CAJ | 학술논문

提出一种神经网络左逆系统与右逆系统协同的解耦控制方法。构造神经网络右逆系统与原系统串联实现对原系统的线性化解耦，构造神经网络左逆系统实现对电机磁链和转速的在线估计，将估计值作为右逆解耦控制所需的反馈量设计闭环线性控制器，完成磁链和转速的解耦控制。仿真试验结果表明，该方法具有良好的解耦控制性能，对转子电阻和负载转矩的变化具有良好的鲁棒性。
제출일충신경망락좌역계통여우역계통협동적해우공제방법。구조신경망락우역계통여원계통천련실현대원계통적선성화해우，구조신경망락좌역계통실현대전궤자련화전속적재선고계，장고계치작위우역해우공제소수적반궤량설계폐배선성공제기，완성자련화전속적해우공제。방진시험결과표명，해방법구유량호적해우공제성능，대전자전조화부재전구적변화구유량호적로봉성。
A decoupling control method which combined neural network left inverse system and right inverse system were proposed. The original system which is cascaded with neural network right system is decoupled. Then, the flex and speed of motor can be estimated by constructing neural network left inverse system as the soft sensor model. Last, the estimate values are given the feedback values of close loop linear controller designed for the coupling system. So the decoupling control is completed. Numerical simulations show that the proposed method has high decoupling control performance and good robustness for rotor resistance variation and load torque variation of induction motor.