二阶随机共振系统的冲击信号检测

A Detection Method of Impact Signals Based on Second-Order Stochastic Resonance

为解决单一峭度指标度量冲击信号出现漏检的问题,提出了一种基于特征系数的二阶欠阻尼随机共振冲击信号检测(CCSR)方法。由于相似度可以定量地表征两个信号的相似程度,故将峭度与相似度结合,构造出新的特征系数指标;然后以特征系数为目标函数,根据网格搜索算法自适应地寻找系统参数,当目标函数取最大值时此参数为最佳参数;最后采用CWRU轴承数据进行轴承内圈故障诊断。仿真结果与实际检测结果表明,CCSR方法能够有效地检测出目标信号,实现轴承内圈磨损故障的精确诊断;在相同条件下,采用特征系数度量冲击信号比峭度检测效果提高了约20%。

A second order detection method of underdamped stochastic resonance impact signals based on the characteristic coefficient（CCSR）is proposed to address the problem of undetection of the impact signal with single kurtosis index.Because the similarity can quantitatively characterize the similarity between two signals,kurtosis and similarity are combined to construct a new index of characteristic coefficient.Then,the characterize coefficient is used as an object function and the grid search algorithm is used to adaptively find the optimal system parameters such that the objective function achieves the maximum.Finally,the bearing inner-race failures are diagnosed by the data of CWRU.Simulation and practical bearing test results show that the CCSR method effectively detects the target signal and achieves bearing diagnosis.A comparison with the kurtosis under the same conditions show that the test results of using characterizing coefficient increases about 20%.