This paper presents a design optimization process for an axial-flow pump impeller; in which geometrical parameters are optimized to increase the efficiency (emη/em) and reduce the net positive suction head required (NPSHr). The design variables evaluated include the hub angle, chord angle, the cascade solidity of the chord, and blade thickness. To identify the relationships between geometrical parameters and efficiency as well as the net positive suction head required, a numerical simulation approach was applied in conjunction with a design of experiments (DOE) and group method of data handling (GMDH)-type neural networks with the meta-model. An integrated approach combining a multi-objective particle swarm optimization (MOPSO) algorithm and mapping method was used to generate Pareto diagram and determine the best optimal solution. The optimized design improved efficiency by 4.24% and reduced the net positive suction head required by approximately 11.68% relative to the initial design. Therefore, this work is expected to improve the performance of prototype axial-flow pumps.
展开▼
