基于调制随机共振的转子故障早期检测

基于调制随机共振的转子故障早期检测

噪声是影响旋转机械早期故障检测的主要因素。只有抑制噪声增强信号、提高信噪比，才能从噪声中提取故障特征信息，为故障诊断特别是早期故障检测提供可靠的依据。文中根据非线性双稳系统在噪声和弱周期信号作用下的周期响应特性，将调制技术与随机共振原理相结合，提出了调制随机共振方法，实现了在较宽的频率范围内从强噪声中检测微弱周期信号。理论分析、数值仿真和实验结果表明，对信号进行调制产生的差频分量可形成低频信号，该低频信号通过双稳系统易产生随机共振，从而使随机共振发生在具有优良频率特性的低频区，能使微弱的故障信号特征突出、明显而易于捕捉。该方法灵活、可靠，在转子系统早期故障检测方面的应用是可行的。

Noise is the main factor influencing incipient rotor machine failure detection. Only when the noise is suppressed and SNR is improved, the failure characteristic information can be picked up from noise supplying the reliable basis for failure detection especially the incipient failure detection. The modulated stochastic resonance method is proposed according to the period response characteristics in nonlinear bistable system caused by noise and weak period signal. The weak periodic signals can be distinguished from strong noise in broaden frequency range with this method. The theoretical analysis, numerical simulation and experiment results manifest that differential frequency component generated by signal modulation can form low frequency signal. The stochastic resonance is easily generated when the low frequency signal acts on the bistable system. The frequency of stochastic resonance is on low frequency range making the characteristics of weak fault signal obviously and easily to capture. The method is flexible, reliable and effective in application of the incipient fault detection in rotor system.