The omnidirectional waterjet propeller, as a lateral thruster or dynamic positioning device, has attracted more and more attentions. Its hydrodynamic characteristics are a key factor in meeting the application requirements. However, there are limited related studies. The numerical simulation of hydrodynamic performance of the omnidirectional waterjet propeller was carried out in this paper. Based on the STAR-CCM+ software, the steady RANS method was applied to investigate the hydrodynamic performance of an omnidirectional waterjet propeller under two conditions, i. e. static water and flowing water. The results show that the hydrodynamic performance of both thrust magnitude and directionality is greatly affected by the magnitude and direction of incoming flow, and the influence is greater when the rotational speed is lower. The research in this paper reveals the thrust loss mechanism of the omnidirectional waterjet propeller. Its hydrodynamic performance should be evaluated according to its working conditions, and the low rotational speed operation should be avoided to ensure that hydrodynamic performance requirements are met.

The motion of ships and marine structures is a nonlinear motion with time series characteristics. The Long Short-Term Memory (LSTM) artificial neural network has the characteristics of memorizing time interval information and processing nonlinear data, which is very suitable for processing such nonlinear motion with time series characteristics. Therefore, LSTM has significant advantages in predicting the very short-term motion response of ships. In this paper, an improved LSTM method for the prediction of very short-term motion response of ships is proposed. This method converts the prediction of ship motion into the prediction of peak and valley values by means of extracting envelopes, which can reduce the data demand of the traditional LSTM model and simplify the complexity of the prediction curve, thereby significantly improving the forecast duration. In this paper, the improved LSTM was used to predict the regular wave curve, irregular wave curve and real ship motion curve. The results show that the improved LSTM prediction method can enlarge the maximum forecast duration of the traditional LSTM model from 6~8 s to about 20 s, and has ideal prediction results for special signals such as abrupt signals, which has high practical value.

The South sea of China is affected by tropical cyclones and typhoons in the western Pacific Ocean, which are prone to double-peak or even multi-peak waves in the form of mixed waves, which are potentially hazardous to the operational safety of marine floating structures. Therefore, based on the potential flow theory and considering the influence of different loading states, a comparative study was conducted on the motion response of FPSO and shuttle tanker side-by-side system under double-peak spectral wave and single-peak spectral wave states. The motion responses of the two hulls in the double-peak spectral wave states of mainly wind waves, wind waves and surge are equal, and mainly surge and the single-peak spectral wind waves state are calculated respectively. The calculation results show that (1) the amplitude of the motion of the two hulls in vertical, horizontal and longitudinal directions increases with the increase of the ratio of the low-frequency energy to the total energy; (2) it is the largest when the surge is dominant, the second largest when the wind waves and surge are equal, and the smallest when the wind waves are dominant; (3) the motion amplitude of both hulls in the double-peak spectral waves is greater than that of the single-peak spectral wind waves under the premise that the total energy is the same, and (4) the motion amplitude of both hulls in the double-peak spectral waves is greater than that of the single-peak spectral wind waves. The study shows that the influence of double-peak spectral waves should be considered in the mooring design and safety planning of FPSO operation system.

A fuzzy adaptive sliding mode based method was designed to control the precise pitching motion of a bionic robotic fish in response to the problem of numerous gait control parameters and imprecise dynamics modelling for the co-propulsion of 3-degree-of-freedom pectoral fins, flexible body and caudal fin. Firstly, based on the given 3-degree-of-freedom pectoral fin and flexible body co-propulsion motion law, the relationship between the fish body thrust/moment and the pectoral fin phase difference and body fluctuation frequency was established by the least-squares method using CFD numerical simulation. Secondly, the designed fuzzy adaptive sliding mode controller was used to achieve the free-diving motion of the robotic fish, taking the smooth curve with the current and the desired depth levels as asymptotes, as the desired motion trajectory. The fuzzy adaptive sliding mode controller was designed to achieve the free diving motion of the machine fish. The fuzzy controller was used to analyse and calculate the adaptive law of fuzzy control rule to compensate the uncertainty term of the dynamics model and the disturbance term of the water current during the movement of the robot fish in real time, and combined with the sliding mode controller to reduce the system jitter to achieve the accurate tracking of the trajectory. Finally, the results of the simulation and the pool experiments show that the robot fish is able to swim between different depths in a fast and smooth manner, and the movement trajectories are smooth with the maximal depth deviation of only 0.08 m, and the steady state error remains at 0.04 m, verifying the effectiveness of the proposed method.

The phenomenon of icing on the surface of superstructure of ships and marine structures is the result of a large number of water droplets impacting on the cold wall surface to form a water film and then accumulating ice. And surely, at a high wind speed more water droplets often collide with vertical structures. The movement and freezing behaviors of water droplets can have a significant effect on the icing process and final icing shape on the surface of the structures. Based on the VOSET gas-liquid interface tracking method coupling with VOF and Level-Set, and Enthalpy-Porosity phase change method, a unidirectional coupling model between water droplets and isolated cold plate was established using the large coefficient method. Simulation of the freezing process of a single water droplet impacting an isolated cold plate was achieved, and the phenomenon of the air entrainment was reproduced. The effects of factors such as water droplet velocity, component surface wettability on the freezing process were analyzed. The process of multiple water droplets impacting a vertical structure surface to form a liquid film and freeze was further simulated. The relevant results can provide technical support for the prediction of the typical component surface icing of superstructure and the study of anti-icing and de-icing methods.

The main design calculation methods applicable to the reinforcement of spherical shell openings were systematically summarized focusing on the simulation of pressure bearing structures in a large-scale deep-sea cold seep environment. Then, as an example, the reinforcement effect was studied based on equal area method, pressure area method, and limit analysis method respectively under the conditions of opening ratio of 0.1-0.6 and pressure of 21 MPa for a ϕ5000 mm spherical shell. Quantitative analysis was conducted on the influence of the reinforcement length of the spherical shell and the reinforcement length of the connecting pipe on the stress distribution of the local structure in the joint zone between the ball and column, and a scaled model was used for verification. The results indicate that the dense reinforcement method needs less reinforced area and is more economical for high-pressure large-opening reinforcement design, and that the increase in the reinforced lengths of the spherical shell and the connecting pipe will effectively reduce the stress concentration factor at the junction and improve the ultimate strength of the local structure. However, the increase in the length of the connecting pipe has a certain marginal effect. On the premise of manufacturing feasibility, the reinforcement length of the spherical shell should be increased as much as possible. The structural stress calculation results based on finite element method are in good agreement with the measured results.

An analytical model of pipe torsional stiffness in clockwise and counterclockwise directions is derived based on the helical winding structural characteristics of steel wires in the unbonded flexible pipe armour layers considering radial contraction and expansion phenomena. Taking a typical unbonded flexible pipe as an example, the bi-directional torsional stiffness analysis is conducted. The results show that the analytical model has a close match with the results of existing numerical model. It can be found that the error is 3.5% in clockwise torsion, and the error in counterclockwise torsion is 4.6%. This paper can provide a useful reference for the design and analysis of the torsional performance of flexible pipes.

High-speed vessels face significant challenges in optimizing bow structures under slamming loads due to uncertainties in load magnitude and spatial distribution. This paper proposes a multi-stage topology optimization method integrating load uncertainty analysis and manufacturing constraints to balance lightweight design and engineering feasibility. Firstly, the uncertain loads are converted into multi-scenario worst-case loading problems. Through an iterative "critical load scenario-topology optimization" process, the critical load positions are dynamically updated. Then, a topology optimization strategy based on the Solid Isotropic Material with Penalization (SIMP) method is employed, incorporating geometric/manufacturing constraints to progressively derive an optimal stiffener layout that meets strength and stiffness requirements. Each iteration retains prior design outcomes and updates worst-case load scenarios to achieve progressive adaptation to uncertain loads. Finally, multiple iterations and geometric reconstruction convert high-density element clusters into manufacturable stiffener configurations. Finite element verification demonstrates that the optimized bow structure exhibits significantly reduced maximum displacement, more uniform multi-scenario responses, and compliance with lightweight and safety requirements. This method effectively addresses the computational burden of double-layer nested optimization, offering a novel approach for structural optimization of high-speed vessel bows under stochastic slamming loads.

Steel/GFRP L-shaped joint and similar components are common in ship manufacture. In this paper, a steel/GFRP L-joint with ±45° groove structure was designed, and the effect of parameters such as glueing length on the performance of the L-joint was investigated through the compression experiments of steel/GFRP L-joints. In the numerical analysis of the steel/GFRP L-joint, the complexity of the ±45° groove structure glueing interface structure and the multi-interface and multi-scale were considered, and the stiffness equivalence method was used to equate the groove structure to a cohesive unit layer of 0 thickness. The failure modes of the L-joint were analyzed in detail concerning the compressive load and displacement by comparing the compression experimental results with the simulation, and the steel/GFRP L-joint exhibits better compressive performance when the bonding length is 100 mm.

The traditional low-order finite element model is usually used to obtain the acoustic scattering field of a submarine structure, and then to evaluate the acoustic stealth performance. However, the traditional finite element method is affected by the numerical pollution effect, and requires very dense mesh to obtain reliable numerical solutions for problems with relatively medium and high frequencies, leading to prohibitive cost in mesh division. In this paper, the overlapping finite element method (OFEM) and Dirichlet-to-Neumann (DtN) mapping technique are combined to construct a coupled numerical model for the acoustic scattering of underwater elastic targets. When constructing local approximations in the OFEM, the virtual nodes are used to generate partition of unity functions, while no degrees of freedom are assigned to these virtual nodes. The novel OFEM can be directly applied to low-order finite element models and achieve higher-order approximations of the unknown variables. Numerical examples show that the OFEM can reduce the numerical error significantly and has broad application prospects in the prediction of underwater acoustic scattering by elastic targets.