A Method Of Rotating Machinery Fault Diagnosis Based On Adaptive And Sparsest Time-frequency Analysis

The fault diagnosis of rotating machinery is actually extract fault characteristic information effectively. Now the fault feature extraction methods based on time-frequency analysis methods are widely used, Such as Empirical Mode Decomposition(EMD), Local Mean Decomposition(LMD), etc. However, there methods still have some problems difficult to solve, therefore, the study of new time-frequency analysis methods is still a hotspot in the field of rotating machinery fault diagnosis. Inspired by EMD and compressed sensing theory, Adaptive and Sparsest Time-Frequency Analysis(ASTFA) method was proposed by THOMAS Y. HOU and Zuoqiang SHI, and it was used to the processing of meteorological data. Based on the characteristic of intrinsic sparse time-frequency distribution of the multi-scale data, ASTFA regards minimum numbers of components as its goal of optimization and decompose the signal in this process of target optimization. Supported by National Natural Science Foundation(No. 51375152), ASTFA method is applied to the fault diagnosis of rotating machinery in this paper.This paper's main research contents are as follows:1. Several time-frequency analysis methods widely used are discussed, and ASTFA method is mainly studied. ASTFA method is proved to be effective through the contrastive analysis of simulation signals; then ASTFA method is introduced into the fault diagnosis of gears and rolling bearing, the experimental results show that ASTFA method is practical in fault diagnosis of rotating machinery.2. Aiming at inaccurate initial value selection and mode mixing of ASTFA, the paper puts forward an Improved Adaptive and Sparsest Time-Frequency Analysis(IASTFA) method. In this method, Particle Swarm Optimization algorithm is used to select the adaptive optimal initial value of ASTFA and frequency masking signal is added into decomposition signal to limit signal bandwidth. Simulated and experimental results prove that the improved method is of great effect.3. Aiming at compound fault diagnosis of rotating machinery, IASTFA method is applied to the single-channel Blind Source Separation, and IASTFA-BSS is proposed to the compound fault diagnosis of rotating machinery. In the decomposition of single-channel signal, underdetermined blind source separation is transformed to positive definite blind source separation through IASTFA method, and then do blind signal separation. The results of simulation and experiment show that ASTFA-BSS method is practical in compound fault diagnosis of rotating machinery.4. IASTFA is used to the feature extraction of rolling bearings, and compared with Quantile Regression-Variable Predictive Mode based Class Discriminate(QRVPMCD), a fault diagnosis method of rolling bearing is proposed based on ASTFA and QRVPMCD. In this method, the vibration signal of rolling bearings is decomposed into several mono-component signals firstly by IASTFA, and then the Hilbert spectrum singular values are extracted from the mono-component signals and formed into fault feature vector, which can be used as input of QRVPMCD for rolling bearing fault diagnosis. The analysis results from different working conditions and failures of roller bearing demonstrate the effectiveness of the proposed method.