基于HVD-AR的斜齿轮故障特征提取及损伤过程分析
发布时间：2019-02-17 12:23
【摘要】：齿轮箱是一种应用面较广的基础性传动部件,主要用于改变转速和传递动力。由于其本身结构复杂,工作环境恶劣等原因,齿轮箱容易出现故障。因此,对齿轮箱进行疲劳损伤监测与诊断是非常必要的。基于振动信号的齿轮箱故障特征提取,其主要任务是从采集到的信号中提取可用的故障特征信息,前提条件是准确分离故障振动信号,根本目的是提取故障特征信息。本文是基于斜齿轮台架疲劳试验,以采集到的斜齿轮整个疲劳寿命周期内的实时振动信号为研究对象,提出有效的振动信号降噪方法和反映齿轮损伤程度的特征指标,并将特征指标的变化趋势应用到斜齿轮的疲劳损伤过程的分析中。本文首先以斜齿轮单齿啮合的简化动力学模型为基础,研究了斜齿轮的振动信号机理。基于信号机理,分析了斜齿轮典型故障及其对应的振动信号特征,并回顾了常用的齿轮故障特征指标。针对斜齿轮振动信号的非线性、非平稳性的特征,本文提出希尔伯特振动分解(Hilbert Vibration Decomposition,HVD)与自回归滤波器(AR Filter)相结合的振动信号降噪方法HVD-AR,其中对HVD的信号重构方法和AR模型阶次确定的相干峭度(Correlation Kurtosis,CK)准则进行了重点讨论,并通过实测点蚀损伤过程振动信号包络谱验证该降噪方法的有效性。根据故障对齿轮振动信号频谱边频带影响的机理,将高低阶啮合频率的边频带分开考虑,提出高低阶啮合频率边频带估计的两个特征指标(4和(4,并将其应用到实测振动信号的分析中。考虑现有的斜齿轮台架试验条件,拟定疲劳寿命试验方案,并对试验结果进行简要分析,同时获取整个寿命过程的振动信号数据,为其后的斜齿轮疲劳损伤过程的分析提供必要的数据支撑。采用斜齿轮的典型损伤过程(齿面点蚀和轮齿断齿)的全寿命振动信号以数据驱动方式验证了HVD-AR信号降噪方法的有效性,并将特征指标(4和(4的变化趋势与常用的特征指标RMS、峭度值、ER和FM4进行对比分析,验证其对于斜齿轮损伤过程分析的实用性。结果表明,HVD-AR能够有效降噪斜齿轮振动信号,特征值(4和(4的变化趋势能够有效反映斜齿轮的疲劳损伤过程,能够及早发现早期的微弱故障。
[Abstract]:Gearbox is a kind of basic transmission component with wide application, which is mainly used to change speed and transfer power. The gearbox is prone to malfunction because of its complicated structure and bad working environment. Therefore, it is necessary to monitor and diagnose the fatigue damage of gearbox. The main task of gearbox fault feature extraction based on vibration signal is to extract the available fault feature information from the collected signal. The precondition is to separate the fault vibration signal accurately, and the fundamental purpose is to extract the fault feature information. Based on the fatigue test of helical gear bench, this paper takes the real time vibration signal of helical gear in the whole fatigue life cycle as the research object, and puts forward an effective noise reduction method of vibration signal and characteristic index to reflect the damage degree of gear. The change trend of characteristic index is applied to the analysis of fatigue damage process of helical gear. In this paper, the vibration signal mechanism of helical gear is studied based on the simplified dynamic model of single tooth meshing of helical gear. Based on the signal mechanism, the typical fault of helical gear and its corresponding vibration signal characteristics are analyzed, and the commonly used gear fault characteristic indexes are reviewed. Aiming at the nonlinearity and nonstationarity of helical gear vibration signal, this paper presents a new method, HVD-AR, which combines Hilbert vibration decomposition (Hilbert Vibration Decomposition,HVD) with autoregressive filter (AR Filter). The signal reconstruction method of HVD and the coherent kurtosis (Correlation Kurtosis,CK) criterion determined by the order of AR model are discussed in detail, and the effectiveness of the method is verified by measuring the envelope spectrum of vibration signal in the process of pitting damage. According to the mechanism of the influence of fault on the frequency band of gear vibration signal, the edge frequency band of high and low order meshing frequency is considered separately, and two characteristic indexes of edge band estimation of high and low order meshing frequency are presented (4 and 4, respectively). It is applied to the analysis of measured vibration signal. Considering the existing helical gear bench test conditions, the fatigue life test scheme is drawn up, and the test results are analyzed briefly, and the vibration signal data of the whole life process are obtained at the same time. It provides necessary data support for the analysis of the fatigue damage process of helical gears. Using the lifetime vibration signal of the typical damage process of helical gear (tooth surface pitting and gear tooth breaking), the effectiveness of HVD-AR signal denoising method is verified by data-driven method. The change trend of characteristic index (4 and 4) is compared with RMS, kurtosis, ER and FM4, to verify its practicability for the analysis of helical gear damage process. The results show that HVD-AR can effectively reduce the vibration signal of helical gears, and the variation trends of eigenvalues (4 and 4) can effectively reflect the fatigue damage process of helical gears, and early weak faults can be detected as early as possible.
【学位授予单位】：太原理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TH132.41
本文编号：2425161
[Abstract]:Gearbox is a kind of basic transmission component with wide application, which is mainly used to change speed and transfer power. The gearbox is prone to malfunction because of its complicated structure and bad working environment. Therefore, it is necessary to monitor and diagnose the fatigue damage of gearbox. The main task of gearbox fault feature extraction based on vibration signal is to extract the available fault feature information from the collected signal. The precondition is to separate the fault vibration signal accurately, and the fundamental purpose is to extract the fault feature information. Based on the fatigue test of helical gear bench, this paper takes the real time vibration signal of helical gear in the whole fatigue life cycle as the research object, and puts forward an effective noise reduction method of vibration signal and characteristic index to reflect the damage degree of gear. The change trend of characteristic index is applied to the analysis of fatigue damage process of helical gear. In this paper, the vibration signal mechanism of helical gear is studied based on the simplified dynamic model of single tooth meshing of helical gear. Based on the signal mechanism, the typical fault of helical gear and its corresponding vibration signal characteristics are analyzed, and the commonly used gear fault characteristic indexes are reviewed. Aiming at the nonlinearity and nonstationarity of helical gear vibration signal, this paper presents a new method, HVD-AR, which combines Hilbert vibration decomposition (Hilbert Vibration Decomposition,HVD) with autoregressive filter (AR Filter). The signal reconstruction method of HVD and the coherent kurtosis (Correlation Kurtosis,CK) criterion determined by the order of AR model are discussed in detail, and the effectiveness of the method is verified by measuring the envelope spectrum of vibration signal in the process of pitting damage. According to the mechanism of the influence of fault on the frequency band of gear vibration signal, the edge frequency band of high and low order meshing frequency is considered separately, and two characteristic indexes of edge band estimation of high and low order meshing frequency are presented (4 and 4, respectively). It is applied to the analysis of measured vibration signal. Considering the existing helical gear bench test conditions, the fatigue life test scheme is drawn up, and the test results are analyzed briefly, and the vibration signal data of the whole life process are obtained at the same time. It provides necessary data support for the analysis of the fatigue damage process of helical gears. Using the lifetime vibration signal of the typical damage process of helical gear (tooth surface pitting and gear tooth breaking), the effectiveness of HVD-AR signal denoising method is verified by data-driven method. The change trend of characteristic index (4 and 4) is compared with RMS, kurtosis, ER and FM4, to verify its practicability for the analysis of helical gear damage process. The results show that HVD-AR can effectively reduce the vibration signal of helical gears, and the variation trends of eigenvalues (4 and 4) can effectively reflect the fatigue damage process of helical gears, and early weak faults can be detected as early as possible.
【学位授予单位】：太原理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TH132.41
【参考文献】
相关期刊论文 前10条
1 吕宏强;武志斐;王铁;谷丰收;;基于AR-MCKD的齿轮点蚀故障特征提取[J];机械传动;2017年03期
2 崔伟成;许爱强;李伟;孟凡磊;;基于拟合误差最小化原则的奇异值分解降噪有效秩阶次确定方法[J];振动与冲击;2017年03期
3 王道勇;王铁;张瑞亮;张瑞;乔金维;;振动光饰齿轮接触疲劳寿命对比试验研究[J];机械传动;2017年01期
4 李宝庆;程军圣;吴占涛;杨宇;;基于ASTFA和PMMFE的齿轮故障诊断方法[J];振动工程学报;2016年05期
5 方军强;周新聪;赵旋;;基于EEMD和分形维数的船用齿轮箱故障诊断[J];船海工程;2016年04期
6 吴定海;张培林;杨望灿;齐蕴光;;基于双树复小波的重叠块阈值降噪方法[J];振动与冲击;2016年10期
7 王鹏飞;胡建中;;基于LMD的包络谱特征值在齿轮箱故障诊断中的应用研究[J];机电工程;2016年05期
8 冯志鹏;秦嗣峰;;基于Hilbert振动分解和高阶能量算子的行星齿轮箱故障诊断研究[J];振动与冲击;2016年05期
9 孟宗;王亚超;胡猛;;基于改进LMD和IED-SampEn的齿轮故障特征提取方法[J];机械工程学报;2016年05期
10 雷建波;;正交面齿轮副动力学仿真及疲劳寿命分析[J];热能动力工程;2016年02期
相关博士学位论文 前1条
1 李文良;计及齿面摩擦的斜齿轮传动动态特性研究[D];哈尔滨工业大学;2013年
，本文编号：2425161
本文链接：https://www.wllwen.com/jixiegongchenglunwen/2425161.html
