摘要
感应电动机参数离辨识是实现高性能变频调速的前提。对基于递推最小二乘(recursive least-squares,RLS)算法的电机参数离线辨识的具体实现进行了详细阐述;基于离线辨识时电机的静止特性,简化了电机数学模型,减少了计算量;采用改进欧拉数值解法,求解巴特沃思数字滤波状态方程,既可获得信号的滤波值,同时又直接解得了信号的1阶和2阶导数值,避免了导数离散化的误差影响,提高了运算精度;运用线性反馈移位寄存器产生伪随机白噪声序列(M序列),作为辨识的激励信号;针对不同功率的电机进行了大量的仿真和实验研究,实验结果表明所提方法稳定性良好,辨识结果的一致性和准确性较高。
The off-line parameters identification of induction motor is the prerequisite condition of realizing high performance frequency control. An off-line parameter identification of induction motor based on recursive least-squares (RLS) algorithm was presented in the dissertation. Based on the motor static character of off-line parameter identification, the method simplified the motor mathematic model and reduced the computation amount. Figuring out the Butterworth digital filter state equation through improved Euler numerical solution, the filtering value of signal was got, and the first-order derivative and the second-order derivative also were got directly. It avoids the error influence from derivative discretization and improves the computation accuracy. The paper adopted white noise, which produced by linear feedback shift register, as the excitation signal. The simulation and experimental results of different power motor are presented to demonstrate the stability, consistency and accuracy of the proposed algorithm.
出处
《中国电机工程学报》
EI
CSCD
北大核心
2011年第18期79-86,共8页
Proceedings of the CSEE
关键词
感应电动机
参数辨识
递推最小二乘
状态变量滤波器
M序列
induction motor
parameter identification
recursive least-squares (RLS)
state variable filter
M sequence