一种采用遗传算法优化离心叶轮结构的方法

A method optimizing centrifugal impeller structure based on genetic algorithm

针对高功重比航空发动机离心叶轮结构减重需求及传统多目标优化难以实现全局优化的特点,以强度计算模块获取的全局样本数据为基础,对代理模型生成的响应面采用多目标遗传算法反复迭代寻优。利用整体-局部的结构优化过程,获得参数化离心叶轮结构设计的全局最优解,并结合应力质量分数与叶尖变形分布的对比,验证了离心叶轮结构优化对结构性能的影响。结果表明:运用Kriging代理模型-遗传算法分步优化,能有效提高离心叶轮结构质量的利用率,改善结构稳定性,获取目标最优结构。

Aiming at the demand of weight reduction for centrifugal impeller of aero-engine with high power to weight ratio,and the difficulty of global optimization through traditional multi-objective optimization,based on the global sample data obtained by the structural strength calculation module,the multi-objective genetic algorithm(MOGA)was used to iteratively optimize the response surface generated by the agent model.The global optimum solution of parametric centrifugal impeller structure design was obtained by the whole-part structural optimization process.The structure stress mass fraction and the tip deformation distribution were compared and the influence of the centrifugal impeller structure optimization on the structural performance was verified.The results show that structural optimization by use of Kriging agent model-genetic algorithm can effectively increase the structural utilization of the parameterized centrifugal impeller,improve structural stability,and obtain the optimal structure of the target.