基于多目标优化的液力变矩器叶形角度设计  

Streamline inlet and outlet angles of torque converter blade design based on multi-objective optimization

作　　者：闫清东[1,2] 李新毅 魏巍 

机构地区：[1]北京理工大学机械与车辆学院,北京100081 [2]北京市电动车协同创新中心,北京100081

出　　处：《华中科技大学学报（自然科学版）》2017年第9期69-75,共7页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(51475041);国防科工局基础产品创新科研基金资助项目(VTDP-2104);北京理工大学国际科技合作专项资助项目(GZ2017035104)

摘　　要：为解决液力变矩器性能优化过程中单一目标优化不能满足实际需求的问题,首先利用一维束流理论对该型号液力变矩器进行特性计算,计算结果与试验结果较为一致.在此基础上,以最高效率、起动变矩比和最大泵轮扭矩系数为目标建立多目标优化模型.利用Matlab与ISIGHT的集成优化,组成以一维束流计算程序为核心的解算器,以叶片入出口角为优化参数,采用变导轮和定导轮两种优化方案,并利用基于存档的小种群遗传算法(AMGA)对该液力变矩器的性能进行优化,获取多目标优化下的非劣解.结果表明:优化后的液力变矩器性能比原型有所提高,同时也可以满足不同应用条件下车辆液力变矩器的性能要求,验证了这种方法的有效性.In order to solve the problem of the single obj ective optimization can′t meet the actual de-mand in the optimization process of the torque converter,the one-dimensional flow theory was adopted to calculate the characteristics of the torque converter.The comparison against experimental data showed a good agreement.On this basis,efficiency,stall torque ratio and max pump torque coeffi-cient were taken as optimization goals,and the multi-obj ective optimization model was built.The op-timization platform was built by integrating Matlab and ISIGHT,in which the one-dimensional beam theory was employed as the core solver.The angle of the blades was taken as design parameters,and the archive-based micro genetic algorithm （AMGA）was utilized to optimize the torque converter per-formance with change stator and no change stator.The results show that method is effective,and the performance of optimized torque converter has increased compared with the original model.The opti-mization results obtained by change stator and no change stator method are all improved in the per-formance indexes of the torque converter,which solves the shortcomings of the single target optimiza-tion can′t meet the different performance requirements of hydraulic torque converter.

关 键 词：液力变矩器 多目标优化 束流理论 非劣解 遗传算法 叶形角度 

分 类 号：TH137.332[机械工程—机械制造及自动化]