The modified beam model uses the interpolated eigenvectors of the 3-D FE model in motion analysis. Linear computations are performed on the three structural models coupled with the 3-D Rankine panel method. In Fig. 11, all responses are shown to be almost identical. The sharp peak of roll motion is observed near the frequency of 1.2 rad/s, which corresponds to the natural frequency of roll motion. The smooth peak of KU-57788 cost roll motion is due to the relationship between the wave and ship length. A small difference between the models is found in the resonant response of the

7th mode near 3.7 rad/s. The difference is acceptable because a resonant response is very sensitive to frequency. Resonant responses to linear and nonlinear wave excitations are compared in the following sections concerning the 6500 TEU and 10,000 TEU containerships. In Fig. 12, the time series of sectional forces in the regular wave are compared. The still water loads are not included. The high-frequency oscillations in the front part of the torsional moment and vertical bending moment are transient motions of 2-node vertical bending and 2-node torsion modes. Good agreement is obtained for both wet mode natural frequencies and

responses to waves. The natural frequency of 2-node vertical bending decreases from 0.92 Hz in dry mode to 0.61 Hz in wet mode. The added mass can be calculated from the wet mode natural frequency. Fig. 13 shows the longitudinal distribution of the sectional forces. It is confirmed MG-132 solubility dmso that the system is balanced in each time step. Fig. 14 shows the time series of normal stresses in the longitudinal BIBW2992 concentration direction. The stress is evaluated on the top at the mid-ship section, the coordinates of which are 30.0 m from AP, 0.0 m from the center line, 2.0 m from the water line. The stress including both quasi-static and dynamic contribution is calculated as follows: equation(73) σx=MyIyz+FzA equation(74)

σx=∑j=7kσxjξjwhere the normal stress of jth mode obtained by eigenvalue analysis of the 3-D FE model. Eq. (73) is used in the beam theory model, and Eq. (74) is used in the modified beam and 3-D FE models. The results show good agreement between the stresses of the different models. In Eq. (74), the stress converges when k=14. If stress is evaluated at the location far from the mid-ship, k must be larger than 14. In order to obtain the converged stress at every location, quasi-static stresses of higher modes should be calculated, which are not included in the coupled-analysis. The most rigorous method is to perform FE analysis with applying all the inertial and external forces. In addition, the mesh of the 3-D FE model should be finer than that for eigenvalue analysis. The stress evaluation is not discussed more than the above because it is too complicated to be fully handled in this study. However, the method for stress evaluation will be thoroughly discussed in the near future because stress evaluation is the final goal of the hydroelastic analysis.