摘要
建立了三相磁保持继电器的虚拟样机模型,基于正交试验结果,使用BP神经网络与NSGA2遗传算法对继电器进行了优化设计。使用ANSYS和Adams/View分别建立三相磁保持继电器的静态和动态模型,进行了试验验证。以电磁系统部分结构的尺寸为变量,以分断电流值和吸合后的磁保持力矩为目标,进行五因素五水平正交试验,建立数据集;使用BP神经网络对数据集进行训练,分别得到两个优化目标的网络函数;使用NSGA2遗传算法对目标函数进行寻优并更新数据集,如此循环,直至最优解满足要求。动态特性仿真表明,优化后样机的分断时间减少了23.89%,吸合时间和弹跳时间分别减少了29.2%和47.05%,为样机的制造提供了参考。
The virtual prototype model of the three-phase magnetic retaining relay was established.Based on the orthogonal test results,the BP neural network and NSGA2 genetic algorithm were used to optimize the design of the relay.ANSYS and Adams/View were used to establish the static and dynamic models of the three-phase magnetic holding relay,and experimental verification was carried out.Taking the size of the part structure of the electromagnetic system as a variable,taking the breaking current value and the magnetic holding moment after pulling as the target,five factors and five levels orthogonal experiments were carried out to establish the data set.BP neural network was used to train the dataset,and the network functions of two optimization objectives were obtained.The NSGA2 genetic algorithm is used to optimize the objective function and update the data set,and so on until the optimal solution meets the requirements.The simulation results show that the breaking time of the optimized prototype is reduced by 23.89%,and the suction time and the bounce time are reduced by 29.2% and 47.05%,respectively,which provides a reference for the manufacture of the prototype.
作者
苏秀苹
范学东
张自有
贾彦庆
SU Xiu-ping;FAN Xue-dong;ZHANG Zi-you;JIA Yan-qing(State Key Laboratory of Reliability and Intelligence of Electrical Equipment,Hebei University of Technology,Tianjin 300130,China;Key Laboratory of Electromagnetic Field and Electrical Apparatus Reliability of Hebei Province,Hebei University of Technology,Tianjin 300130,China;Zhejiang Lihui Electric Appliance Co.,LTD.,Wenzhou Zhejiang 325609,China)
出处
《计算机仿真》
2024年第8期268-275,共8页
Computer Simulation
关键词
磁保持继电器
虚拟样机技术
正交试验
神经网络
遗传算法
优化设计
Magnetic holding relay
Virtual prototype technology
Orthogonal test
Neural network
Genetic algorithm
Optimization design
