Time-frequency Graph Research Articles

Fault Diagnosis Method for Vacuum Contactor Based on Time-Frequency Graph Optimization Technique and ShuffleNetV2.

This paper presents a fault diagnosis method for a vacuum contactor using the generalized Stockwell transform (GST) of vibration signals. The objective is to solve the problem of low diagnostic performance efficiency caused by the inadequate feature extraction capability and the redundant pixels in the graph background. The proposed method is based on the time-frequency graph optimization technique and ShuffleNetV2 network. Firstly, vibration signals in different states are collected and converted into GST time-frequency graphs. Secondly, multi-resolution GST time-frequency graphs are generated to cover signal characteristics in all frequency bands by adjusting the GST Gaussian window width factor λ. The OTSU algorithm is then combined to crop the energy concentration area, and the size of these time-frequency graphs is optimized by 68.86%. Finally, considering the advantages of the channel split and channel shuffle methods, the ShuffleNetV2 network is adopted to improve the feature learning ability and identify fault categories. In this paper, the CKJ5-400/1140 vacuum contactor is taken as the test object. The fault recognition accuracy reaches 99.74%, and the single iteration time of model training is reduced by 19.42%.

Research on microseismic signal identification through data fusion

The present study proposes a double-branch classification network, DPNet (Double Path Net), for the classification and identification of microseismic and blasting signals based on multimodal feature extraction. The vibration signals’ one-dimensional spectrogram and two-dimensional wavelet time–frequency graph are inputted into the double branch network. Subsequently, convolutional neural networks and ResNet are employed to extract the one-dimensional frequency features and two-dimensional time–frequency features of the vibration signals, respectively. Experimental results demonstrate that our proposed method achieves outstanding classification performance with an accuracy of 97.34% for microseismic signals and blasting signals. This research not only provides innovative solutions to practical problems but also introduces a novel idea of multimodal feature extraction at a theoretical level. By successfully applying it to efficiently classify complex signals in mining engineering, we offer a feasible solution that holds promising prospects for practical applications in this field.

A Novel Intelligent Fault Diagnosis Method for Bearings with Multi-Source Data and Improved GASA.

In recent years, single-source-data-based deep learning methods have made considerable strides in the field of fault diagnosis. Nevertheless, the extraction of useful information from multi-source data remains a challenge. In this paper, we propose a novel approach called the Genetic Simulated Annealing Optimization (GASA) method with a multi-source data convolutional neural network (MSCNN) for the fault diagnosis of rolling bearing. This method aims to identify bearing faults more accurately and make full use of multi-source data. Initially, the bearing vibration signal is transformed into a time-frequency graph using the continuous wavelet transform (CWT) and the signal is integrated with the motor current signal and fed into the network model. Then, a GASA-MSCNN fault diagnosis method is established to better capture the crucial information within the signal and identify various bearing health conditions. Finally, a rolling bearing dataset under different noisy environments is employed to validate the robustness of the proposed model. The experimental results demonstrate that the proposed method is capable of accurately identifying various types of rolling bearing faults, with an accuracy rate reaching up to 98% or higher even in variable noise environments. The experiments reveal that the new method significantly improves fault detection accuracy.

Fault Diagnosis of Planetary Gear Train Crack Based on DC-DRSN

To solve the problem that the existing planetary gear train fault diagnosis methods have, namely their low diagnostic accuracy under low signal-to-noise ratio (SNR), a fault diagnosis method based on a double channel-deep residual shrinkage network (DC-DRSN) is proposed. The short-time Fourier transform (STFT) is used to convert the original vibration signal into a two-dimensional time-frequency graph, which effectively enhances the ability to express information. A DC-DRSN model is constructed, and the optimal number of residual shrinkage modules is determined by combining the diagnostic characteristics with different noises, which effectively improves the accuracy and anti-noise ability of fault diagnosis. The results of bearing and planetary gear train crack fault diagnosis show that the diagnosis method based on DC-DRSN has higher diagnostic accuracy while realizing fault diagnosis, which is better than other deep learning diagnosis methods. At the same time, the method can adapt to fault diagnosis in different noise environments, and has good expression ability and generalization ability.

Convolutional Neural Network with Attention Mechanism and Visual Vibration Signal Analysis for Bearing Fault Diagnosis.

Bearings, as widely employed supporting components, frequently work in challenging working conditions, leading to diverse fault types. Traditional methods for diagnosing bearing faults primarily center on time-frequency analysis, but this often requires expert experience for accurate fault identification. Conversely, intelligent fault recognition and classification methods frequently lack interpretability. To address this challenge, this paper introduces a convolutional neural network with an attention mechanism method, denoted as CBAM-CNN, for bearing fault diagnosis. This approach incorporates an attention mechanism, creating a Convolutional Block Attention Module (CBAM), to enhance the fault feature extraction capability of the network in the time-frequency domain. In addition, the proposed method integrates a weight visualization module known as the Gradient-Weighted Class Activation Map (Grad-CAM), enhancing the interpretability of the convolutional neural network by generating visual heatmaps on fault time-frequency graphs. The experimental results demonstrate that utilizing the dataset employed in this study, the CBAM-CNN achieves an accuracy of 99.81%, outperforming the Base-CNN with enhanced convergence speed. Furthermore, the analysis of attention weights reveals that this method exhibits distinct focus of attention under various fault types and degrees. The interpretability experiments indicate that the CBAM module balances the weight allocation, emphasizing signal frequency distribution rather than amplitude distribution. Consequently, this mitigates the impact of the signal amplitude on the diagnostic model to some extent.

Fault diagnosis of rotating machinery using novel self-attention mechanism TCN with soft thresholding method

Rotating machinery (e.g. rolling bearings and gearboxes) is usually operated in high-risk and vulnerable environments such as time-varying loads and poor lubrication. Timely assessment of the operational status of rotating machinery is crucial to prevent damage caused by potential failure and shutdown, which significantly enhances the reliability of mechanical systems, prolongs the service life of critical components in rotating machinery, and minimizes unnecessary maintenance costs. In this regard, in this paper, a novel approach named self-attention mechanism combining time convolutional network with soft thresholding algorithm (SAM-TCN-ST) is proposed for fault intelligent recognition of rotating machinery. Specifically, the vibration signals are transformed into time-frequency graphs with distinct features utilizing the continuous wavelet transform, and then the proposed SAM-TCN-ST algorithm is employed for capturing essential data characteristics and classification performance. Eventually, datasets from rolling bearings and gearboxes are used to verify the accuracy and effectiveness of the proposed method compared with state-of-the-art benchmark networks such as pure TCN, convolutional neural networks and long short-term memory models. Experimental results demonstrate that the recognition accuracy rate of the proposed SAM-TCN-ST is higher than that obtained from the benchmark methods. This research presents an intelligent and viable solution for achieving real-time monitoring of the status and detecting faults in rotating machinery, thereby expectedly enhancing the reliability of mechanical systems. Consequently, the proposed SAM-TCN-ST algorithm holds significant potential for applications in prognostic and health management practices related to rotating machinery.

A novel framework for the diagnosis of Parkinson’s disease using transfer learning with RESNET50 and SVM classifier

<span>Most Parkinson’s patients exhibit voice disorders in the early phases of the condition. Recent study on Parkinson’s disease (PD) has concentrated on identifying speech problems from pronunciation of vowels with people affected by this disease. Proposed algorithm offers analysis of time-frequency images using transfer learning methods with support vector machine (SVM) for classification of PD affected with healthy controls using residual network 50. PD morphology is preserved in 2-dimensional time-frequency graphs <br /> that were helpful in implementing the proposed approach. The technique employs a hybrid HT/Wigner-Ville distribution to convert one-dimensional PD soundtracks to two-dimensional time-frequency (TF) graphs. After implementation of the proposed approach, classification of healthy and unhealthy controls is done using a pre-trained ResNet50 and the result is further improved through transfer learning. The features are extracted by passing preprocessed 2-D time-frequency diagrams through ResNet50’s FC1000 layer and trained using a binary nonlinear SVM classifier. The training process with 5-fold cross-validation (CV) got accuracy of 95.07% and in testing; it reached 92.13%, attaining better results.</span>

Automated and Adaptive Ridge Extraction for Rotating Machinery Fault Detection

A ridge in a time-frequency graph (TFG) describes the relationship of a signal component's instantaneous frequencies over time. Accurate ridge extraction from TFGs is beneficial for assessing machine health conditions without rotational speed measurement. This article proposes a new automated and adaptive ridge extraction (AARE) method. The AARE develops an adaptive edge detection strategy to avoid excessive interferences when searching for a ridge. Besides, the AARE creates a balance between exploring peak amplitudes and guaranteeing a continuous curve through an adaptive core function, which is constructed entirely based on the instantaneous characteristics of the analyzed signal. The unique advantage of the proposed method is that it dispenses with the tuning of parameters and runs automatically. Thus, human intervention is minimized. Gear and bearing vibration signals collected under variable speed conditions are applied to investigate and demonstrate the performance of AARE. In addition, some challenging cases are analyzed and discussed. Results show that AARE has a superior performance in ridge extraction compared with the reported approaches.

A Long-Term Recurrent Convolutional Network for SNR Estimation of Cone-Shaped Target

Signal-to-Noise Ratio (SNR) is an important prior information for many micro-motion echo processing approaches and techniques. Estimating SNR in advance can effectively enhance the performance of such technologies. This letter proposes a Long Term Recurrent Convolutional Network (LRCN) based SNR estimation method for cone-shaped target. First, mathematical expression of the echo is derived by studying the cone-shaped target micro-motion model, and hence the characteristics of the echo and its Short-Time Fourier Transform (STFT) are analyzed. Second, the motivation for the proposed are analyzed, and the echo is converted as an time-frequency graph (TFG) with STFT, so that an estimation technique based on LRCN (a hybrid of convolutional neural network (CNN) and recurrent neural networks (RNN)) is designed to estimate the SNR. Experiments indicate that the proposed method outperforms other already existing methods with better accuracy and robustness.

A deep convolutional neural network model with two-stream feature fusion and cross-load adaptive characteristics for fault diagnosis

Research aimed at diagnosing rolling bearing faults is of great significance to the health management of equipment. In order to solve the problem that rolling bearings are faced with variable operating conditions and the fault features collected are single in actual operation, a new lightweight deep convolution neural network model called FC-CLDCNN, composed of a convolution pooling dropout group with two-stream feature fusion and cross-load adaptive characteristics, is proposed for rolling bearing fault diagnosis. First, the original vibration signal is transformed into a one-dimensional frequency domain signal and a two-dimensional time-frequency graph by fast Fourier transform and continuous wavelet transform. Then, the one-dimensional frequency domain signal and two-dimensional time-frequency diagram are input into the two channels of the model respectively to extract and recognize the one-dimensional and two-dimensional features. Finally, the one-dimensional and two-dimensional features are combined in the fusion layer, and the fault types are classified in the softmax layer. FC-CLDCNN has the characteristics of two-stream feature fusion, which can give full consideration to the characteristics of rolling bearing fault data, so as to achieve efficient and accurate identification. The Case Western Reserve University (CWRU) dataset is used for training and testing, and it is proved that the proposed model has high classification accuracy and excellent adaptability across loads. The Machinery Failure Prevention Technology (MFPT) dataset was used to validate the excellent diagnostic performance and generalization of the proposed model.

Transformer Vibration Analysis Based on Double Branch Convolutional Neural Network

The power transformer is one of the important pieces of equipment in the power grid system, and its normal operation is related to the safety and reliability of the whole power system. There are many factors influencing transformer vibration in operation, and its characteristics are complex, so it is difficult to be directly used for transformer state analysis. This paper proposes a method for vibration signal analysis based on a continuous wavelet time-frequency graph. The segmented samples of transformer vibration signals are selected by the time-domain sample segmentation method, and the segmented time sequence samples are transformed by continuous wavelet transform to obtain a two-dimensional time-frequency graph. The time-frequency graph is input into the two-branch convolutional neural network, and the transformer state classification is given based on the features extracted from the network. The simulation analysis on transformer vibration data measured by multiple measuring points shows that the proposed method has an average recognition accuracy of 98.3%. The work in this paper can provide a reference for the vibration analysis of the transformer.

Rotating speed estimation of spinning objects based on rotational Doppler effect

Rotational Doppler effect is an important phenomenon when the vortex electromagnetic wave carrying orbital angular momentum is used to detect a rotating target. Compared with the traditional plane wave case, rotational Doppler effect enables the vortex electromagnetic wave to detect the spin motion along the rotation axis of target. However, there are still some blind zones when the integer orbital angular momentum beams are used to detect specific spinning objects. To expand the application scope of detection scheme based on the rotational Doppler effect, according to the time-frequency analysis, in this paper we study the method of estimating the rotation speed of spinning objects under the normal incidence and oblique incidence of fractional orbital angular momentum beams. Firstly, based on the ideal scattering point model, the echo models are derived under the normal incidence and oblique incidence of integer orbital angular momentum beam and fractional orbital angular momentum beam, respectively, as well as theoretical time-frequency curves. Then taking the three-dimensional practical object for example, the echo under the incidence of fractional orbital angular momentum beams and its time-frequency graph are achieved by using the momentum method and short-time Fourier transform. The time-frequency ridge and its fluctuation period are extracted from the time-frequency graph, thereby estimating the spinning speed of target. The results show that the fractional orbital angular momentum beams can be used to estimate the rotation speed of spinning objects effectively, whether it is normal incidence or oblique incidence, thereby solving the problem about the detection blind zone of integer orbital angular momentum beams, and improving the applicability in detecting spin motion.

Ball screw fault diagnosis based on continuous wavelet transform and two-dimensional convolution neural network

Due to extreme operating conditions such as high-speed and heavy loads, ball screws are prone to damages, that affect the accuracy and operational safety of the mechanical equipment. As strong background noise and weak fault characteristics, it is difficult to capture the inherent fault state only depending on the time-domain or frequency-domain information from the vibration signal. In this paper, a fault diagnosis method for the ball screw based on continuous wavelet transform (CWT) and two-dimensional convolutional neural network (2DCNN) is proposed. The noise-reducing vibration signal is obtained via CWT. The time-frequency graph of the noise reduction signal can more comprehensively reflect the fault information of the ball screw. The time-frequency graph is used as the input to train and test the 2DCNN. Finally, diagnosis results of different types of faults reveal that the proposed CWT-2DCNN fault diagnosis method can achieve an average recognition rate of 99.67%. Compared with one-dimensional convolutional neural network (1DCNN) and traditional BP neural network, the proposed method has fast network convergence and high recognition accuracy. Time-frequency graphs of the noise-reduced signal used as fault features for classification can effectively avoid the problem of uncertainty due to the manual extraction of features. The proposed method has high application potential in the field of ball screw pair fault diagnosis.

Low SNR Multi-Emitter Signal Sorting and Recognition Method Based on Low-Order Cyclic Statistics CWD Time-Frequency Images and the YOLOv5 Deep Learning Model.

It is difficult for traditional signal-recognition methods to effectively classify and identify multiple emitter signals in a low SNR environment. This paper proposes a multi-emitter signal-feature-sorting and recognition method based on low-order cyclic statistics CWD time-frequency images and the YOLOv5 deep network model, which can quickly dissociate, label, and sort the multi-emitter signal features in the time-frequency domain under a low SNR environment. First, the denoised signal is extracted based on the low-order cyclic statistics of the typical modulation types of radiation source signals. Second, the time-frequency graph of multisource signals was obtained through CWD time-frequency analysis. The cyclic frequency was controlled to balance the noise suppression effect and operation time to achieve noise suppression of multisource signals at a low SNR. Finally, the YOLOv5s deep network model is used as a classifier to sort and identify the received signals from multiple radiation sources. The method proposed in this paper has high real-time performance. It can identify the radiation source signals of different modulation types with high accuracy under the condition of a low SNR.

Analysis of the Relationship Between Motor Imagery and Age-Related Fatigue for CNN Classification of the EEG Data.

BackgroundThe aging of the world population poses a major health challenge, and brain–computer interface (BCI) technology has the potential to provide assistance and rehabilitation for the elderly.ObjectivesThis study aimed to investigate the electroencephalogram (EEG) characteristics during motor imagery by comparing young and elderly, and study Convolutional Neural Networks (CNNs) classification for the elderly population in terms of fatigue analysis in both frontal and parietal regions.MethodsA total of 20 healthy individuals participated in the study, including 10 young and 10 older adults. All participants completed the left- and right-hand motor imagery experiment. The energy changes in the motor imagery process were analyzed using time–frequency graphs and quantified event-related desynchronization (ERD) values. The fatigue level of the motor imagery was assessed by two indicators: (θ + α)/β and θ/β, and fatigue-sensitive channels were distinguished from the parietal region of the brain. Then, rhythm entropy was introduced to analyze the complexity of the cognitive activity. The phase-lock values related to the parietal and frontal lobes were calculated, and their temporal synchronization was discussed. Finally, the motor imagery EEG data was classified by CNNs, and the accuracy was discussed based on the analysis results.ResultFor the young and elderly, ERD was observed in C3 and C4 channels, and their fatigue-sensitive channels in the parietal region were slightly different. During the experiment, the rhythm entropy of the frontal lobe showed a decreasing trend with time for most of the young subjects, while there was an increasing trend for most of the older ones. Using the CNN classification method, the elderly achieved around 70% of the average classification accuracy, which is almost the same for the young adults.ConclusionCompared with the young adults, the elderly are less affected by the level of cognitive fatigue during motor imagery, but the classification accuracy of motor imagery data in the elderly may be slightly lower than that in young persons. At the same time, the deep learning method also provides a potentially feasible option for the application of motor-imagery BCI (MI-BCI) in the elderly by considering the ERD and fatigue phenomenon together.

Identifying microseismic events using a dual-channel CNN with wavelet packets decomposition coefficients

Microseismic monitoring is a widely used technique in coal mine safe production. Microseismic events classification is one of the most important steps in microseismic monitoring. CNN has been proven to be the potential tool to identify microseismic events automatically. When the noise is serious and the signal is weak, the recognition ability of CNN is limited. Some scholars use denoising methods, such as wavelet transform, to enhance the signal-to-noise ratio (SNR) of the input data. However, the effective signal will be broken if the incorrect threshold parameters are set in this wavelet transform-based denoising method. In this paper, we intend to make the data volume of the time-frequency graph obtained by WPT as the input of CNN. But it may lead to a dimensional disaster if taking the time-frequency domain signals as input directly. To utilize the effective information extraction and reduce the data dimension, we propose a dual-channel CNN model by combining time domain information and wavelet packet decomposition coefficients (T-WPD CNN). The wavelet packet decomposition coefficients highlight the characteristics of the signal and suppress the characteristics of noise. In addition, the coefficients have the same dimension as the original signal. These advantages are useful for enhancing the performance of CNN. Two field datasets are used to test the network. One from Australia and another from Jiayang coal mine, Sichuan, China. The feature map of the convolutional layer with different inputs are shown to illustrate the influence effect of raw time domain signal and WPD coefficient on the classification performance. The final results show that the T-WPD CNN is superior to the traditional CNN method in accuracy and robustness.

Deep transfer learning-based vehicle classification by asphalt pavement vibration

Deep transfer learning (TL) has great potential for a wide range of applications in civil engineering. This work aims to propose a deep transfer learning-based method for vehicle classification by asphalt pavement vibration. This work first used the pavement vibration IoT monitoring system to collect raw vibration signals and performed the wavelet transform to obtain denoised vibration signals. The vibration signals were then represented in two different ways, including the time-domain graph and the time-frequency graph. Finally, two deep transfer learning-based methods, namely Method Ⅰ (Time-domain & TL) and Method Ⅱ (Time-frequency & TL), were applied for vehicle classification according to the two different representations of vibration signals. The results show that the CNN model had a satisfactory performance in both methods with the accuracy of Method Ⅰ exceeding 0.94 and Method Ⅱ exceeding 0.95. The CNN model in Method Ⅱ performed better in the accuracy metrics with considering label imbalance, but worse in the accuracy metrics without considering label imbalance than that in Method Ⅰ. The differences between these two methods have been investigated and discussed in detail in terms of input types, accuracy metrics, and application prospects. The CNN model with deep transfer learning could be an effective, accurate, and reliable technique for vehicle classification based on asphalt pavement vibration.

Recognition Method of Pipeline Weld Defects Based on Auxiliary Classifier Generative Adversarial Networks

This study aims to address the problem of poorly generated data quality and slow network training speed in data augmentation of 1-dimensional time domain signals. A data augmentation method for the generative adversarial network based on the time-frequency graph of magnetic flux leakage (MFL) signals is proposed. First, the original MFL is converted into a time-frequency signal graph, which is suitable for the input data of the conventional neural network. Second, to produce an adversarial network combined with auxiliary classification, the label of the original MFL signal is used as the input of this network for data augmentation. Compared with a 1-dimensional time domain signal data augmentation method, the proposed method effectively reduces the number of network training parameters and simultaneously solves the problem of slow training speed in traditional methods. Finally, the proposed method is applied to the identification of MFL signals collected by a pipeline company. The experimental results show that the proposed method can produce higher-quality data and effectively improve the accuracy of weld defect magnetic leakage signal recognition and the speed of network training, while the accuracy rate of magnetic leakage signal recognition is increased from 94.5% of the traditional method to 99.5%, which is feasible for practical engineering applications.

Micro-Doppler Extraction of Radar Targets With Translational Motion Based on Spatial Transformer Network

The micro-Doppler (MD) signature, an effect that is induced by micro-motion, characterizes the rich motion information of targets, making it a valuable tool for target classification and recognition. However, the MD structure is always destroyed for radar targets with translational motion. In this letter, we propose an end–to–end translational motion compensation network based on a spatial transformer network (STN). First, the effect of the translational motion on the radar echo and its time-frequency graph (TFG) is modeled, so that translational motion compensation is regarded as an image spatial transformation task. Then, a two-branch convolutional network based on residual modules is designed to locate the transformation parameters, and the TFG containing only MD is obtained using a grid generator and sampler. Finally, the simulation results demonstrate that the proposed algorithm has high accuracy and stability.

Parkinson’s Disease Identification using Deep Neural Network with RESNET50

Recent Parkinson's disease (PD) research has focused on recognizing vocal defects from people's prolonged vowel phonations or running speech since 90% of Parkinson's patients demonstrate vocal dysfunction in the early stages of the illness. This research provides a hybrid analysis of time and frequency and deep learning techniques for PD signal categorization based on ResNet50. The recommended strategy eliminates manual procedures to perform feature extraction in machine learning. 2D time-frequency graphs give frequency and energy information while retaining PD morphology. The method transforms 1D PD recordings into 2D time-frequency diagrams using hybrid HT/Wigner-Ville distribution (WVD). We obtained 91.04% accuracy in five-fold cross-validation and 86.86% in testing using RESNET50. F1-score achieved 0.89186. The suggested approach is more accurate than state-of-the-art models.