基于自适应最稀疏时频分析的机械故障诊断方法

发布时间：2018-07-05 05:42

本文选题：自适应最稀疏时频分析 + 广义解调时频分析　；参考：《湖南大学》2016年博士论文

【摘要】：机械设备状态监测和故障诊断对于保证设备健康运行具有重要的作用。随着科学技术的发展,现代机械设备越来越复杂,对机械故障诊断新技术和新方法的研究具有重要的意义。机械故障诊断技术的核心是故障特征的提取,设备运行中产生的振动信号包含了大量的设备状态信息,因此,基于振动的诊断方法是非常有效的机械故障诊断方法。然而,机械振动信号比较复杂,常表现为非平稳、多分量、多调制特征,且经常会受到噪声干扰。因此,研究一种合适的能够准确提取故障特征信息的信号分析处理方法具有重要的实际意义。自适应时频分析方法在信号分解中根据信号本身的特性自动选择基函数或参数,能够有效提取机械故障振动信号的本质特征,因此在机械故障诊断中得到了广泛的应用。自适应最稀疏时频分析(Adaptive and Sparsest Time-Frequency Analysis,ASTFA)方法是一种新的自适应时频分析方法,它以分解得到的单分量个数最少为优化目标,以单分量的瞬时频率具有物理意义为约束条件,通过求解最优化问题将信号自适应地分解为若干个内禀模态函数之和。ASTFA方法结合了EMD(Empirical Mode Decomposition,EMD)方法中将信号自适应地分解为若干个内禀模态函数之和与稀疏分解中在过完备字典库中寻优以获得信号稀疏分解的优点。论文在国家自然科学基金项目(编号:51375152)的资助下,对ASTFA方法的理论进行了研究与完善,并将ASTFA方法应用于机械设备故障诊断。论文的主要研究工作有:(1)对ASTFA方法进行了研究,将ASTFA方法与EMD、LCD(Local Characteristic-scale Decomposition,LCD)、LMD(Local Mean Decomposition,LMD)方法进行了对比,其分解结果的正交性、精确性等评价指标优于EMD、LCD、LMD方法;对ASTFA方法的分解能力进行了研究;对ASTFA方法的抗模态混叠性能进行了研究,仿真和转子故障信号的分析验证了ASTFA方法在抑制模态混淆方面的优势。(2)从瞬时频率计算对ASTFA方法进行理论解释。ASTFA方法与EMD、LCD、LMD等方法具有理论共性,都是将复杂信号自适应地分解为具有调幅部分和调频部分乘积形式的单分量信号之和。但是,ASTFA方法可以直接计算信号的瞬时频率,不用通过归一化的方法得到纯调频信号,再进行瞬时频率计算,避免了极值点处存在的波动和估计误差,具有明显的优越性。(3)从滤波器设计对ASTFA方法进行解释。ASTFA方法将信号由时间域转化到相位域,基于数据本身相位函数的驱动来设计自适应滤波器,即:通过选择合适的初始相位函数和平滑度控制参数来设计符合要求的滤波器,从而实现信号的自适应分解。(4)针对ASTFA方法初始相位函数的选择问题,提出了基于一维精确搜索的ASTFA方法和基于遗传算法(Genetic Algorithm,GA)的ASTFA方法。两种改进的ASTFA方法都是基于初始相位函数的最优选择而提出,解决了原始ASTFA方法中初始相位函数选择问题。(5)研究了ASTFA方法在齿轮故障诊断中的应用,提出了基于ASTFA的瞬时幅值谱和瞬时频率谱方法;提出了基于ASTFA的多尺度模糊熵偏均值(Partial Mean MFE,PMMFE)的齿轮故障诊断方法;研究了ASTFA方法在齿轮箱复合故障诊断中的应用;研究了ASTFA方法在行星齿轮箱故障诊断中的应用。(6)研究了ASTFA方法在变速工况下机械设备故障诊断中的应用。提出了基于ASTFA的阶次方法,基于ASTFA的阶次方法能够准确提取变速齿轮的故障特征信息;提出了基于ASTFA的广义解调方法,ASTFA方法可以提供广义解调方法所需的解调相位函数,解决了广义解调时频分析方法中相位函数的选择问题,基于ASTFA的广义解调方法能够有效提取变速工况下的滚动轴承故障特征信息。
[Abstract]:The state monitoring and fault diagnosis of mechanical equipment plays an important role in ensuring the healthy operation of the equipment. With the development of science and technology, modern machinery and equipment are becoming more and more complex. It is of great significance to study the new technology and new methods of mechanical fault diagnosis. The core of the mechanical fault diagnosis technology is the extraction of the fault features and the operation of the equipment. The vibration signals produced contain a lot of equipment state information. Therefore, the vibration based diagnosis method is a very effective method of mechanical fault diagnosis. However, the mechanical vibration signals are complex, often characterized by non stationary, multicomponent and multi modulation features, and often suffer from noise interference. Therefore, a suitable method can be used to accurately extract the vibration signals. The signal analysis and processing method of fault feature information is of great practical significance. The adaptive time frequency analysis method can automatically select the basic function or parameters according to the characteristics of the signal itself in the signal decomposition. It can effectively extract the essential characteristics of the mechanical fault vibration signal, so it has been widely used in the mechanical fault diagnosis. The Adaptive and Sparsest Time-Frequency Analysis (ASTFA) method is a new adaptive time-frequency analysis method. It takes the least number of single components of the decomposition as the optimization target, and the single component instantaneous frequency has physical meaning as the constraint condition. The signal is decomposed to the signal adaptively by solving the optimization problem. Several intrinsic modal functions and.ASTFA methods combine the EMD (Empirical Mode Decomposition, EMD) method to decompose the signal adaptively into the sum of several intrinsic modal functions and in the sparse decomposition in the overcomplete dictionary library to obtain the advantages of the signal sparse decomposition. 5152) the theory of ASTFA method is studied and perfected, and the ASTFA method is applied to the fault diagnosis of mechanical equipment. The main research work of this paper is as follows: (1) the method of ASTFA is studied, and the method of ASTFA is carried out with EMD, LCD (Local Characteristic-scale Decomposition, LCD), LMD (Local). In contrast, the evaluation indexes such as orthogonality and accuracy of the decomposition results are superior to the EMD, LCD and LMD methods; the decomposition ability of the ASTFA method is studied; the anti modal aliasing performance of the ASTFA method is studied. The simulation and the analysis of the rotor fault signal verify the advantage of the ASTFA method in the suppression of modal confusion. (2) from the instantaneous frequency meter. The theoretical explanation of the ASTFA method is based on the theoretical generality of the.ASTFA method and EMD, LCD, LMD and so on. All of these are the sum of a single component signal which adaptively decomposes the complex signal into the amplitude modulation part and the FM part product. However, the ASTFA method can directly calculate the instantaneous frequency of the signal without the normalization method. FM signals, then calculate the instantaneous frequency, avoid the fluctuation and estimation error at the extreme point, and have obvious superiority. (3) from the design of the filter design to the ASTFA method, the.ASTFA method transforms the signal from the time domain to the phase domain, and the adaptive filter is designed based on the phase function of the data itself, that is, through the selection of the filter. Select the appropriate initial phase function and the skid control parameter to design the filter that conforms to the requirement, thus realizing the adaptive decomposition of the signal. (4) in view of the selection of the initial phase function of the ASTFA method, the ASTFA method based on one dimensional precise search and the ASTFA method based on the genetic algorithm (Genetic Algorithm, GA) are proposed. The two improved methods are improved. The ASTFA method is based on the optimal selection of initial phase function, and solves the initial phase function selection problem in the original ASTFA method. (5) the application of ASTFA method in gear fault diagnosis is studied. The instantaneous amplitude spectrum and instantaneous frequency spectrum method based on ASTFA are proposed, and the multiscale fuzzy entropy partial mean (Pa) based on ASTFA (Pa) is proposed. The gear fault diagnosis method of rtial Mean MFE, PMMFE), the application of ASTFA method in the complex fault diagnosis of gear box, and the application of ASTFA in the fault diagnosis of the planetary gearbox. (6) the application of the ASTFA method in the fault diagnosis of the mechanical equipment under the variable speed condition is studied. The order method based on ASTFA, based on ASTF, is proposed. The order method of A can accurately extract the fault feature information of the transmission gear. A generalized demodulation method based on ASTFA is proposed. The ASTFA method can provide the demodulation phase function required by the generalized demodulation method, and solve the problem of the selection of the phase function in the generalized demodulation time frequency analysis method. The generalized demodulation method based on the ASTFA can be effectively extracted. Fault feature information of rolling bearing under variable speed conditions.
【学位授予单位】：湖南大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TH17

【相似文献】