Search

The Capsule Network (CapsNet) has been shown to have significant advantages in improving the accuracy of bearing fault identification. Nevertheless, the CapsNet faces challenges in identifying the type of bearing fault under nonstationary and noisy conditions. These challenges arise from the distinctive nature of its dynamic routing algorithm and the use of fixed single-scale kernels. To address these challenges, a multi-scale spatial–temporal capsule network (MSCN) based on sequence encoding is proposed for bearing fault identification under nonstationary and noisy environments. A spatial–temporal sequence encoding module focuses on feature correlations at various times and positions. Dilated convolution-based multiscale capsule layer (MCaps) is designed to capture spatial–temporal features at different scales. MCaps establishes connections between various layers, enhancing the comprehension and interpretation of spatial–temporal features. Furthermore, the Bhattacharyya coefficient is introduced into the dynamic routing to compare the similarity between capsules. The validity of the model is verified through comparative experiments, and the results show that MSCN has significant advantages over traditional methods. [ABSTRACT FROM AUTHOR]

In recent years, deep learning-based bearing fault diagnosis methods have made significant achievements. However, these methods only work with single faults and cannot diagnose compound faults because compound fault data is often unavailable in practice. To address this problem, this paper proposes a zero-shot learning-based bearing fault diagnosis method for compound defects. The proposed method utilizes an autoencoder network to observe the attributes of single faults and then estimates the attributes of compound faults. Afterward, a mapping from the data space to the attribute space is established to predict the attribute output of the data. The attribute output is then compared with prior attributes to determine the type of bearing fault. Verification experiments were conducted on HUST bearing dataset. The experimental results showed that the proposed method achieved a high accuracy of 75.64 % in diagnosing compound bearing faults. [ABSTRACT FROM AUTHOR]

Mathematical modeling of bearing faults, worn tool holder taper contact interface, and unbalance are presented and integrated into a digital dynamic model of spindles in Part I of this paper. These faults lead to changes in preload and dynamic stiffness over time, consequently resulting in observable vibrations. This paper predicts the vibrations of a spindle at a particular measurement location by simulating the presence of a specific fault or multiple faults during spindle rotation. The vibration spectra generated by the digital spindle model at the spindle speed and its harmonics, the changes in the natural frequencies, and dynamic stiffnesses are correlated to faults with experimental validations. The simulated vibration spectrums are later used in training an artificial neural network for fault condition monitoring presented in Part III of the paper. [ABSTRACT FROM AUTHOR]

The spindle determines the performance of machine tools; hence, monitoring its health is essential to maintain machining productivity and avoid costly downtimes. The magnitudes and locations of wear and cracks in the bearing balls and races gradually develop which are difficult to detect. This article presents a physics-based digital model of the spindle with bearing faults, worn contact interface between the shaft and tool holder, and spindle imbalance. The wear of races and balls is considered in the bearing model. The worn taper contact interface and the spindle imbalance are included in the digital model. The spindle's dynamic model is used to simulate the vibrations at any location in the spindle assembly where sensors can be mounted for online monitoring. The wear type and bearing location are correlated with the frequency spectrum of vibrations at operating speeds. The proposed fault models are used to analyze the critical signal features and experimentally validated by the frequency extracted from a damaged spindle in Part II. The physics-based digital model is used to train data analytic models to detect spindle faults in Part III. [ABSTRACT FROM AUTHOR]

In recent years, various deep learning techniques have been utilized for dealing with bearing fault diagnosis. Although these methods have achieved remarkable accomplishments, they are challenging to apply in practice due to the difficulty in collecting fully labeled data in the industry. This paper proposes a solution to this issue by developing a generalized approach for simulating the vibration signals of bearings to establish data for deep learning. The proposed solution is based on insights from the classical physical model of bearings combined with an experimental method to estimate the natural frequency parameter and damping coefficient for simulation. The proposed simulation method can generate sufficiently high-quality data for the training of deep learning models. An efficient domain adaptation (DA) method is introduced to create a diagnostic model that can manage the domain shift between simulated training data and real-world testing data. Results from experiments conducted on the HUST bearing dataset indicate the proposed approach achieves a high performance in fault classification with up to 99.75% accuracy. Compared with other simulation methods on classical machine learning and DA models, the proposed approach exhibits superior performance in fault diagnosis and shows great promise for real-world industrial applications. [ABSTRACT FROM AUTHOR]

Abstract The Capsule Network (CapsNet) has been shown to have significant advantages in improving the accuracy of bearing fault identification. Nevertheless, the CapsNet faces challenges in identifying the type of bearing fault under nonstationary and noisy conditions. These challenges arise from the distinctive nature of its dynamic routing algorithm and the use of fixed single-scale kernels. To address these challenges, a multi-scale spatial–temporal capsule network (MSCN) based on sequence encoding is proposed for bearing fault identification under nonstationary and noisy environments. A spatial–temporal sequence encoding module focuses on feature correlations at various times and positions. Dilated convolution-based multiscale capsule layer (MCaps) is designed to capture spatial–temporal features at different scales. MCaps establishes connections between various layers, enhancing the comprehension and interpretation of spatial–temporal features. Furthermore, the Bhattacharyya coefficient is introduced into the dynamic routing to compare the similarity between capsules. The validity of the model is verified through comparative experiments, and the results show that MSCN has significant advantages over traditional methods.

Different from the current rolling bearing fault diagnosis which mainly focuses on the qualitative analysis stage, a quantitative fault diagnosis method for rolling bearings based on image classification is proposed. The overlapping sampling method is used to enhance the one‑dimensional time series data, and then the Markov transition field (MTF) method is used to convert the one‑dimensional time series data into two‑dimensional images, which provide two‑dimensional image samples for inputting into the neural network model and retain the time‑domain information. The ResNeXt and ResNeSt modified residual networks with fine‑tuning processing based on transfer learning are built and trained to classify fault images and realize fault diagnosis. The confusion matrix method and t‑distributed stochastic neighbor embedding (t‑SNE) visualization method are used to carry out experiments. The results show that the proposed method for rolling bearing fault diagnosis can realize the quantitative diagnosis of multi‑working condition rolling bearing fault, and has higher diagnosis accuracy and faster training speed. [ABSTRACT FROM AUTHOR]

Statistical features and wavelet based fault detection are attempted to find computationally less complex, low-memory, and power for real-time implementation. The mean absolute value (MAV), simple sign integral (SSI), waveform length (WL), slope sign change, and zero crossing are extracted from the vibration signal, phase current signal-1, and phase current signal-2. The extracted features are combined varyingly to obtain 31 combinations and classified using a decision tree, k-nearest neighbor {k-NN}, and support vector machine. The identified features {MAV, SSI, WL} performed better with vibration and combined current signals, with an average accuracy of 99.8% and 99.5% with the k-NN classifier, respectively. Wavelet has shown an accuracy of 98%, and the Alexnet method obtained an average accuracy of 97.5% using a combined current signal, which is less than the time domain features-based machine learning approach. In addition, simple time-domain features require memory of 9.6 MB times less than wavelets and 4.18MB times less than Alexnet. The time domain-based technique requires a computation time of 30.21 minutes less than Alexnet and 53.54 minutes less than wavelets. Experimentally, the effectiveness of identified minimal features is verified using an induction motor current signal and achieved 100% accuracy with {MAV, SSI, WL}.

Electrical equipment plays a vital role in industry. Among various electrical equipment, induction motors are quite commonly used in many industrial applications. One of the most common faults that occurs in induction motors is bearing fault. In this article, bearing fault is diagnosed in an induction motor using vibration signals with the help of a simple Artificial Intelligence (AI)-based model. Because, the vibration signals are not dependent on the motor type, simple to measure, cost effective and yields good results. In the proposed system, accurate prediction of bearing condition is carried out using Bayesian optimization-based ensemble classifier (BOEC). The performance of the BOEC-based bearing fault diagnosis system is compared with other conventional techniques and the comparison results confirm the superior performance of the proposed system. Also, the accuracy obtained from the BOEC-based bearing fault diagnosis system is 99.97%. To verify the effectiveness of the proposed system, a hardware prototype is set up in the laboratory and bearing conditions of various induction motors are analyzed. [ABSTRACT FROM AUTHOR]

A bearing fault is one of the major causes of rotating machinery faults. However, in real industrial scenarios, the harsh and complex environment makes it very difficult to collect sufficient fault data. Due to this limitation, most of the current methods cannot accurately identify the fault type in cases with limited data, so timely maintenance cannot be conducted. In order to solve this problem, a bearing fault diagnosis method based on the fractional order Siamese deep residual shrinkage network (FO-SDRSN) is proposed in this paper. After data collection, all kinds of vibration data are first converted into two-dimensional time series feature maps, and these feature maps are divided into the same or different types of fault sample pairs. Then, a Siamese network based on the deep residual shrinkage network (DRSN) is used to extract the features of the fault sample pairs, and the fault type is determined according to the features. After that, the contrastive loss function and diagnostic loss function of the sample pairs are combined, and the network parameters are continuously optimized using the fractional order momentum gradient descent method to reduce the loss function. This improves the accuracy of fault diagnosis with a small sample training dataset. Finally, four small sample datasets are used to verify the effectiveness of the proposed method. The results show that the FO-SDRSN method is superior to other advanced methods in terms of training accuracy and stability under small sample conditions. [ABSTRACT FROM AUTHOR]

Artificial intelligence is gaining traction in bearing fault detection and diagnosis. Generally, signal processing and feature selection are carried out to facilitate the fault classification process; however, classification accuracy tends to degrade under limited training data. In this paper, various artificial intelligence (AI) classification models are studied and compared using raw vibration data without signal processing and feature engineering. A Cosine k-Nearest Neighbours (CosKNN)- based classification model is optimized by integrating a Segmentive Mechanism, resulting in an overall classification F1-score improvement to 90.8% compared to the original classifier's 76.9%. The comparative findings show that the proposed model is suitable for circumstances with limited availability of training data, signal processing tools, and feature engineering tuning. [ABSTRACT FROM AUTHOR]

This manuscript proposes a hybrid technique for fault detection and classification in a three-phase induction motor (IM). The proposed hybrid technique combines enhanced ladder spherical evaluation (ELSE) and recurrent neural networks (RNN); hence, it is called the ELSE–RNN method. The main goal of the manuscript is to detect and categorize the faults that occur in the IM. The key objective of the proposed method is to enhance the performance of supercapacitor (SC) storage technology and fuzzy-tuned proportional–integral (PI) supervision over conventional control. The main contribution of this paper is developing an effective fault diagnosis method for three-phase IMs. The presence of stator, rotor, winding, and bearing faults is employed using signal processing techniques, such as the Hilbert transform and SIFT. The proposed ELSE–RNN technique is utilized with the end goal of detecting and classifying faults. Here, the proposed ELSE–RNN technique recognizes motors’ healthy or unhealthy conditions in many situations to distinguish faults for protection. The proposed ELSE–RNN technique reduces the complexity of detecting and classifying faults with a validated system and increases the system's accuracy. The ELSE–RNN technique is implemented in MATLAB, and its performance is compared to existing techniques. [ABSTRACT FROM AUTHOR]

Purpose: The most common cause of CNC machine failure is the bearing faults that influence the performance of the manufacturing system. The detection and the diagnosis of bearing defects are crucial to the reliable functioning of revolving machinery. This paper suggests an intelligent vibration-based condition and fault diagnostic technique for the identification of bearing faults in CNC machine. Investigational vibration data obtained for various bearings and operational requirements were analyzed to create a structure for the monitoring and classification of bearing defects to determine the health of the machine. Methods: Fault diagnosis was made using Hybrid Signal Decomposition (HSD) methodology for the decomposition of the vibration signal. Vibration features derived from the received decomposed raw signal were chosen using the key component review to eliminate redundant features using Principal Component Analysis (PCA). Subsequently, these related features were input into Discrete AdaBoost and Gentle AdaBoost and compared to traditional Machine Learning (ML) algorithms for the classification of various bearing faults in CNC machine tools. Results: The proposed Gentle AdaBoost is outperformed with 100% classification accuracy by Discrete AdaBoost and other machine learning algorithms. Experimental outcomes suggest that the proposed approach has an immense capacity to prevent unplanned and unnecessary device shutdowns due to loss of bearings in the CNC machine. Conclusion: The success rate achieved using Gentle AdaBoost outperformed Discrete AdaBoost and other traditional machine learning algorithms, the performance rate obtained for defining various bearing states dependent on vibration signals. [ABSTRACT FROM AUTHOR]

In industrial settings, analyzing vibration signals from rolling bearings is a key method for diagnosing faults in rotating machinery. However, these faults often appear as a series of transient events with quickly changing frequencies, which makes it difficult for traditional time-frequency analysis techniques to accurately capture these characteristics. Methods like synchrosqueezing transmute and multi-synchrosqueezing transmute struggle with this task. In this study, we introduce a new method called the Wavelet-Enhanced Transient-Extracting Transform (WTET). Unlike traditional approaches that rely on instantaneous frequency in a time-domain model, WTET uses a group delay (GD) estimator in a frequency-domain model. This allows for more accurate characterization of bearing and rub impact fault signals. Additionally, WTET is better at resisting noise compared to the traditional wavelet transform because it isolates only the central coefficients in the time-frequency domain. WTET also retains the ability to reconstruct the original signal, which makes it effective at extracting different signal modes. Our analysis shows that WTET offers improved diagnostic capability for detecting bearing faults.

Abstract This study presents an enhanced envelope detection technique implemented on a field-programmable gate array (FPGA) to diagnose bearing faults in rotating machinery. Bearing faults frequently result in machinery breakdowns, incurring substantial downtime and maintenance expenses. In our approach, we employ the Teager energy operator (TEO) to extract the vibration signal envelope. Subsequently, we subject the envelope signal to the fast Fourier transform (FFT) to generate the envelope spectrum of the vibration signal. Finally, we further refine the envelope spectrum using TEO for a second time, resulting in a pronounced fault peak that facilitates early fault detection. We evaluate the effectiveness and performance of the proposed method using two distinct types of bearing vibration signals, one being simulated and the other measured. Our findings reveal that the suggested approach outperforms traditional envelope detection methods, leading to a substantially enhanced fault diagnosis capability. For instance, when we assess the characteristic frequency ratio (FCFR) for faults in the inner and outer rings of the bearing using the proposed method, we observe that the FCFR values are significantly elevated, ranging from 160 to 330% higher compared to the analysis performed by the TEO and HT methods. Consequently, this indicates that the proposed approach has the ability to detect faults at an earlier stage than other methods. Furthermore, the FPGA-based implementation makes it suitable for critical industrial applications where rapid fault detection is essential to prevent catastrophic failures.

Diagnosing bearing faults is crucial for maintaining, ensuring reliability, and extending the lifespan of rotary machines. This process helps prevent unexpected downtime in industries, ultimately reducing economic losses caused by the failure of rotary machines. Timely diagnosis of bearing faults is crucial to prevent catastrophic breakdowns, minimize maintenance expenses, and ensure uninterrupted productivity. With industries evolving rapidly and machines operating in increasingly diverse conditions, traditional fault detection methods face limitations. Despite extensive research in recent decades, there is an ongoing need for further advancements to enhance existing fault diagnosis techniques. This study addresses these challenges by utilizing advanced machine learning algorithms Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), Recurrent Neural Network (RNN), Gated Recurrent Unit Network (GRU), Bidirectional LSTM, and for precise bearing fault diagnosis. Leveraging the CWRU dataset encompassing diverse fault classes and machine conditions, a comprehensive data preprocessing pipeline was executed to clean, normalize, and augment the dataset, ensuring model readiness and enhancing performance. Performance analysis revealed the proposed models achieving remarkable accuracies on the CWRU dataset. The CNN and LSTM models attained accuracies of 95%, while the RNN and GRU models achieved accuracies of $97\%$ . Additionally, the Bidirectional LSTM model yielded an accuracy of 96%. These results signify substantial advancements in bearing fault diagnosis, emphasizing the models’ efficacy in accurately detecting and categorizing faults within the 10 classes of the CWRU dataset. The findings underscore the potential of advanced machine learning techniques in revolutionizing fault diagnosis for rotary machines, addressing the persistent need for more robust and accurate diagnostic methodologies.

Rotating machines require bearings to operate smoothly. However, wear, misalignment, and poor lubrication can degrade bearings over time. Fault diagnosis models identify and classify bearing faults. A fault diagnosis model trained in a specific working condition may not perform well in different working conditions. Real-world datasets are mixed with various work environment conditions; therefore, validating a model using different working conditions datasets is necessary. In this study, raw vibrational accelerometer data of variable working conditions is preprocessed using the window length and stride method to generate a data format suitable for evaluating the proposed model. This model employs the Transfer learning-based VGG16 model as the feature extractor and random forest as the classifier, and it has proven to be highly effective. This proposed fault diagnosis model adapts to different work environments and enhances fault classification at variable working conditions. The performance of the proposed model is evaluated using various metrics such as confusion matrix heatmap, t-SNE plot, precision-recall curve and learning curve. Results obtained from these metrics indicate that this model performs well compared to others. The overall accuracy of the model is 99.90%, and both the training and testing of this model are fast. It is evident from the learning curve evaluation that this model is free from over- or under-fitting issues. Overall, this model is reliable and suitable for classifying bearing faults at different working conditions and can be useable for real world purposes.

This study presents an enhanced envelope detection technique implemented on a field-programmable gate array (FPGA) to diagnose bearing faults in rotating machinery. Bearing faults frequently result in machinery breakdowns, incurring substantial downtime and maintenance expenses. In our approach, we employ the Teager energy operator (TEO) to extract the vibration signal envelope. Subsequently, we subject the envelope signal to the fast Fourier transform (FFT) to generate the envelope spectrum of the vibration signal. Finally, we further refine the envelope spectrum using TEO for a second time, resulting in a pronounced fault peak that facilitates early fault detection. We evaluate the effectiveness and performance of the proposed method using two distinct types of bearing vibration signals, one being simulated and the other measured. Our findings reveal that the suggested approach outperforms traditional envelope detection methods, leading to a substantially enhanced fault diagnosis capability. For instance, when we assess the characteristic frequency ratio (FCFR) for faults in the inner and outer rings of the bearing using the proposed method, we observe that the FCFR values are significantly elevated, ranging from 160 to 330% higher compared to the analysis performed by the TEO and HT methods. Consequently, this indicates that the proposed approach has the ability to detect faults at an earlier stage than other methods. Furthermore, the FPGA-based implementation makes it suitable for critical industrial applications where rapid fault detection is essential to prevent catastrophic failures. [ABSTRACT FROM AUTHOR]

The study introduces an Intelligent Diagnosis Framework (IDF) optimized using the Grasshopper Optimization Algorithm (GOA), an advanced swarm intelligence method, to enhance the precision of bearing defect diagnosis in electrical machinery. This area is vital for the energy sector and IoT manufacturing, but the evolving designs of electric motors add complexity to fault identification. Machine learning offers potential solutions but faces challenges due to computational intensity and the need for fine-tuning hyperparameters. The optimized framework, named GOA-IDF, is rigorously tested using experimental bearing fault data from the CWRU database, focusing on the 12,000 drive end and fan end datasets. Compared to existing machine learning algorithms, GOA-IDF shows superior diagnostic capabilities, especially in processing high-frequency data that are susceptible to noise interference. This research confirms that GOA-IDF excels in accurately categorizing faults and operates with increased computational efficiency. This advancement is a significant contribution to fault diagnosis in electrical motors. It suggests that integrating intelligent frameworks with meta-heuristic optimization techniques can greatly improve the standards of health monitoring and maintenance in the electrical machinery domain. [ABSTRACT FROM AUTHOR]

The use of neural network models to monitor bearing vibration signals can easily be affected by noise, which leads to a decrease in the model test accuracy. Therefore, the existence of noise problems increases the requirements for non-linear mapping capability and robustness of deep neural network (DNN) models. In order to deal with the noise problem, the concept of qubit neurons is introduced into a deep learning stacked autoencoder (SAE) model, and a quantum stacked autoencoder (QSAE) model based on qubits and quantum gates is proposed. The properties of SAE layer-by-layer coding and the arithmetic of qubit neurons are combined in the QSAE. The quantum state signal is taken as the input signal to the encoder and the coding activation function and coding weight matrix are redefined by quantum-controlled non-gates and quantum revolving gates, so that the quantum state signal can be coded layer by layer. Experimental results show that the QSAE can train and diagnose noisy experimental data and maintain high test accuracy in an anti-attack test. This shows that the QSAE has non-linear mapping capability and robustness. [ABSTRACT FROM AUTHOR]

As the field of fault diagnosis in electrical machines has significantly attracted the interest of the research community in recent years, several methods have arisen in the literature. Also, raw data signals can be acquired easily nowadays, and, thus, machine learning (ML) and deep learning (DL) are candidate tools for effective diagnosis. At the same time, a challenging task is to identify the presence and type of a bearing fault under noisy conditions, especially when relevant faults are at their incipient stage. Since, in real-world applications and especially in industrial processes, electrical machines operate in constantly noisy environments, a key to an effective approach lies in the preprocessing stage adopted. In this work, an evaluation study is conducted to find the most suitable signal preprocessing techniques and the most effective model for fault diagnosis of 16 conditions/classes, from a low-workload (computational burden) perspective using a well-known dataset. More specifically, the reliability and resiliency of conventional ML and DL models is investigated here, towards rolling bearing fault detection, simulating data that correspond to noisy industrial environments. Diverse preprocessing methods are applied in order to study the performance of different training methods from the feature extraction perspective. These feature extraction methods include statistical features in time-domain analysis (TDA); wavelet packet decomposition (WPD); continuous wavelet transform (CWT); and signal-to-image conversion (SIC), utilizing raw vibration signals acquired under varying load conditions. The noise effect is examined and thoroughly commented on. Finally, the paper provides accumulated usual practices in the sense of preferred preprocessing methods and training models under different load and noise conditions. [ABSTRACT FROM AUTHOR]

Feature selection is a difficult but highly important preliminary step for bearings remaining useful life (RUL) estimation. To avoid the weights setting problem in hybrid metric, this work devotes to conduct feature selection by using a single metric. Due to noise and outliers, an existing feature selection metric, called monotonicity, used for estimating bearings RUL, requires data smoothing processing before adequate implementation. Such a smoothing process may remove significant part of meaningful information from data. To overcome this issue, a mixture distribution analysis-based feature selection metric is proposed. Moreover, based on this new metric, a feature selection approach for bearings RUL estimation is proposed. Numerical experiments benchmarking the proposed method and the existing metric monotonicity method on available real datasets highlight its effectiveness.Article Highlights: A feature selection method for bearings remaining useful life estimation Retrieval the underlying trend of vibration signal Address the drawbacks of smoothing methods that remove meaningful information from data [ABSTRACT FROM AUTHOR]

Abstract Feature selection is a difficult but highly important preliminary step for bearings remaining useful life (RUL) estimation. To avoid the weights setting problem in hybrid metric, this work devotes to conduct feature selection by using a single metric. Due to noise and outliers, an existing feature selection metric, called monotonicity, used for estimating bearings RUL, requires data smoothing processing before adequate implementation. Such a smoothing process may remove significant part of meaningful information from data. To overcome this issue, a mixture distribution analysis-based feature selection metric is proposed. Moreover, based on this new metric, a feature selection approach for bearings RUL estimation is proposed. Numerical experiments benchmarking the proposed method and the existing metric monotonicity method on available real datasets highlight its effectiveness.

Since the rolling bearing fault signal captured by a vibration sensor contains a large amount of background noise, fault features cannot be accurately extracted. To address this problem, a rolling bearing fault feature extraction algorithm based on improved pelican optimization algorithm (IPOA)–variable modal decomposition (VMD) and multipoint optimal minimum entropy deconvolution adjustment (MOMEDA) methods is proposed. Firstly, the pelican optimization algorithm (POA) was improved using a reverse learning strategy for dimensional-by-dimensional lens imaging and circle mapping, and the optimization performance of IPOA was verified. Secondly, the kurtosis-square envelope Gini coefficient criterion was used to select the optimal modal components from the decomposed components of the signal, and MOMEDA was used to process the optimal modal components in order to obtain the optimal deconvolution signal. Finally, the Teager energy operator (TEO) was employed to demodulate and analyze the optimally deconvoluted signal in order to enhance the transient shock component of the original fault signal. The effectiveness of the proposed method was verified using simulated and actual signals. The results showed that the proposed method can accurately extract failure characteristics in the presence of strong background noise interference. [ABSTRACT FROM AUTHOR]

Purpose: Electrical motors are among the most widely used equipment components across many industries. Therefore, monitoring electrical motors for the early detection of faults is essential for uninterrupted production and can save time and money for manufacturers. Bearing faults are one of the most frequently encountered fault types in induction motors. Although standard offline datasets and algorithm designs have been studied extensively in recent years, they lack a full-scale IoT-based monitoring system design for data collection and deployment in the field. In this paper, outer race, inner race, and ball faults of the bearings of a three-phase (1.5 kW, 6 poles) induction motor are studied using vibration signals and an IoT-based monitoring system. Methods: The vibration signals are collected on a new testbed where the Edge processing units can process the signals and also transfer the sensor data over an IoT-based system. Raw data signals, such as vibration, current, and torque, are preprocessed at the Edge. Features are transferred to a database with Message Queuing Telemetry Transport (MQTT) for long-term storage. Only the vibration signals are analyzed for the detection of bearing faults. The monitoring of vibration signals of the motor can be implemented online or offline locally at the Edge with the implemented IoT system. A Convolution Neural Network (CNN)-based deep learning algorithm is utilized to establish a data-driven condition monitoring AI model. The vibration signals are converted into spectrograms using Short Time Fourier Transform (STFT), and then the CNN model is trained. Results: The proposed IoT-based monitoring system solution, combined with the AI method, can successfully process the sensor signals from the motor. The model can distinguish a healthy bearing from three types of faulty bearings with an average accuracy rate of 95%. Conclusion: In this study, unlike the majority of the studies in the literature focusing on algorithm development with standard datasets for motor faults, a new induction motor condition monitoring system based on IoT technologies is designed and implemented. The system is scalable and can be deployed to monitor a fleet of motor equipment in manufacturing. Based on the needs, the AI model can run both at the Edge and at the server side with data analytics tools to monitor the condition of induction motor faults. [ABSTRACT FROM AUTHOR]

Anomaly detection (AD) algorithms can be instrumental in industrial scenarios to enhance the detection of potentially serious problems at a very early stage. Of course, the "Industry 4.0" revolution is fostering the implementation of intelligent data-driven decisions in industry based on increasingly efficient machine learning (ML) algorithms. Most well-known AD methods use a supervised learning approach focusing on fault classification. They assume the availability of labeled data for both normal and anomalous classes. However, in many industrial environments, a labeled set of anomalous data instances is more challenging to obtain than a labeled set of normal data. Hence, this work implements an unsupervised approach based on two different methods using a typical benchmark bearing-fault dataset. The first method relies on the manual extraction of typical vibration metrics provided as input to an ML algorithm. The second one is based on a deep learning (DL) approach, automatically learning latent representation from raw data. The performance metrics demonstrate that both approaches can distinguish the state of a bearing from normal to faulty. DL methodology proves a higher accuracy rate in recognizing faults and a better ability to provide information about the fault size. [ABSTRACT FROM AUTHOR]

Summary: The condition monitoring of squirrel cage induction motors (SCIMs) is vital for uninterrupted production and minimum downtime. Early fault detection can boost output with minimum effort. This article combines the application of transfer learning and convolution neural network (TL‐CNN) for developing an efficient model for bearing and rotor broken bars damage identification in SCIMs. A simple technique for the 1‐D current signal‐to‐image conversion is also proposed to provide input to the proposed deep learning‐based TL‐CNN technique. The proposed approach embodies the advantages of TL and CNN for effective fault identification in SCIMs. The developed technique has classified faults efficiently with an average accuracy of 99.40%. The complete analysis and data collection have been done on the experimental set‐up with a 5 kW SCIM and LabVIEW‐based data acquisition system. The propounded fault detection model has been created in python with the help of packages like Keras and TensorFlow. [ABSTRACT FROM AUTHOR]

A bearing fault is one of the major causes of rotating machinery faults. However, in real industrial scenarios, the harsh and complex environment makes it very difficult to collect sufficient fault data. Due to this limitation, most of the current methods cannot accurately identify the fault type in cases with limited data, so timely maintenance cannot be conducted. In order to solve this problem, a bearing fault diagnosis method based on the fractional order Siamese deep residual shrinkage network (FO-SDRSN) is proposed in this paper. After data collection, all kinds of vibration data are first converted into two-dimensional time series feature maps, and these feature maps are divided into the same or different types of fault sample pairs. Then, a Siamese network based on the deep residual shrinkage network (DRSN) is used to extract the features of the fault sample pairs, and the fault type is determined according to the features. After that, the contrastive loss function and diagnostic loss function of the sample pairs are combined, and the network parameters are continuously optimized using the fractional order momentum gradient descent method to reduce the loss function. This improves the accuracy of fault diagnosis with a small sample training dataset. Finally, four small sample datasets are used to verify the effectiveness of the proposed method. The results show that the FO-SDRSN method is superior to other advanced methods in terms of training accuracy and stability under small sample conditions.

Abstract Objectives The rapid growth of machine learning methods has led to an increase in the demand for data. For bearing fault diagnosis, the data acquisition is time-consuming with complicated processes. Existing datasets are only focused on only one type of bearing, which limits real-world applications. Therefore, the objective of this work is to propose a diverse dataset for ball bearing fault diagnosis based on vibration. Data description In this work, we introduce a practical dataset named HUST bearing, which provides a large set of vibration data on different ball bearings. This dataset contains 99 raw vibration signals of 6 types of defects (inner crack, outer crack, ball crack, and their 2-combinations) on 5 types of bearing (6204, 6205, 6206, 6207, and 6208) at 3 working conditions (0 W, 200 W, and 400 W). Each vibration signal is sampled at a rate of 51,200 samples per second for 10 s. The data acquisition system is elaborately designed with high reliability.

A key problem in the fault diagnosis of rolling element bearings is the extraction of features of repetitive transients from vibration signals. The accurate evaluation of maximizing spectral sparsity under complex interference conditions for measuring the periodicity of transients is typically difficult to implement. Accordingly, a novel periodicity measurement approach was designed for time waveforms. According to the Robin Hood criteria, the Gini index of a sinusoidal signal has a stable low sparsity. The periodic modulation of cyclo-stationary impulses can be represented by several sinusoidal harmonics based on envelope autocorrelation and bandpass filtering. Thus, this low sparsity of Gini index can be used to evaluate the periodic strength of modulation components. Finally, a sequential feature evaluation method is developed to extract periodic impulses accurately. The proposed method is tested on simulation and bearing fault datasets and compared with the state-of-art methods so to assess its effectiveness. • Sinusoidal signal has stable low sparsity of Gini index. • The ASE and band-pass filtering is used to extract the periodic modulation of signal. • The periodicity measure of impulses is designed by the low sparsity of Gini index for filtered ASE. • A Gini-assisted Autogram is proposed to improve the accuracy of blind demodulation method. [ABSTRACT FROM AUTHOR]

Predictive maintenance in machines aims to anticipate failures. In rotating machines, the component that suffers the most wear and tear is the bearings. Currently, based on the Industry 4.0 paradigm, advances have been made in obtaining data, specifically, vibration signals that can be used to predict deterioration using various techniques. In this study, we have applied vibration analysis to obtain features that can be used in an optimal Machine Learning model using a public dataset from CWRU, widely used in research, which contains data on bearing failures. The main objective of this research is to detect bearing failures using a minimum set of observations and selecting the minimum number of features. To achieve this, frequency domain vibration analysis, combined with envelope analysis, is utilized as an effective method for detecting bearing failures. The results were further improved by incorporating an optimal bandwidth determined using the kurtogram. When the results of the envelope analysis are applied to various machine learning models, using the calculated amplitudes as predictors, the Kernel Naive Bayes model achieved an accuracy of 94.4%. Meanwhile, the Decision Tree (Fine Tree) and KNN (Fine KNN) models demonstrate exceptional accuracy, achieving a perfect accuracy rate of 100%.

Induction motors are becoming crucial components in numerous industries. The daily usage of induction motors creates the demand for proper maintenance and slight fault detection to avoid serious damage to the induction motor and the shutdown of industries. Among the various kinds of faults in induction motors, bearing faults, broken rotor bar faults, and short-circuit insulation faults are the most common. Thus, detection and classification of these faults in initial stage are attracting great attention. There are conventional methods for detecting such faults, such as the vibration method for bearing faults, the self-organizing map in the case of broken rotor bar faults, and motor current signature analysis for short-circuit insulation faults. From an industrial point of view, diagnosis methods that can classify all these major faults are required. However, reports on the detection and classification of these faults in initial stage using common diagnosis methods are scarce. In this paper, all three kinds of notable faults in an induction motor were artificially induced, and diagnoses using motor stator current spectral features and the rotation speed of the motor were performed. The diagnosis was accomplished using an auto-tunable and arbitrary featured support vector machine algorithm. Although the faults were minor, a high accuracy rate was obtained. The capability to classify the faults and the high diagnosis accuracy prove the robustness and high sensitivity of the method, enabling its practical applications in industries.

In the fault diagnosis of the existing permanent magnet synchronous motor, the characteristics of the fault are often extracted based on the vibration acceleration signal. The acquisition of vibration acceleration requires the additional installation of expensive sensors, and the diagnosis effect is greatly affected by the surrounding environment. On the other hand, at this time, due to the influence of the load side, the fault characteristics of the bearing and other faults are easily submerged, resulting in a diagnosis failure or even a misdiagnosis. This paper proposes a bearing fault diagnosis method based on the motor speed signal. Combined with the CNN, the feature extraction and analysis of the rotational speed signal are carried out, and an improved algorithm is proposed by combining the artificially selected eigenvalues in the frequency domain. The experimental results show that this method can still complete the diagnosis of bearing faults well in the presence of misalignment fault interference, which shows the potential of deep learning technology represented by CNN in the use of rotational speed signals to diagnose various types of motor faults and provide experimental and theoretical basis for it.

With increasing customer demand, industry 4.0 gained a lot of interest, which is based on smart factories. In smart factories, robotic components are vulnerable to failure due to various industrial operations such as assembly, manufacturing, and product handling. Timely fault detection and diagnosis (FDD) is important to keep the industrial operation smooth. Previously, only the unloaded-based FDD algorithms were considered for the industrial robotic system. In the industrial environment, the robot is working under various working conditions such as speeds, loads, and motions. Hence, to reduce the domain discrepancy between the lab scale and the real working environment, we conducted experimentations under various working conditions. For that purpose, an extensive experimental setup is prepared to perform a series of various experiments mimicking the real environmental condition. In addition, in previous research work, various machine learning (ML) and deep learning (DL) approaches were proposed for robotic arm component fault detection. However, various issues are related to the DL and ML approaches. The ML models are problem-specific, and complex in computations. The DL model needs a huge amount of data. The DL model is composed of various layers that have not been thoroughly explored; as a result, the fault detection model lacks a comprehensive explanation. To overcome these issues, the transfer learning (TL) model is considered with the diverse experimental scenarios. The main contribution is to increase the generalization capabilities of the robotic PHM in the context of previously available research work. For that purpose, the VGG16 model is used because of its autonomous feature extractions for fault classification. The data are collected under a variety of different operating conditions such as loadings, speeds, and motion patterns. The 1D signal is converted to a 2D signal (scalogram) to perform the TL model. The proposed approach shows effective fault detection performance and has the capabilities of generalization under variable working conditions. [ABSTRACT FROM AUTHOR]

The rolling bearing is an important part of mechanical equipment, and its performance significantly affects the quality and life of the mechanical equipment. This article uses the integrated fiber Bragg grating resonant structure sensor excited by periodic micro-shocks caused by micro faults to realize the extraction of information relating to potential faults. Because the fault signal is weak and can easily be interfered with by ambient noise, in order to extract the effective signal, this article determines the autoregressive model of bearing vibration by the final prediction error criterion and the recursive least squares estimation algorithm. The augmented state space model is established based on the autoregressive model. A Kalman filter is used to reduce the noise interference, and then the reduction noisy signal is analyzed by power spectrum and improved autocorrelation envelope spectrum to realize the detection of bearing faults. Through data analysis and method comparison, the proposed improved autocorrelation envelope spectrum analysis can directly extract the bearing fault frequency, which is superior to other methods such as cepstral analysis. [ABSTRACT FROM AUTHOR]

Copyright of Experimental Technology & Management is the property of Experimental Technology & Management Editorial Office and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

As the field of fault diagnosis in electrical machines has significantly attracted the interest of the research community in recent years, several methods have arisen in the literature. Also, raw data signals can be acquired easily nowadays, and, thus, machine learning (ML) and deep learning (DL) are candidate tools for effective diagnosis. At the same time, a challenging task is to identify the presence and type of a bearing fault under noisy conditions, especially when relevant faults are at their incipient stage. Since, in real-world applications and especially in industrial processes, electrical machines operate in constantly noisy environments, a key to an effective approach lies in the preprocessing stage adopted. In this work, an evaluation study is conducted to find the most suitable signal preprocessing techniques and the most effective model for fault diagnosis of 16 conditions/classes, from a low-workload (computational burden) perspective using a well-known dataset. More specifically, the reliability and resiliency of conventional ML and DL models is investigated here, towards rolling bearing fault detection, simulating data that correspond to noisy industrial environments. Diverse preprocessing methods are applied in order to study the performance of different training methods from the feature extraction perspective. These feature extraction methods include statistical features in time-domain analysis (TDA); wavelet packet decomposition (WPD); continuous wavelet transform (CWT); and signal-to-image conversion (SIC), utilizing raw vibration signals acquired under varying load conditions. The noise effect is examined and thoroughly commented on. Finally, the paper provides accumulated usual practices in the sense of preferred preprocessing methods and training models under different load and noise conditions.

Variational mode decomposition (VMD) has been widely applied in sensors. However, the mode number and balance parameter seriously limit VMD application. To solve this problem, this study proposes a novel method, which combines an improved energy fluctuation index (IEFI) and modified VMD (MVMD). In the proposed method, IEFI provided better performance to resist interference from random impulses by considering the periodicity of fault feature components. Consequently, it is applied to determine the initial center frequency for MVMD, which fixed the problem of the mode number. Moreover, a novel balance parameter search strategy, which can adaptively determine the optimal balance parameter, is combined with MVMD whose stop condition is replaced by kurtosis to extract the fault feature. Simulation results indicated that the proposed method does well in detecting the feature of a periodic impulse signal from the signal polluted by some interference impulses. Moreover, the bearing fault diagnosis results demonstrate that the proposed method can accurately detect bearing fault features. Furthermore, the method was validated with bearing fault data. The results showed that the method can accurately extract the fault characteristics of the impulse signal and achieve fault diagnosis.

In this paper, a series of tests were conducted on the bearings of induction motors to investigate vibration signal analysis-based diagnosis of bearing faults, and a thorough analysis was also conducted. In the engineering field, the kurtosis coefficient of vibration acceleration and the root mean square of vibration velocity, as well as resonant demodulated spectrum analysis of vibration acceleration, have been widely used for bearing fault diagnosis. These are integrated in almost any commercially available device for diagnosing bearing faults. However, the unsuitable use of these devices results in many false diagnoses. In light of this, they were selected as research objects and were investigated experimentally. In three induction motors, faults of different severity in the bearing outer race and cage were modeled for tests, and the corresponding results were used to evaluate the performance of the selected diagnosis methods. Some vague information in engineering was clarified, and some instructive rules were outlined to improve the bearing fault diagnosis performance. Taking the kurtosis coefficient of vibration acceleration (Ku) as an example, in engineering, Ku = 4 is generally taken as the diagnostic threshold of bearing faults. This means the following rule applies: if Ku ≤ 4, the bearing is healthy; otherwise, the bearing is faulty. However, the test results in this paper show that even if Ku ≤ 4, the bearing might be faulty; if Ku > 4, the bearing is indeed faulty. Therefore, the diagnostic rule should be improved as follows: if Ku > 4, the bearing is faulty (which can be assured), and if Ku ≤ 4, the status of the bearing is still undetermined. Thus, this paper can be helpful for researchers to gain an experimental understanding of the selected diagnosis methods and provides some improved rules on their use for reducing false diagnoses. [ABSTRACT FROM AUTHOR]

The domain of fault detection has seen tremendous growth in recent years. Because of the growing demand for uninterrupted operations in different sectors, prognostics and health management (PHM) is a key enabling technology to achieve this target. Bearings are an essential component of a motor. The PHM of bearing is crucial for uninterrupted operation. Conventional artificial intelligence techniques require feature extraction and selection for fault detection. This process often restricts the performance of such approaches. Deep learning enables autonomous feature extraction and selection. Given the advantages of deep learning, this article presents a transfer learning–based method for bearing fault detection. The pretrained ResNetV2 model is used as a base model to develop an effective fault detection strategy for bearing faults. The different bearing faults, including the outer race fault, inner race fault, and ball defect, are included in developing an effective fault detection model. The necessity for manual feature extraction and selection has been reduced by the proposed method. Additionally, a straightforward 1D to 2D data conversion has been suggested, altogether eliminating the requirement for manual feature extraction and selection. Different performance metrics are estimated to confirm the efficacy of the proposed strategy, and the results show that the proposed technique effectively detected bearing faults. [ABSTRACT FROM AUTHOR]

Because of the operating environment and load, the main fault form of flywheel bearing is the friction fault between the cage and the rolling elements, which often lead to an increase in the friction torque of the bearing and even to the failure of the flywheel. However, due to the complex mechanism of the friction fault, the characteristic frequencies often used to indicate cage failure are not obvious, which makes it difficult to monitor and quantitatively judge such faults. Therefore, this paper studies the mechanism of the friction fault of the flywheel bearing cage and establishes its fault feature identification method. Firstly, the basic dynamic model of the bearing is established in this paper, and the friction between the cage and the rolling elements is simulated by the variable stiffness. The influence law of the bearing vibration response reveals the relationship between the periodic fluctuation of cage-rolling element friction failure and the bearing load. After analyzing the envelope spectrum of the vibration data, it was found that when a friction fault occurred between the cage and the rolling element, the rotation frequency component of the cage modulated the rotational frequency component of the rolling element, that is, the side frequency components appeared on both sides of the characteristic frequency of the rolling element (with the characteristic frequency of the cage as the interval). In addition, the modulation frequency components of the cage and rolling element changed with the severity of the fault. Then, a modulation sideband ratio method based on envelope spectrum was proposed to qualitatively diagnose the severity of the cage-rolling element friction faults. Finally, the effectiveness of the presented method was verified by experiments. [ABSTRACT FROM AUTHOR]

• Industrial robot's servomotor bearing fault detection is presented. • Simple motion and welding motion of industrial robots is considered. • Currents signals of servomotor are converted to scalogram images for analysis. • Transfer learning expunged model training from scratch and streamlined operation. In consequence of their superior performance and durability, industrial robots have enjoyed widespread adoption across a variety of industries. However, despite their sturdy build, they are susceptible to malfunction. The servomotor is a fundamental component of industrial robots, and to ensure smooth and uninterrupted functioning, it is essential to detect any defects it may develop. Although research has addressed methods for detecting bearing failure, diagnosis of a servomotor bearing failure in the industrial robot remains difficult and requires intensive research. In this paper, a novel method for detecting servomotor bearing defects in the industrial robot is provided by integrating knowledge transfer via transfer learning. Initially, current signals of the servomotor are transformed to scalogram images. This processed data is utilized to build the model for fault detection. Applying transfer learning eliminates model training from scratch and streamlined operations. The purported approach shows an average accuracy of more than 99 %. [ABSTRACT FROM AUTHOR]

Since the rolling bearing fault signal captured by a vibration sensor contains a large amount of background noise, fault features cannot be accurately extracted. To address this problem, a rolling bearing fault feature extraction algorithm based on improved pelican optimization algorithm (IPOA)–variable modal decomposition (VMD) and multipoint optimal minimum entropy deconvolution adjustment (MOMEDA) methods is proposed. Firstly, the pelican optimization algorithm (POA) was improved using a reverse learning strategy for dimensional-by-dimensional lens imaging and circle mapping, and the optimization performance of IPOA was verified. Secondly, the kurtosis-square envelope Gini coefficient criterion was used to select the optimal modal components from the decomposed components of the signal, and MOMEDA was used to process the optimal modal components in order to obtain the optimal deconvolution signal. Finally, the Teager energy operator (TEO) was employed to demodulate and analyze the optimally deconvoluted signal in order to enhance the transient shock component of the original fault signal. The effectiveness of the proposed method was verified using simulated and actual signals. The results showed that the proposed method can accurately extract failure characteristics in the presence of strong background noise interference.

This paper proposes a method of eliminating common-mode voltage (CMV) and shaft voltage in dual three-phase motors using the carrier phase shift of space vector pulse width modulation (SVPWM). SVPWM for the dual three-phase motor is simply implemented using two traditional three-phase SVPWM structures. The representative work of the study is to eliminate the shaft voltage by applying the carrier phase shift method due to the cancellation of CMV. It was verified that the proposed method could reduce the shaft voltage with a simple technique, rather than applying a complex algorithm to reduce it. Also, the motor current and torque behaviors are analyzed to confirm that the proposed method does not affect the motor characteristics. The proposed method has been analyzed based on an equivalent circuit model and it has been verified through the experiment that the shaft voltage is simply eliminated by applying the proposed method.