基于遗传算法的超磁致伸缩执行器优化

Optimization for giant magnetostrictive actuator based on genetic algorithm

提出了超磁致伸缩执行器(GMA)优化设计模型,并应用多目标遗传算法对超磁致伸缩执行器进行优化设计.模型优化目标包括:减少执行器导磁回路磁阻使Terfenol-D棒上磁场强度高;空心线圈产生的磁场强度高;Terfenol-D棒上磁场均匀;线圈的效率系数大和线圈与Terfenol-D棒间气隙小.优化变量包括:Terfenol-D棒的尺寸、执行器导磁回路结构尺寸、导磁回路材料的磁导率和线圈结构.根据设计要求选取变量范围,利用非支配排序遗传算法(NSGA)在整个参数空间内搜索,得到执行器的主要参数,应用有限元软件ANSYS分析验证了该结构设计的合理性,并试制了执行器.

The optimization design model of giant magnetostrictive actuator （GMA） was presented and the multi-object genetic algorithm was applied for the optimization design of GMA. The optimal objects of the model included： maximization for magnetic field density of Terfenol-D rod by reducing the magnetic reluctance, maximization for magnetic field density of coil, homogeneous magnetic field of Terfenol-D rod, maximization for efficiency coefficient of coil and minimization for gap between Terfenol-D rod and coil. The optimization variables included： the size of Terfenol-D rod, the structure of magnetic flux path of actuator, the permeability of parts and the structure of coil. The domain of optimal variables was determined according to the demand for application. The fittest parameters of GMA were obtained using non-dominated sorting genetic algorithm （NSGA） by the search in parameter space. The application of finite element software ANSYS demonstrated the rationality of optimization structure of GMA and the GMA was fabricated.