Abstract: Aiming at the problem of multivariable, nonlinearity and strong coupling of the permanent magnet synchronous motor(PMSM), a strategy of inverse system identification which is independent of precise mathematical model and parameters based on support vector machines(SVM) is proposed. The dynamic decoupling control of PMSM is researched based on multivariable nonlinear control inverse system theory. To deal with direct inverse control open-loop system with poor robustness and inverse modeling error of SVM, a parameter self-tuning PID(Proportional Integral Differential) closed-loop controller based on cloud model rule inference is designed. The simulation results confirm that the cloud model PID control based on SVM inverse system incorporates the merits of model-free learning, strong anti-interference capability. While realizing the dynamic decoupling of the excitation component of stator current and rotor speed, the system has high-precision speed tracking feature and excellent dynamic and static performance.

Key words: permanent magnet synchronous motor(PMSM), inverse system method, SVM, decoupling control, cloud model PID control