变转速下旋转机械瞬态特征表示和提取与故障诊断研究
发布时间：2018-08-23 20:06
【摘要】：旋转机械在机械设备中占有举足轻重的地位，它们大多数为生产企业中的关键设备，因此，保证旋转机械的安全可靠运行对企业和国民经济有重要的意义。轴承、齿轮等旋转零部件在机械系统中起着支撑、定位和传递动力等作用，是旋转设备的关键部件，当它们出现局部故障时，在运行中振动信号中会出现瞬态冲击响应成分，对旋转机械瞬态特征的表示和提取关系到故障诊断的可靠性与准确性，是机械设备故障诊断的关键问题。本文以轴承和齿轮两种典型的零部件为研究对象，针对变转速下旋转机械瞬态特征表示与提取的问题提出了基于时频特征融合的时频特征表示方法、时频特征重采样方法、周期特征的极坐标同步增强检测方法和变转速下瞬态特征的自动检测方法，并分别就有关问题进行了深入的理论研究和应用研究。 介绍了轴承和齿轮常见的失效形式以及滚动轴承的运动学，并进行了轴承的局部故障实验设置，对恒定转速和变转速的振动信号进行了采集，保证理论研究工作都建立在实验验证的基础上。 现有的两类时频分析方法，线性时频分析和双线性时频分析，线性时频分析没有交叉项但聚集性差，双线性时频分析聚集性好但存在交叉项，基于此基础提出了一种基于线性时频分析和双线性时频分析融合的策略，此方法类似于逻辑与，对时频特征融合，得到了一种新的时频特征表征方法。对仿真信号以及轴承和齿轮实测振动信号的应用验证了该方法对轴承和齿轮故障特征提取的有效性和适用性。 在变转速下，瞬态成分不再是等周期的，而是和转速有关。在实际应用中，对于变转速下振动信号的处理，阶比分析是一种非常有效的手段，对时域信号重采样使其在阶域上满足傅里叶分析的要求。由于旋转部件局部故障引起的瞬态成分在高频处趋向于振荡阻尼振动，对瞬态成分重采样会不可避免的引起失真。考虑到瞬态成分在时频域上是能量分布，不存在明显的振荡，因此提出了时频特征重采样方法，，其本质是将对信号时域上的重采样扩展到时频域上对时频特征值的重采样，可以将非等周期的时频特征转化成等周期的时频特征。 对于变转速下地局部故障，通常会存在某一瞬态角度特征周期，而且通过时频特征重采样，瞬态特征将会等周期的表示在角度频率域，然而，这仍然很难获得角频域上的角度特征周期和故障类型之间的关系。本文提出了周期特征的极坐标同步增强检测方法。将时频特征重采样和周期特征的极坐标同步增强应用到轴承的外圈、内圈和滚动体三种典型故障中，可以根据极坐标上的特征分布有效的识别故障类型。 针对变转速下机械设备故障诊断问题，提出了一种变转速下旋转机械瞬态特征的自动检测方法，是对时频特征融合策略、时频特征重采样和周期特征的极坐标同步增强的结合。首先，对变转下测得的振动信号应用Wigner-ville分布和小波尺度谱这两种不同的时频分析方法，在这两种时频表示的基础上应用融合策略，可以将变转下的瞬态特征很好的表示在时频面上。然后，对得到的时频特征矩阵应用时频特征重采样，将非等周期的瞬态特征变成了角频域上的等周期瞬态特征。最后，应用周期特征的极坐标同步增强，可以得到变转下振动信号瞬态特征的增强表示。仿真信号以及轴承实测振动信号的应用验证了该方法的有效性和适用性。 本论文依托于“变工况下旋转设备轻微局部故障的特征增强检测与诊断方法研究”国家自然科学基金青年基金项目（批准号：50905121）。
[Abstract]:Rotating machinery plays an important role in machinery and equipment. Most of them are key equipment in manufacturing enterprises. Therefore, it is of great significance for enterprises and national economy to ensure the safe and reliable operation of rotating machinery. The key components of the equipment, when they have local faults, will appear transient impulse response components in the vibration signals in operation. The representation and extraction of the transient characteristics of rotating machinery is related to the reliability and accuracy of fault diagnosis, and is the key problem of mechanical equipment fault diagnosis. Aiming at the problem of instantaneous feature representation and extraction of rotating machinery under variable speed, the time-frequency feature representation method based on time-frequency feature fusion, the time-frequency feature resampling method, the polar coordinate synchronization enhancement detection method of periodic feature and the automatic detection method of instantaneous feature under variable speed are proposed. Deep theoretical research and applied research.
The common failure modes of bearings and gears and the kinematics of rolling bearings are introduced. The local fault experiment settings of bearings are carried out. The vibration signals of constant and variable speed are collected to ensure that the theoretical research work is based on the experimental verification.
There are two kinds of time-frequency analysis methods, linear time-frequency analysis and bilinear time-frequency analysis. Linear time-frequency analysis has no cross-terms but poor aggregation. Bilinear time-frequency analysis has good aggregation but cross-terms. Based on this, a fusion strategy based on linear time-frequency analysis and bilinear time-frequency analysis is proposed. This method is similar to logic and bilinear time-frequency analysis. A new time-frequency feature representation method is proposed for time-frequency feature fusion. The validity and applicability of this method for bearing and gear fault feature extraction are verified by the application of simulation signal and bearing and gear vibration signal.
In practical application, order analysis is a very effective method for vibration signal processing under variable rotational speed. The time domain signal is resampled to meet the requirements of Fourier analysis in order domain. The resampling of transient components will inevitably lead to distortion when they tend to oscillate and damp vibration at high frequencies. Considering that the transient components are energy-distributed in time-frequency domain and there is no obvious oscillation, a time-frequency characteristic resampling method is proposed. The essence of this method is to extend the resampling of signals in time-domain to the time-frequency domain and to the time-frequency characteristic values. The resampling method can transform the non equal period time-frequency characteristics into equal cycle time-frequency characteristics.
For local faults with variable speed, there is usually a transient angle characteristic period, and by resampling the time-frequency characteristics, the transient characteristics will be expressed in the angle-frequency domain. However, it is still difficult to obtain the relationship between the angle characteristic period and the fault type in the angle-frequency domain. The time-frequency feature resampling and periodic feature polar coordinate synchronization enhancement are applied to three typical faults of bearing, i.e. outer ring, inner ring and rolling element. Fault types can be effectively identified according to the distribution of features in polar coordinates.
Aiming at the fault diagnosis of rotating machinery under variable rotating speed, an automatic detection method for the transient characteristics of rotating machinery under variable rotating speed is proposed, which combines time-frequency feature fusion strategy, time-frequency feature resampling and polar coordinate synchronization enhancement of periodic features. Firstly, Wigner-ville distribution and wavelet ruler are applied to the vibration signals measured under variable rotating speed. Based on these two time-frequency representations, the fusion strategy can be used to represent the transient characteristics in the time-frequency plane. Then, the time-frequency characteristic resampling is applied to the obtained time-frequency characteristic matrix, and the non-periodic transient characteristics are transformed into the equal-periodic transient characteristics in the angular-frequency domain. Finally, the polar coordinate synchronous enhancement of periodic characteristics can be used to enhance the transient characteristics of vibration signals under variable rotation. The effectiveness and applicability of this method are verified by the application of simulation signals and bearing vibration signals.
This paper relies on the project of National Natural Science Foundation of China (Grant No. 50905121) of "Research on Feature Enhanced Detection and Diagnosis Method for Slight Local Faults of Rotating Equipment under Variable Working Conditions".
【学位授予单位】：苏州大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TH165.3
本文编号：2199782
[Abstract]:Rotating machinery plays an important role in machinery and equipment. Most of them are key equipment in manufacturing enterprises. Therefore, it is of great significance for enterprises and national economy to ensure the safe and reliable operation of rotating machinery. The key components of the equipment, when they have local faults, will appear transient impulse response components in the vibration signals in operation. The representation and extraction of the transient characteristics of rotating machinery is related to the reliability and accuracy of fault diagnosis, and is the key problem of mechanical equipment fault diagnosis. Aiming at the problem of instantaneous feature representation and extraction of rotating machinery under variable speed, the time-frequency feature representation method based on time-frequency feature fusion, the time-frequency feature resampling method, the polar coordinate synchronization enhancement detection method of periodic feature and the automatic detection method of instantaneous feature under variable speed are proposed. Deep theoretical research and applied research.
The common failure modes of bearings and gears and the kinematics of rolling bearings are introduced. The local fault experiment settings of bearings are carried out. The vibration signals of constant and variable speed are collected to ensure that the theoretical research work is based on the experimental verification.
There are two kinds of time-frequency analysis methods, linear time-frequency analysis and bilinear time-frequency analysis. Linear time-frequency analysis has no cross-terms but poor aggregation. Bilinear time-frequency analysis has good aggregation but cross-terms. Based on this, a fusion strategy based on linear time-frequency analysis and bilinear time-frequency analysis is proposed. This method is similar to logic and bilinear time-frequency analysis. A new time-frequency feature representation method is proposed for time-frequency feature fusion. The validity and applicability of this method for bearing and gear fault feature extraction are verified by the application of simulation signal and bearing and gear vibration signal.
In practical application, order analysis is a very effective method for vibration signal processing under variable rotational speed. The time domain signal is resampled to meet the requirements of Fourier analysis in order domain. The resampling of transient components will inevitably lead to distortion when they tend to oscillate and damp vibration at high frequencies. Considering that the transient components are energy-distributed in time-frequency domain and there is no obvious oscillation, a time-frequency characteristic resampling method is proposed. The essence of this method is to extend the resampling of signals in time-domain to the time-frequency domain and to the time-frequency characteristic values. The resampling method can transform the non equal period time-frequency characteristics into equal cycle time-frequency characteristics.
For local faults with variable speed, there is usually a transient angle characteristic period, and by resampling the time-frequency characteristics, the transient characteristics will be expressed in the angle-frequency domain. However, it is still difficult to obtain the relationship between the angle characteristic period and the fault type in the angle-frequency domain. The time-frequency feature resampling and periodic feature polar coordinate synchronization enhancement are applied to three typical faults of bearing, i.e. outer ring, inner ring and rolling element. Fault types can be effectively identified according to the distribution of features in polar coordinates.
Aiming at the fault diagnosis of rotating machinery under variable rotating speed, an automatic detection method for the transient characteristics of rotating machinery under variable rotating speed is proposed, which combines time-frequency feature fusion strategy, time-frequency feature resampling and polar coordinate synchronization enhancement of periodic features. Firstly, Wigner-ville distribution and wavelet ruler are applied to the vibration signals measured under variable rotating speed. Based on these two time-frequency representations, the fusion strategy can be used to represent the transient characteristics in the time-frequency plane. Then, the time-frequency characteristic resampling is applied to the obtained time-frequency characteristic matrix, and the non-periodic transient characteristics are transformed into the equal-periodic transient characteristics in the angular-frequency domain. Finally, the polar coordinate synchronous enhancement of periodic characteristics can be used to enhance the transient characteristics of vibration signals under variable rotation. The effectiveness and applicability of this method are verified by the application of simulation signals and bearing vibration signals.
This paper relies on the project of National Natural Science Foundation of China (Grant No. 50905121) of "Research on Feature Enhanced Detection and Diagnosis Method for Slight Local Faults of Rotating Equipment under Variable Working Conditions".
【学位授予单位】：苏州大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TH165.3
【引证文献】
相关博士学位论文 前1条
1 郭瑜;基于时—频分析的虚拟式旋转机械特征分析仪系统的研究[D];重庆大学;2003年
相关硕士学位论文 前2条
1 陈祥芹;振动信号多元统计分析特征提取及传动系统关键部件故障诊断应用[D];苏州大学;2014年
2 李刚;基于虚拟仪器技术的逆变器监测系统设计[D];青岛大学;2014年
本文编号：2199782
本文链接：https://www.wllwen.com/kejilunwen/jixiegongcheng/2199782.html
