Original Vibration Signal Research Articles

Rotor Fault Diagnosis Method Based on VMD Symmetrical Polar Image and Fuzzy Neural Network

Rotor fault diagnosis has attracted much attention due to its difficulties such as non-stationarity of fault signals, difficulty in fault feature extraction and low diagnostic accuracy of small samples. In order to extract fault feature information of rotors more effectively and to improve fault diagnosis precision, this paper proposed a fault diagnosis method based on variational mode decomposition (VMD) symmetrical polar image and fuzzy neural network. Firstly, the original rotor vibration signal is decomposed by using the VMD method and the relevant parameter selection algorithm of the VMD method is also proposed. Secondly, the intrinsic mode functions (IMF), which are sensitive to the signal characteristics, are selected for signal reconstruction based on a comprehensive evaluation factor method. As well, the reconstructed signal is transformed into a two-dimensional snowflake image through using the symmetrical polar coordinate method. Finally, the image features are extracted by the gray level co-occurrence matrix to form the state feature vector, which is input into the fuzzy neural network to realize the rotor fault diagnosis. Through the analysis of measured signals, the experimental results show that the proposed method can reach a higher recognition rate of 98% and the k-cross-validation experiment is used to demonstrate the robustness of the fuzzy neural network, and the average recognition accuracy of this experiment is 99.2%. Compared with some similar methods, the proposed method still has the highest fault recognition precision 98.4%, and the smallest standard deviation 0.5477.

Anti‐noise diesel engine misfire diagnosis using a multi‐scale CNN‐LSTM neural network with denoising module

AbstractCurrently, accuracy of existing diesel engine fault diagnosis methods under strong noise and generalisation performance between different noise levels are still limited. A novel multi‐scale CNN‐LSTM neural network (MSCNN‐LSTMNet) is proposed with a residual‐CNN denoising module for anti‐noise diesel engine misfire diagnosis. First, a residual‐CNN module is designed for denoising the original vibration signal measured from the diesel engine cylinder and residual loss for constructing a new loss function is utilised. Considering the essential characteristics of measured vibration signals at different scales, a multi‐scale convolutional NN (CNN) block is designed to realize multi‐scale feature extraction. Specifically, multiple convolution layers with different branches and different convolution kernel sizes are utilised to extract different time scales features, enhancing the robustness of the model. On this basis, the LSTM is utilised to further extract sequential features for improving anti‐noise and generalisation performances. The effectiveness of MSCNN‐LSTMNet is validated by experimental results of both one‐ and hybrid‐cylinder diesel engine misfires diagnosis under various noise levels and working conditions. The results demonstrate that MSCNN‐LSTMNet achieved much better anti‐noise and generalisation performances than the existing methods. Under strong noise conditions (−10 dB signal‐to‐noise ratio) for four datasets, MSCNN‐LSTMNet obtained 97.561% average accuracy, while average accuracy for random forest, deep neural network, CNN and MSCNNNet were 73.828%, 79.544%, 82.247%, and 89.741%, respectively. Moreover, for 11 noise generalisation tasks between different noise levels, MSCNN‐LSTMNet obtained at least 96.679%, 97.849%, 98.892%, and 94.010% accuracy on the four datasets, which are much higher than those of the existing methods.

Zero-Shot Compound Fault Diagnosis Method Based on Semantic Learning and Discriminative Features

Compound fault identification has always been a challenge in bearing fault diagnosis. Existing learning-based compound fault diagnosis methods require numerous labeled compound fault samples. However, it is impractical to collect numerous samples for each compound fault in industrial scenarios. Based on this, a Zero-shot compound fault diagnosis method based on semantic learning and discriminative features is proposed in this paper. Firstly, to solve the absence of fault semantic vectors, we construct novel single and compound fault semantics, which contain both manually constructed low-dimensional semantics from the original vibration signal and high-dimensional semantics learning from the Convolutional Autoencoder. Then a convolution neural network-based feature extractor is designed, and an adaptive edge center loss is defined on the feature extractor to maximize the inter-category distance and minimize the intra-category distance. Finally, the semantic vectors of the fault are embedded into the feature space, and the Euclidean distance is used to measure the distance between the fault features and the semantic vectors to identify the category of unknown compound fault. The proposed method is validated on a self-built testbed. The results showed that the accuracy of compound fault identification reached 78.74% in the absence of compound fault samples.

Multiscale Noise Reduction Attention Network for Aeroengine Bearing Fault Diagnosis

Aero-engine has a complex mechanical structure and its working condition are varying widely. Its fault signals are thus modulated through complex nonlinear transfer paths and influenced by non-Gaussian noises. However, popular data-driven multi-scale diagnostic model does not sufficiently consider the embedded noises, and consequently keeps the noise content in the advanced discriminative features, which decreases the diagnostic accuracy. Therefore, a multi-scale attention network with adaptive noise reduction (MANANR) is proposed in this paper for aero-engine bearing fault diagnosis. The MANANR firstly divides the original vibration signal into different scales via the average method of adjacent points. Then, two stacked multi-scale noise reduction modules B1 and B2 is designed for noise reduction. The core strategy behind B1 and B2 is the threshold noise reduction (TNR) which removes the noises from multi-scale convolution features adaptively. Based on the physical principles that global average pooling (GAP) and max pooling (MAP) can extract periodic and impulsive characteristics of the fault signals respectively, the threshold value of TRN is thus constructed through fusing GAP and MAP outputs. Furthermore, attention mechanism is employed to enhance the discriminative capability of multi-scale features by globally capturing the relations among different scales and channels. Finally, one two-layer classifier is introduced to confirm bearing fault patterns. Experimental results demonstrate that the proposed method has the feature-level noise reduction property, and more importantly achieves satisfying intelligent diagnosis precision for aero-engine bearing with the minimum peeling area of 0.5 mm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> under the continuous acceleration condition from 12000 rpm to 12550 rpm, which outperforms state-of-the-art intelligent diagnostic methods.

Fault diagnosis of printing press bearing based on deformable convolution residual neural network

&lt;abstract&gt; &lt;p&gt;Rolling bear is a major critical component of rotating machinery, as its working condition affects the performance of the equipment. As a result, the condition monitoring and fault diagnosis of bearings get more and more attention. However, the strong background noise makes it difficult to extract the bearing fault features exactly. Furthermore, regular gradient disappearance and overfit appear in traditional network model training. Therefore, taking the printing press bearings as the research object, an intelligent fault diagnosis method based on strong background noise is proposed. This method integrates frequency slice wavelet transform (FSWT), deformable convolution and residual neural network together, and realizes the high-precision fault diagnosis of the printing press bearings. First, FSWT is used to preprocess the original vibration signal to obtain bearing fault features in the time and frequency domain, reconstruct the signal in any frequency band and describe local features accurately. Second, the ResNet is selected as the base network, and the two-dimensional time-frequency diagrams (TFD) obtained by preprocessing are used as input. For the model that has a poor ability to extract subtle features under strong background noise, the deformable convolution layer is introduced to reconstruct the convolution layer of ResNet, called deformable convolution residual neural network (DC-ResNet). Finally, the effectiveness of this method is verified by using the data sets collected under experimental conditions and actual working conditions for fault diagnosis of the printing press. The results show that the DC-ResNet can classify different bearing faults under strong background noise, and the accuracy and stability are greatly improved, which the accuracy meets 93.90%. The intelligent fault diagnosis with high-precision of printing press bearings under complex working conditions is realized by the proposed method.&lt;/p&gt; &lt;/abstract&gt;

Research on frequency band sparse optimization algorithm in health index of rolling bearing

SummaryThe construction of health indicators of rolling bearing is significant for the monitoring of its health condition and residual life prediction. The main work is based on the frequency band sparse optimization algorithm in the rolling bearing health index. On the one hand, the index construction method based on machine learning model has poor interpretability; on the other hand, there are many interference components in the original vibration signal, extracting statistical features directly based on the original signal as a poor health indicator. For the above problems, this paper proposes a rolling bearing health index construction method based on sparse frequency band characterization. First, the signal including each independent sub‐band has been obtained by wavelet packet decomposition. Second, a sparse model is constructed by combining the sub‐band amplitudes and matrices decomposed by the wavelet packet in the fault state and the normal state, so as to select the frequency band sensitive to the fault information. Thereafter, the spectral amplitude sum of sensitive frequency band (SASSFB) is calculated as a health indicator that reflects the degradation information during the complete life operation of the rolling bearing. Finally, the effectiveness of the method is verified by utilizing the published bearing accelerated life experimental data set.

Damage Location Diagnosis of Frame Structure Based on a Novel Convolutional Neural Network

In the case of strong noise, when the damage occurs at different locations of the frame structure, the fault vibration signals generated are relatively close. It is difficult to accurately diagnose the specific location of the damage by using the traditional convolution neural network method. In order to solve this problem, this paper proposes a novel convolutional neural network. The method first uses wavelet decomposition and reconstruction to filter out the noise signal in the original vibration signal, then uses CEEMDAN (Complete Ensemble Empirical Mode Decomposition with Adaptive Noise Analysis) to decompose the filtered signal to highlight the feature information in the filtered signal. Finally, a convolution neural network combined with WDCNN (First Layer Wide Convolution Kernel Deep Convolution Neural Network) and LSTM (Long Short-Term Memory Network) is used to achieve the accurate classification of the signal, so as to achieve the accurate diagnosis of the damage location of the frame structure. Taking the four-story steel structure frame of Columbia University as the research object, the fault diagnosis method proposed in this paper is used to carry out experimental research under strong noise conditions. The experimental results show that the accuracy of the fault diagnosis method proposed in this paper can reach 99.97% when the signal-to-noise ratio is −4 dB and the objective function value is reduced to 10−4. Therefore, the fault diagnosis method proposed in this paper has a high accuracy in the strong noise interference environment; it can realize a high precision diagnosis of the damage location of the frame structure under a strong noise environment. The contribution and innovation of this paper is to propose a novel fault diagnosis method based on the convolutional neural network, which solves the problem of accurate damage location diagnosis of frame structures under strong noise environment.

Prediction of Fluid Accumulation in Engine Disk Cavity Based on Encoder-Decoder

A neural network model based on the encoder and decoder is proposed for aero-engine fluid accumulation. The model uses the attention mechanism to adjust the output weight of the encoder and then uses the decoder to predict the fluid volume. The vibration eigenvalues were obtained by fast Fourier transform of the original vibration signals, and the eigenvalues of vibration data were used as the input model to predict the volume of fluid in the disk cavity. To evaluate the effectiveness of the model, an aero-engine disk cavity fluid simulation experiment was built, and the data obtained from the experiment were used to test the method and the model. The prediction accuracy of the proposed model for the volume of engine disk cavity fluid was over 95%, which has an important engineering value for the reliable operation of an aero-engine.

Multisensor information fusion method for intelligent fault diagnosis of reciprocating compressor in shale gas development

To address the problems of the poor feature extraction ability and weak data generalization ability of traditional fault diagnosis methods in reciprocating shale gas compressor fault diagnosis applications, in this study, a fault diagnosis method for reciprocating shale gas was developed. This method uses a novel optimized learning method, free energy in persistent contrastive divergence, in deep belief network learning and training. It solves the problem of the deep belief network classification ability degradation in long-term training. The root mean square error is used as the fitness function to search for the optimal parameter combination of the DBN network by using the sparrow search algorithm. At the same time, the learning rate and batch size of the deep belief network, which have a large impact on the training error, are selected for optimization. Then, the original vibration signal is preprocessed by calculating 13 different time domain indicators, and feature-level data and decision-level data are fused in a parallel superposition method to obtain a fused time domain index dataset. Finally, combined with the powerful adaptive feature extraction and nonlinear mapping ability of deep learning, the constructed sample dataset is input to the deep belief network for training, and the deep belief network based on reciprocating shale gas compressor fault diagnosis model is established.

Vibration Image Representations for Fault Diagnosis of Rotating Machines: A Review

Rotating machine vibration signals typically represent a large collection of responses from various sources in a machine, along with some background noise. This makes it challenging to precisely utilise the collected vibration signals for machine fault diagnosis. Much of the research in this area has focused on computing certain features of the original vibration signal in the time domain, frequency domain, and time–frequency domain, which can sufficiently describe the signal in essence. Yet, computing useful features from noisy fault signals, including measurement errors, needs expert prior knowledge and human labour. The past two decades have seen rapid developments in the application of feature-learning or representation-learning techniques that can automatically learn representations of time series vibration datasets to address this problem. These include supervised learning techniques with known data classes and unsupervised learning or clustering techniques with data classes or class boundaries that are not obtainable. More recent developments in the field of computer vision have led to a renewed interest in transforming the 1D time series vibration signal into a 2D image, which can often offer discriminative descriptions of vibration signals. Several forms of features can be learned from the vibration images, including shape, colour, texture, pixel intensity, etc. Given its high performance in fault diagnosis, the image representation of vibration signals is receiving growing attention from researchers. In this paper, we review the works associated with vibration image representation-based fault detection and diagnosis for rotating machines in order to chart the progress in this field. We present the first comprehensive survey of this topic by summarising and categorising existing vibration image representation techniques based on their characteristics and the processing domain of the vibration signal. In addition, we also analyse the application of these techniques in rotating machine fault detection and classification. Finally, we briefly outline future research directions based on the reviewed works.

Cross-domain intelligent fault diagnosis of rolling bearing based on distance metric transfer learning

Rolling bearings are present ubiquitously in mechanical equipment, timely fault diagnosis has great significance in guaranteeing the safety of mechanical operation. In real world industrial applications, the distribution of training dataset (source domain) and testing dataset (target domain) is often different and varies with operating environment, which may lead to performance degradation. In this study, a cross-domain fault diagnosis of rolling bearing method based on distance metric transfer learning (DMTL) and wavelet packet decomposition (WPD) is proposed. The Mahalanobis distance is adopted for learning the intrinsic similarity or dissimilarity between instances and learned by simultaneously maximizing the intra-class distances and minimizing the inter-class distances for target domain. The features of source domain and target domain are first extracted from original vibration signals by WPD which is a powerful tool in dealing with non-stationary signals and can provide meticulous analysis. Then, the DMTL model is adopted to eliminate the error propagation across different components, which can weaken the weight of low-quality instances and enhance the weight of high-quality samples. Finally, the k-nearest neighbor (KNN) classifier is applied to accomplish the cross-domain intelligent fault-type classification. The superiority and effectiveness of the proposed fault diagnosis model is validated by two diagnosis cases. The experimental results demonstrated that the proposed method performs better than other compared methods in recognizing various fault types and has the capability in handling the complex cross-domain adaptation scenarios.

Fault diagnosis of planetary gears based on intrinsic feature extraction and deep transfer learning

The planetary gearbox is a key transmission apparatus used to change speed and torque. The planetary gear is one of the most failure-prone components in a planetary gearbox. Due to the complexity of working environments, collected vibration signals contain a lot of noise and interference; fault characteristic frequencies are usually submerged or even lost. Thus, feature extraction from the vibration signal is beneficial to subsequent fault diagnosis. As a fault identification approach that has been increasingly popular in the field of fault diagnosis, deep learning requires a large number of samples to train the model. Insufficient samples lead to low diagnostic accuracy for deep learning models. This paper proposes a novel fault diagnosis approach for planetary gears based on intrinsic feature extraction and deep transfer learning. The original vibration signals are decomposed into a series of band-limited intrinsic mode functions (BLIMFs) by variational mode decomposition. BLIMF with the most apparent fault characteristics is selected to generate two-dimensional time-frequency maps by continuous wavelet transform. The preprocessed time-frequency maps are adopted as the input of the pretrained VGG16 model. The bottom layers are frozen, and the top layers are fine-tuned to achieve fault diagnosis for planetary gears. Applications to planetary gear datasets verify the superiority of the proposed method.

Logistic-ELM: a novel fault diagnosis method for rolling bearings

The fault diagnosis of rolling bearings is a critical technique to realize predictive maintenance for mechanical condition monitoring. In real industrial systems, the main challenges for the fault diagnosis of rolling bearings pertain to the accuracy and real-time requirements. Most existing methods focus on ensuring the accuracy, and the real-time requirement is often neglected. In this paper, considering both requirements, we propose a novel fast fault diagnosis method for rolling bearings, based on extreme learning machine (ELM) and logistic mapping, named logistic-ELM. First, we identify 14 kinds of time-domain features from the original vibration signals according to mechanical vibration principles and adopt the sequential forward selection strategy to select optimal features from them to ensure the basic predictive accuracy and efficiency. Next, we propose the logistic-ELM for fast fault classification, where the biases in ELM are omitted and the random input weights are replaced by the chaotic logistic mapping sequence which involves a higher uncorrelation to obtain more accurate results with fewer hidden neurons. We conduct extensive experiments on the rolling bearing vibration signal dataset of the Case Western Reserve University bearing data centre. The experimental results show that the proposed approach outperforms existing state-of-the-art comparison methods in terms of the predictive accuracy, and the highest accuracies are 100%, 99.71%, 98%, 100%, 100%, and 100%, respectively, in seven separate sub data environments. Moreover, in terms of the runtime cost, the experimental results indicate that the proposed logistic-ELM can predict the fault in 40 ms with a high accuracy, up to 21-1858 times more rapid than existing methods based on support vector machine, convolutional neural network and multi-scale entropy. Other experiments of fault diagnosis of the rolling bearings under four different loads also indicate that the logistic-ELM can adapt to different operation conditions with high efficiency. The relevant code is publicly available at https://github.com/TAN-OpenLab/logistic-ELM .

Fault Diagnosis Method Based on AUPLMD and RTSMWPE for a Reciprocating Compressor Valve.

In order to effectively extract the key feature information hidden in the original vibration signal, this paper proposes a fault feature extraction method combining adaptive uniform phase local mean decomposition (AUPLMD) and refined time-shift multiscale weighted permutation entropy (RTSMWPE). The proposed method focuses on two aspects: solving the serious modal aliasing problem of local mean decomposition (LMD) and the dependence of permutation entropy on the length of the original time series. First, by adding a sine wave with a uniform phase as a masking signal, adaptively selecting the amplitude of the added sine wave, the optimal decomposition result is screened by the orthogonality and the signal is reconstructed based on the kurtosis value to remove the signal noise. Secondly, in the RTSMWPE method, the fault feature extraction is realized by considering the signal amplitude information and replacing the traditional coarse-grained multi-scale method with a time-shifted multi-scale method. Finally, the proposed method is applied to the analysis of the experimental data of the reciprocating compressor valve; the analysis results demonstrate the effectiveness of the proposed method.

Bearing Fault Diagnosis of Hot-Rolling Mill Utilizing Intelligent Optimized Self-Adaptive Deep Belief Network with Limited Samples.

Given the complexity of the operating conditions of rolling bearings in the actual rolling process of a hot mill and the difficulty in collecting data pertinent to fault bearings comprehensively, this paper proposes an approach that diagnoses the faults of a rolling mill bearing by employing the improved sparrow search algorithm deep belief network (ISAA-DBN) with limited data samples. First, the fast spectral kurtosis approach is adopted to convert the non-stationary original vibration signals collected by the acceleration sensors installed at the axial and radial ends of the rolling mill bearings into two-dimensional (2D) spectral kurtosis time–frequency images with higher feature recognition, and the principal component analysis (PCA) technique is used to decrease the dimension of the data in order to achieve a high diagnosis rate with a limited number of samples. Subsequently, the sparrow search algorithm (SSA) is used to realize the intelligent optimized self-adaptive function of a deep belief network (DBN). Furthermore, the firefly disturbance algorithm is employed to improve the spatial search capability and robustness of SSA-DBN in order to achieve better performance of the ISSA-DBN method. Finally, the proposed approach is experimentally compared to other approaches used for diagnosis. The results show that the proposed approach not only retains the useful features of the data through dimension reduction but also improves the efficiency of the diagnosis and achieves the highest diagnosis accuracy with limited data samples. In addition, the optimal position of the sensor for diagnosing rolling mill roll faults is identified.

Heterogeneous selective ensemble learning model for mill load parameters forecasting by using multiscale mechanical frequency spectrum

A ball mill is a heavy mechanical device and its safe operation affects the entire grinding process. Mill load is a key index in the optimum operation of the grinding process, but it cannot be measured directly. In industrial practice, operational experts normally estimate its value based on their experiences and the mechanical signals produced by the ball mill. In this paper, we proposed a heterogeneous selective ensemble method by using a multiscale mechanical frequency spectrum. The multicomponent adaptive decomposition algorithm is first used to decompose the original shell vibration and acoustic signals into sub-signals with different timescales. Then, selective ensemble (SEN) kernel projection to latent structure algorithm is used to model the spectral data of these sub-signals. Furthermore, the latent features of multiscale spectra are extracted to construct SEN models based on fuzzy inference. Finally, the two types of heterogeneous SEN models are fused by using information entropy. The main contribution of this study is that the proposed soft-sensing model has a dual-layer ensemble structure that can fuse multi-source information in different mechanical sub-signals with physical meaning. Moreover, the proposed model can simulate the fuzzy cognitive behavior of domain experts in the mineral grinding process. The effectiveness of the method is verified by the shell vibration and acoustical data of a laboratory-scale ball mill.

Automatic Transmission Bearing Fault Diagnosis Based on Comprehensive Index Method and Convolutional Neural Network

Rolling-element bearing fault diagnosis has some problems in the applied environment, such as low signal-to-noise ratio, weak feature extraction, low efficiency of feature learning and the complex structure of diagnosis models. A fault diagnosis method based on the comprehensive index method, complete ensemble empirical mode decomposition with adaptive noise independent component analysis (CEEMDANICA) and two-dimensional convolutional neural network (TDCNN) is proposed. Firstly, the original vibration signal of the bearing is preprocessed by CEEMDANICA, and the ICA components with different frequencies are obtained. Secondly, the ICA components are selected as the sample set by using multiscale permutation entropy, correlation coefficient, kurtosis and box dimension. Finally, the sample set are trained and tested by a DCNN model to realize the fault diagnosis of different bearing fault types. In order to verify the reliability of the method, a bearing fault vibration monitoring platform for an electric vehicle two-speed automatic transmission was built to collect the bearing vibration signals of multiple fault types under different working conditions. The diagnostic accuracy of several deep learning models is compared. The results show that the proposed method can realize the single and compound fault diagnosis of rolling-element bearings in an automatic transmission, with a high degree of accuracy.

Intelligent fault diagnosis of rolling mills based on dual attention- guided deep learning method under imbalanced data conditions

As an important link in the steel production chain, the health of the rolling mill directly affects the steel production. Therefore, the study of rolling mill fault diagnosis methods is of great significance to improve the continuity, reliability and safety of production. However, in the case of uneven data distribution, in order to improve the recognition performance, the traditional fault diagnosis method has developed the deep network architecture of convolutional neural network, which is not easy to obtain accurate fault characteristics and it is difficult to achieve better recognition accuracy. Aiming at these problems, we propose a rolling mill fault diagnosis method based on time–frequency image and Dual Attention-guided Feature Enhancement Network (DAFEN). First of all, the original one-dimensional vibration signal is converted into two-dimensional time–frequency images and used as the input of the network, and then the DAFAE is designed to analyze and integrate all convolutional features to complete the fault identification, in order to verify the superiority of the proposed method, we verified based on balanced datasets and imbalanced datasets, and our model was at least 0.71% and 1.43% higher than the highest accuracy fault classification results of other advanced CNN models.

A Two-Stage Method for Weak Feature Extraction of Rolling Bearing Combining Cyclic Wiener Filter with Improved Enhanced Envelope Spectrum

Due to the interference of various strong background signals, it is often difficult to extract effective features by using conventional methods such as envelope spectrum analysis when early weak fault arises in rolling bearing. Inspired by the current two main research directions of weak fault diagnosis of rolling bearing, that is, the enhancement of impulse features of faulty vibration signal through vibration analysis and the selection of fault information sensitive frequency band for further envelope spectrum analysis, and based on the second-order cyclostationary characteristic of the vibration signal of faulty bearing, a two-stage method for weak feature extraction of rolling bearing combining cyclic Wiener filter with improved enhanced envelope spectrum (IEES) is proposed in the paper. Firstly, the original vibration signal of the rolling bearing’s early weak fault is handled by cyclic Wiener filter exploiting the spectral coherence (SCoh) theory and the noise components are filtered out. Then, SCoh is applied on the filtered signal. Subsequently, an IEES method obtained by integrating over the selected fault information sensitive spectral frequency band of the SCoh spectral is used to extract the fault features. The innovation of the proposed method is to fully excavate the advantages of cyclic Wiener filter and IEES simultaneously. The feasibility of the proposed method is verified by simulation firstly, and vibration signals collected from accelerated bearing degradation tests and engineering machines are used to further verify its effectiveness. Additionally, its superiority over the other state-of-the-art methods is also compared.

A Novel Attentional Feature Fusion with Inception Based on Capsule Network and Application to the Fault Diagnosis of Bearing with Small Data Samples

Fault diagnosis of bearing with small data samples is always a research hotspot in the field of bearing fault diagnosis. To solve the problem, a convolutional block attention module (CBAM)-based attentional feature fusion with an inception module based on a capsule network (Capsnet) is proposed in the paper. Firstly, the original vibration signal is decomposed into multiple intrinsic mode function (IMF) sub-signals by the ensemble empirical mode decomposition (EEMD), and then the original vibration signal and the corresponding former four order IMF sub-signals are input into the inception modules to extract the features. Secondly, these features are concatenated and optimized by the CBAM. Finally, the selected sensitive features are fed into the Capsnet to diagnose the faults. Through the multifaceted experiment analysis on fault diagnosis of bearing with small data samples, the diagnosis results demonstrate that the proposed attentional feature fusion with inception based on Capsnet not only diagnoses the fault of bearing with small data samples, but also is superior to other feature fusion methods, such as feature fusion with inception based on Capsnet and attentional feature fusion with inception based on CNN, etc., and other single diagnosis models such as Capsnet with CBAM and inception, and CNN with CBAM and inception.