Life Prediction of Rolling Bearing Based on Optimal Time-Frequency Spectrum and DenseNet-ALSTM.

To address the challenges faced in the prediction of rolling bearing life, where temporal signals are affected by noise, making fault feature extraction difficult and resulting in low prediction accuracy, a method based on optimal time-frequency spectra and the DenseNet-ALSTM network is proposed. Firstly, a signal reconstruction method is introduced to enhance vibration signals. This involves using the CEEMDAN deconvolution method combined with the Teager energy operator for signal reconstruction, aiming to denoise the signals and highlight fault impacts. Subsequently, a method based on the snake optimizer (SO) is proposed to optimize the generalized S-transform (GST) time-frequency spectra of the enhanced signals, obtaining the optimal time-frequency spectra. Finally, all sample data are transformed into the optimal time-frequency spectrum set and input into the DenseNet-ALSTM network for life prediction. The comparison experiment and ablation experiment show that the proposed method has high prediction accuracy and ideal prediction performance. The optimization terms used in different contexts in this paper are due to different optimization methods, specifically the CEEMDAN method.

A Fast Sparse Decomposition Based on the Teager Energy Operator in Extraction of Weak Fault Signals.

In order to diagnose an incipient fault in rotating machinery under complicated conditions, a fast sparse decomposition based on the Teager energy operator (TEO) is proposed in this paper. In this proposed method, firstly, the TEO is employed to enhance the envelope of the impulses, which is more sensitive to frequency and can eliminate the low-frequency harmonic component and noise; secondly, a smoothing filtering algorithm was adopted to suppress the noise in the TEO envelope; thirdly, the fault signal was reconstructed by multiplication of the filtered TEO envelope and the original fault signal; finally, sparse decomposition was used based on a generalized S-transform (GST) to obtain the sparse representation of the signal. The proposed preprocessing method using the filtered TEO can overcome the interference of high-frequency noise while maintaining the structure of fault impulses, which helps the processed signal perform better on sparse decomposition; sparse decomposition based on GST was used to represent the fault signal more quickly and more accurately. Simulation and application prove that the proposed method has good accuracy and efficiency, especially in conditions of very low SNR, such as impulses with anSNR of -8.75 dB that are submerged by noise of the same amplitude.

Weak Fault Feature Extraction Method Based on Improved Stochastic Resonance.

Aiming at the problems of early weak fault feature extraction of bearings in rotating machinery, an improved stochastic resonance (SR) is proposed combined with the advantage of SR to enhance weak characteristic signals with noise energy. Firstly, according to the characteristics of the large parameters of the actual fault signal, the amplitude transform coefficient and frequency transform coefficient are introduced to convert the large parameter signal into small parameter signal which can be processed by SR, and the relationship of second-order parameters are introduced. Secondly, a comprehensive evaluation index (CEI) consisted of power spectrum kurtosis, correlation coefficient, structural similarity, root mean square error, and approximate entropy, is constructed through BP neural network. Moreover, this CEI is adopted as fitness function to search the optimal damping coefficient and amplitude transform coefficient with adaptive weight particle swarm optimization (PSO). Finally, according to the improved optimal SR system, the weak fault feature can be extracted. The simulation and experiment verify the effectiveness of the proposed method compared with traditional second-order general scale transform adaptive SR.

Empirical Variational Mode Decomposition Based on Binary Tree Algorithm.

Aiming at non-stationary signals with complex components, the performance of a variational mode decomposition (VMD) algorithm is seriously affected by the key parameters such as the number of modes K, the quadratic penalty parameter α and the update step τ. In order to solve this problem, an adaptive empirical variational mode decomposition (EVMD) method based on a binary tree model is proposed in this paper, which can not only effectively solve the problem of VMD parameter selection, but also effectively reduce the computational complexity of searching the optimal VMD parameters using intelligent optimization algorithm. Firstly, the signal noise ratio (SNR) and refined composite multi-scale dispersion entropy (RCMDE) of the decomposed signal are calculated. The RCMDE is used as the setting basis of the α, and the SNR is used as the parameter value of the τ. Then, the signal is decomposed into two components based on the binary tree mode. Before decomposing, the α and τ need to be reset according to the SNR and MDE of the new signal. Finally, the cycle iteration termination condition composed of the least squares mutual information and reconstruction error of the components determines whether to continue the decomposition. The components with large least squares mutual information (LSMI) are combined, and the LSMI threshold is set as 0.8. The simulation and experimental results indicate that the proposed empirical VMD algorithm can decompose the non-stationary signals adaptively, with lower complexity, which is O(n2), good decomposition effect and strong robustness.

Early fault feature extraction of bearings based on Teager energy operator and optimal VMD.

As the fault shock component in vibration signals is extremely sparse and weak, it is difficult to extract the fault features when large-scale, low-speed and heavy-duty mechanical equipment is in the early stage of failure. To solve this problem, an early fault feature extraction method based on the Teager energy operator, combined with optimal variational mode decomposition (VMD) is presented in this study. First, the Teager energy operator was used to strengthen the weak shock component of the original signal. Next, a logistic-sine complex chaotic mapping with variable dimensions was constructed to enhance the global search ability and convergence speed of the pigeon-inspired optimization (PIO) algorithm, which is named the variable dimension chaotic pigeon-inspired optimization (VDCPIO) algorithm. Then, the VDCPIO algorithm is used to search for the optimal combination value of key parameters of VMD. The enhanced vibration signal is decomposed into a set of intrinsic mode functions (IMFs) by the optimized VMD, and then kurtosis for every IMF and mean kurtosis of all IMFs are extracted. According to the average kurtosis, several IMFs, whose kurtosis value is greater than the average kurtosis value, are selected to reconstruct a new signal. Then, envelope spectrum analysis of the reconstructed signal is carried out to extract the early fault features. Finally, experimental verification of the method was performed using the simulated signal and measured signal from a rolling bearing; the experimental results indicate that the method presented in this paper is more effective to extract the early fault features of this kind of mechanical equipment.

Sparse decomposition method based on time-frequency spectrum segmentation for fault signals in rotating machinery.

The impulse signal in large rotating machinery with damage fault is sparse, weak, coupled, and even nonperiodic in intermittent operation. To extract this complex signal is a key topic in machinery fault diagnosis. Sparse decomposition (SD) has excellent adaptability in describing arbitrary complex signals based on over-complete dictionary. However, the pursuit speed of best atom is a serious drawback. To alleviate this, a method of sparse decomposition based on time-frequency spectrum segmentation (SD-TFSS) is introduced. Generalized S transform (GST) provides the capability to show the distribution of vibration signals, but the resolution is susceptible to noise, multiresolution generalized S-transform (MGST) is developed to generate multiresolution time-frequency spectrums. Then, spectrums fusion with an appropriate threshold is adopted to acquire multiresolution binary spectrums and produce an optimal binary spectrum. From this optimal binary spectrum, all the connectivity areas are extracted and marked by spectrum segmentation. Thus, an optimal library can be constructed by selecting the optimal atoms of every connectivity area, and the signal can be expressed with this library. We conduct simulations and experiments demonstrating that the proposed method performs well with lower pursuit complexity, higher decomposition efficiency, and better approximation precision.