基于Root-MUSIC法的轧辊偏心参数估计（英文）

Roll eccentricity parameter estimation based on Root-MUSIC

针对轧辊偏心信号是混杂在各种随机干扰中的复杂高频周期信号,因FFT法对信号分析的局限性,提出了一种Root-MUSIC(Multiple Signal Classification)法和Prony法相结合的轧辊偏心信号估计新方法。利用Root-MUSIC法准确估计出偏心谐波的频率及谐波的个数,同时使用Prony方法估计出偏心信号的各次谐波幅值和相位。仿真结果验证了可行性和有效性,在信噪比较低的情况下仍能准确地同时估计出偏心谐波的频率、幅值及相位,尤其在频率分辨率和抗噪声上比FFT法具有优越性。

The roll eccentricity signal is complex and high frequency periodic signal which is mixed with all kinds of random disturbances.Because of the limitation of the FFT method to signal analysis,a new method which is the combination of the Root-Multiple Signal Classification(Root-MUSIC)and the Prony method is presented.The frequency and the number of eccentricity harmonics are estimated by the Root-MUSIC method,and simultaneously the amplitude and phase of eccentricity harmonics are estimated by the Prony method.The simulation results show that in the case of low SNR(Signal Noise Ratio),the frequency,amplitude,and phase of eccentricity harmonics can still be estimated.It is superior to FFT in frequency resolution and noise rejection.