In order to avoid solving the problems of obtaining feedback gain matrix in traditional model reference adaptive system (MRAS), the asynchronous motor mathematical model was established in stator α-β coordinate system based on direct torque control (DTC) theory, discrete space vector modulation (DSVM)-DTC system and stator current equation were gained according to this model and then the equation was transformed into neural element with five input nodes, and an output node as stator current adaptive identification modules. According to speed identification formula, the speed of motor was real-time estimated and transferred back to adjustable model whose parameters were modified and made close to the real value. Finally, the simulation was made in the Matlab/Sinulink using parameters of 37 kW asynchronous motors. The results show that the identified speed can accurately track the actual speed and perform good dynamic response. When system parameters change, the system has good robustness and is easy for application.%针对一般消噪法对噪声非均匀分布心电信号消噪存在的不足,提出基于奇异值分解和小波阈值消噪相结合的消噪方法.该方法利用矩阵的奇异值分解将噪声非均匀分布的心电信号正交分解为噪声分布相对均匀的分量,在正交子空间中对每个分量进行小波阈值消噪,重构消噪后分量,得到消噪后的心电信号.研究结果表明:本方法有效地克服了因噪声分布不均匀而造成的小波阈值选择矛盾的缺点,有效地消除了大噪声区域的噪声,又完好保存小噪声区域的心电特征信息,且消噪后的信号与无噪信号之间的欧氏距离最小.
展开▼
