推进技术 ›› 2006, Vol. 27 ›› Issue (3): 230-233.

• • 上一篇    下一篇

基于多目标遗传算法的多级轴流压气机优化设计

丁伟,刘波,曹志鹏,陈云永   

  1. 西北工业大学翼型叶栅空气动力学国防重点实验室 陕西西安710072;西北工业大学翼型叶栅空气动力学国防重点实验室 陕西西安710072;西北工业大学翼型叶栅空气动力学国防重点实验室 陕西西安710072;西北工业大学翼型叶栅空气动力学国防重点实验室 陕西西安710072
  • 发布日期:2021-08-15
  • 基金资助:
    航空基金(04C53002);航空支撑基金(04B53007)

Optimization design of multistage axial-flow compressor using multiobjective genetic algorithm

  1. National Defence Key Lab.of Airfoil and Cascade Aero-dynamics,Northwestern Polytechnical Univ.,Xi’an 710072,China;National Defence Key Lab.of Airfoil and Cascade Aero-dynamics,Northwestern Polytechnical Univ.,Xi’an 710072,China;National Defence Key Lab.of Airfoil and Cascade Aero-dynamics,Northwestern Polytechnical Univ.,Xi’an 710072,China;National Defence Key Lab.of Airfoil and Cascade Aero-dynamics,Northwestern Polytechnical Univ.,Xi’an 710072,China
  • Published:2021-08-15

摘要: 建立了一种多级轴流压气机的多目标优化设计方法。采用流线曲率法计算压气机气动性能,结合多目标遗传算法来进行性能参数优化。针对一台两级轴流压气机,选定优化设计的目标是最大的总压比和最高的总绝热效率。将转子尾缘的稠度和相对气流角以及静子尾缘的稠度和气流角作为设计变量。通过优化设计得到一组在两个目标上均优于初始设计的Pareto最优解,对典型的Pareto最优解和初始设计进行分析、比较,证明了该优化设计方法的有效性。

关键词: 轴流压气机;多目标遗传算法+;流线曲率法+

Abstract: A multiobjective design optimization method for multistage compressors was developed.Performances of compressors were evaluated by the streamline curvature method,multiobjective genetic Algorithm was used to handle multiobjective design optimization problems.A multiobjective optimization for a two-stage axial-flow compressor design was performed for maximization of the overall adiabatic efficiency and the total pressure ratio.Relative flow angles and solidities at the rotor trailing edges,and absolute flow angles and solidities at the stator trailing edges are considered as design parameters.The present method obtained a set of Pareto-optimal solutions which outperformed the baseline design in both objectives.Detailed observation and analysis of the Pareto-optimal designs confirms the validity of the multiobjective design optimization method.

Key words: Axial-flow compressor;Multiobjective genetic algorithm~+;Streamline curvature method~+