基于希尔伯特—黄变换和支持向量机的齿轮箱故障诊断研究

发布时间：2018-04-06 02:06

本文选题：齿轮箱　切入点：希尔伯特-黄变换(HHT)　出处：《中北大学》2011年硕士论文

【摘要】：齿轮箱作为机械设备中一个重要的组成部分,对其进行状态检测和故障诊断具有很强的现实意义。本文通过对齿轮箱常见故障进行模拟实验,利用希尔伯特变换法(HHT)对测得的故障信号进行特征值提取,进而利用支持向量机(SVM)的方法对齿轮箱故障状态进行识别,得到了较好的效果。齿轮箱故障诊断主要包括诊断信息的获取,故障特征值的提取和模式识别三个部分。其中故障特征的提取和状态识别是齿轮箱故障诊断的关键。当齿轮箱发生故障时,其振动信号往往表现为非平稳性,本文提出的希尔伯特黄变换法中的EMD分解法是基于信号的局部时间特征尺度,具有很强的自适应性,可以将信号分解为有限个内禀模态函数(IMF)之和,每个IMF分量分别包括了不同时间特征尺度大小的成分,其尺度依次由小到大,因此,每个IMF分量包含了从高到低不同频率段信号成分。本文将EMD方法引入齿轮箱故障诊断,选取故障信息明显的IMF分量,进而求得边际谱图,实现了齿轮箱故障初步诊断。支持向量机方法是统计学习的一种,是在统计学习理论基础上发展起来一种新的机器学习方法。目前,支持向量机已经成为解决非线性分类问题的一种强有力的工具。它具有对经验的依赖小,能够获得全局最优解以及良好的泛化性能等特点,已被广泛应用于模式识别中。本文将EMD能量特征提取法和支持向量机相结合应用于齿轮箱故障诊断识别中,实现了对齿轮箱故障状态准确的诊断识别。
[Abstract]:As an important part of mechanical equipment, gearbox has a strong practical significance for condition detection and fault diagnosis.In this paper, through simulating the common faults of the gearbox, using Hilbert transform (HHT) to extract the eigenvalue of the measured fault signal, and then using the support vector machine (SVM) method to identify the fault state of the gearbox.Good results were obtained.Gearbox fault diagnosis includes three parts: obtaining diagnosis information, extracting fault eigenvalue and pattern recognition.Fault feature extraction and state recognition are the key of gearbox fault diagnosis.When the gearbox fails, the vibration signal is usually non-stationary. The EMD decomposition method of Hilbert-Huang transform is based on the local time characteristic scale of the signal and has strong adaptability.The signal can be decomposed into the sum of a finite intrinsic mode function. Each IMF component consists of components of different temporal characteristic scales, whose scales range from small to large, so,Each IMF component contains signal components at different frequencies from high to low.In this paper, the EMD method is introduced into the gearbox fault diagnosis, and the obvious IMF component of the fault information is selected, then the marginal spectrum is obtained, and the primary diagnosis of the gearbox fault is realized.Support vector machine (SVM) is a new machine learning method based on statistical learning theory.At present, support vector machine (SVM) has become a powerful tool to solve nonlinear classification problem.It has been widely used in pattern recognition due to its small dependence on experience, its ability to obtain global optimal solutions and its good generalization performance.In this paper, the EMD energy feature extraction method and support vector machine are combined in the gearbox fault diagnosis and identification, and the accurate diagnosis and recognition of the gearbox fault state is realized.
【学位授予单位】：中北大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：TH165.3

【引证文献】