基于信号共振稀疏分解的转子早期碰摩故障诊断方法

摘要

提出了基于信号共振稀疏分解的转子早期碰摩故障诊断方法,该方法用信号共振稀疏分解从转子系统振动信号中提取早期碰摩冲击信号.与常规的基于频带划分的信号分解方法不同,信号共振稀疏分解方法根据信号中各成分品质因子的不同,将信号分解成高共振分量和低共振分量.当转子出现早期碰摩故障时,振动信号由以转频及谐波为主要成分的周期信号、包含转子故障信息的瞬态冲击信号以及噪声组成.周期信号为窄带信号,具有高的品质因子,可分解为高共振分量；瞬态冲击信号为宽带信号,具有低的品质因子,可分解为低共振分量.利用信号共振稀疏分解方法从转子早期碰摩信号中提取冲击成分,根据冲击的周期可进行转子早期碰摩故障诊断.算法仿真和应用实例验证了该方法从转子系统中提取早期碰摩冲击信号的有效性.%A method for early rub—impact diagnosis of rotors by using resonance — based sparse signal decomposition was proposed herein. The proposed method used resonance—based sparse signal decomposition to extract impulses from vibration signals of early rub —impact in a rotor system. Different from the traditional signal decomposition way based on frequency band, resonance — based sparse signal decomposition method decomposed a signal into high resonance component and low resonance component based on the Q—factor. When early rub — impact occurred in a rotor system,the vibration signals were composed of periodic signal with rotor frequency component and its harmonics, transient impulse signal with rotor fault information and noise. Periodic signal belonged to narrowband signal and had a high Q—factor,and could be decomposed into high resonance component;while transient impulse signal belonged to wideband signal and had a low Q —factor,and could be decomposed into low resonance component. Impulses were drawn from rub—impact signal of a rotor by resonance—based sparse signal decomposition, and the fault diagnosis of early rub —impact in a rotor system was carried on according to the impulse period. Simulation and application example show that the proposed method is effective in extracting impulse signal form rotor systems.