Raw Time Domain Research Articles

A comparative study of principal component analysis and kernel principal component analysis for photogrammetric shape-based turbine blade damage analysis

Photogrammetry is popular as a non-contact full-field data acquisition technique for machine condition monitoring purposes. Two popular implementations thereof are Digital Image Correlation (DIC) and 3D Point Tracking (3DPT). Since these two approaches require surface preparation in the form of applying speckle patterns or discrete markers, their use is negated in several industrial applications, such as the online condition monitoring and/or assessment of horizontal axis wind turbines. A shape-based photogrammetric approach that focuses on analysing changes in the form of a rotating blade’s boundary contour is a viable alternative that does not require any surface preparation. 3D shapes of a blade at a particular instance can be extracted from a pair of stereoscopic images of the blade. Shape Principal Component Descriptors (SPCDs) determined from applying Principal Component Analysis (PCA) to Fourier coefficients of chain-code shape signatures of the blade contour can be determined. These can be regarded as indicators of the form of the shape, and as the blade rotates, variations in the SPCDs can be investigated to better understand the dynamics of the blades. This paper focuses on the application and performance evaluation of data reduction and classification techniques for a novel shape-based photogrammetry condition monitoring technique. Investigations are conducted to show that applying data reduction methods to raw time domain SPCDs can result in successful classification of differently damaged blades. Post-processing strategies are developed to ensure a more robust shape based condition monitoring tool for analysing turbine blades. Kernel Principal Component Analysis (KPCA) is implemented and it is shown that it out-performs the conventional PCA in terms of classifying different blade damage modes. The feasibility of using multi-domain statistical features as feature vectors to which PCA or KPCA is applied for classification purposes is also investigated. Results indicating how well differently damaged blades can be distinguished are provided, and it is clearly illustrated that multi-domain statistical features are more robust to noise contamination in the signals compared to using the raw SPCDs data.

A Novel Deep Learning Method for Underwater Target Recognition Based on Res-Dense Convolutional Neural Network with Attention Mechanism

Long-range underwater targets must be accurately and quickly identified for both defense and civil purposes. However, the performance of an underwater acoustic target recognition (UATR) system can be significantly affected by factors such as lack of data and ship working conditions. As the marine environment is very complex, UATR relies heavily on feature engineering, and manually extracted features are occasionally ineffective in the statistical model. In this paper, an end-to-end model of UATR based on a convolutional neural network and attention mechanism is proposed. Using raw time domain data as input, the network model combines residual neural networks and densely connected convolutional neural networks to take full advantage of both. Based on this, a channel attention mechanism and a temporal attention mechanism are added to extract the information in the channel dimension and the temporal dimension. After testing the measured four types of ship-radiated noise dataset in experiments, the results show that the proposed method achieves the highest correct recognition rate of 97.69% under different working conditions and outperforms other deep learning methods.

Deep order-wavelet convolutional variational autoencoder for fault identification of rolling bearing under fluctuating speed conditions

Because of the complex operating environment of high-end industrial machinery, rolling bearing is generally operated at fluctuating working conditions such as variable speeds or loads, thus enables fault feature information is not obvious. That said, bearing fault identification under fluctuating working conditions are recognized as a very challenging problem. Deep learning blazes a valid route to address this issue by right of strong self-learning performance. Nevertheless, the performance of traditional deep learning model will degrade in the face of the fluctuating data with a sharp rising and heavy external interference. Therefore, to overcome this limitation, this study proposes a novel method named deep order-wavelet convolutional variational autoencoder (DOWCVAE) to identify bearing faults under fluctuating speed conditions, which can improve feature learning ability of a plain convolutional variational autoencoder (CVAE). Within this approach, an improved energy-order analysis with frequency-weighted energy operator (FWEO) is firstly presented to convert the raw time-domain vibration signal into the resampled angle-domain signal to relieve the influence of speed fluctuating and acquire the enhanced order spectrum data. Afterwards, wavelet kernel convolutional block (WKCB) with anti-symmetric real Laplace wavelet (ARLW) is constructed to extract the latent feature information closely related to equipment states from the enhanced order spectrum data via the stacked way layer by layer, which is capable of further promoting learning performance of overall network model and improve its generalizability. In addition, a high-efficiency intelligent optimization algorithm termed as multi-objective gray wolf optimizer (MOGWO) is introduced for choosing automatically optimal wavelet parameters of DOWCVAE model and avoiding negative impact posed by artificially adjusting parameter. Ultimately, the learned latent features are loaded to the softmax classifier to achieve automatic identification of different bearing health states and provide comprehensive diagnosis result. The analysis results from two experiment cases testify the effectiveness of our approach. Quantitatively, average identification accuracy of the proposed approach can reach 99% above, which shows its competitive advantages and is more satisfying as compared to some representative deep learning methods.

A Deep-Learning-Based Multi-Modal Sensor Fusion Approach for Detection of Equipment Faults

Condition monitoring is a part of the predictive maintenance approach applied to detect and prevent unexpected equipment failures by monitoring machine conditions. Early detection of equipment failures in industrial systems can greatly reduce scrap and financial losses. Developed sensor data acquisition technologies allow for digitally generating and storing many types of sensor data. Data-driven computational models allow the extraction of information about the machine’s state from acquired sensor data. The outstanding generalization capabilities of deep learning models have enabled them to play a significant role as a data-driven computational fault model in equipment condition monitoring. A challenge of fault detection applications is that single-sensor data can be insufficient in performance to detect equipment anomalies. Furthermore, data in different domains can reveal more prominent features depending on the fault type, but may not always be obvious. To address this issue, this paper proposes a multi-modal sensor fusion-based deep learning model to detect equipment faults by fusing information not only from different sensors but also from different signal domains. The effectiveness of the model’s fault detection capability is shown by utilizing the most commonly encountered equipment types in the industry, such as electric motors. Two different sensor types’ raw time domain and frequency domain data are utilized. The raw data from the vibration and current sensors are transformed into time-frequency images using short-time Fourier transform (STFT). Then, time-frequency images and raw time series data were supplied to the designed deep learning model to detect failures. The results showed that the fusion of multi-modal sensor data using the proposed model can be advantageous in equipment fault detection.

Breast tumor detection by 1D-convolutional neural network based on ultra-wide-band microwave technology

Ultra-wide band (UWB) microwave is a promising technology for non-invasive detection of breast tumors due to its low cost and the absence of ionizing radiation. However, breast heterogeneity is still the main technical barrier of microwave imaging at present. It is challenging to identify whether there are tumor responses from the detected microwave signals because they contain many high-magnitude skin and fibroglandular scattering signals. To this end, a 1D-convolution neural network (CNN) is proposed to detect the presence of breast tumors directly from raw time-domain microwave signals, avoiding complex microwave image reconstruction and time-consuming feature engineering. The designed network has 10 weight layers; it is composed of a stack of convolution-rectified linear unit-batch normalization layers, and every two stacks are followed by a maxpooling layer. At the end of the network are two fully connected layers to complete the detection task. With the proposed network, the detection accuracies on two datasets simulated by an electromagnetic solver reach 98.20% and 95.97%. The accuracy of experimental verification based on the practical UWB detection system remains good, reaching 94.66%. Finally, through the t-distributed stochastic neighbor embedding technique, the results of feature visualization prove the effectiveness of the trained CNN model. The evaluation results indicate the superior performance of the proposed approach for UWB microwave-based breast tumor detection.

Intelligent Fault Diagnosis for Inertial Measurement Unit through Deep Residual Convolutional Neural Network and Short-Time Fourier Transform

An Inertial Measurement Unit (IMU) is a significant component of a spacecraft, and its fault diagnosis results directly affect the spacecraft’s stability and reliability. In recent years, deep learning-based fault diagnosis methods have made great achievements; however, some problems such as how to extract effective fault features and how to promote the training process of deep networks are still to be solved. Therefore, in this study, a novel intelligent fault diagnosis approach combining a deep residual convolutional neural network (CNN) and a data preprocessing algorithm is proposed. Firstly, the short-time Fourier transform (STFT) is adopted to transform the raw time domain data into time–frequency images so the useful information and features can be extracted. Then, the Z-score normalization and data augmentation strategies are both explored and exploited to facilitate the training of the subsequent deep model. Furthermore, a modified CNN-based deep diagnosis model, which utilizes the Parameter Rectified Linear Unit (PReLU) as activation functions and residual blocks, automatically learns fault features and classifies fault types. Finally, the experiment’s results indicate that the proposed method has good fault features’ extraction ability and performs better than other baseline models in terms of classification accuracy.

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.

Bearing Fault Diagnosis Using Multidomain Fusion-Based Vibration Imaging and Multitask Learning.

Statistical features extraction from bearing fault signals requires a substantial level of knowledge and domain expertise. Furthermore, existing feature extraction techniques are mostly confined to selective feature extraction methods namely, time-domain, frequency-domain, or time-frequency domain statistical parameters. Vibration signals of bearing fault are highly non-linear and non-stationary making it cumbersome to extract relevant information for existing methodologies. This process even became more complicated when the bearing operates at variable speeds and load conditions. To address these challenges, this study develops an autonomous diagnostic system that combines signal-to-image transformation techniques for multi-domain information with convolutional neural network (CNN)-aided multitask learning (MTL). To address variable operating conditions, a composite color image is created by fusing information from multi-domains, such as the raw time-domain signal, the spectrum of the time-domain signal, and the envelope spectrum of the time-frequency analysis. This 2-D composite image, named multi-domain fusion-based vibration imaging (MDFVI), is highly effective in generating a unique pattern even with variable speeds and loads. Following that, these MDFVI images are fed to the proposed MTL-based CNN architecture to identify faults in variable speed and health conditions concurrently. The proposed method is tested on two benchmark datasets from the bearing experiment. The experimental results suggested that the proposed method outperformed state-of-the-arts in both datasets.

Prototype design for bidirectional control of stepper motor using features of brain signals and soft computing tools

The brain-computer interface (BCI) plays a significant role in supporting specially-abled people to control devices with the brain signals or electroencephalogram (EEG). The proper functioning of BCI systems requires classification algorithms to distinguish between different tasks based on the features extracted from EEG. Motivated by the requirement of designing a reliable BCI with reduced memory and computational cost, this work proposes an EEG controlled stepper motor drive-based aiding device. The drive system is executed in real-time based on the processing and classification of EEG. The scheme initiates with the extraction of discriminatory features from raw time-domain EEG signals using higher-order statistics (HOS) and phase locking value (PLV).Further, with extracted feature vector, the present study contemplates the operation of backpropagation neural network (BPNN), k-Nearest neighbours (k-NN), and support vector machine (SVM). Signal classification by SVM in conjugation with nonlinear principal component analysis (NLPCA) is implemented to reduce the feature vector dimension. Average classification accuracy of 82.70% and 80.46% is achieved using NLPCA with SVM for PLV and HOS features. The classifier output is utilized to spin the stepper motor clockwise and anticlockwise as needed. The control of stepper motor uses the classifier output and hence the brain signals are implemented on a laboratory-developed digital test bench comprising of TI TMS320F28379D launchpad DSP board and MITSUMI stepper motor. The validation of the proposed scheme for different signals reflects its effectiveness in combining the software and hardware aspects required for realizing an actual BCI system for real-time settings.

Generative adversarial learning for improved data efficiency in underwater target classification

In the realms of the ocean, it becomes a formidable task to detect and classify the passive acoustic targets from the convoluted acoustic mixture confronted by the sonar frontend. Though the advances in deep learning driven by enormity of data and computational infrastructure have resulted in a tremendous leap in performance across various domains, passive sonar target recognition still remains an elusive task for the acoustic as well as signal processing communities. Various channel related artifacts together with the inherent difficulty in obtaining annotated data limit the target records required for training these massive supervised networks so as to yield an optimal performance. This demands models that can generalize well beyond the often sparse training instances. In order to address this issue, generative frameworks can be utilized to model the causal attributes of the target signature so that the network becomes tolerant to the distortions induced by the ambient noise and channel artifacts. This paper exploits the generative modelling capability of an Auxiliary Classifier Generative Adversarial Network (ACGAN) to construct a data-efficient underwater target classifier. These class-conditioned frameworks based on unsupervised representation learning can model the true data distribution using the latent attributes of the training data. In order to make the causal factors of variation more explicit, the raw time domain samples are transformed into joint time–frequency representations using filterbanks initialized at different perceptual scales. Experimental evaluation of the proposed system on target instances collected from diverse locations of the Indian Ocean yields promising results in terms of data efficiency, class confidence and classification accuracy.

Cluster-based acoustic emission signal processing and loading rate effects study of nanoindentation on thin film stack structures

This paper presents a high-resolution, in-situ material testing system that integrates acoustic emission (AE) testing with a nanoindentation system for crack generation and detection in thin film stack structures. This is used to find the critical contact load during wafer probing of crack-sensitive backend-of-line (BEOL) structures in semiconductor integrated circuits. Scanning electron microscopy (SEM) and load–displacement curve analysis were used to confirm the formation and propagation of cracks in the multilayer structures. In order to improve the manual classification performance and understand the physical meaning of AE signals, this paper introduces a machine learning based signal processing approach based on a k-means clustering algorithm applied on collected AE signals. To obtain the optimal number of k-means clusters, Davies–Bouldin, Dunn, and Silhouette indices were calculated, and the individual ratings were cumulated based on a voting scheme. Multiple feature extraction methods, including raw time-domain AE signals, conventional AE extracted parameters, short-term signal energy, and representation features learned by the autoencoder, were used and evaluated by manually labeled clusters and binary confusion matrices. A supervised learning technique, the k-nearest neighbors algorithm, was also utilized on different AE signal datasets using different loading rates to further investigate the damage processes during nanoindentation and the physical meaning of different AE signals. The influences of loading rates on AE signals have been investigated, and loading rate effects on the critical load were observed – higher loading rates led to higher critical loads. This integrated test system and signal processing approach provides a high-resolution mechanical testing platform for studying and enabling automatic crack detection in wafer probing.

Single and Multi-label Fault Classification in rotors from unprocessed multi-sensor data through deep and parallel CNN architectures

An attempt has been made in this study to address two issues related to fault classification in machinery. The first concerns the need of pre-processing raw data collected by sensors and the other is identification in cases where more than one fault exist simultaneously, i.e. where multi-label faults are present. We describe a Convolution Neural Network (CNN) based Deep architecture for classification of Single faults from raw time domain data and a similar but shallower Parallel Multiple Binary Classifier Network for Multi-label Fault Classification. Data from previously conducted experiments on a Rotor-Bearing Fault Simulator are used. In the case of Single-Label fault classification, Support Vector Machines (SVMs), Clustering, Artificial Neural Networks (ANNs) and other algorithms have been used in the past . These procedures require the raw time-domain data collected by sensors, to be first processed and handcrafted into parameters like Fast Fourier Transform (FFT) coefficients, Statistical Moments, etc. before being fed as inputs. Usage of ANN with raw time domain data is inadequate as it suffers from the vanishing gradient problem. We propose a Deep Learning Multi-channel Convolutional Neural Network (McCNN) architecture here, which eliminates this problem and employs the RGB image analogy for channelizing raw time-domain vibration signals from various sub-systems of the rotor-bearing installation. It is shown that this architecture effectively works on raw time-domain data and recognizes all kinds of Single-Label Faults. For Multi-Label faults also, which generally get classified as erroneous Single-Label type through routine codes, the parallel architecture is demonstrated to give good results.

Extreme learning Machine-based classifier for fault diagnosis of rotating Machinery using a residual network and continuous wavelet transform

Effective fault diagnosis of rotating machinery is essential for the predictive maintenance of modern industries. In this study, a novel framework that combines a residual network (ResNet) as a backbone and an extreme learning machine (ELM) as a classifier (RNELM) is proposed to diagnose faults of rotating machinery. Firstly, continuous wavelet transform (CWT) is used to convert the raw time-domain signal into time–frequency domain images. Subsequently, the ResNet backbone in the framework extracts advanced features from the images for the ELM classifier, substantially improving the fault diagnosis performance. The performance of the framework is compared with four other methods using four evaluation metrics on datasets from Case Western Reserve University (CWRU), laboratory results and industrial applications. The experimental results show that the RNELM achieves outstanding results on the test samples of the three datasets, demonstrating the excellent performance and practicability of the proposed framework for fault diagnosis.

Artificial neural network–based fault diagnosis for induction motors under similar, interpolated and extrapolated operating conditions

The diagnosis of mechanical and electrical faults of induction motors (IMs) has been performed using artificial neural networks (ANN) for similar, interpolated and extrapolated operating speeds. The current and vibration signals of faulty and healthy IMs measured from a Machinery Fault Simulator are used in this work. In total, ten different IM fault conditions have been considered: four mechanical faults (bearing fault, unbalanced rotor, misaligned rotor, and bowed rotor), five electrical faults (broken rotor bar, phase unbalanced fault with two severity levels, and stator winding fault with two severity levels), and one healthy motor condition. An ANN model is developed in which raw time domain data of faulty IMs are used and the fault diagnosis is then performed for the motor’s various operating conditions. Initially, diagnosis is performed to predict and classify the motor faults, for the same operating conditions for which we trained ANN. The diagnosis is then extended for interpolated and extrapolated speeds in order to accomplish the diagnosis when data are not available at all the required operating speeds. From the results, it is found that the present ANN-based diagnosis is effective in the same speed case for various operating conditions (seven speeds as well as three loads). In addition, the diagnosis is found to be satisfactory for all interpolated and extrapolated speed cases. It is also observed that the present IM fault diagnosis is better in the interpolation speed cases than the extrapolation speed cases.

Line spectrum extraction based on autoassociative neural networks.

Line spectrum is an important feature for the detection and classification of underwater targets. This letter presents a method for extracting the line spectrum submerged in underwater ambient noise through autoassociative neural networks (AANN). Compared with the traditional methods, the proposed method based on AANN can directly enhance the line spectrum from the raw time-domain noise data without relying on prior information and spectral features. Moreover, the proposed method can suppress the background noise while extracting the line spectrum. Both the numerical simulation and experimental data test results demonstrate that the proposed method provides a good ability to extract the line spectrum from the strong background noise.

A Novel Gas Recognition and Concentration Detection Algorithm for Artificial Olfaction

A novel gas recognition and concentration detection algorithm consisting of a dynamic wavelet convolutional neural network (DWCNN) and a many-to-many long short-term memory-recurrent neural network (LSTM-RNN), respectively, is proposed as a replacement for the traditional data processing algorithm in artificial olfaction. The proposed DWCNN gas recognition algorithm does not require gas signal preprocessing, and it directly converts the raw time-domain gas signal data to 64 * 64 2-D gray images as the input layer of a convolutional neural network (CNN). The experiments show that the recognition accuracy of CO, H2, and the gas mixture of CO and H2 is nearly 100%, and the many-to-many LSTM-RNN algorithm requires only a few labeled data from the steady-state values of the gas sensor array signals. In addition, comparisons with other neural network multilayer perceptrons (MLPs), gated recurrent unit (GRU) algorithms, and conventional algorithms, such as the Bayesian ridge, support vector machines (SVMs), decision tree, k-nearest neighbor (KNN), random forest, AdaBoost, gradient-boosting decision tree (GBDT), and bagging, revealed that the algorithm can obtain a higher concentration detection accuracy, which was evaluated using two different kernel functions for the kernel principal component analysis (KPCA) dimensionality reduction: polynomial and rbf. The experimental results demonstrated that the proposed many-to-many LSTM gas concentration detection model outperformed the abovementioned algorithms and can more accurately estimate the concentration of different gases while using less labeled data.

Spectral features based convolutional neural network for accurate and prompt identification of schizophrenic patients.

Schizophrenia is a fatal mental disorder, which affects millions of people globally by the disturbance in their thinking, feeling and behaviour. In the age of the internet of things assisted with cloud computing and machine learning techniques, the computer-aided diagnosis of schizophrenia is essentially required to provide its patients with an opportunity to own a better quality of life. In this context, the present paper proposes a spectral features based convolutional neural network (CNN) model for accurate identification of schizophrenic patients using spectral analysis of multichannel EEG signals in real-time. This model processes acquired EEG signals with filtering, segmentation and conversion into frequency domain. Then, given frequency domain segments are divided into six distinct spectral bands like delta, theta-1, theta-2, alpha, beta and gamma. The spectral features including mean spectral amplitude, spectral power and Hjorth descriptors (Activity, Mobility and Complexity) are extracted from each band. These features are independently fed to the proposed spectral features-based CNN and long short-term memory network (LSTM) models for classification. This work also makes use of raw time-domain and frequency-domain EEG segments for classification using temporal CNN and spectral CNN models of same architectures respectively. The overall analysis of simulation results of all models exhibits that the proposed spectral features based CNN model is an efficient technique for accurate and prompt identification of schizophrenic patients among healthy individuals with average classification accuracies of 94.08% and 98.56% for two different datasets with optimally small classification time.

A hybrid learning strategy for structural damage detection

Over the past decades, several methods for structural health monitoring have been developed and employed in various practical applications. Some of these techniques aimed to use raw dynamic measurements to detect damage or structural changes. Desirably, structural health monitoring systems should rely on computational tools capable of evaluating the information acquired from the structure continuously, in real time. However, most damage detection techniques fail to identify novelties automatically (e.g. damage, abnormal behaviors, and among others), rendering human decisions necessary. Recent studies have shown that the use of statistical parameters extracted directly from raw time domain data, such as acceleration measurements, could provide more sensitive responses to damage with less computational effort. In addition, machine learning techniques have never been more in trend than nowadays. In this context, this article proposes an original approach based on the combination of statistical indicators—to characterize acceleration measurements in the time domain—and computational intelligence techniques to detect damage. The methodology consists in the combined use of supervised (artificial neural networks) and unsupervised (k-means clustering) learning classification methods for the construction of a hybrid classifier. The objective is to detect not only structural states already known but also dynamic behaviors that have not been identified yet, that is, novelties. The main purpose is to allow a real-time structural integrity monitoring, providing responses in an automatic and continuous way while the structure is under operation. The robustness of the proposed approach is evaluated using data obtained from numerical simulations and experimental tests performed in laboratory and in situ. Results achieved so far attest a promising performance of the hybrid classifier.

Artificial neural network based fault diagnostics for three phase induction motors under similar operating conditions

This paper describes an Artificial Neural Network (ANN) based fault diagnosis methodology for Induction Motors (IM) operating under the same conditions for various speeds and loads. In this study, ten different IM fault conditions are considered. We considered five mechanical faults (bearing fault, unbalanced rotor, misaligned rotor, bowed rotor, rotor with broken bar), four electrical faults (phase unbalance fault with two levels of severity, stator winding fault with two levels of severity), and one healthy motor condition. The current and vibration signals were considered in this work as these signals are generally considered to be the most efficient for the detection of mechanical and electrical faults in IM when used simultaneously. A machine fault simulator was used for the generation of vibration and current signals from different fault conditions. An ANN model was developed in which raw time domain vibration (in three directions) as well as current (in three phases) data are used simultaneously as input and then the fault diagnosis (training and testing) is performed. In this work, the fault diagnosis was attempted when testing was done for the same operating conditions as training. The developed fault diagnosis methods were found to be robust for various operating conditions (speeds and loads) of the IM.

Intelligent fault diagnosis of rotating machinery via wavelet transform, generative adversarial nets and convolutional neural network

Abstract The fault detection of rotating machinery systems especially its typical components such as bearings and gears is of special importance for maintaining machine systems working normally and safely. However, due to the change of working conditions, the disturbance of environment noise, the weakness of early features and various unseen compound failure modes, it is quite hard to achieve high-accuracy intelligent failure monitoring task of rotating machinery using existing intelligent fault diagnosis approaches in real industrial applications. In the paper, a novel and high-accuracy fault detection approach named WT-GAN-CNN for rotating machinery is presented based on Wavelet Transform (WT), Generative Adversarial Nets (GANs) and convolutional neural network (CNN). The proposed WT-GAN-CNN approach includes three parts. To begin with, WT is employed for extracting time-frequency image features from one-dimension raw time domain signals. Secondly, GANs are used to generate more training image samples. Finally, the built CNN model is used to accomplish the fault detection of rotating machinery by the original training time-frequency images and the generated fake training time-frequency images. Two experiment studies are implemented to assess the effectiveness of our proposed approach and the results demonstrate it is higher in testing accuracy than other intelligent failure detection approaches in the literatures even in the interference of strong environment noise or when working conditions are changed. Furthermore, its result in the stability of testing accuracy is also quite excellent.