摘要
永磁同步电机(PMSM)实际运行过程中由于外界环境温度磁场变化以及自身磁路饱和,会引起电机参数变化,导致电机控制性能下降。提出了一种变步长的自适应线性(adaline)神经网络对电机进行多参数在线辨识,在学习过程中使用自适应于误差的步长因子,并在权值更新公式中增加了动量项,从而加快了PMSM多参数在线辨识的收敛速度,提高了辨识精度,进而提高永磁同步电机的控制性能。通过建立数学模型,用计算机对辨识方法进行了仿真研究。仿真结果表明,对比传统的Adaline神经网络,该算法收敛速度更快,并且具有稳态误差小,鲁棒性好的特点。
In the actual operation of permanent magnet synchronous motor(PMSM),due to the change of temperature and magnetic field in the external environment and the saturation of its own magnetic circuit,the motor parameters will change,resulting in the decline of motor control performance.An adaptive linear(Adaline)neural network with variable step size is proposed for online identification of multi-parameters of PMSM.In the learning process,a step factor adaptive to error is used,and momentum term is add to the weight update formula,which speeds up the convergence speed of online identification of multi-parameters of PMSM,improves the identification accuracy,and improves the control performance of PMSM.By establishing a mathematical model,the identification method is simulated by computer.The simulation results show that compared with the traditional Adaline neural network,the algorithm has faster convergence speed,smaller steady-state error and robustness.
作者
申中一
吕刚震
王建祥
Shen Zhongyi;Lyu Gangzhen;Wang Jianxiang(School of Mechanical Engineering,University of Shanghai for Science and Technology・Shanghai 200093,China)
出处
《电子测量技术》
2019年第23期85-90,共6页
Electronic Measurement Technology
关键词
永磁同步电机
参数辨识
神经网络
变步长
permanent magnet synchronous motor(PMSM)
parameter identification
neural network
variable step size
