摘要
提出一种利用自动搜峰的瞬时频率估计技术来实现旋转机械自适应多阶比分析(Adaptive Multiple Order Tracking,AMOT)的新方法。首先,通过时频分析得到振动信号的时频分布,根据频率峰值坐标自动选取搜峰起始点,自适应搜索出不同阶比分量的时频峰值。其次,利用最小二乘法将不同频率分量进行拟合实现瞬时频率估计,然后根据参考分量计算出重采样的鉴相时标对原始信号进行重采样,最后通过FFT变换实现阶比分析。该方法通过瞬时频率估计能够自动识别出所有阶比分量,实现优中选优,避免了传统算法中人为直观选取一个分量进行遮掩滤波提取分量的方法,减少了人为选取分量及起始点造成的误差,具有自适应性。并且无需同步采集转速信号,大大简化了应用条件,同时减少了人为因素,提高了分析精度,为旋转机械故障诊断提供了新方法。仿真实验和应用实例验证了方法的有效性。
A new method of adaptive multi-order tracking （AMOT） in rotating machinery was proposed based on the instantaneous frequency estimation with an adaptive peak search algorithm. The time-frequency distribution was obtained by using a certain time-frequency analysis method. According to the frequency peak coordinate, the starting point of peak search was selected automatically and time-frequency peaks of different orders were searched adaptively. The different frequency components were fitted to achieve the instantaneous frequency estimation by using the least square polynomial, and then the resampling time was calculated according to the reference component to resample the original signal. Finally multi-order tracking spectrums were obtained by FFT transform. By virtue of the method of instantaneous frequency estimation, the proposed method can automatically identify all-order components and choose the best one, avoiding the human error owing to choosing reference component and starting point subjectively in traditional algorithms. The method can identify the components of different orders without sampling synchronously the speed signal, greatly simplify the application conditions and improve the accuracy of analysis. It provides a new method for fault diagnosis of rotating machinery. The method was proved to be effective by simulation examples and practical applications.
出处
《振动与冲击》
EI
CSCD
北大核心
2015年第18期19-24,32,共7页
Journal of Vibration and Shock
基金
龙源电力集团股份有限公司科技项目
关键词
故障诊断
自适应多阶比分析
瞬时频率
角域重采样
fault diagnosis
adaptive multiple-order tracking
instantaneous frequency
angle domain resampling
