双边永磁同步直线电机随机模型系统辨识.

In order to solve the problem of dynamic change of discrete transfer function model parameters of double-sided permanent magnet synchronous linear motor and low identification accuracy of traditional methods, a system identification method based on stochastic model was proposed. The system characteristics in the motion process were analyzed, and the Box-Jenkins model with random disturbance term was used to identify the parameters of the dynamic system transfer function model. By injecting different frequency inverse M sequence current signals, the system response was fully stimulated, and system identification results under different experimental conditions were compared and analyzed. The prediction error method was selected to define the variance cost function, and the Levenberg Marquardt algorithm was used to iteratively optimize the parameters of the transfer function model. The optimal identification model was obtained by comparing and analyzing the system identification under different sampling frequencies, different injection current amplitudes and different motor speed. The results show that when sampling frequency is 1 000 Hz, injection current amplitude is 2. 0 A and motor speed is 50 mm/ s, the optimal identification simulation output matching degree is 94. 81% and the square difference is 7. 84 × 10- 4. [ABSTRACT FROM AUTHOR]