Characteristics Of Vibration Signals Research Articles

Abnormal State Detection of OLTC Based on Improved Fuzzy C-means Clustering

An accurate extraction of vibration signal characteristics of an on-load tap changer (OLTC) during contact switching can effectively help detect its abnormal state. Therefore, an improved fuzzy C-means clustering method for abnormal state detection of the OLTC contact is proposed. First, the wavelet packet and singular spectrum analysis are used to denoise the vibration signal generated by the moving and static contacts of the OLTC. Then, the Hilbert-Huang transform that is optimized by the ensemble empirical mode decomposition (EEMD) is used to decompose the vibration signal and extract the boundary spectrum features. Finally, the gray wolf algorithm-based fuzzy C-means clustering is used to denoise the signal and determine the abnormal states of the OLTC contact. An analysis of the experimental data shows that the proposed secondary denoising method has a better denoising effect compared to the single denoising method. The EEMD can improve the modal aliasing effect, and the improved fuzzy C-means clustering can effectively identify the abnormal state of the OLTC contacts. The analysis results of field measured data further verify the effectiveness of the proposed method and provide a reference for the abnormal state detection of the OLTC.

Squared envelope sparsification via blind deconvolution and its application to railway axle bearing diagnostics

A sparse squared envelope is crucial for efficient and accurate diagnosis of bearing faults. Blind deconvolution (BD) is a well-established sparse feature enhancement method for the diagnostics of rolling bearings. Traditional BD methods, such as minimum entropy deconvolution (MED), are susceptible to random transients, making it difficult to enhance fault features of rolling bearings subject to strong random shocks. Deconvolution methods that take the fault characteristic frequency (or fault impulse period) of interest as an algorithm input parameter, such as maximum second-order cyclostationarity blind deconvolution (CYCBD), can alleviate this deficiency. However, bearing fault features are difficult to be enhanced by these methods when the specified characteristic frequency deviates from the actual value greatly. To overcome these problems, the modified smoothness index (MSI) of the squared envelope is proposed as the objective function of the deconvolution method, and a new BD method is developed to achieve a sparse squared envelope for fault diagnosis of rolling bearings. Furthermore, the methodology is extended to the frequency domain, and another new BD method that utilizes the MSI of the squared envelope spectrum as the objective function is established to achieve a sparse squared envelope spectrum for bearing diagnostics. These two proposed BD methods are robust to random transients and do not require characteristic frequency or impulse period as an input parameter for feature enhancement. The performance of the two proposed BD methods is verified on experimental datasets from two different railway axle bearing test rigs and compared with the state-of-the-art deconvolution methods. The results show that the two proposed methods can effectively enhance repetitive transient features in noisy vibration signals and accurately diagnose different faults of railway axle bearings.

Deep stacked pinball transfer matrix machine with its application in roller bearing fault diagnosis

At present, the matrix-based classification methods have played an extremely important role to mechanical fault diagnosis. However, the traditional matrix-based classification methods are mostly shallow classifiers, which are difficult to obtain the deep-seated sensitive features of vibration signals. To learn and extract low-rank structure information, a new deep stacked pinball transfer matrix machine (DSPTMM) is proposed to enhance the performance of traditional shallow matrix-based classifiers, which takes the stacking generalization principle as the main idea. In DSPTMM, a pinball transfer module (PTM) is constructed as the basic module of deep stacking network, in which the pinball loss is used to achieve an enhanced noise robustness. Specifically, PTM performs the function to obtain the weak prediction of the previous module, and the original results of previous weak prediction are modified by random projection and then will be input as a new feature set in the next module. Therefore, DSPTMM has a better classification performance and can modeled without enough annotation samples. Extensive experiments are carried out on two different datasets of roller bearing, and the results of experiment show that the proposed DSPTMM can use limited samples to establish the accurate model.

Identification of wear mechanisms of main bearings of marine diesel engine using recurrence plot based on CNN model

Wear failure of the main bearings on marine diesel engines is difficult to diagnose online, especially to identify the wear mechanisms. To solve the problem, the recurrence plot (RP) was used to characterize the nonlinear characteristics of vibration signals. It was combined with convolutional neural networks (CNN), and the RP-CNN method was proposed. The new method was also compared with six common machine learning methods. Results show that the diagnostic accuracy of the RP-CNN method is more than 96% under five typical working conditions, which is better than the other methods. The confusion matrix shows that the normal state and the three mechanisms of fatigue wear, abrasive wear, and adhesive wear can be effectively identified.

Fault Diagnosing of Cycloidal Gear Reducer Using Statistical Features of Vibration Signal and Multifractal Spectra.

The article presents a method for diagnosing cycloidal gear damage on a laboratory stand. The damage was simulated by removing the sliding sleeves from two adjacent external pins of the cycloidal gearbox. Damage to the sliding sleeves may occur under operating conditions and can lead to the destruction of the gear unit. Hence, early detection is essential. Signals from torque sensors, rotational speed sensors and vibration acceleration sensors of input and output shafts for various rotational speeds and transmission loads were recorded. The frequency analysis of these signals was carried out. Due to the fluctuation of the rotational speed, the frequency spectrum gives an approximate picture and is not useful in detecting this type of damage. The statistical characteristics of the signal were determined. However, only statistical moments of higher orders, such as kurtosis, are sensitive to the tested damage. Therefore, the use of multifractal analysis of the vibration signal using the wavelet leader method (WLMF) was considered. Then log-cumulants of the multifractal spectrum were selected as the new signal features.

Self-supervised feature extraction via time–frequency contrast for intelligent fault diagnosis of rotating machinery

Data-driven intelligent fault diagnosis requires a large amount of data. However, collecting sufficient labeled data from the field is generally difficult because mechanical devices are not allowed to operate under fault conditions for long periods of time. In addition, the working conditions of rotating machinery are often subject to change, which makes it hard for fault diagnosis. In this paper, a self-supervised Contrastive Learning model combining Frequency-domain and Time-domain information (CLFT) is proposed for pre-training. This model fully considers the time–frequency characteristics of vibration signals and reduces the dependence on the training data. Based on CLFT, a two-stage fault diagnosis pipeline applicable to the rotating machinery is constructed, which uses pre-training and fine-tuning strategies to accomplish specific diagnostic tasks. Experiments indicate that CLFT can effectively extract generalization features from unlabeled vibration signals, which makes the fault diagnosis method perform well under the condition of few samples and variable working conditions.

An anti-noise fault diagnosis approach for rolling bearings based on multiscale CNN-LSTM and a deep residual learning model

Bearing fault vibration signals collected in real engineering cases often contain environmental noise which can easily mask the fault type characteristics of vibration signals, making it difficult to determine the corresponding fault type when traditional deep learning methods are used for fault diagnosis. To solve the above problem, a neural network model named multiscale CNN-LSTM (convolutional neural network-long short-term memory) and a deep residual learning model was designed, which combines a multiscale wide CNN-LSTM module and a deep residual module for rolling bearing fault diagnosis. In this model, a wide convolution kernel CNN-LSTM structure with different convolution scales is used to extract a variety of different types of frequency and sequential features from vibration signals. It is worth noting that the wide convolution kernel CNN-LSTM structure not only has stronger feature extraction performance compared with the common convolution layer but can also reduce the interference of high-frequency noise. Moreover, the deep residual module with a wide convolution kernel CNN-LSTM structure is used to further improve the feature expression ability of the proposed model. The above algorithm enables the proposed model to better extract the fault features hidden in the noise signal. When compared with some state-of-the-art methods, the experimental results showed that this model has better anti-noise performance and better generalization ability for rolling bearing fault diagnosis.

Early Fault Diagnosis of Rolling Bearing Based on Threshold Acquisition U-Net

Considering the problem that the early fault signal of rolling bearing is easily interfered with by background information, such as noise, and it is difficult to extract fault features, a method of rolling bearing early fault diagnosis based on the threshold acquisition U-Net (TA-UNet) is proposed. First, to improve the feature extraction ability of U-Net, the channel spatial threshold acquisition network (CS-TAN) and the dilated convolution module (DCM) based on different dilated rate combinations are introduced into the U-Net to construct the TA-UNet. Among them, the CS-TAN can adaptively learn the threshold, reduce the interference of noise in the signal, and the DCM can improve the multi-scale feature extraction ability of the network. Then, the TA-UNet is used for early fault diagnosis, and the method is divided into two steps: The model training phase and the vibration signal fault feature extraction phase. In the first step, additive gaussian white noise is added to the vibration signal to obtain the noise-added vibration signal, and the TA-UNet is trained to learn how to denoise the noise-added vibration signal. In the second step, the trained TA-UNet is used to extract the fault features of vibration signals and diagnose the early fault types of rolling bearing. The two-step method solves the problem that U-Net, as a supervised neural network, needs corresponding labeled data to be trained, as it realizes the fault diagnosis of unlabeled data. The feature extraction capability of the TA-UNet is evaluated by denoising the simulated signal of rolling bearing. The effectiveness of the proposed diagnostic method is demonstrated by the early fault diagnosis of open-source datasets.

Investigation of Cutting Path Effect on Spindle Vibration and AE Signal Features for Tool Wear Monitoring in Micro Milling

In the development of a tool wear monitoring system in milling, the complex cutting path always brings challenges to the system’s reliability in the production line. The cutting path effect on the acoustic emission (AE) and vibration signals during the micro milling processes was investigated in this study by implementing three types of cutting paths in a micro milling experiment. To generate the data for analysis, an experiment was conducted on a micro milling research platform using an AE sensor and an accelerometer installed on a fixture attached to the spindle housing. To demonstrate the effect of the cutting path on the performance in the monitoring of tool wear, a simple linear classifier is proposed, along with the signal features generated from the different signal lengths and the bandwidth size in the frequency domain. The results show that the signal features generated from the cutting of a straight line, the corner of the square path, and the circle path are different from each other. The increase in the signal length to generate features, which will reduce the corner effect, could improve the performance of the developed monitoring system. However, the results suggest that avoiding the complex cutting path for feature generation might be a better strategy for developing a micro milling tool wear monitoring system.

Health State Identification Method of Nuclear Power Main Circulating Pump Based on EEMD and OQGA-SVM

Main circulation pump is the only high-speed rotating equipment in primary loop of nuclear power plant. Its function is to ensure the normal operation of primary loop system by controlling the circulating flow of reactor coolant. In order to ensure long-term healthy operation of nuclear power main circulating pump, a method for identifying the health states of nuclear power main circulating pump based on ensemble empirical mode decomposition (EEMD) and support vector machine optimized by optimized quantum genetic algorithm (OQGA-SVM) is proposed. Vibration signal of main circulating pump is decomposed by EEMD. Vibration signal characteristics of nuclear power main circulating pump in healthy state and different fault states are analyzed and target characteristic indexes are put forward. Then, health state identification model of main circulation pump of OQGA-SVM is established, and target characteristic indexes are used as input parameter of the model. Finally, combined with experimental data, the model analysis and validation show that the health state identification method of nuclear power main circulating pump based on EEMD-OQGA-SVM can accurately and effectively identify the states of main circulation pump, has a higher identification accuracy than EEMD-SVM method and is more efficient and accurate than EEMD-QGA-SVM method.

A Novel Ensemble Learning-Based Multisensor Information Fusion Method for Rolling Bearing Fault Diagnosis

It is meaningful to learn fault-related information from multisensor signals automatically and provide accurate diagnostic results. To further improve the multisensor information fusion effect, reduce the discrepancy between the real value and predicted value, and improve the rolling bearing fault diagnosis accuracy, in this article, a novel ensemble learning-based multisensor information fusion method is proposed. First, a multiscale convolutional neural network (MSCNN) is constructed as a base learning model to learn multiscale features of raw vibration signals. Second, based on the ensemble learning framework, a multibranch MSCNN is built to realize the simultaneous extraction of multiple sensor signal features and output the decision score of each sensor. Then, considering the confidence of multiple different sensors, fuzzy rank is utilized to fuse the decision scores of each sensor, minimizing the discrepancy between the real value and predicted value. Finally, the usefulness and robustness of the proposed method are validated on two different types of rolling bearing datasets. The dataset and code are available under request and the contact e-mail is pantc2006@163.com.

Nonstationary feature extraction based on stochastic resonance and its application in rolling bearing fault diagnosis under strong noise background.

When the load and speed of rotating machinery change, the vibration signal of rolling bearing presents an obvious nonstationary characteristic. Stochastic resonance (SR) mainly is convenient to analyze the stationary feature of vibration signals with high signal-to-noise ratio. However, it is difficult for SR to extract the nonstationary feature of rolling bearings under strong noise background. For one thing, the frequency change of nonstationary signals makes the occurrence of SR very difficult. For another, the features of rolling bearings are large parameters and further prevent the SR method from performing well. Therefore, combined with order analysis (OA), adaptive frequency-shift SR is presented in this paper. To solve the problem of frequency change, OA is used to convert the nonstationary feature into stationary feature, which resamples the nonstationary signal in the time domain to stationary signal in the angular domain. To solve the other problem, the frequency-shift method based on Fourier transform is adopted to move the fault feature frequency to low frequency, and thus SR is more likely to occur under small parameter conditions. The simulated and experimental results indicate that not only the amplitude of fault feature but also the signal-to-noise ratio is significantly improved. These demonstrate that the fault features of rolling bearing in variable speed conditions are extracted successfully.

Bearing Fault Diagnosis Based on Multi-Scale CNN and Bidirectional GRU

Finding a reliable approach to detect bearing faults is crucial, as the most common rotating machine defects occur in its bearings. A convolutional neural network can automatically extract the local features of the mechanical vibration signal and classify the patterns. Nevertheless, these types of networks suffer from the extraction of the global feature of the input signal as they utilize only one scale on their input. This paper presents a method to overcome the above weakness by employing a combination of three parallel convolutional neural networks with different filter lengths. In addition, a bidirectional gated recurrent unit is utilized to extract global features. The CWRU-bearing dataset is used to prove the performance of the proposed method. The results show the high accuracy of the proposed method even in the presence of noise.

Fault Feature of Gearbox Vibration Signals Based on Morphological Filter Dynamic Convolution Autoencoder

Vibration signals are widely used in fault diagnosis of rotating machinery. Although deep neural networks (DNNs) have been widely used in feature learning of vibration signals, they have not shown good performance in the extraction of impulsive features and noise filtering. In this study, a new DNN, morphological filter dynamic convolutional autoencoder (MF-DCAE), is proposed for fault feature extraction from vibration signals. First, a multiscale morphological filter layer embedded with a convolutional autoencoder (CAE) is proposed to extract fault features of vibration signals. Kurtosis is used to fuse the feature information extracted by morphological operators with different scales. Dynamic convolution is used to adjust the weight of each convolution kernel for adaptive feature extraction. Finally, residual learning is used to connect the encoder and decoder to obtain good feature learning of vibration signal. The effectiveness of MF-DCAE is verified on two gearbox testbeds. The results show that MF-DCAE can perform both noise reduction and feature learning on vibration signals in an unsupervised learning manner. The comparison results show that the feature extraction performance of MF-DCAE is better than that of the state-of-the-art DNNs.

Research on State Recognition Technology of Elevator Traction Machine Based on Modulation Feature Extraction

Vibration signal analysis of the traction machine is an important part of the current rotating machinery state recognition technology, and its feature extraction is the most critical step. In this study, the time-frequency characteristics of the vibration of the traction machine under different elevator running directions, running speeds and load weights are analyzed. The novel demodulation method based on time-frequency analysis and principal component analysis (DPCA) is used to extract the periodic modulated wave signal. In order to compare different influence of background noise and unknown frequency influence, the Fast Fourier Transform (FFT) and Short Time Fourier Transform (STFT) methods are used to extract the characteristics of the traction machine vibration signal, respectively. Under different load conditions, it is difficult to observe the obvious differences and similarities of the vibration signals of the traction machine by time-frequency method. However, the DPCA demodulation method provides a guarantee for the reliability and accuracy of the state identification of the traction machine.

Spiking Neural Networks for Structural Health Monitoring

This paper presents the first implementation of a spiking neural network (SNN) for the extraction of cepstral coefficients in structural health monitoring (SHM) applications and demonstrates the possibilities of neuromorphic computing in this field. In this regard, we show that spiking neural networks can be effectively used to extract cepstral coefficients as features of vibration signals of structures in their operational conditions. We demonstrate that the neural cepstral coefficients extracted by the network can be successfully used for anomaly detection. To address the power efficiency of sensor nodes, related to both processing and transmission, affecting the applicability of the proposed approach, we implement the algorithm on specialised neuromorphic hardware (Intel ® Loihi architecture) and benchmark the results using numerical and experimental data of degradation in the form of stiffness change of a single degree of freedom system excited by Gaussian white noise. The work is expected to open a new direction of SHM applications towards non-Von Neumann computing through a neuromorphic approach.

Intelligent Identification of Cavitation State of Centrifugal Pump Based on Support Vector Machine

In order to perform intelligent identification of the various stages of cavitation development, a micro high-speed centrifugal pump was used as a research object for vibration signal analysis and feature extraction for normal, incipient cavitation, cavitation and severely cavitated operating states of the pump at different temperatures (25 °C, 50 °C and 70 °C), based on support vector machines to classify and identify the eigenvalues in different cavitation states. The results of the study showed that the highest recognition rate of the individual eigenvalues of the time domain signals, followed by time frequency domain signals and finally frequency domain signals, was achieved in the binary classification of whether cavitation occurred or not. In the multi-classification recognition of the cavitation state, the eigenvalues of the time domain signals of the four monitoring points, the time frequency domain signals of the monitoring points in the X-direction of the inlet pipe and the Y-direction of the inlet pipe are combined, and the combined eigenvalues can achieve a multi-classification recognition rate of more than 94% for the cavitation state at different temperatures, which is highly accurate for the recognition of the cavitation state of centrifugal pumps.

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.

Vibration Characteristics of Corn Combine Harvester with the Time-Varying Mass System under Non-Stationary Random Vibration

In field harvesting conditions, the non-stationary random vibration characteristics of the harvester are rarely considered, and the results of vibration frequency calculated by different time–frequency transformation methods are different. In this paper, the harvester’s vibration characteristics under the time-varying mass were studied, and the correlation between vibration frequency and modal frequency was analyzed. Firstly, under the conditions of time-varying mass (field harvesting conditions) and non-time-varying mass (empty running condition), the non-stationarity characteristics of vibration signals at 16 measurement points of a combined corn harvester frame were studied. Then, fast Fourier transform (FFT), short-time Fourier transform (STFT), and continuous wavelet transform (CWT) were used to calculate the vibration frequency distribution characteristics of the corn harvester. Finally, based on the EFDD (enhanced frequency domain decomposition) algorithm, the correlation between the primary vibration frequency and the operating mode frequency is studied. The results show that the mean, variance, and maximum difference of the vibration amplitude under harvesting conditions (mass time-varying system) are 0.10, 26.5, and 1.0, respectively, at different harvesting periods (0~10 s, 10~20 s, 20~30 s). The harvesting conditions’ vibration signals conform to the characteristics of non-stationary randomness. The FFT algorithm is used to obtain more dense vibration frequencies, while the frequencies based on STFT and CWT algorithms are sparse. The correlation between the FFT method and the EFDD algorithm is 0.98, and the correlation between the STFT, CWT, and the EFDD algorithm is 0.99 and 0.98. Therefore, the primary frequency of the STFT methods is closer to the modal frequency. Our research laid the foundation for further study and application of mass time-varying combined harvester system non-stationary random vibration modal frequency identification and vibration control.

A power information guided-variational mode decomposition (PIVMD) and its application to fault diagnosis of rolling bearing

As a powerful tool of mode component extraction, variational mode decomposition (VMD) has found its wide application in fault diagnosis. However, VMD still suffers from an unsolved problem. To be specific, how to adaptively determine the decomposition number, the initial center frequencies and the variable penalty factors in VMD according to the characteristics of vibration signal is still an unsolved problem. To address this problem, the authors propose a novel mode component extraction method, namely, the power information guided-variational mode decomposition (PIVMD). The essential idea of the proposed method is using the enveloped auto-regressive (AR) power spectrum of raw vibration signal to determine the VMD parameters (decomposition number, initial center frequencies and variable penalty factors). After determining these VMD parameters, an iterative search is conducted to obtain the optimal solution for constructed variational model. A noise reduction envelope spectrum is further used to enhance the weak fault features contained in the extracted mode component. A set of simulation signals representing the faulty characteristics of rolling bearing with inner/outer race faults is analyzed by the proposed PIVMD method to test its mode extraction ability. Then, an open dataset from Case Western Reserve University and a measured dataset from a steel mill in Ma Steel Co. as well as an experimental dataset collected from the authors' laboratory are analyzed to verify the fault diagnosis capability of the proposed PIVMD method. Meanwhile, the analyzed results are compared with those obtained from other mode decomposition methods to highlight the superiority of the proposed PIVMD method in terms of extracting mode components with rich fault information.