汽车变速器齿轮和轴承混合故障诊断方法及实验研究

发布时间：2018-04-15 13:56

本文选题：变速器 + 齿轮和轴承　；参考：《华南理工大学》2016年硕士论文

【摘要】：汽车变速器中故障率最高的零部件为齿轮和滚动轴承,因此汽车变速器的故障诊断主要是对齿轮和滚动轴承故障的诊断。汽车变速器发生故障时,有时不止一个零部件出现故障,而表现为混合故障。当汽车变速器存在齿轮和轴承混合故障时,其振动信号相互耦合难以分离,实际采集到的信号受噪声干扰导致故障信号信噪比低,且齿轮和轴承的冲击型故障容易淹没在齿轮的平稳型故障中,齿轮冲击型故障和轴承冲击型故障在时域表现出相近的冲击调制现象,加大了故障分离和分析的难度。所以进行混合故障的故障诊断具有重大的工程应用价值。根据汽车变速器中主要失效部件齿轮和轴承的故障类型、故障成因、振动机理以及故障信号振动特性,分别建立了表征齿轮平稳型故障、齿轮冲击型故障以及轴承冲击型故障的振动信号模型。采用匹配追踪与相关滤波融合的方法对滚动轴承故障特征进行提取,使用二阶单位脉冲响应函数构造冲击调制字典,应用相关滤波法自适应地识别系统中的多阶固有频率和阻尼比,降低了匹配字典的冗余度;通过合理的分段,有效地提高了稀疏系数求解的精度和速度。将匹配追踪方法与最小熵解卷积方法进行对比,仿真和滚动轴承内圈故障试验表明,两种方法均能有效地诊断滚动轴承故障,但是在强背景噪声下和变速器实测信号中,最小熵解卷积方法对滚动轴承单一故障的诊断能力不如匹配追踪方法,并且最小熵解卷积方法不能提取齿轮稳态调制成分,对混合故障诊断存在局限性。在上述研究的基础上,提出了一种基于匹配追踪的汽车变速器齿轮和轴承混合故障诊断流程。首先构造稳态调制字典进行匹配追踪分离稳态调制成分,再对剩余信号进行相关滤波,识别系统的固有频率和阻尼比来构造优化的冲击调制字典,运用分段匹配追踪的方法来分离齿轮和轴承混合冲击调制成分,最后结合希尔伯特选带解调方法实现混合故障诊断。通过混合故障的仿真和汽车变速器混合故障试验,验证了所提诊断流程和采用方法的有效性和通用性。
[Abstract]:The components with the highest failure rate in automobile transmission are gear and rolling bearing, so the fault diagnosis of automobile transmission is mainly the fault diagnosis of gear and rolling bearing.When the automobile transmission breaks down, sometimes more than one component malfunction, but the performance is the mixed fault.When there is a hybrid fault of gear and bearing in automobile transmission, it is difficult to separate the vibration signals from each other. The signal to noise ratio of the signals collected from the transmission is low due to the noise interference.The impact fault of gear and bearing is easily submerged in the stationary fault of gear, and the impact fault of gear and bearing has similar impact modulation phenomenon in time domain, which increases the difficulty of fault separation and analysis.Therefore, the fault diagnosis of mixed faults has great engineering application value.According to the fault types, causes, vibration mechanism and vibration characteristics of gears and bearings, the stationary faults of gears are established respectively.Vibration signal model of gear impact fault and bearing impact fault.The fault feature of rolling bearing is extracted by matching tracing and correlation filtering, and the second order unit pulse response function is used to construct the impulse modulation dictionary.The correlation filtering method is used to identify the multi-order natural frequency and damping ratio of the system adaptively, which reduces the redundancy of the matching dictionary, and improves the accuracy and speed of the sparse coefficient solution effectively by reasonable segmentation.The matching tracing method is compared with the minimum entropy deconvolution method. Simulation and rolling bearing inner ring fault test show that both methods can effectively diagnose the rolling bearing fault, but in the strong background noise and transmission measured signal,The ability of minimum entropy deconvolution method for single fault diagnosis of rolling bearing is not as good as that of matching tracking method, and the minimum entropy deconvolution method can not extract the steady state modulation component of gear, which is limited to hybrid fault diagnosis.On the basis of the above research, a hybrid fault diagnosis process of gear and bearing of automobile transmission based on matching tracking is proposed.Firstly, the steady-state modulation dictionary is constructed to match and trace the steady-state modulation components, and then the residual signals are filtered by correlation to identify the natural frequency and damping ratio of the system to construct the optimized impulse modulation dictionary.A piecewise matching tracking method is used to separate the mixed impulse modulation components from the gear and bearing. Finally, the hybrid fault diagnosis is realized with Hilbert band selection demodulation method.The effectiveness and generality of the proposed diagnosis flow and the method are verified by the simulation of the hybrid fault and the hybrid fault test of the vehicle transmission.
【学位授予单位】：华南理工大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：U472.9

【相似文献】