Multi-Objective Automated Optimization of Centrifugal Impeller Using Genetic Algorithm
A development and application of an automated optimization method for aerodynamic design of centrifugal impeller blades has been presented in this paper. A Non-uniform mutation and Pareto tournament and Fitness-sharing techniques based Multi-Objective Genetic Algorithm (MOGA) has been developed. The fast speed to convergence and well ability to search the Pareto front of the MOGA has been demonstrated through single-objective and multiobjective function tests. By introducing the MOGA, a three-dimensional reconstruction system for centrifugal impeller blades using non-uniforM rational B-spline (NURBS) and a commercial software NUMECA, an aerodynamic automated optimization design system has been established. To a centrifugal impeller, the maximization of the absolute total pressure ratio and the isentropic efficiency has been taken as the design targets. The Pareto solutions have been obtained by using the present optimization technique. Through analysis and comparison the optimized design and the initial design, the validity and feasibility of the developed optimization design system is confirmed. The optimized results showed the performance of the optimized impeller has been improved.
centrifugal impeller genetic algorithms NURBS automated optimization
Wenbin Zhang Xiaomin Liu
School of Energy & Power Engineering, Xian Jiaotong University Xian Jiaotong University, Xian, 710049, China
国际会议
The Fourth International Symposium on Fluid Machinery and Fluid Engineering(第四届流体机械与流体工程国际会议)
北京
英文
130-136
2008-11-25（万方平台首次上网日期，不代表论文的发表时间）