【作者】 李海龙；

【导师】 程军圣； 张劲；

【作者基本信息】 湖南大学 ， 机械工程， 2012， 硕士

【摘要】 齿轮是机械设备中重要的连接和传动部件，在机械设备运行过程中发挥有重要的作用。但由于齿轮本身结构的特点，是特别易受损害和出现故障的零部件，如果不及时发现故障，将会给整个生产和社会造成很大的损失。可见齿轮故障诊断研究的重要性。齿轮故障诊断的关键是从齿轮的振动信号中提取故障特征，信号分析与处理是提取故障特征最常用的方法。然而，齿轮发生故障时，其振动信号大多是非平稳、非线性的时变信号，这就要选择合适的信号分析方法。本文正是针对齿轮故障振动信号的非平稳性及其多为多分量的调制信号之和等特性，将一种新的自适应时频分析方法—局部特征尺度分解方法(Local characteristic-scaledecomposition，简称LCD)引入到齿轮故障诊断中，进一步将该方法与倒频谱、能量矩、双谱等方法相结合应用于齿轮故障诊断中，并取得了较好的分析效果。本文的主要研究内容如下：1、论文提出了一种新的自适应时频分析方法—局部特征尺度分解方法(Localcharacteristic-scale decomposition，简称LCD)，通过对其理论本身进行剖析研究和对仿真信号进行分析，指出了LCD方法有一定的优越性。同时，LCD方法也有一定的缺陷和不足，论文对其进行了改进，并且改进后的LCD方法对信号进行分解得到的分量有更好的光滑性。2、针对齿轮故障振动信号大多数为若干的调幅调频信号之和这一特点，将基于B样条函数的局部特征尺度分解方法(B spline-based Localcharacteristic-scale decomposition，简称BLCD)和倒频谱应用于齿轮故障诊断中，有效的提取了故障齿轮的故障特征。3、把基于三次样条函数的局部特征尺度分解方法和能量矩相结合应用于齿轮的故障诊断，通过对正常齿轮和断齿齿轮进行分析，验证了该方法的有效性。4、把基于有理样条函数的局部特征尺度分解方法(RLCD)和双谱应用于齿轮的故障诊断中，从双谱图中有效的提取了故障齿轮的故障特征。这也为齿轮的故障诊断提供了一种新的方法。更多还原

【Abstract】 Gear is the mechanical equipment indispensable connection and powertransmission parts.It has the vital role in the machinery and equipment. But because ofthe complexity of the structure of the gear, working conditions and human factors, thegear is particularly vulnerable to damage and malfunction parts. If the damage geardose not be found in time, it will cause great loss to the entire production and social.So the gear fault diagnosis is very important.The key of gear fault diagnosis is extracted the gear vibration signal fault feature.Signal analysis and processing is the most commonly used method of extracting faultfeatures. However, the gear fault vibration signals are mostly unstable, nonlinear andtime varying signal, so it is important to choose a suitable signal analysis method. Asgear vibration signal is closely related to the gear state, and which often contains agreat deal of fault information, for the non-stationarity, modulation andmulti-component properties of gear fault signal, in this paper, a new kind ofself-adaptive time-frequency analysis method—Local characteristic scaledecomposition (LCD) is applied to gear fault diagnosis. The paper is also introducedcespstrum method, Energy moment method and wavelet energy spectrum method. TheLCD method and the several methods are used in gear vibration signal analysis in thispaper. The experimental results are showed the effectiveness of these methods.The main research contents of this paper are as follows:1.In this paper, a new kind of self-adaptive time-frequency analysis method—Local characteristic scale decomposition (LCD) is proposed. By using LCDmethod,each complicated signal can be decomposed into a number of componentswhose instantaneous frequencies own physical meaning. So the analysis results showthat the LCD method is effective.But the LCD method also has some flaws andshortcomings.With the studying of the LCD method,some improvement isproposed.When used the improved LCD method to decompose the complicatedsignals, it is found that the components are more smoothness.2. Aiming at the characteristic that gear fault vibration signal is composed ofseveral AM(amplitude modulation)-FM (frequency modulation) components, the Bspline-based Local characteristic-scale decomposition method (BLCD) and cespstrummethod are applied to gear fault diagnosis. The analysis results show that BLCD method can be applied to the gear fault diagnosis effectively.3. The spline-based Local characteristic-scale decomposition method and energymoment are combined and applied to gear fault diagnosis. With the analysis of thenormal gear and the fault gear,the results show that this method can be applied to thegear fault diagnosis effectively.4. The gear vibration signals are decomposed adaptively into some intrinsic scalecomponents by using the rational spline-based local characteristic-scaledecomposition (RLCD) method. The experimental results show that based on theRLCD method and bispectrum method can be effectively applied to gear faultdiagnosis.It is also proposed a new method to analyse the gear vibration signals.更多还原