부산대학교 도서관

This research aims to adopt the reverse engineering and 3D printing technology to imitate a model of water-jet propulsion and to conduct a parameter design. Using 'Load Cell' and 'Laser Doppler Velocimetry (LDV)', this study carried out to measure the velocity distribution of nozzle, the flow rate and the estimation of thrust. An efficient and systematic way is used to investigate the robust design based on Taguchi method that is one of the designs of experiments (DOE). There are many factors influencing the output efficiency and thrust on the performance of the water-jet propulsion system. In this study, three control factors, such as 'number of rotor blades', 'number of stator blades' and 'stator placement angle' are discussed. The signal-to-noise ratio (S/N) of the experimental data is used as the index of parameter design, and the contribution of each factor and the mutual influence of each factor are discussed through the analysis of variation. The work successfully proposes a modeling by using the Radial Basis Function Network (RBFN) in ANN; the optimum factor-level combination is obtained by the adaptive particle swarm optimization (APSO) in order to get the parameter design of the best thrust of waterjet propulsion.