二级减速器故障系统建模及SVD-MMSE劣化评估  

Failure System Modeling and SVD‑MMSE Degradation Evaluation for Two Stage Reducer

作　　者：解开泰 章翔峰[1] 周建星[1] 余满华 王胜男 姚俊 张旭龙 XIE Kaitai;ZHANG Xiangfeng;ZHOU Jianxing;YU Manhua;WANG Shengnan;YAO Jun;ZHANG Xulong(School of Mechanical Engineering,Xinjiang University Urumqi,830047,China)

机构地区：[1]新疆大学机械工程学院,乌鲁木齐830047

出　　处：《振动．测试与诊断》2024年第3期580-588,624,共10页Journal of Vibration,Measurement & Diagnosis

基　　金：国家自然科学基金资助项目(51865054);新疆维吾尔自治区天山英才培育资助项目。

摘　　要：为检测故障齿轮劣化程度并进行有效的程度评估,通过有限元法建立含有正常、裂纹和断齿等3种齿轮状态的二级直齿轮减速器系统模型。首先,分别计算3种状态的齿轮时变啮合刚度,并综合考虑轴承支撑刚度,得到了3种不同状态下的轴承振动响应;其次,引入多元多尺度样本熵(multivariate multiscale sample entropy,简称MMSE)对故障齿轮的劣化程度进行分析;最后,引进奇异值分解(singular value decomposition,简称SVD)算法进行预处理,以达到更好的诊断效果来综合评定故障齿轮生命周期的劣化程度。结果表明:齿轮发生故障时,主要导致时频域信号发生转频调制,时域存在有规律的冲击,频域出现边频带,且分布在输入轴的转频及其倍频和啮频及其倍频处;随着故障程度的增加,劣化越发明显,频率成分也发生改变,致使MMSE值也随之变化,且整体呈单调递减趋势;SVD-MMSE算法能有效地对齿轮故障程度进行判别,降低了噪声对于劣化程度检测准确性的影响。In order to detect the deterioration degree of faulty gear and evaluate the deterioration degree effectively,a two-stage spur gear reducer system model with three kinds of gear states(normal,cracked and broken)is established by finite element method.The time-varying meshing stiffness of the gear in three different states is calculated respectively,and the bearing support stiffness is considered comprehensively.On this basis,the bearing vibration responses in three different states are obtained.Then multivariate multiscale sample entropy(MMSE)is introduced to analyze the deterioration degree of the faulty gear.However,the gear signal is highly jumbled and noisy,so singular value decomposition(SVD)algorithm is introduced for preprocessing in order to achieve better diagnostic effect and comprehensively assess the deterioration degree of the life cycle of the faulty gear.The results show that the gear fault mainly leads to frequency variation in the time-frequency domain,regular impact in the time domain,and the presence of side frequency band in the frequency domain,which are distributed around the modulation and doubling of the input shaft's rotational frequency.With the increase of the degree of failure,the deterioration becomes more obvious,and the frequency component changes,resulting in the change of MMSE value,and the overall trend is monotonously decreasing.SVD-MMSE algorithm can effectively distinguish the degree of gear failure,and this method can better reduce the impact of noise on the detection accuracy of deterioration degree.The analysis results show that SVD-MMSE is effective and feasible to diagnose the deterioration degree of gear life cycle,and provides reference and basis for the state monitoring of gear box in the future.

关 键 词：性能劣化 有限元分析 时变啮合刚度 奇异值分解 多元多尺度样本熵 

分 类 号：TH132[机械工程—机械制造及自动化]