【摘要】 局部特征尺度分解(local characteristic-scale decomposition,LCD)方法是最近新提出的一种先进的自适应时频分析方法。由于LCD方法中均值曲线插值点的属性主要由相邻两同类极值点的属性决定,不能很好地体现数据的整体变化趋势,从而可能引起分解精度降低。根据上述问题,提出了基于Newton插值的局部特征尺度分解(newton interpolation based local characteristic-scale decomposition,NILCD)方法。该方法采用Newton插值取代LCD中的线性插值,均值曲线的插值点由相邻的三个同类极值点构成的Newton插值多项式计算产生,改进LCD。通过仿真信号将NILCD与LCD方法进行分析对比,结果证实了NILCD在提高分量正交性和精确性等方面具有一定的优越性。并将NILCD方法应用于转子碰摩故障的诊断,实验结果表明了新方法的有效性。更多还原

【Abstract】 A novel non-stationary signal method i.e.Newton interpolation based local characteristic-scale decomposition( NILCD) was proposed for improving the local characteristic-scale decomposition( LCD) method.It is a new advanced self-adaptive time-frequency analysis method.The property of mean curve interpolation point is mainly decided by adjacent similar extremum point which can’t properly reflect the overall trend of signal,thereby the decomposition precision is lowered.According to the above problems,and in order to improve the LCD method,Newton interpolation was used in NILCD to replace linear interpolation in LCD,and the mean curve interpolation points were computed by the Newton interpolation polynomial which was generated by three adjacent similar extremum points.The simulation experiments were used to compare NILCD with LCD.The results confirm that NILCD is more efficient in improving the orthogonality and veracity in components than LCD.Finally,the proposed method was applied to diagnose the rotor with rub-impact fault successfully which indicated the effectiveness of NILCD.更多还原