一阶线性系统随机共振在转子轴故障诊断中的应用研究

Application of a first-order linear system's stochastic resonance in fault diagnosis of rotor shaft

针对机械系统早期微弱故障信号难识别诊断问题,提出一阶线性系统调参广义随机共振的特征提取方法,该方法基于调节一阶线性系统参数,可以得到信噪比取极大值的广义随机共振现象。为得到清晰的特征信号,以可辨识性为优化目标,给出了系统参数、信号频率、采样频率等参数之间的选择关系。滑动轴承试验台上转子轴的早期微弱故障模拟实验,验证了此方法的有效性。

Aiming at recognizing a weak fault signal in its early stage, a feature extraction method of a signal based on the parameter-adjusted stochastic resonance （SR） of a first-order linear system in a broad sense was presented here. This method was based on a generalized stochastic resonance phenomenon that the output signal-to-noise ratio （SNR） reaches a maximum by tuning parameters of a first-order linear system. To acquire a more distinct characteristic signal, the selecting relations among system parameters, signal frequencies and sampling rates were explained taking the best signal reognition as an optimization objective. Finally, the analog test rotor shaft＇s early weak fault obtained on a test-bed of sliding bearings verified the effectiveness of this method.