon precursor, and carbonized at a low temperature (540 ℃) (denoted as NPC-540). NPC-540 possesses high specific surface area (1509.5 m²/g) and large pore volume, showing a good adsorption performance (243.9 mg/g), which is higher than conventional market available Powder Active Carbon (PAC) (226.2 mg/g). Experiments show that the adsorption process is monolayer adsorption. p-Nitrophenol (PNP)/NPC-540 system indicates the spontaneous and endothermic nature of the process, and the adsorption of PNP on NPC-540 involves both physisorption and chemisorption. Besides, kinetics of the adsorption process follows the pseudo-second-order kinetics model. Moreover, salt ionic strength is beneficial for adsorption. When the concentration of NaCl is increased from 0.02 mol/L to 0.35 mol/L, its capacity increases about 57 mg/g. More importantly, NPC-540 can be regenerated easily and remain high capacities (200 mg/g) after three cycles. Therefore, NPC-540 has great potential in organics removal.

A summary of recent researches and achievements in the field of two types of ship wind-assisted propulsion devices-rotor sails and arc-shaped sails are summarized. The literature on the subject is reviewed. The scopes and the methodologies of the works are studied. The contributions of these works to the field are analyzes. The review indicates that the aerodynamic performance and its influencing factors of rotor sails and arc-shaped sails have been researched comprehensively. The development of research methods such as model testing and numerical calculations has reached a mature stage, with numerical calculations generally demonstrating high accuracy in most cases. Further in-depth research on the numerical calculation of aerodynamic performance for rotor sails at high Reynolds numbers is warranted.

When measuring empty ship weight and center of gravity position in ship inclining tests, and, further, calculating the integrity and damage stability of the ship, the density of water is one of the essential parameters. But the density of water is changing because of tides. The effect of this phenomenon is analyzed. The research is conducted on four states: high tide to low tide, low tide to high tide, middle tide returning to middle tide after low tide, and middle tide returning to middle tide after high tide. The specific impact of density changes related to the tides states on the results of ship tilt tests are examed.

This paper proposes a distributed energy storage control strategy for the coordination of distributed energy storage in ship integrated electric propulsion systems containing AC and DC bus-bars. Simulation models of the parts of the ship’s integrated electric propulsion system are built, including power generation system, energy storage system, etc.; A distributed energy storage control strategy is designed for the system, which achieves the automation of the energy storage system by collecting information such as bus voltage and energy storage status; A system simulation model of the ship’s integrated electric propulsion system is constructed and operated. The simulation results show that the fluctuation of DC bus voltage is less than 2%. The system’s AC and DC bus voltage can meet the operational requirements. The distributed energy storage management method proposed in this paper can effectively manage energy storage systems and achieve stable operation.

A fuel consumption prediction method based on high-density operating condition recognition is proposed to address the complexity of multiple factors affecting ship navigation fuel consumption. Spearman correlation analysis is used to explore the degree of correlation between various feature factors and the fuel consumption. The features with high correlation degree are selected to characterize the engine operating conditions; Cluster analysis is performed on the selected features. Time axis projection and kernel density estimation techniques are introduced to accurately identify typical operating conditions by calculating the probability density distribution of each operating condition in the time dimension; A Long Short Term Memory (LSTM) network type fuel consumption prediction model for specific scenarios is constructed based on typical operating conditions, achieving a synergetic combination of operating condition recognition and prediction modeling. Experimental results show that the LSTM model with clustering technology can significantly improve the prediction accuracy compared to the model without clustering. The root mean square error decreases from 2.42 to 1.52; the average absolute error decreases from 2.12 to 1.78, and the coefficient of determination R² increases from 0.52 to 0.71. The introduction of feature screening and operating condition partitioning enables LSTM to more accurately capture the time series patterns of fuel consumption under similar operating conditions.

Resistance optimization is carried out for a low-speed, full-formed twin-skeg ship at the target draft and target speed. Based on Computational Fluid Dynamics (CFD) simulations, the surface pressure distribution, wave-making distribution, and flow field around the hull of the parent ship are analyzed, and resistance optimization directions and strategies are formulated accordingly. Under constraints such as displacement and hard points, modifications are made to the bow entrance angle, bow shape, and the length of run body at the stern. Additionally, the fullness of the skeg is increased, the skeg shoulder is reduced, and the run angle is decreased. Comparative analysis reveals that the modified design can significantly improve the free-surface wave-making, result in more uniform surface pressure distribution at the bow, and avoid noticeable flow separation near the stern. The total resistance can be reduced by more than 5%, achieving the goal of resistance reduction and energy savings.

In order to improve the operational management level and asset utilization efficiency of container asset management enterprises, considering the uncertainty of the behavior in returning container upon lease termination and the influence of other internal and external factors, a predictive analysis is conducted on the behavior of returning container upon lease termination. An intelligent prediction model is developed based on customer historical data and comprehensive variables such as regional container return limits and monthly container return limits of ports for particular container types. The container return volume is predicted through specific calculation methods. The practical application results have proved that the model can perform prediction and analysis of action of returning container upon lease termination, assist container asset management enterprises in predicting the future evolution trend of lease termination, help them adjust management and operation strategies in a timely manner, and improve operational management efficiency and asset utilization.

This study examines the decline of the U.S. shipbuilding industry since World War II (WWII), analyzes its key influencing factors, and evaluates recent policies aimed at revitalization. The research reveals: Post-WWII, U.S. commercial shipbuilding relied on the Jones Act (1920) to retain domestic markets, while prioritizing military contracts. This imbalance led to stagnation in civilian sectors. High labor costs, lagging technological innovation, and the unintended consequences of protectionist policies (e.g., the Jones Act) have hindered industrial competitiveness in recent decades. Recent measures include the Fair Act for Offshore Workers in America (2023), Section 301 investigations against foreign competitors, the introduction of “shipbuilding managers,” and the establishment of a White House Office of Shipbuilding. However, these policies struggle to reverse industrial hollowing-out due to entrenched militarization and technological path dependence. The study concludes that the U.S. shipbuilding dilemma reflects a fundamental tension between strategic security and market efficiency, offering cautionary insights for China’s shipbuilding development.

This paper Explores the application of RAG (Retrieval-Augmented Generation) technology in building a large-scale shipping professional knowledge base, providing support for the intelligent development of the shipping industry. The principle of RAG technology is introduced systematically. The limitations of traditional RAG knowledge bases in text slicing, retrieval accuracy, and global understanding are analyzed. Advanced solutions such as introducing reordering models, database MPC (Multi-Process Communication) Server architecture, and super large context models are proposed to establish a shipping professional knowledge base. The research results indicate that the construction of a shipping professional knowledge base based on the RAG big model can significantly enhance the information retrieval and knowledge generation capabilities in the shipping professional field, providing strong support for the digital transformation of the shipping industry.

The linear engineering project plan is a strip-shaped design drawing with a large aspect ratio. It requires dozens of sheets when printed on the standard A-series size paper, which means a large amount of manual labor. To improve the drafting efficiency of the linear engineering, programs are used to optimize the time-consuming framing and layout operations. First, the drawing sheet format and scale are configured, and the length and width of the print area are defined. Then, the centerline is segmented and the drawing simplified according to the stake number with coordinates exported sequentially by chainage. The drawing can be segmented into sheets by using two layout methods: Fixed-size (constrained length and width) or Adaptive-length (width-constrained only). Seamless joint of adjacent sheets’ print areas is guaranteed. The alignment azimuth is determined by the start and the end stake points of each drawing sheet, while the geometric center is calculated based on the stake point set. The alignment azimuth and the geometric center are used to adjust the position of the print, making it vertically centered from left to right within the viewport in paper space. Compared to traditional methods, new method improves workflow efficiency by nearly two orders.

In order to deal with the shortcomings of traditional data quality governance methods in terms of efficiency, coverage and adaptability, an automatic method of data quality rule recommendation based on intelligent algorithms is introduced. By constructing a dynamically evolving enterprise-level data asset feature repository, the accuracy rule recommendation algorithm based on similarity calculation and the timeliness rule prediction algorithm based on lineage analysis are combined to realize the whole process automation from metadata extraction, rule generation to recommendation execution. The core innovations include: establishing a multi-dimensional feature extraction mechanism to support unified modeling of structured attributes, business semantics, and data kinship; a dual-algorithm collaborative recommendation strategy design to solve the structural quality problem and the timeliness constraint problem respectively. Development of compact recommendation engine components to achieve seamless integration with existing data center systems. The research results have been applied to a group enterprise’s big data platform and supply chain collaborative business platform. The results show that the proposed method can shorten the average generation time of a single rule to less than 10 seconds, and the overall configuration efficiency is about 5 times higher than that of the traditional manual method, thus promoting the evolution of data quality governance from “human experience-driven ” to “ intelligent algorithm-driven ” which has good practicabilityand generalization possibility.

To address the safety risks of leakage, phase imbalance, and explosion associated with ammonia fuel injection operations, for the injection scenario of a certain type of ammonia powered tugboat tanker truck, a collaborative plan for process optimization and risk prevention is developed. Based on CFD (Computational Fluid Dynamics) numerical simulation (SST k-ω turbulence model), the diffusion law of ammonia leakage is quantified, and a three-level dynamic safety control zone is designed to solve the multiphase flow balance problem of liquid ammonia injection. By means of a nitrogen dual loop purging system with real-time monitoring technology based on laser absorption spectroscopy, a leak prevention and control closed-loop mechanism is formed. The research results indicate that this scheme can effectively control the risks of leakage, explosion, and low temperature during the process of liquid ammonia refueling, providing a highly applicable and safe technological paradigm for ammonia fuel ship refueling, and supporting the low-carbon transformation of the shipping industry.

A collision risk assessment model for fishing vessels is established and applied to the development and construction of a collision risk assessment and safety management system based on Automatic Identification System (AIS) data. In risk assessment the data of accident causation factors are deeply explored, and the AIS ship data center reconstruct. The fishing boat collision risk is predicted and risk causation assessment is performed by using the Analytic Hierarchy Process and Fuzzy Mathematics methods. The results of the actual ship tests show that the established collision risk assessment model for fishing boats has strong practicality, high estimation accuracy, and can provide theoretical data and technical support for the causal assessment and safety management of risk factors in fishing boat navigation and operations.

Regions in Shanghai are classified based on their characteristics in terms of usage—production area, living quarter or ecological zone. The land use distribution pattern and its evolution characteristic in Shanghai in the period of 2010—2020 are studied from two aspects of spatial layout and quantitative structure using the land use transition matrix and land use dynamic degree. The research shows the facts as follows: The change of space distribution in Shanghai was significant in the period of 2010—2020. The area of ecological zone was relatively small, mainly covered by Shanghai Jiuduansha Wetland National Nature Reserve. However, it had expanded and further expansion is expected. The area of ecological production space and production ecological space reduced in the process of development, and the living and production space expanded. Among four types of spaces, the transition from production ecological space to living production space accounted for the largest proportion, the conversion from ecological production space to production ecological space follow. The ecological space had been transferred out to some extent in the 10 years, but the amount of transfer in is larger, so the overall area has increased. As for the periods from 2010 to 2015 and from 2015 to 2020, the ecological space area showed increasing trends in both periods. The area of ecological production space and production ecological space continued to decrease, while the area of living and production space continued to increase. The intensity of overall land use change was at a low level in both periods.

Current application status of data-driven artificial intelligence technology in the shipping industry is systematically analyzed, and the disruptive impact of this technology’s application on the industry’s employment pattern is deeply explored. Focusing on phenomena such as job substitution, job creation, skill redefinition, and organizational restructuring, this paper aims to revealing the risks of structural unemployment, job mismatch, and the impact of these factors on traditional human resource systems. Countermeasures are proposed from two levels: enterprise microcosmic response and national policy support. At the enterprise level, large shipping companies should build a “three transformations” response framework of “digitization of job position setting, professionalization of employee, and flexibilization of organization”, and adhere to the core logic of “people-oriented, digital intelligence first”; At the national level, taking China as an example, it is recommended to include the development of shipping professionals in top-level planning, and improve relevant laws and regulations as well as industry-education integration mechanisms to provide institutional support for enterprise transformation.