Fusion Domain-Adaptation CNN Driven by Images and Vibration Signals for Fault Diagnosis of Gearbox Cross-Working Conditions

The vibration signal of gearboxes contains abundant fault information, which can be used for condition monitoring. However, vibration signal is ineffective for some non-structural failures. In order to resolve this dilemma, infrared thermal images are introduced to combine with vibration signals via fusion domain-adaptation convolutional neural network (FDACNN), which can diagnose both structural and non-structural failures under various working conditions. First, the measured raw signals are converted into frequency and squared envelope spectrum to characterize the health states of the gearbox. Second, the sequences of the frequency and squared envelope spectrum are arranged into two-dimensional format, which are combined with infrared thermal images to form fusion data. Finally, the adversarial network is introduced to realize the state recognition of structural and non-structural faults in the unlabeled target domain. An experiment of gearbox test rigs was used for effectiveness validation by measuring both vibration and infrared thermal images. The results suggest that the proposed FDACNN method performs best in cross-domain fault diagnosis of gearboxes via multi-source heterogeneous data compared with the other four methods.

Median line-gram and its application in the fault diagnosis of rolling bearing

The level selection of frequency band division structure relies on previous information in most gram approaches that capture the optimal demodulation frequency band (ODFB). When an improper level is specified in these approaches, the fault characteristic information contained in the produced ODFB may be insufficient. This research proposes a unique approach termed median line-gram (MELgram) to tackle the level selection problem. To divide the frequency domain signal into a series of demodulation frequency bands, a spectrum median line segmentation model based on Akima interpolation is first created. The level and boundary of the segmentation model can be adaptively identified by this means. Second, the acquired frequency bands are quantized using the negative entropy index, and the ODFB is defined as the frequency band with the largest value. Third, the envelope spectrum is used to determine the ODFB characteristic frequency to pinpoint the bearing fault location. Finally, both simulation and experimental signal analysis are used to demonstrate the efficiency of the suggested method. Furthermore, the suggested method extracts more defect feature information from the ODFB than existing methods.

Rolling Bearing Fault Diagnosis Based on Component Screening Vector Local Characteristic-Scale Decomposition

The fault vibration signal of a bearing has nonstationary and nonlinear characteristics and can be regarded as the combination of multiple amplitude- and frequency-modulation components. The envelope of a single component contains the fault characteristics of a bearing. Local characteristic-scale decomposition (LCD) can decompose the vibration signal into a series of multiple intrinsic scale components. Some components can clearly reflect the running state of a bearing, and fault diagnosis is conducted according to the envelope spectrum. However, the conventional LCD takes a single-channel signal as the research object, which cannot fully reflect the characteristic information of the rotor, and the analysis results based on different channel signals of the same section will be inconsistent. To solve this problem, based on full vector spectrum technology, the homologous dual-channel information is fused. A vector LCD method based on cross-correlation coefficient component selection is given, and a simulation analysis is completed. The effectiveness of the proposed method is verified by simulated signals and experimental signals of a bearing, which provides a method for bearing feature extraction and fault diagnosis.

Fault diagnosis of rolling bearings based on improved direct fast iterative filtering and spectral amplitude modulation

As the transient impulse components in early fault signals are weak and easily buried by strong background noise, the fault features of rolling bearings are difficult to be extracted effectively. Focusing on this issue, a novel method based on improved direct fast iterative filtering and spectral amplitude modulation (IDFIF-SAM) is presented for detecting the early fault of rolling bearings. First, the ratio of the average crest factor of autocorrelation envelope spectrum to the average envelope entropy is taken as the fitness function to search the optimal parameters of direct fast iterative filtering (DFIF) adaptively via particle swarm optimization (PSO). Then, the efficient kurtosis entropy (EKE) index is being employed to choose the suitable components to reconstruct the signal. Finally, the reconstructed signal is subjected to spectral amplitude modulation (SAM) to strengthen the impulse features. The superiority of improved direct fast iterative filtering (IDFIF) over fixed-parameter DFIF, fast iterative filtering (FIF), and hard thresholding fast iterative filtering (HTFIF) is clarified through the simulated signal. Moreover, the comparative experimental analysis shows that the proposed IDFIF-SAM method can identify the early fault feature of rolling bearings more effectively.

Research on gearbox impact feature extraction method based on improved ESMD

A gearbox vibration signal contains non-linear impact characteristics and the significant feature information tends to be overwhelmed by other interference components, which make it difficult to extract the typical fault features fully and effectively. Aiming at the key issue of how to effectively extract the impact characteristics, a fault diagnosis method based on improved extreme symmetric mode decomposition (ESMD) and a support vector machine (SVM) is proposed in this paper. The vibration signal is adaptively decomposed into multiple intrinsic mode function (IMF) components by the improved ESMD and then a certain number of components are selected with the maximum kurtosis-envelope spectrum index. The singular spectral entropy, energy entropy and permutation entropy of each component are applied to construct the feature vector set, in which the dimensionality of the set is reduced with the distance separability criterion. Finally, the dimension-reduced feature vector set is input into the SVM for pattern recognition. Dynamic simulation and experimental gearbox research show that the improved ESMD method can extract and identify gearbox fault information effectively.

Weak Fault Feature Extraction of Rolling Bearing Based on SVMD and Improved MOMEDA

In order to solve the problem that it is very difficult to extract fault features directly from the weak impact component of early fault signal of rolling bearing, a method combining continuous variational mode decomposition (SVMD) with modified MOMEDA based on Teager energy operator is proposed. Firstly, the low resonance impulse component in the fault signal is separated from the harmonic component and noise by SVMD, and then the Teager energy operator is used to enhance the impulse feature in the low resonance component to ensure that the accurate fault period is selected by the MOMOEDA algorithm. After further noise reduction by MOMEDA, the envelope spectrum of the signal is analyzed, and finally the fault location is determined. The results of simulation and experimental data show that this method can accurately and effectively extract the characteristic frequency of rolling bearing weak fault.

Compound Fault Diagnosis of Rolling Bearing Based on ACMD, Gini Index Fusion and AO-LSTM

Due to the symmetry of the rolling bearing structure and the rotating operation mode, it will cause the coupling modulation phenomenon when it is damaged in multiple places at the same time, which makes it difficult to accurately identify all kinds of faults. For such problems, a compound fault diagnosis method based on adaptive chirp mode decomposition (ACMD), Gini index fusion and long short-term memory (LSTM) neural network optimized by Aquila Optimizer (AO) is proposed. Firstly, a series of IMF components are obtained by decomposing the vibration signal by means of ACMD, and the required components are selected by using the correlation coefficient method. Then, the Gini index of the square envelope (GISE) and the Gini index of the square envelope spectrum (GISES) of each component are calculated, respectively, and they are fused to construct a highly dimensional feature matrix. Then, with the aim of solving the problem of difficult selection of LSTM hyperparameters, the AO-LSTM model is constructed. Finally, the feature matrix is divided into a training set and a test set. The training set is input into the model for training, and then the training network is used to predict the test set, and outputs diagnostic results. The simulation and experimental results show that the proposed method can achieve higher accuracy and stronger robustness, compared with the existing intelligent diagnosis methods for bearing compound faults.