Fault Feature Extraction Research Articles

A multi-step loss meta-learning method based on multi-scale feature extraction for few-shot fault diagnosis

Existing deep learning (DL) algorithms are based on a large amount of training data and they face challenges in effectively extracting fault features when dealing with few-shot fault diagnoses. Model-agnostic meta-learning (MAML) also faces some challenges, including the limited capability of the basic convolutional neural network (CNN) with a single convolutional kernel to extract fault features comprehensively, as well as the instability of model training due to the inner and outer double-layer loops. To address these issues, this paper presents a multi-step loss meta-learning method based on multi-scale feature extraction (MFEML). Firstly, an improved multi-scale feature extraction module (IMFEM) is designed to solve the problem of the insufficient feature extraction capability of the CNN. Secondly, the multi-step loss is used to reconstruct the meta-loss to address the issue of MAML training instability. Finally, the experimental results of two datasets demonstrate the effectiveness of the MFEML.

A deep residual neural network model for synchronous motor fault diagnostics

Synchronous motors play a significant role in a wide range of industrial applications. Their reliable operation is paramount. Any faults in synchronous motors can lead to costly downtime, decreased productivity, and potential safety hazards. By accurately diagnosing and classifying faults, we can proactively address issues before they escalate, ensuring the smooth operation of synchronous motors and minimizing the risk of equipment failure. The accurate diagnosis and fault detection in synchronous motors pose a significant challenge in their operation and maintenance. In the existing models, the feature data at various depths are not thoroughly extracted to maximize their feature extraction capability. Additionally, they employ a single support vector machine to make the final decision on the output. The single support vector machine may not consistently produce more accurate outcomes. Therefore, this paper proposes a novel fault diagnosis model based on a deep residual neural network and multiple support vector machines to diagnose mechanical and electrical faults of synchronous motors. The proposed model improves upon existing fault diagnosis models in two key aspects. Firstly, by employing a deep neural network, the model is able to effectively process and extract fault features from the motor fault dataset, capturing more nuanced information that may be missed by existing models. Secondly, the use of multiple support vector machines enhances the decision-making capability of the model, allowing for more accurate fault diagnosis. By combining these two aspects, the proposed model achieves superior diagnostic performance compared to single support vector machine-based models. Our proposed model has been rigorously evaluated using mechanical and electrical fault datasets, and the results of experimental tests clearly demonstrate its superior diagnostic performance when compared to existing fault diagnosis models. The synergy of deep neural networks and multiple support vector machines not only improves fault detection accuracy but also enhances the robustness and generalizability of the model, making it a valuable tool for real-world industrial applications.

Incipient Bearing Fault Extraction based on an Adaptive Multi-stage Noise Reduction Method

Considering the strong nonlinear and non-stationary characteristics of rolling bearing vibration signals, this paper proposes a multi-stage noise reduction method using adaptive variational mode decomposition and modulation signal bispectrum (AVMD-MSB) to extract the fault features of rolling bearings. Firstly, the AVMD is employed to adaptively select VMD parameters K and α and decompose the signal into a series of Intrinsic mode functions (IMFs), which allows an adaptive selection of the parameters of VMD. Then, all IMF components are reconstructed with weights according to the index of correlation kurtosis to avoid accidental omission of the IMFs containing important fault information. Finally, MSB is implemented to further suppress residual noises and interference components in the signal, precisely extract the bearing fault features. Numerical simulation and case study show that the AVMD-MSB is more advantageous in extracting fault characteristics from rolling bearing vibration signals compared with AVMD-Envelope and conventional VMD-MSB.

Vibration signals reconstruction with diffusion model for bearing fault diagnosis

Rotating bearing plays an important role in many mechanical equipment operation, such as satellite in orbit. There are in urgent need of constructing effective prognosis and diagnosis system oriented rotating bearing. Traditional bearing fault diagnosis methods enhance fault feature with sparsity-assisted prior, which are not suitable for complex vibration signals. Meanwhile, it is difficult to extract fault feature from under-sampled vibration signals. In this paper, we firstly utilize generation diffusion model to learn probability distribution of bearing fault data. As a deep generation prior, we combine the diffusion model and reconstruction problem with regularization term. The unsupervised learning method is not restricted to specific measurement matrix. The experiment in Machinery Failure Prevention Technology (MFPT) Dataset verifies the effectiveness of our algorithm. Compared with the sparse model (L1-norm) and hierarchical hyper-Laplacian prior induced model (HHLP), our method achieves better reconstruction signal to noise ratio (SNR) performance and maintains the fault frequency information from different under-sampled data.

Rolling bearing fault diagnosis based on multi-source data fusion

In this paper, a fault diagnosis method is proposed based on multi-source data fusion to address the inaccurate results caused by single evidence used for decision-making. First, empirical mode decomposition is performed using sensor data to extract fault features, and a fault feature matrix and a diagnostic matrix are established. Then the deviation vector is defined to obtain the difference between the diagnostic sample and the fault sample, and then the basic probability assignment is obtained for each diagnostic sample. Finally, the Dempster combination rule is used for fusion. The effectiveness and accuracy of this method in fault diagnosis of rolling bearings have been verified through examples of rolling bearings.

A Fault Diagnosis Method for Rectifier-Filter Circuit Integrating EEMD Algorithm and Transformer Network

Rectifier-filter circuit, as a critical component of the drive circuit in instrumentation and control systems of nuclear power plants, can convert the 50 Hz AC into the smooth DC. Thus, it plays a vital role in the power control of reactors. However, the weak waveform anomalies of soft faults in the rectifier-filter circuit make fault feature extraction difficult. Therefore, in this article, an ensemble empirical modal decomposition (EEMD) algorithm is employed to decompose the signal mode components in the monitor data of the rectifier-filter circuit. The weak waveform anomalies are indirectly enhanced by the IMF and residual components. Subsequently, the Transformer network is utilized to construct the feature extractor. With the advantage of multi-head attention (MHA) mechanism in the Transformer network, the multi-directional, multi-scale, and highly sensitive long-range time-dependent features in the EEMD feature data are extracted. Then, a deep Softmax classifier is adopted to reduce the dimensionality and diagnose the soft faults of the rectifier-filter circuit. Finally, a fault simulation model of the rectifier-filter circuit is constructed and the condition monitor data are collected. The effectiveness and diagnosis accuracy of the proposed method are verified by a real case experiment and some comparative methods.

Spot scanning imaging calibration method based on deviation model for wafer inspection

The spot scanning defect detection system has good detection performance for wafer defects. However, due to the existence of system assembly errors, the wafer defect inspection system could not scan following the planned trace, which resulted in distortion of the reconstructed image and had a greater impact on defect feature extraction and accurate defect location. The system deviation calibration and adjustment can further facilitate the accuracy of the system position and image reconstruction. To address this issue, the causes and shapes of imaging distortion were analyzed and simulated, and a simple and effective calibration method based on a scanning deviation model for the spot scanning defect detection system was proposed. First, a scanning deviation model was established following the characteristics of the spot scanning system, and different degrees of reconstructed image distortion were simulated and analyzed as the system deviation changes. Then, an 8-inch wafer with a grid was used as the calibrator and scanned using the developed spot scanning system. Serveral grid corners in the reconstructed image are selected and used as the calibration feature points. The constraint function is constructed using the equidistant features between adjacent grid points, and then the quasi-Newton method is used to optimize the parameters of the function to obtain the best deviation estimation of the spot scanning system. To prove the validity of the estimated deviation parameters, another patterned wafer was scanned and the distortion image was used for reconstruction. The experiment demonstrated that an accurate and distortion-free wafer image can be reconstructed using the calibrated deviation parameters, and the root mean square error of the measured distance between two points of interest was only 0.61 pixels, which proved the effectiveness of the proposed calibration method.

A novel method for fault diagnosis of fluid end of drilling pump under complex working conditions

Accurate fault diagnosis of drilling pump under complex operating conditions poses a significant challenge in drilling operations. This paper addresses the difficulty of extracting fault features from the timing signals of the drilling pump's fluid end under complex working conditions by proposing a method for generating images to represent multidimensional timing signals. Additionally, a multiscale deep recursive inverse residual neural network (MDRIRNN) is introduced to achieve fault diagnosis of the fluid end under such conditions. Firstly, a multidimensional relative position matrix (MRPM) is proposed, which transforms one-dimensional timing signals into three-dimensional images using a 3D data structure. This approach effectively distinguishes the features of different timing signals by adding dimensions. Next, a recursive inverse residual block is designed, and multiscale learning is incorporated to form the MDRIRNN. This model enables the extraction of feature information from the multidimensional images at different scales, thereby enhancing the fault diagnosis process. Subsequently, a fault diagnosis experiment is conducted, demonstrating an average diagnostic accuracy of 96.14 %. Furthermore, the adaptability of the proposed method is validated using the MFPT dataset, achieving a diagnostic accuracy of 99.88 %. This novel method provides a prospecting approach for equipment fault diagnosis under complex operating conditions.

Fault Diagnosis Method of Flexible Converter Valve Equipment Based on Ensemble Empirical Mode Decomposition and Temporal Convolutional Networks

The flexible converter valve is a crucial component of the flexible transmission system, and its proper functioning is directly related to the power system's reliability. This study proposes a method for diagnosing faults in flexible converter valve equipment based on ensemble empirical mode decomposition and temporal convolutional networks. The method involves measuring voltage signal data in the power submodule of the flexible converter valve, decomposing and reconstructing the voltage signal using ensemble empirical mode decomposition to extract frequency variation patterns and fault features. Subsequently, a temporal convolutional network is introduced, and a device fault diagnosis model is constructed by learning the evolution law of voltage signals in time series. The experimental results demonstrate that the proposed method has high fault diagnosis accuracy and robustness with an average F1-score of 89.58% and an average area under the curve (AUC) of 94.38%, which are higher than other methods by at least 1.39% and 1.03%.

Open-circuit fault diagnosis method of inverter in wind turbine yaw system based on GADF image coding

Considering the concealment of inverter open-circuit faults in yaw system of wind turbines, a method for diagnosing open-circuit fault of yaw system inverter is proposed based on Gramian Angular Difference Field (GADF) image coding in this article. Firstly, the current vector phase in the vicinity of relatively high yaw speed is collected as monitoring data. Secondly, considering the insufficient fault feature extraction ability of Convolutional Neural Network (CNN) structures with single dimensional data input during model training, the one-dimensional current vector phase is encoded into two-dimensional image by GADF image coding and an effective CNN fault diagnosis model is obtained with fewer fault samples. Finally, by comparing with actual monitoring data of wind turbine, the effectiveness of yaw system simulation model is verified. The proposed method can achieve identification and localization of open-circuit fault for single and two power devices, which is found to have better anti-noise interference effect and the results are not affected by yaw angle and mechanical torque. It is approved that the proposed method is effective with a simulation on the RT-LAB platform.

Rolling Bearing Fault Feature Extraction Method based on SSA-VMD and MOMEDA

To address the challenge of extracting bearing fault features, this study proposes a new rolling bearing fault feature extraction method based on the Sparrow Search Algorithm (SSA) to optimize Variational Mode Decomposition (VMD) and Multipoint Optimal Minimum Entropy Deconvolution with Convolution Adjustment (MOMEDA). Firstly, SSA is employed to identify optimal parameters in VMD, followed by the utilization of correlation coefficients and kurtosis to filter relevant Intrinsic Mode Function (IMF) components. Subsequently, MOMEDA is applied to denoise the reconstructed signal, mitigating the interference caused by pulse fault signals. Finally, the envelope spectrum analysis is conducted on the denoised signal. Experimental results demonstrate the efficacy of the proposed method in extracting fault features and mitigating noise interference.

Helicopter Planet Gear Rim Crack Diagnosis and Trending Using Cepstrum Editing Enhanced with Deconvolution.

Detecting gear rim fatigue cracks using vibration signal analysis is often a challenging task, which typically requires a series of signal processing steps to detect and enhance fault features. This task becomes even harder in helicopter planetary gearboxes due to the complex interactions between different gear sets and the presence of vibration from sources other than the planetary gear set. In this paper, we propose an effectual processing algorithm to isolate and enhance rim crack features and to trend crack growth in planet gears. The algorithm is based on using cepstrum editing (or liftering) of the hunting-tooth synchronous averaged signals (angular domain) to extract harmonics and sidebands of the planet gears and low-pass filtering and minimum entropy deconvolution (MED) to enhance extracted fault features. The algorithm has been successfully applied to a vibration dataset collected from a planet gear rim crack propagation test undertaken in the Helicopter Transmission Test Facility (HTTF) at DSTG Melbourne. In this test, a seeded notch generated by an electric discharge machine (EDM) was used to initiate a fatigue crack that propagated through the gear rim body over 94 load cycles. The proposed algorithm demonstrated a successful isolation of incipient fault features and provided a reliable trending capability to monitor crack progression. Results of a comparative analysis showed that the proposed algorithm outperformed the traditional signal processing approach.

An adaptive threshold-selected symplectic geometry mode decomposition for application to multi-modulation complex fault signals

The signals obtained from complex mechanical systems are characterized by multilevel modulation and strong noise, which can lead to difficulties in fault feature extraction. Symplectic geometry mode decomposition (SGMD) proves to be a valid approach for decomposing signals. However, inaccurate threshold selection in the iterative decomposition process can compromise the quality of fault diagnosis results. To address the shortcomings of SGMD, this paper proposes adaptive SGMD with adaptive threshold selection for fault diagnosis. Based on minimum dispersion entropy indicators, correlation coefficient and stopping thresholds are adaptively chosen using the proposed enhanced dung beetle optimizer algorithm. Then the optimal symplectic geometry component (SGC) is filtered based on the value of the integrated indicators after decomposition. Finally, the optimal SGC is analyzed by envelope demodulation to extract gear fault information. Through simulation and experimental analysis, this method surpasses SGMD and other signal decomposition methods in the aspect of fault feature extraction and noise robustness. Additionally, the results indicate an increase in feature energy ratio by 2.14%–9.85% compared to SGMD. The paper demonstrates that the proposed method extracts the fault feature frequencies of gears more effectively in complex mechanical systems.

Rotating Machinery Fault Diagnosis under Time–Varying Speed Conditions Based on Adaptive Identification of Order Structure

Rotating machinery fault diagnosis is of key significance for ensuring safe and efficient operation of various industrial equipment. However, under nonstationary operating conditions, the fault–induced characteristic frequencies are often time–varying. Conventional Fourier spectrum analysis is not suitable for revealing time–varying details, and nonstationary fault feature extraction methods are still in desperate need. Order spectrum can reveal the rotational–speed–related time–varying frequency components as spectral peaks in order domain, thus facilitating fault feature extraction under time–varying speed conditions. However, the speed–unrelated frequency components are still nonstationary after angular–domain resampling, thus causing wide–band features and interferences in the order spectrum. To overcome such a drawback, this work proposes a rotating machinery fault diagnosis method based on adaptive separation of time–varying components and order feature extraction. Firstly, the rotational speed is estimated by the multi–order probabilistic approach (MOPA), thus eliminating the inconvenience of installing measurement equipment. Secondly, adaptive separation of the time–varying frequency component is achieved through time–varying filtering and surrogate test. It effectively eliminates interference from irrelevant components and noise. Finally, a high–resolution order spectrum is constructed based on the average amplitude envelope of each mono–component. It does not involve Fourier transform or angular–domain resampling, thus avoiding spectral leakage and resampling errors. By identifying the fault–related spectral peaks in the constructed order spectrum, accurate fault diagnosis can be achieved. The Rényi entropy values of the proposed order spectrum are significantly lower than those of the traditional order spectrum. This result verifies the effective energy concentration and high resolution of the proposed order spectrum. The results of both numerical simulation and lab experiments confirm the effectiveness of the proposed method in accurately presenting the time–varying frequency components for rotating machinery diagnosing faults.

Diagnosis of Al-CFRTP TA-FSLW defect using acoustic emission signal based on SPWVD and ResNet

As one of the important development directions in modern railway and automobile, lightweight requires reliable connection technology of dissimilar materials. Thermal-assisted friction stir lap welding (TA-FSLW) is an available welding technology for the joining of dissimilar materials, which can solve the problems of physical and chemical properties in the welding process of dissimilar materials. Unfortunately, defects may occur during the welding process. This paper proposes a defect diagnosis method based on smooth pseudo Wigner-Ville distribution (SPWVD) and ResNet18 for Al-CFRTP TA-FSLW. Firstly, the acoustic emission (AE) signal and the temperature of the weld nugget are acquired by the AE sensor and thermography camera in the FSLW process. Then, the time–frequency domain features of AE signal are extracted by SPWVD, and the maximum temperature of the weld nugget is extracted by exponential moving average (EMA) from thermography. Moreover, the dataset is established by SPWVD features, maximum temperature of welding nugget, and several time–frequency features. Finally, the ResNet18-attention network is employed to classify welding defects and stages. The Al-CFRTP TA-FSLW experiments are carried out based on the proposed method. Experimental result indicates that the SPWVD spectrum can effectively extract the feature frequency of deep and shallow hole defects in the welding process. In the section of classification, the accuracy of the dataset with SPWVD features is improved by 2% to 4% compared to other methods, which demonstrates that the SPWVD feature is superior in defect feature extraction.

Small-sample fault diagnosis study of rolling bearings based on a residual parameterised convolutional capsule network

Although intelligent fault diagnosis has achieved good results, the application in practical engineering scenarios is still unsatisfactory due to the lack of sufficient fault signals to support the training of the diagnosis methods and the difficulty of extracting sensitive fault features from the original signals. To address the problem that small-sample fault data limit the diagnostic performance of traditional neural networks, a multi-scale residual parametric convolutional capsule network (MRCCCN) for small-sample bearing fault diagnosis is proposed. In the MRCCCN, the input fault information is averaged and segmented multiple times and then the initial features of the multi-segmented input are extracted by residual parameterised convolution. Then, the multi-branch features are fused and fed into an improved parametric capsule network to further extract fault features and store feature information using dynamic routing. The performance of the MRCCCN is validated using the Case Western Reserve University (CWRU) rolling bearing dataset and the Paderborn University rolling bearing dataset of vibration signals and compared with some advanced deep learning methods. The comparison results show that the proposed MRCCCN is able to accurately diagnose faults under small-sample conditions and still has significant diagnostic performance in small-sample variable noise tests.

A Novel Noise-Aided Fault Feature Extraction Using Stochastic Resonance in a Nonlinear System and Its Application

A Novel Noise-Aided Fault Feature Extraction Using Stochastic Resonance in a Nonlinear System and Its Application

Ternary Precursor Centrifuge Rolling Bearing Fault Diagnosis Based on Adaptive Sample Length Adjustment of 1DCNN-SeNet

To address the issues of uneven sample lengths in the centrifuge machine bearings of the ternary precursor, inaccurate fault feature extraction, and insensitivity of important feature channels in rolling bearings, a rolling bearing fault diagnosis method based on adaptive sample length adjustment of one-dimensional convolutional neural network (1DCNN) and squeeze-and-excitation network (SeNet) is proposed. Firstly, by controlling the cumulative variance contribution rate in the principal component analysis algorithm, adaptive adjustment of sample length is achieved, reducing data with uneven sample lengths to the same dimensionality for various classes. Then, the 1DCNN extracts local features from bearing signals through one-dimensional convolution-pooling operations, while the SeNet network introduces a channel attention mechanism which can adaptively adjust the importance between different channels. Finally, the 1DCNN-SeNet model is compared with four classic models through experimental analysis on the CWRU bearing dataset. The experimental results indicate that the proposed method exhibits high diagnostic accuracy in rolling bearings, demonstrating good adaptability and generalization capabilities.

Intelligent Fault Diagnosis of Rolling Bearing Based on Gramian Angular Difference Field and Improved Dual Attention Residual Network.

With the rapid development of smart manufacturing, data-driven deep learning (DL) methods are widely used for bearing fault diagnosis. Aiming at the problem of model training crashes when data are imbalanced and the difficulty of traditional signal analysis methods in effectively extracting fault features, this paper proposes an intelligent fault diagnosis method of rolling bearings based on Gramian Angular Difference Field (GADF) and Improved Dual Attention Residual Network (IDARN). The original vibration signals are encoded as 2D-GADF feature images for network input; the residual structures will incorporate dual attention mechanism to enhance the integration ability of the features, while the group normalization (GN) method is introduced to overcome the bias caused by data discrepancies; and then the model is trained to complete the classification of faults. In order to verify the superiority of the proposed method, the data obtained from Case Western Reserve University (CWRU) bearing data and bearing fault experimental equipment were compared with other popular DL methods, and the proposed model performed optimally. The method eventually achieved an average identification accuracy of 99.2% and 97.9% on two different types of datasets, respectively.

A novel gearbox local fault feature extraction method based on quality coefficient and dictionary learning

The performance of sparse decomposition is directly determined by the similarity between impact atoms and the actual fault impact waveform. The shift-invariant K-singular value decomposition (K-SVD) dictionary learning algorithm is capable of training impact patterns from vibration signals collected by sensors to construct impact atoms, thereby extracting fault impact components from the vibration signals. However, the impact pattern training using the shift-invariant K-SVD algorithm is influenced by the presence of harmonics and white noise in the gear transmission system vibration signals. To solve the above problems, a novel gearbox local fault feature extraction method based on the quality coefficient and dictionary learning is proposed in this paper. Firstly, the original signal is decomposed into a series of intrinsic mode functions (IMFs) by empirical mode decomposition. Then, a new quality coefficient is proposed by comprehensively considering the intensity of the impact, harmonics and noise components in each IMF, as well as the degree of correlation with the original signal. The IMF with the largest quality coefficient is used to train the impact pattern and solve the sparse coefficients. Finally, the orthogonal matching pursuit algorithm is adopted to solve the sparse coefficients, which are used to reconstruct the fault impact response signal from the dictionary. Simulation and experimental analysis demonstrate the superior performance of the proposed method compared to other state-of-the-art methods.