Blind source separation based vibration mode identification

In this paper, a novel method for linear normal mode (LNM) identification based on blind source separation (BSS) is introduced. Modal coordinates are considered as a specific case of sources that have certain time structure. This structure makes modal coordinates identifiable by many BSS algorithms. However, algorithms based on second-order statistics are particularly suited for extracting LNMs of a vibration system. Two well-known BSS algorithms are considered. First, algorithm for multiple unknown signals extraction (AMUSE) is used to illustrate the similarity with Ibrahim time domain (ITD) modal identification method. Second, second order blind identification (SOBI) is used to demonstrate noise robustness of BSS-based mode shape extraction. Numerical simulations and experimental results from these BSS algorithms and ITD method are presented.     

高速列车非平稳振动信号盲源分离方法及应用

高速列车具有若干时变激励源,传统的时频分析方法只能对观测的混合振动的总体强度分布、时频域结构加以分析,不能分离出与各振源对应的信号分量从而明晰振源状态与故障特征。盲源分离是一种可行的分析方法,但由于高速列车振动信号具有时变振源数目、时变信号长度、受车速调制的变频非平稳等特征,传统的盲源分离方法不适用。为了提高高速列车非平稳信号的盲源分离效果,基于自适应滤波理论提出全局最优信噪比盲源分离新方法,并对其可分离性的判别依据进行论证。新方法的有效性经仿真计算和实测数据分析得到验证。研究表明：新方法对高速列车时变非平稳信号的盲源分离效果优于传统的基于非线性函数的盲源分离方法和基于高阶累积量的盲源分离方法。     

Principal component analysis and blind source separation of modulated sources for electro-mechanical systems diagnostic

Under the only hypothesis of independent sources, blind source separation (BSS) consists of recovering these sources from several observed mixtures of them. As it extracts the contributions of the sources independently of the propagation medium, this approach is usually used when it is too difficult to modelise the transfer from the sources to the sensors. In that way, BSS is a promising tool for non-destructive machine condition monitoring by vibration analysis. Principal component analysis (PCA) is applied as a first step in the separation procedure to filter out the noise and whiten the observations. The crucial point in PCA and BSS methods remains that the observations are generally assumed to be noise-free or corrupted with spatially white noises. However, vibration signals issued from electro-mechanical systems as rotating machine vibration may be severely corrupted with spatially correlated noises and therefore the signal subspace will not be correctly estimated with PCA.This paper extends a robust-to-noise technique earlier developed for the separation of rotating machine signals. It exploited spectral matrices of delayed observations to eliminate the noise influence. In this paper, we focus on the modulated sources and prove that the proposed PCA is available to denoise such sources as well as sinusoidal ones. Finally, performance of the algorithm is investigated with experimental vibration data issued from a complex electro-mechanical system.     

航空发动机转子振动信号的分离测试技术

在传统谱分析方法的基础上,尝试应用盲源分离技术对飞机发动机振动信号进行振源分离.首先,介绍了发动机振动信号的基本处理方法和常见的发动机故障类型及特征,引入了盲源分离理论并讨论了其在航空发动机振动信号处理中应用的可行性.然后,对某型涡扇发动机振动过大的现象进行了故障诊断分析.最后,应用FastICA和JADE算法对检测的振动信号进行分析,分离出了发动机的振源信号.这说明发动机振动信号分析采用盲源分离与谱分析相结合的技术可以有效分离振源信号,提高故障诊断的准确性.     

Source separation of diesel engine vibration based on the empirical mode decomposition and independent component analysis

Vibration signals from diesel engine contain many different components mainly caused by combustion and mechanism operations,several blind source separation techniques are available for decomposing the signal into its components in the case of multichannel measurements,such as independent component analysis(ICA).However,the source separation of vibration signal from single-channel is impossible.In order to study the source separation from single-channel signal for the purpose of source extraction,the combination method of empirical mode decomposition(EMD) and ICA is proposed in diesel engine signal processing.The performance of the described methods of EMD-wavelet and EMD-ICA in vibration signal application is compared,and the results show that EMD-ICA method outperforms the other,and overcomes the drawback of ICA in the case of single-channel measurement.The independent source signal components can be separated and identified effectively from one-channel measurement by EMD-ICA.Hence,EMD-ICA improves the extraction and identification abilities of source signals from diesel engine vibration measurements.     

盲分离与Hilbert-Huang结合应用于航空发动机振动信号分析

在航空发动机故障诊断中,首要任务是分析故障信号提取故障特征。针对航空发动机非平稳振动信号,提出了利用盲分离(BSS)获得发动机的振源信号,结合Hilbert-Huang变换(HHT)对振源信号进行时频分析提取故障特征的方法。首先利用仿真信号验证了此方法的有效性,然后分析了某航空涡扇发动机空中停车故障并与直接应用HHT分析的结果进行比较,证实了盲分离与HHT的结合能更准确地提取航空发动机非平稳故障特征。     

基于盲源分离技术的故障特征信号分离方法

信号采集过程中，传感器测量到的信号是实际振动信号在此测量方向的投影值，由于其他不相干振源的影响，测量信号由多个振动信号成分组成。在分析多振源信号混合模型的基础上，采用盲源分离技术分离不同的振源信号，讨论分离结果的广义初等相等性质的影响，研究估计振源数目的方法和选取测量信号的方法，利用二阶特征矩阵联合近似对角化算法，从测量信号中分离故障特征源信号。该算法可减小信号采集不当造成的影响，有效提高特征信号的提取。     

基于鲁棒预白化的机械噪声信号盲分离

机械噪声故障诊断的难度在于实际检测的噪声是多个设备或零部件噪声信号的混合,信噪比低,基于二阶统计量盲源分离算法的故障噪声诊断技术,利用二阶协方差矩阵的联合对角化,从测量噪声中分离出感兴趣故障噪声进而提取特征,但该算法抗干扰噪声性能差。本文利用多个协方差矩阵平滑滤波后的矩阵进行白化,进一步提高了抗干扰噪声能力,在样本数据较少时仍能实现较好的盲源分离效果,仿真实验证实了该算法的有效性。     

Separation of time variant vibration sources by short time coherent output power

This effort describes the use of time variant coherence causality based analysis to separate the effects of nonstationary time variant vibration excitation sources. A time variant coherence function using the Short Time Fourier Transform (STFT) is first discussed. The concept of a time variant coherent output power for source separation of systems with time variant transfer functions is developed. A parametric study is performed to examine the coherent output power separation capabilities with respect to the data processing parameters. The study guided the selection of the time-frequency processing parameters judged to provide a suitable compromise between the time event localization and output amplitude source separation. The time variant coherent output power is then applied to separate the effects of the two possible excitation sources on the prototype vibration isolation floor. The application was a subscale prototype isolation floor for a proposed vibration sensitive equipment site adjacent to a busy freight rail line. The moving train created time variant transmission paths. As there was a direct line of sight between the prototype floor and the rail line there was an airborne acoustic excitation path in addition to a ground path. The short time coherent output power was applied to separate prototype isolation floor vibration into respective components related to the two candidate sources. The analysis and discussion of the results focuses upon the interpretation and issues in such a complicated realistic environment. Ultimately the application was successful providing an explanation as to why the observed vibration isolation was degraded at higher frequencies.     

Application of the blind source separation method to feature extraction of machine sound signals

As the result of vibration emission in air, a machine sound signal carries important information about the working condition of machinery. But in practice, the sound signal is typically received with a very low signal-to-noise ratio. To obtain features of the original sound signal, uncorrelated sound signals must be removed and the wavelet coefficients related to fault condition must be retrieved. In this paper, the blind source separation technique is used to recover the wavelet coefficients of a monitored source from complex observed signals. Since in the proposed blind source separation (BSS) algorithms it is generally assumed that the number of sources is known, the Gerschgorin disk estimator method is introduced to determine the number of sound sources before applying the BSS method. This method can estimate the number of sound sources under non-Gaussian and non-white noise conditions. Then, the partial singular value analysis method is used to select these significant observations for BSS analysis. This method ensures that signals are separated with the smallest distortion. Afterwards, the time-frequency separation algorithm, converted to a suitable BSS algorithm for the separation of a non-stationary signal, is introduced. The transfer channel between observations and sources and the wavelet coefficients of the source signals can be blindly identified via this algorithm. The reconstructed wavelet coefficients can be used for diagnosis. Finally, the separation results obtained from the observed signals recorded in a semi-anechoic chamber demonstrate the effectiveness of the presented methods .