چكيده به لاتين
In this thesis, first, based on linear potential theory, two-dimensional models are presented that has been used to investigate the free surface forced vibration in arbitrary fliud-filled tank with a horizontal or vertical Baffle. In addition, detailed three-dimensional hydro-elastic analysis of a flexible vertical cylindrical tank, cylinderical tank resting on elastic fundasion and 3d rectangular tank in which the free surface of the fluid is covered by a three-layer piezoelectric floating panel is peresented. The problem formulation is based on the linear water wave theory, the classical (Kirchhoff/Sanders) thin plate and shell models, Maxwell's equations of electrodynamics, Stokes’ transformation, and eigen-function expansions in cylindrical (or catesina) coordinates. The control action is achieved by combined volume displacement and volume velocity feedbacks (VDF, VVF) implemented in a second-order active damping (AD) compensator (or PID controller) via competent evolutionary heuristic optimization techniques that systematically tune the controller gain parameters while constraining the floating panel displacement and control voltage. The uncontrolled and controlled transient responses of the coupled hydro-elastic system under various external disturbances (i.e., a harmonic base excitation, a real seismic event (lateral and bi-directional), a severe launch vehicle liftoff event, and a distributed impulsive transverse load on the floating panel) are calculated by means of Durbin’s numerical inverse Laplace transform scheme. Moreover, the free vibration characteristics of the coupled fluid/structure interaction (FSI) system are briefly studied. The superior performance of the proposed active floating roof control configuration in effective suppression of the key hydro-elastic parameters (panel displacement, and shell displacements/stresses) is demonstrated. It is also presneted that, in the current FSI control problems, the Multi-objective Particle Swarm Optimization (MOPSO)-based ADC outperforms the Non-dominated Sorting Genetic Algorithm (NSGA-II)-based method, in terms of convergence rate and computational effort. Limiting cases are examined and the precision of results is verified by comparisons with the existing data as well as with the results produced by a commercial finite element package.
