基于二阶非对称随机共振的轴承故障特征提取

Feature Extraction of Bearing Fault based on Second-Order Asymmetric Stochastic Resonance

在强噪声背景下,针对故障信号特征提取困难这一问题,提出了二阶非对称随机共振方法.该方法首先以输出信噪比作为人群搜索算法的目标函数,然后对二阶非对称随机共振模型的参数同时优化,最后将优化的参数代入随机共振模型实现微弱特征信号的增强与提取.仿真和工程数据表明,二阶非对称随机共振方法能有效提取微弱特征信号,实现轴承故障诊断.同时,与传统随机共振方法相比,二阶非对称随机共振的滤波性能更加优越.

A second-order asymmetric stochastic resonance method was proposed to extractthe characteristic signals in the heavy background noise.The signal to noise ration(SNR)of output signal was set as the objection function of seeker optimization algorithm(SOA),and seeker optimization algorithm was used to optimize the parameters of the second-order asymmetric stochastic resonance model.The optimized parameters were substituted into the stochastic resonance model to enhance and extract the weak feature signals.The simulation data and the real fault data of a bearing show that the second-order asymmetric stochastic resonance method can effectively detect weak characteristic signals and realize fault diagnosis of bearings,and its filtering performance is better than traditional stochastic resonance method.