Real-time prediction of hull girder loads is of great significance for the safety of ship structures. Some scholars have used neural network technology to investigate hull girder load real-time prediction methods based on motion monitoring data. With the development of deep learning technology, a variety of recurrent neural networks have been proposed; however, there is still a lack of systematic comparative analysis on the prediction performance of different networks. In addition, the real motion monitoring data inevitably contains noise, and the effect of data noise has not been fully considered in previous studies. In this paper, four different recurrent neural network models are comparatively investigated, and the effect of different levels of noise on the prediction accuracy of various load components is systematically analyzed. It is found that the GRU network is suitable for predicting the torsional moment and horizontal bending moment, and the LSTM network is suitable for predicting the vertical bending moment. Although filtering has been applied to the original noise data, the prediction accuracy still decreased as the noise level increased. The prediction accuracy of the vertical bending moment and horizontal bending moment is higher than that of the torsional moment.

A bulbous bow is a typical ship structure. Due to the influence of the bulbous bow, complex flow separation and gas capture phenomena may appear during the water entry of ship-like sections. In this paper, experimental and numerical studies on the water entry of a ship-like section with an obvious bulbous bow are carried out. Two thin plates are installed at both ends of the test model to ensure that the flow field during the impact process is approximately two-dimensional. The free-fall drop test is carried out in the test rig equipped with guide rails. By changing drop heights, impact pressure on the model surface with different initial impact velocities is measured. A numerical model for simulating the water entry of the ship-like section is established by using the Computational Fluid Dynamics (CFD) method, based on the Navier-Stokes equations. Reasonable time steps and mesh size are determined by convergence analysis. Four different flow models are used in the numerical analysis. It is found that the K-Epsilon turbulence model can present the most reasonable numerical prediction by comparing numerical results with the experimental data. Furthermore, the influence of the bulbous bow on the impact loads is numerically studied by using the validated numerical model. It suggests that the bulbous bow has little effect on the impact force acting on the bow-flared area but, in the position near the bulbous bow, the pressure will be affected by the second slamming and the air cushion.

The bow-flared section may be simplified in the prediction of slamming loads and whipping responses of ships. However, the difference of hydrodynamic characteristics between the water entry of the simplified sections and that of the original section has not been well documented. In this study, the water entry of several different bow-flared sections was numerically investigated using the computational fluid dynamics method based on Reynolds-averaged Navier–Stokes equations. The motion of the grid around the section was realized using the overset mesh method. Reasonable grid size and time step were determined through convergence studies. The application of the numerical method in the water entry of bow-flared sections was validated by comparing the present predictions with previous numerical and experimental results. Through a comparative study on the water entry of one original section and three simplified sections, the influences of simplification of the bow-flared section on hydrodynamic characteristics, free surface evolution, pressure field, and impact force were investigated and are discussed here.