带有扰动补偿的永磁同步电机自适应预测控制  

Adaptive Predictive Control of Permanent Magnet Synchronous Motor with Disturbance Compensation

作　　者：刘达[1,2] 李木国[2] 

机构地区：[1]大连理工大学电子信息与电气工程学部,辽宁大连116023 [2]大连理工大学海岸与近海工程国家重点实验室,辽宁大连116023

出　　处：《电机与控制应用》2012年第1期31-35,共5页Electric machines & control application

基　　金：国家自然科学基金(50879098);中央高校基本科研业务费专项资金资助(DUT10JR14);辽宁省教育厅高校科研计划项目(LS2010032)

摘　　要：为保证永磁同步电机(PMSM)控制系统的稳定性和鲁棒性,提出一种基于小波神经网络的自适应Smith预测控制器(WNNSPC),通过对模型误差的补偿,消除模型不确定性的影响,保证了预测输出的准确性,同时针对电机运行中突加负载等的扰动,采用小波神经网络进行前馈补偿,有效抑制了各类扰动对系统的影响,而且具有良好的动、静态品质。仿真结果证明了提出方法的有效性。An adaptive Smith predictive controller based on wavelet neural network （WNN） was proposed to guarantee the stability and robustness of the PMSM control system. It could eliminate the influence of model uncer- tainties by compensating the model error and make a accurate predictive output. Feed forward compensation using WNN was also utilized to reject kind of disturbances like a sudden load, which effectively restrains the affect of dis- turbances and provides good static and dynamic performance. The simulation results proved the validity of the pro- posed control scheme.

关 键 词：永磁同步电机 SMITH预估器 小波神经网络 鲁棒性 

分 类 号：TM351[电气工程—电机]