Time-frequency Analysis Method Research Articles

Local time-reassigned synchrosqueezing transform and its application in bearing fault characteristic extraction

Rolling bearing fault signals and other types of impulse signals present transient broadband and fast time-varying characteristics in the time–frequency plane, which makes accurately locating the time–frequency trajectory of the signal a challenging task. In this paper, a new time–frequency analysis method is proposed to achieve a high energy-concentrated time–frequency representation for impulsive-like signals. At first, the points on the time–frequency ridges are detected to construct a fixed point set as the reassignment target of the fuzzy energy. Then the local blurred energy is directly assigned to the nearest ridge according to the reassignment direction of time–frequency points. The proposed method ensures a high concentration of energy and supports perfect signal reconstruction. The results of numerical validation prove that the new time–frequency analysis method has excellent noise robustness and energy concentration capability. Finally, the application of bearing fault diagnosis under multiple operating conditions confirms the excellence in the fault diagnosis field.

Transformer winding mechanical fault diagnosis method based on closing transient acoustic vibration signal

The current conventional transformer winding fault diagnosis method mainly extracts the vibration signal through the time-frequency analysis method, which leads to poor diagnosis due to the lack of analysis of the transformer winding state. In this regard, a transformer winding mechanical fault diagnosis method based on the closing transient acoustic vibration signal is proposed. The closing vibration signals of the transformer in the normal and loose winding states are collected separately, the transformer winding force situation is analysed, and the transformer fault signal characteristics are extracted and combined with a fault classifier to construct a fault diagnosis model. In the experiments, the proposed method is verified for fault identification. The experimental results show that the proposed method has a high correct rate when used to identify different types of transformer faults, and has a more desirable fault diagnosis performance.

Optimization of Gearbox Fault Detection Method Based on Deep Residual Neural Network Algorithm.

Because of its long running time, complex working environment, and for other reasons, a gear is prone to failure, and early failure is difficult to detect by direct observation; therefore, fault diagnosis of gears is very necessary. Neural network algorithms have been widely used to realize gear fault diagnosis, but the structure of the neural network model is complicated, the training time is long and the model is not easy to converge. To solve the above problems and combine the advantages of the ResNeXt50 model in the extraction of image features, this paper proposes a gearbox fault detection method that integrates the convolutional block attention module (CBAM). Firstly, the CBAM is embedded in the ResNeXt50 network to enhance the extraction of image channels and spatial features. Secondly, the different time-frequency analysis method was compared and analyzed, and the method with the better effect was selected to convert the one-dimensional vibration signal in the open data set of the gearbox into a two-dimensional image, eliminating the influence of the redundant background noise, and took it as the input of the model for training. Finally, the accuracy and the average training time of the model were obtained by entering the test set into the model, and the results were compared with four other classical convolutional neural network models. The results show that the proposed method performs well both in fault identification accuracy and average training time under two working conditions, and it also provides some references for existing gear failure diagnosis research.

PROGRAM FOR DETERMINING THE INFORMATIVE PARAMETERS OF SURFACE ELECTROMYOGRAPHIC SIGNALS

The article analyzes and calculates informative parameters of surface electromyographic signals (sEMS), which can be used to control biotechnical systems, as well as to diagnose the state of the musculoskeletal system. The analysis parameters in the time and frequency-time domains of the signal are considered. A program has been developed for calculating the informative indicators of the signal in the indicated areas. The program is implemented in the LabVIEW environment. To analyze the sEMG signal in the time domain, using the developed program, such indicators as Integral EMG, Average amplitude change, Wavelength, Simple quadratic integral, Absolute value of the 3rd time moment, and others were calculated; and to describe the signal spectrum by methods of time-frequency analysis, the average frequency of the spectrum (mean power frequency-MPF), the median frequency of the spectrum (median Frequency-MF), root mean square (RMS), power density spectrum (PDS), half width - the width of the spectrum at half maximum amplitude (HW). To test the program, files of real sEMQ signals were used. The calculated parameters of the sEMG analysis in the time and frequency-time domains make it possible to non-invasively and objectively assess the state of the musculoskeletal system. Keywords: Surface electromyographic signals, biotechnical systems, time-frequency analysis, Labvıew software.

Hybrid Method with Parallel-Factor Theory, a Support Vector Machine, and Particle Filter Optimization for Intelligent Machinery Failure Identification

Here, a novel hybrid method of intelligent fault identification within complex mechanical systems was proposed using parallel-factor (PARAFAC) theory and adaptive particle swarm optimization (APSO) for a support vector machine (SVM). The parallel-factor multi-scale analysis theory was studied to reconstruct tensor feature information based on a three-dimensional matrix for time, frequency, and spatial vectors. A multi-scale wavelet analysis was used to transform the original multi-channel experimental data acquired from a gearbox into a three-dimensional feature matrix of the multi-level structure. The optimal correspondence among the two-dimensional feature signals in the frequency and time domains for the different fault modes was established by the PARAFAC theory. An intelligent APSO algorithm was developed to obtain the optimal parameter structures of an SVM classifier. A comparison with the existing time–frequency analysis method showed that the proposed hybrid PARAFAC-PSO-SVM diagnosis model effectively eliminated the redundant information in the multi-dimensional tensor features but retained the important components. The PARAFAC-APSO-SVM hybrid diagnostic model achieved fast, accurate, and simple fault-classification and identification results, and could provide theoretical support for the application of the PARAFAC theory to complex mechanical fault diagnosis.

Time-synchroextracting of generalized S-transform and its application in fault identification

Seismic signals are typical non-stationary signals, and time-frequency analysis method is a powerful tool for processing non-stationary signals, which is also widely used in seismic signal processing. S-transform is a frequently used time-frequency analysis method. However, the time and frequency resolution of S-transform is limited due to Heisenberg uncertainty principle. The post-processing method of time-frequency analysis can solve this problem by further calculation on the basis of the original time-frequency spectrum of the signal. In order to improve the time and frequency resolution, we propose a new time-frequency analysis post-processing method - time synchroextracting of generalized S-transform algorithm. We first deduce the group delay operator expression of the generalized S-transform, and then remove a large amount of fuzzy energy by extracting only the time-frequency coefficients at the group delay operator in the time-frequency spectrum to improve the time-frequency resolution. The comparative analysis of synthetic signals shows that the proposed algorithm has better time-frequency aggregation and can more accurately describe the time-frequency characteristics of seismic signals. We apply the proposed method to the fault identification of the field seismic data, and the results show that the coherent attribute slice extracted by the time-synchroextracting of the generalized S-transform can well describe the fault.

Structural instantaneous frequency identification based on synchrosqueezing fractional Fourier transform

The fractional Fourier transform can transform the signal or function into any intermediate domain between time and frequency domains called fractional Fourier domain, in which the non-stationary signals can be processed. To be a time-frequency analysis tool, the fractional Fourier transform may face the same ridge line blurred problem for noisy signals as other time-frequency analysis methods. To improve the instantaneous frequency identification accuracy of fractional Fourier transform, this paper proposes the synchrosqueezing fractional Fourier transform by using the mechanism that can rearrange the energy in the time-frequency plane and make the frequencies more concentrated on the real frequency of the signal. The synchrosqueezing fractional Fourier transform and its corresponding inverse transformation are theoretically derived. Then the corresponding instantaneous frequency identification based on synchrosqueezing fractional Fourier transform is formulated. The accuracy and effectiveness of proposed instantaneous frequency method is verified by a numerical example of three-degree-of-freedom damped time-varying structural system and experiments of moving vehicle-beam system with simply supported and continuously supported beam tested in the laboratory. Compared with the fractional Fourier transform without synchrosqueezing transform, synchrosqueezing wavelet transform and Fourier synchrosqueezing transform, it is demonstrated that the synchrosqueezing fractional Fourier transform can effectively improve the time-frequency distribution divergence and enhance the accuracy of the instantaneous frequency identification. The method has a certain anti-noise performance. Due to quick calculation by means of traditional fast Fourier transform and superior orthogonal transform feature, the proposed synchrosqueezing fractional Fourier transform can be used as an effective time-frequency signal processing tool for multi-component non-stationary signals.

Multi-attribute Probabilistic Neural Network Reservoir Prediction Based on Mudstone Acoustic Time Difference Decompaction Correction

In order to accurately predict reservoirs with similar physical properties between sandstone and surrounding rocks, this study takes Carboniferous-Permian system in southeastern Ordos Basin as an example, and puts forward a multi-attribute probabilistic neural network prediction method based on acoustic time difference compaction correction. Firstly, the time-frequency analysis method is used to divide the frequency of the original acoustic time difference curve, and the low-frequency long-trend correction of all wells is carried out based on the low-frequency components of standard wells. Then, it is frequency-divided and fused with the original high-frequency components to form new acoustic time difference data, and the long-term trend decompaction correction of acoustic time difference is completed, so that the overlapping area of acoustic time difference between reservoir and surrounding rock is reduced. Multi-attribute probabilistic neural network is an algorithm based on sampling points, before reservoir prediction, inversion parameters must be optimized to save calculation time and reduce prediction error. There is a high correlation between longitudinal velocity and lithologic change trend on the inversion profile of sound velocity. The comparison results of various inversions show that the velocity profile of multi-attribute probabilistic neural network has the highest coincidence rate and high resolution with the velocity curve of logging acoustic wave. The example shows that this method has high accuracy in predicting reservoir thickness, especially for the strata with little difference in physical properties between reservoir and surrounding rock.

Less is More: A Novel Feature Extraction Method for Heart Sound Classification via Fractal Transformation.

Cardiovascular diseases (CVDs) are the leading cause of death globally. Heart sound signal analysis plays an important role in clinical detection and physical examination of CVDs. In recent years, auxiliary diagnosis technology of CVDs based on the detection of heart sound signals has become a research hotspot. The detection of abnormal heart sounds can provide important clinical information to help doctors diagnose and treat heart disease. We propose a new set of fractal features - fractal dimension (FD) - as the representation for classification and a Support Vector Machine (SVM) as the classification model. The whole process of the method includes cutting heart sounds, feature extraction, and classification of abnormal heart sounds. We compare the classification results of the heart sound waveform (time domain) and the spectrum (frequency domain) based on fractal features. Finally, according to the better classification results, we choose the fractal features that are most conducive for classification to obtain better classification performance. The features we propose outperform the widely used features significantly (p < .05 by one-tailed z-test) with a much lower dimension.Clinical relevance-The heart sound classification model based on fractal provides a new time-frequency analysis method for heart sound signals. A new effective mechanism is proposed to explore the relationship between the heart sound acoustic properties and the pathology of CVDs. As a non-invasive diagnostic method, this work could supply an idea for the preliminary screening of cardiac abnormalities through heart sounds.

Synchroextracting frequency synchronous chirplet transform for fault diagnosis of rotating machinery under varying speed conditions

The fault diagnosis of rotating machine is essential to maintain its operational safety and avoid catastrophic accidents. The vibration signals collected from the varying speed rotating machinery are non-stationary, and time–frequency analysis (TFA) is a feasible method for varying speed fault diagnosis by revealing time-varying instantaneous frequency (IF) information in signals. However, most conventional TFA methods are not specifically designed for rotating machinery vibration signals and may not be able to handle these signals, especially in the presence of noise. Therefore, this paper develops a unique TFA method designated as synchroextracting frequency synchronous chirplet transform (SEFSCT) for vibration signal analysis and fault diagnosis of rotating machinery. In the proposed method, the frequency synchronous chirplet transform (FSCT) that utilizes the frequency synchronous chirp rate is first introduced, which takes fully into account the intrinsic proportional relationship of time-varying IFs of the signal. Then, to further concentrate the time–frequency representation (TFR) of FSCT, the synchroextracting operator is constructed based on the Gaussian modulated linear chirp model and the SEFSCT is naturally developed by integrating the FSCT and synchroextracting operator. With the proposed SEFSCT, a high-quality TFR can be generated, thus the time-varying IFs and mechanical failure can be accurately identified. The SEFSCT is employed to deal with synthetic and actual signals, and the results illustrate its efficacy in handling non-stationary signals and diagnosing the mechanical failure.

Research on Diesel Engine Fault Status Identification Method Based on Synchro Squeezing S-Transform and Vision Transformer.

The reliability and safety of diesel engines gradually decrease with the increase in running time, leading to frequent failures. To address the problem that it is difficult for the traditional fault status identification methods to identify diesel engine faults accurately, a diesel engine fault status identification method based on synchro squeezing S-transform (SSST) and vision transformer (ViT) is proposed. This method can effectively combine the advantages of the SSST method in processing non-linear and non-smooth signals with the powerful image classification capability of ViT. The vibration signals reflecting the diesel engine status are collected by sensors. To solve the problems of low time-frequency resolution and weak energy aggregation in traditional signal time-frequency analysis methods, the SSST method is used to convert the vibration signals into two-dimensional time-frequency maps; the ViT model is used to extract time-frequency image features for training to achieve diesel engine status assessment. Pre-set fault experiments are carried out using the diesel engine condition monitoring experimental bench, and the proposed method is compared with three traditional methods, namely, ST-ViT, SSST-2DCNN and FFT spectrum-1DCNN. The experimental results show that the overall fault status identification accuracy in the public dataset and the actual laboratory data reaches 98.31% and 95.67%, respectively, providing a new idea for diesel engine fault status identification.

Wind Turbine Planetary Gearbox Fault Diagnosis via Proportion-Extracting Synchrosqueezing Chirplet Transform

Wind turbine planetary gearboxes usually work under time-varying conditions, leading to nonstationary vibration signals. These signals often consist of multiple time-varying components with close instantaneous frequencies. Therefore, high-quality time-frequency analysis (TFA) is needed to extract the time-frequency feature from such nonstationary signals for fault diagnosis. However, it is difficult to obtain high-quality time-frequency representations (TFRs) through conventional TFA methods due to low resolution and time-frequency blurs. To address this issue, we propose a new TFA method termed the proportion-extracting synchrosqueezing chirplet transform (PESCT). Firstly, the proportion-extracting chirplet transform (PECT) is employed to generate high-resolution underlying TFRs. Then, the energy concentration of the underlying TFRs is enhanced via the synchrosqueezing transform. Finally, wind turbine planetary gearbox fault can be diagnosed by analysis of the dominant time-varying components revealed by the concentrated TFRs with high resolution. The proposed PESCT is suitable for achieving high-quality TFRs for complicated nonstationary signals. Numerical and experimental analyses validate the effectiveness of the PESCT in characterizing the nonstationary signals from wind turbine planetary gearboxes.

Feature Extraction of a Non-Stationary Seismic–Acoustic Signal Using a High-Resolution Dyadic Spectrogram

Using a novel mathematical tool called the Te-gram, researchers analyzed the energy distribution of frequency components in the scale–frequency plane. Through this analysis, a frequency band of approximately 12 Hz is identified, which can be isolated without distorting its constituent frequencies. This band, along with others, remained inseparable through conventional time–frequency analysis methods. The Te-gram successfully addresses this knowledge gap, providing multi-sensitivity in the frequency domain and effectively attenuating cross-term energy. The Daubechies 45 wavelet function was employed due to its exceptional 150 dB attenuation in the rejection band. The validation process encompassed three stages: pre-, during-, and post-seismic activity. The utilized signal corresponds to the 19 September 2017 earthquake, occurring between the states of Morelos and Puebla, Mexico. The results showcased the impressive ability of the Te-gram to surpass expectations in terms of sensitivity and energy distribution within the frequency domain. The Te-gram outperformed the procedures documented in the existing literature. On the other hand, the results show a frequency band between 0.7 Hz and 1.75 Hz, which is named the planet Earth noise.

Synchrosqueezing Transform Based on Frequency-Domain Gaussian-Modulated Linear Chirp Model for Seismic Time–Frequency Analysis

The synchrosqueezing transform (SST) has attracted much attention as a post-processing technique since it was proposed. In recent years, improvements to SST have been made. However, the existing methods are mainly based on the time-domain signal model, and the weak frequency modulation assumption for the components composing the signal is always taken into account. Thus, the signals characterized by a rapidly changing instantaneous frequency (IF) may fail to be adequately tackled. To address this problem, the paper presents a novel seismic time–frequency analysis method via synchrosqueezing transform where a frequency-domain Gaussian modulated linear chirp model is utilized to deduce the SST. The group delay (GD) rather than the IF estimator is implemented to compute an estimation of the ridge. Furthermore, a new synchrosqueezing operator is constructed to rearrange the energy around the ridge. A synthetic example verifies the efficiency and robustness of the proposed SST method, which generates better results than some classic time–frequency analysis (TFA) approaches, e.g., short-time Fourier transform (STFT) and STFT-based SST (FSST). A field dataset further demonstrates this method’s potential in the delineation of subsurface geological structures.

A novel epileptic seizure prediction method based on synchroextracting transform and 1-dimensional convolutional neural network

Background and objectiveEpilepsy is a serious brain disorder affecting more than 50 million people worldwide. If epileptic seizures can be predicted in advance, patients can take measures to avoid unfortunate consequences. Important approaches for epileptic seizure predictions are often signal transformation and classification using electroencephalography (EEG) signals. A time-frequency (TF) transformation, such as the short-term Fourier transform (STFT), has been widely used over many years but curtailed by the Heisenberg uncertainty principle. This research focuses on decomposing epileptic EEG signals with a higher resolution so that an epileptic seizure can be predicted accurately before its episodes. MethodsThis study applies a synchroextracting transformation (SET) and singular value decomposition (SET-SVD) to improve the time-frequency resolution. The SET is a more energy-concentrated TF representation than classical TF analysis methods. ResultsThe pre-seizure classification method employing a 1-dimensional convolutional neural network (1D-CNN) reached an accuracy of 99.71% (the CHB-MIT database) and 100% (the Bonn University database). The experiments on the CHB-MIT show that the accuracy, sensitivity and specificity from the SET-SVD method, compared with the results of the STFT, are increased by 8.12%, 6.24% and 13.91%, respectively. In addition, a multi-layer perceptron (MLP) was also used as a classifier. Its experimental results also show that the SET-SVD generates a higher accuracy, sensitivity and specificity by 5.0%, 2.41% and 11.42% than the STFT, respectively. ConclusionsThe results of two classification methods (the MLP and 1D-CNN) show that the SET-SVD has the capacity to extract more accurate information than the STFT. The 1D-CNN model is suitable for a fast and accurate patient-specific EEG classification.

Rotating machinery fault-induced vibration signal modulation effects: A review with mechanisms, extraction methods and applications for diagnosis

Rotating machinery faults can induce characteristic modulation effects in a vibration signal, and their diagnosis can thus be conducted by extracting fault-induced modulation features. To be specific, such a diagnostic strategy typically involves three main aspects, i.e., fault-induced signal modulation mechanisms, modulation feature extraction methods and applications for diagnosis. To date, the research works on these three aspects all achieved great progresses. A systematic review is thus urgently needed to summarize these achievements, and guide their future directions and developments. This paper aims to fill these gaps and address the needs. Three kinds of typical modulation effects induced by transmission elements are firstly reviewed and summarized. For each kind of modulation effects, a representative phenomenological model is given to formulate and illustrate the relationship between fault-induced modulation patterns and characteristic spectral distributions. As the primary tools to extract signal modulation features, different kinds of time-frequency analysis and signal decomposition methods are then systematically reviewed, including their up-to-date developments, classifications and application scenarios. Some representative works of vibration signal modulation feature extraction-based fault diagnosis are finally introduced along with several typical examples. Based on current research achievements, two potential topics for the future developments of vibration-based diagnostic techniques are the quantitative study of fault feature evolution patterns and the development of more robust and efficient feature extraction algorithms.

An Adaptive Parameterized Domain Mapping Method and Its Application in Wheel-Rail Coupled Fault Diagnosis for Rail Vehicles.

The rapid development of cities in recent years has increased the operational pressure of rail vehicles, and due to the characteristics of rail vehicles, including harsh operating environment, frequent starting and braking, resulting in rails and wheels being prone to rail corrugation, polygons, flat scars and other faults. These faults are coupled in actual operation, leading to the deterioration of the wheel-rail contact relationship and causing harm to driving safety. Hence, the accurate detection of wheel-rail coupled faults will improve the safety of rail vehicles' operation. The dynamic modeling of rail vehicles is carried out to establish the character models of wheel-rail faults including rail corrugation, polygonization and flat scars to explore the coupling relationship and characteristics under variable speed conditions and to obtain the vertical acceleration of the axle box. An APDM time-frequency analysis method is proposed in this paper based on the PDMF adopting Rényi entropy as the evaluation index and employing a WOA to optimize the parameter set. The number of iterations of the WOA adopted in this paper is decreased by 26% and 23%, respectively, compared with PSO and SSA, which means that the WOA performs at faster convergence speed and with a more accurate Rényi entropy value. Additionally, TFR obtained using APDM realizes the localization and extraction of the coupled fault characteristics under rail vehicles' variable speed working conditions with higher energy concentration and stronger noise resistance corresponding to prominent ability of fault diagnosis. Finally, the effectiveness of the proposed method is verified using simulation and experimental results that prove the engineering application value of the proposed method.

Attenuated post-movement beta rebound reflects psychomotor alterations in major depressive disorder during a simple visuomotor task: a MEG study

BackgroundPsychomotor alterations are a common symptom in patients with major depressive disorder (MDD). The primary motor cortex (M1) plays a vital role in the mechanism of psychomotor alterations. Post-movement beta rebound (PMBR) in the sensorimotor cortex is abnormal in patients with motor abnormalities. However, the changes in M1 beta rebound in patients with MDD remain unclear. This study aimed to primarily explore the relationship between psychomotor alterations and PMBR in MDD.MethodsOne hundred thirty-two subjects were enrolled in the study, comprising 65 healthy controls (HCs) and 67 MDD patients. All participants performed a simple right-hand visuomotor task during MEG scanning. PMBR was measured in the left M1 at the source reconstruction level with the time–frequency analysis method. Retardation factor scores and neurocognitive test performance, including the Digit Symbol Substitution Test (DSST), the Making Test Part A (TMT-A), and the Verbal Fluency Test (VFT), were used to measure psychomotor functions. Pearson correlation analyses were used to assess relationships between PMBR and psychomotor alterations in MDD.ResultsThe MDD group showed worse neurocognitive performance than the HC group in all three neurocognitive tests. The PMBR was diminished in patients with MDD compared to HCs. In a group of MDD patients, the reduced PMBR was negatively correlated with retardation factor scores. Further, there was a positive correlation between the PMBR and DSST scores. PMBR is negatively associated with the TMT-A scores.ConclusionOur findings suggested that the attenuated PMBR in M1 could illustrate the psychomotor disturbance in MDD, possibly contributing to clinical psychomotor symptoms and deficits of cognitive functions.

Deep learning-based adaptive mode decomposition and instantaneous frequency estimation for vibration signal

Under time-varying environments and loads, vibration monitoring data of structures exhibit nonstationarity with multiple frequency components, and may have obvious nonlinearity under some extreme loads. The time–frequency analysis method is ideal for analyzing these nonlinear and nonstationary signals. However, the vibration signals of large-scale structures typically contain a variety of frequency components, making it challenging for effective mode decomposition without the interference of false mono-components by existing adaptive mode decomposition methods. False mono-components will severely decrease the accuracy of instantaneous frequency. To address the mode mixing and false mono-components in mode decomposition, a data-driven approach based on deep learning for signal adaptive mode decomposition and instantaneous frequency estimation is proposed in this paper. By transforming the time–frequency analysis problem into an optimization task of a deep neural network, built-in redundant adaptive filters with sparse regularization are designed, then the accurate intrinsic mode functions and time-varying frequencies can be estimated. The proposed method is verified by the numerical nonstationary signal and vibration signal of an experimental long cable model. The results demonstrate the superior capability of the proposed method in decomposing a signal into mono-components adaptively without false mono-components and achieving high accuracy of corresponding instantaneous frequencies. From the examples, an accuracy of 98.52% is achieved for a noise-free signal with spectral overlap. The proposed method is more robust to noise, and the instantaneous frequency identification error is 4.90%, which is much lower than EEMD and VMD under a signal-to-noise ratio of 5 dB.

Research of surface oxidation defects in copper alloy wire arc additive manufacturing based on time-frequency analysis and deep learning method

In this paper a novel method was proposed for monitoring the defects of surface oxidation in copper alloy wire arc additive manufacturing (WAAM) which using voltage sensor. Due to the good fluidity of liquid copper alloy, the dimension of copper alloy WAAM process is difficult to control especially in the height direction. This will lead to error accumulation with the layers increasing during WAAM period. The phenomenon of error accumulation will cause continuous changes in the distance between the welding wire and substrate, result in inappropriate gas protection which will lead to surface oxidation defects happen. To monitoring this defect，time-frequency analysis method was used to compare the voltage signals between normal and anomaly WAAM process. The result show that the voltage signal of abnormal weld bead appears fluctuations in its peak value, and the voltage waveform of the normal WAAM process is relatively smooth. Furthermore, continuous wavelet transform (CWT) method is used to convert one-dimensional time-series voltage signals into two-dimensional time-frequency images, which is combined with deep learning method to establish an online monitoring model to monitor the surface oxidation defects in the process of copper alloy WAAM. The accuracy of the proposed model can reach 95.83%, and t-SNE(t-distributed stochastic neighbor embedding) algorithm is used to visualize the performance of the model. The result indicates that the voltage signal characteristics of normal and abnormal WAAM processes were significantly distinguished.