摘要
为改善常规基于常值切换的滑模转速观测器的低速性能,提出采用修正的指数趋近滑模和扩展卡尔曼滤波结合的方式进行速度估计。针对直接转矩控制低速时转矩脉动大的问题,采用基于模糊网络与PI控制结合的转矩角调节器来改进控制效果。仿真结果表明提出的速度估计器较常规滑模观测器的动态响应快,速度估计更为准确。改进的SVM-DTC方案的转矩角计算合理,转矩响应脉动减少,性能优于常规恒定PI系数的SVM-DTC。
To improve the low speed performance of the conventional sliding-mode speed observer with constantswitching scheme, a speed estimator based on modified exponent-approaching sliding-mode and extended Kalman filter was proposed. Due to the large torque ripple of direct torque control in low speed range, a torque angle adjustor based on fuzzy neural network and PI control was adopted to improve the control effect. The simulation results confirm that the proposed speed estimator responses more rapidly and estimates more accurately in comparison with the conventional sliding-mode observer. The proposed modified SVM-DTC method calculates the torque angle reasonably, reduces the torque ripple obviously and has better performance than those of the conventional SVM-DTC method with constant PI coefficients.
出处
《中国电机工程学报》
EI
CSCD
北大核心
2007年第3期28-34,共7页
Proceedings of the CSEE
关键词
直接转矩控制
空间矢量调制
模糊神经网络
滑模控制
扩展卡尔曼滤波
irect torque control
space vector modulation
fuzzy neural network
sliding-mode control
extended Kalman filter
作者简介
李君（1978-），男，博士研究生，研究方向为电动汽车的电机驱动控制系统，Ijhust@21cn．com；
李毓洲（1978-），男，博士研究生，研究方向为电机调速与控制。
