自适应时变梳状滤波与FCM变转速齿轮故障识别

Fault Recognition for Gears with Variable Rotational Speed Using ATVCF and FCM

针对在变转速下齿轮的故障特征提取与故障识别,提出一种基于自适应时变梳状滤波(Adaptive Time-varying Comb Filtering,ATVCF)与模糊C均值(Fuzzy C-means,FCM)聚类的故障识别方法。该方法先用ATVCF方法对变转速下的齿轮振动信号进行时变滤波,并计算滤波信号的均方根值和样本熵,以此构建特征向量矩阵;然后通过FCM对训练样本的特征向量矩阵进行分析,获取各状态下的聚类中心;最后通过计算测试样本与各聚类中心的欧式氏距离来识别测试样本的状态。采用风力机传动系统模拟故障试验台(Wind Turbine Drivetrain Diagnostic Simulator,WTDDS)的齿轮不同状态数据验证了该方法识别变转速下齿轮故障的有效性;同时,通过与直接FCM和经验模态分解(Empirical Mode Decomposition,EMD)-FCM方法的聚类效果对比,凸显了该方法的优越性。

Aiming at the fault feature extraction and fault recognition of gears with variable rotational speed,a fault recognition method based on adaptive time-varying comb filter(ATVCF)and fuzzy C-means(FCM)clustering method is proposed.In this method,ATVCF is utilized to filter the vibration signal of the gear with variable rotational speed,and the root mean square(RMS)and sample entropy(SE)of the filtered signal are calculated to construct the eigenvector matrix.Then,FCM is employed to analyze the eigenvector matrix of training samples,and the clustering center of each state is obtained.Finally,the Euclidean distance between the test sample and each cluster center is calculated to identify the state of the test sample.The validity of the proposed method is verified by the data of different gear states in WTDDS test-bed.Meanwhile,by comparing the clustering results with direct FCM and EMD-FCM,the superiority of the proposed method is highlighted.