Search

The Transformable-Craft (T-Craft) is an innovative vessel serving as a connector between sea base and beachheads. When operatingat seas, the T-Craft is a surface effect ship (SES) with compartmented air cushions. To analysis the seakeeping performance of the T-Craft SES, a high efficient 2D/2.5D analytical model has been developed by combining the one dimensional waveequation for solving aerodynamics of pressurized cushion air with the STF/2.5D method for solving hydrodynamics of demihulls at low/high speeds. To enhance the computational efficiency, the entire model is linearized except the dynamics of air leakage, which is strongly nonlinear and inappropriate for linearization. Results obtained from the proposed seakeeping analysis model show reasonable agreement with experimental data, while the proposed model hascompetitiveness in the computational efficiency as compared with some 3D models from public works.

This thesis presents the results from the computational fluid dynamics simulations of surface effect ship model tests. The model tests being simulated are of a generic T-Craft model running in calm seas through a range of Froude numbers and in two head seas cases with regular waves. Simulations were created using CD-adapco's STAR-CCM+ and feature incompressible water, compressible air, pitch and heave degrees of freedom, and the volume of fluid interface-capturing scheme. The seals are represented with rigid approximations and the air cushion fans are modeled using constant momentum sources. Drag data, cushion pressure data, and free surface elevation contours are presented for the calm seas cases while drag, pressure, heave, and roll data are presented for the head seas cases. The calm seas cases are modeled both with no viscosity and with viscosity and turbulence. All simulations returned rather accurate estimations of the free surface response, ship motions, and body forces. The largest source of error is believed to be due to the rigid seal approximations. While the wake's amplitude is smaller when viscosity is neglected, both viscous and inviscid simulations' estimations of the free surface qualitatively match video footage from the model tests. It was found that shear drag accounts for about a quarter of the total drag in the model test simulations with viscosity, which is a large source of error in inviscid simulations. Adding the shear drag calculated using the ITTC-1957 friction coefficient line to the total drag from the inviscid simulation gives the total drag from the viscous simulations within a 6% difference. Master of Science