摘要
随着永磁电机运行过程中转速的升高,电机的内部参数会随着磁路饱和逐渐发生改变,其中电感参数对电机的稳态和动态运行性能影响较大。本文基于模型参考自适应的理论基础,提出了加入神经网络算法对d、q轴电感进行辨识的算法,并采用mish函数作为激活函数应用到神经网络算法中,并将辨识结果应用到电机矢量控制中,仿真结果验证了该激活函数的有效性,在不影响电机系统正常运行的同时,提升了辨识过程中的收敛速度。
As the speed of permanent magnet motors increases during operation,the internal parameters gradually change due to magnetic circuit saturation,and inductance parameters significantly impact both steady-state and dynamic operational performance.Therefore,building upon the theoretical foundation of model reference adaptive control,this paper proposes an algorithm for identifying d-and q-axis inductances by integrating a neural network approach.The Mish function is employed as the activation function for the neural network algorithm.The identification results are then applied to motor vector control,with simulation results confirming the effectiveness of the activation function in improving convergence speed during parameter identification without disrupting normal motor system operation.
作者
刘钰
李欣燃
邵慧威
Liu Yu;Li Xinran;Shao Huiwei
出处
《时代汽车》
2024年第19期7-9,共3页
Auto Time
基金
新能源汽车用永磁同步电机弱磁控制技术研究(QNL202111)。
关键词
参数辨识
神经网络
mish函数
收敛速度
Parameter Identification
Neural network
Mish Function
Convergence Rate
作者简介
刘钰(1993-),女,硕士,讲师,研究方向为牵引电机及其控制技术。
