| dc.description.abstract | For ships with open cross sections, such as container vessels, it is of primary importance to accurately estimate torsional moment (TM) and horizontal bending moment (HBM) as well as vertical load effects. In recent decades, the increase of the size of open ships makes their hydroelastic effects in the response more important. Especially, the lowest natural frequency for large open ships is mostly likely linked to the coupled-horizontal-torsional modes. Consequently, the influence of the lateral and vertical vibrations on the wave-induced load effects should be evaluated in an appropriate manner. At present, there are few numerical tools available, which are able to predict the nonlinear springing and whipping responses in horizontal bending and torsional modes. With regard to the predictions of the vertical vibrations, the reliability of numerical tools still needs further validation, because the excitation source for springing and whipping in vertical bending is incomplete. Experimental study can provide useful guide for the practical design of open ships and also the development of reliable numerical tools.
This thesis mainly deals with an experimental investigation of wave-induced nonlinear load effects, to contribute to the understanding of hull girder behavior in terms of vertical bending moment (VBM), HBM, and TM. Model tests of four open ships have been carried out and compared in this thesis. The models include two backbone models – one model with pronounced bow flare, an 8600-TEU Ultra-Large Container Ship (ULCS) model, and a 13000-TEU ULCS model. Numerical calculations are carried out mainly to compare the wave-frequency load effects, while some time-domain simulations are also conducted to validate the high-frequency responses in vertical bending. New observations are made during model tests that have not been accounted for in current numerical codes.
In the first part of this thesis, the experimental and numerical methods applied in the investigation are discussed. Three numerical tools are first introduced, which are based on either strip theory or panel method. Experiments of the four flexible ship models are then described, respectively. And uncertainties associated with measured results are identified and quantified. Next, data analysis methods used for the investigation are presented. The second part of this thesis consists of nine papers. The following issues related to the main contributions are presented.
For the wave-frequency load effects on the initial backbone model, the numerical results based on the two-dimensional (2D) strip and three-dimensional (3D) panel methods are consistent with the experimental results in regular and irregular waves. Furthermore, the tank wall effects on the wave-induced load effects are clearly observed during measurements, and the numerical calculations, which consider the existence of tank walls, reflect the trend correctly.
For the same backbone model, calculated hydroelastic properties, such as natural frequencies and modal shapes, are compared with the experimental data. The hull girder vibrations in bending and torsion are investigated systematically both in regular and irregular waves. The springing response in torsional mode could be significantly excited in large amplitude in regular waves, and their contributions are dominant in short irregular waves. In long irregular waves, the extreme sectional load effects including VBM, HBM, and TM are amplified greatly due to high-frequency vibrations. Furthermore, vertical vibrations initiating in the hogging condition are observed, and the hogging VBM is found to be significantly larger than the sagging VBM for this particular backbone model.
A new bow with pronounced bow flare is then installed on the initial backbone model to compare the influence of bow flare slamming on the lateral and vertical vibrations. Significant increase in lateral vibrations is observed for the modified backbone model than for the initial backbone model.
Because the line profile of the backbone model is unrealistic, at least for container ships in service, the unexpected discovery that the hogging VBM is much larger than the sagging VBM under certain wave conditions is less persuasive for the practical design of open ships. The test results of the 8600-TEU and 13000-TEU models in head seas are investigated systematically to show solid evidence that the nonlinear hogging VBM may be larger for realistic vessels. The spectral analysis in irregular waves shows that the second harmonics in VBM increase the hogging VBM and decrease the sagging VBM, which is consistent with the results in regular waves. It is found that the hogging VBM is more likely to be larger than the sagging VBM in relatively short waves.
The above discussions about the symmetry of VBM are mainly limited to head seas. The 13000-TEU model is also tested in oblique seas. The influences of heading angle and wave scattering on nonlinear VBM under different conditions are further discussed. The results under a heading angle of 30 deg are similar to the results in head seas, and extreme hogging VBM is observed. For a heading angle of 60 deg, the sagging VBM is significantly larger than the hogging VBM.
Numerical calculations of the vertical vibrations are validated against the measured results. For the backbone model, the second-order springing in vertical bending is well predicted, while the vertical vibrations in severe seas are significantly underestimated. With regard to the 8600-TEU vessel, the numerical results in regular waves are compared with the experimental data. The high-order harmonic responses in the VBM amidships are focused on. Although relatively poor agreements are obtained for the second and third harmonics, the fourth and fifth harmonic responses are better predicted. In general, the dominant high-frequency vibrations are always related to the fourth and fifth harmonic responses. As for the 13000-TEU vessel, the influence of the vertical vibrations on the vertical bending moments is investigated under three most unfavorable conditions. The high-frequency vibrations are slightly underestimated. | nb_NO |
