现代电机优化方法综述  

A Review of State-of-the-art Electrical Machine Optimization Methods

作　　者：花为[1] 顾佳磊 HUA Wei;GU Jia-lei(School of Electrical Engineering,Southeast University,Nanjing 210096)

机构地区：[1]东南大学电气工程学院,南京210096

出　　处：《导航与控制》2021年第5期13-25,共13页Navigation and Control

基　　金：国家自然科学基金重大项目(编号:51991381)。

摘　　要：电机在装备制造、国防、航天、工业生产和人们的日常生活中得到了广泛的应用,随着转速和功率密度的提升,高性能电机的需求迅速增长,使得高效的电机优化设计方法显得尤为重要。首先,针对基于有限元法的大规模电机优化耗时严重的固有缺陷,总结了有效减少优化时间的最新方法。其次,对于工作点不确定的电机,指出了电机优化设计应当考虑整个驾驶周期以提高整体效率。接着,介绍了诸如温度场、应力场和转子动力学等影响电机性能的多物理场特性,归纳了电机多物理场建模优化方法。再者,考虑到制造过程和实际运行中不可避免的不确定性因素,总结了电机鲁棒性优化设计方法。最后,对电机优化目前有待解决的问题和未来的发展趋势进行了展望。Electrical machines have been widely applied in equipment manufacturing, national defense, aerospace industry, industrial production and daily life. With the improvement of rotation speed and power density, the demand for high performance electrical machines is increasing sharply, which makes the efficient machine design optimization process extremely important. Firstly, a review of some state-of-the-art methods is presented in this paper which can effectively reduce computational consumption caused by large-scale finite element analysis(FEA). The computational burden is the main drawback of FE-based machine optimization. Secondly, for machines with uncertain operating points, it reveals that the optimization process should take the actual drive cycle into consideration to improve the whole drive-cycle efficiency. Thirdly, multi-physics modeling and optimization methods are covered to consider the effect for machine performance from other physics characteristics, such as thermal field, stress field, and rotor dynamics. Meanwhile, robust design optimization method is presented with the consideration of unavoidable uncertainties in machine processing and actual operation. Finally, the problems to be solved and the future development trend of machine optimization are prospected.

关 键 词：电机优化设计 代理模型 高效计算有限元法 磁网络模型 驾驶周期 多物理场优化 鲁棒性优化 

分 类 号：TP351[自动化与计算机技术—计算机系统结构]