Rolling Bearing Fault Detection and Classification of Train Bogie Systems Based on Deep Learning Algorithms

Autocorrelated Envelopes for early fault detection of rolling bearings

The scarcity of rolling bearing fault data and the difficulty in annotating fault types prompt us to employ unsupervised fault detection methods. However, existing unsupervised fault detection methods rely on manual feature extraction or use indirect approaches such as reconstruction error or density approximation for fault detection. To address this issue, this paper proposes an unsupervised rolling bearing fault detection method based on a Normalizing flow model, namely the Real-Valued Non-Volume Preserving (Real NVP) model. Compared to existing methods, this approach utilizes vibration data from normal bearings for model training. It transforms the complex distribution of normal bearing data into a simple prior distribution through the Real NVP transformation. Subsequently, fault detection indicators are constructed within this simple prior distribution to achieve rolling bearing fault detection. Experiments on the benchmark bearing dataset demonstrate that the proposed method can effectively identify the fault states of rolling bearings and achieve fault detection results superior to other fault detection methods.

The incipient fault identification of rolling bearings is of great significance in avoiding the occurrence of malignant accidents in rotating machinery. However, at early stages the fault-related features are weak and easily contaminated by environmental noise, making them difficult to identify by traditional methods. Hence, in this paper, a new optimized Fourier spectrum decomposition method, termed bandwidth Fourier decomposition (BFD), is proposed for early fault detection in rolling bearings. Firstly, in the BFD method, the vibration signal is adaptively decomposed into sparse narrow-band sub-signals in the frequency domain through bandwidth optimization. In order to improve the performance of spectrum decomposition, a new bandwidth estimation method and an improved variable initialization strategy are proposed on the basis of spectral energy distribution. Then, the obtained sub-signals are converted into time-domain bandwidth mode functions (BMFs) by inverse Fourier transform. After that, the fault characteristic frequency ratio (FCFR) is introduced to select the effective component from the decomposition results. Finally, the bearing faults are identified by matching the envelope spectrum with the defect frequency of the theoretical calculation. To verify the validity of the proposed method, simulation and experimental analysis are carried out in this paper. Preliminary results indicate that the proposed BFD can effectively enhance the recognition of incipient faults in rolling bearings. The superiority of the proposed BFD is also demonstrated by comparing it with ensemble empirical mode decomposition (EEMD), variational mode decomposition (VMD) and an improved kurtogram method.

A robust construction of normalized CNN for online intelligent condition monitoring of rolling bearings considering variable working conditions and sources

For the purpose of analyzing the application of back propagation (BP) neural network model in retinal vascular segmentation of color fundus images, in this study, the traditional BP neural network was first improved by adopting the additional momentum method. Second, adaptive histogram, morphological background subtraction, Gauss preprocessing matching filter and Heisen matrix were used to enhance the image and extract the features. Third, the retinal vascular segmentation algorithm for color fundus images was constructed based on optimized BP neural network. Finally, the DRIVE and MESSIDOR data sets of color fundus images were introduced to compare the proposed algorithm with the convolutional neural network (CNN) and pulse-coupled neural network (PCNN) algorithms in terms of performance. In addition, the three algorithms were also compared in terms of the sensitivity (Se), specificity (Sp), accuracy (Ac), average operation time for each image and F1 value. The results show that the BP neural network algorithm proposed in this study shows obvious advantage over the other two algorithms in the segmentation of color fundus images. In the DRIVE data set, the Se (81.37%), Sp (90.55%) and Ac (95.82%) of BP algorithm are the highest among the three; in the MESSIDOR data set, the Se (85.22%), Sp (91.08%) and Ac (96.16%) of BP algorithm are also highest; in the DRIVE and MESSIDOR data sets, the operation time of BP algorithm is (28.46 ± 3.19 ms; 24.73 ± 4.81 ms, respectively), which are significantly less than the other two algorithms. Besides, the F1 value of the proposed algorithm is obviously higher than that the other two algorithms. As a result, it is concluded that compared with the CNN algorithm and the PCNN algorithm, the proposed algorithm is more effective in the retinal vascular segmentation of color fundus images; the capillaries have better connectivity, and the proposed algorithm can improve the segmentation Ac while reducing the operation time.

The real-time monitoring of events in an industrial plant is vital, to monitor the actual conditions of operation of the machinery responsible for the manufacturing process. A predictive maintenance program includes condition monitoring of the rotating machinery, to anticipate possible conditions of failure. To increase the operational reliability it is thus necessary an efficient tool to analyze and monitor the equipments, in real-time, and enabling the detection of e.g. incipient faults in bearings. To fulfill these requirements some innovations have become frequent, namely the inclusion of vibration sensors or stator current sensors. These innovations enable the development of new design methodologies that take into account the ease of future modifications, upgrades, and replacement of the monitored machine, as well as expansion of the monitoring system. This paper presents the development, implementation and testing of an instrument for vibration monitoring, as a possible solution to embed in industrial environment. The digital control system is based on an FPGA, and its configuration with an open hardware design tool is described. Special focus is given to the area of fault detection in rolling bearings.

Rolling bearings are the crucial parts in rotating machines and its fault detection is indispensable for ensuring operational reliability of entire mechanical system. To accurately extract fault-induced fault features from vibration signal, this article proposes a hierarchical frequency-domain sparsity-based algorithm (HFDSA). Concretely, the algorithm is established based on the frequency-domain sparsity property of bearing fault features, in which a customized penalty function incorporating weighting strategy is introduced for hierarchically penalizing sparse coefficients. In addition, an iterative convergence algorithm is deduced to solve optimization problem in the proposed method HFDSA and parameter selection method is discussed in detail. Finally, the effectiveness of the HFDSA is verified through the simulation analysis and case studies. The results demonstrate that the HFDSA can efficiently extract fault transients in noisy vibration signal and achieve fault detection of rolling bearings.

The idea of safety region was introduced into the rolling bearing condition monitoring. The safety region estimation and the state identification of the rolling bearing operational were performed by the comprehensive utilization of Empirical Mode Decomposition (EMD), Principal Component Analysis (PCA), and the Least Square Support Vector Machine (LSSVM). The collected vibration data was segmented according to a certain time interval, and then the Intrinsic Mode Functions (IMFs) of each piece of the data were obtained by EMD. The control limits of two statistical variables extracted by PCA were presented as state characteristics. The safety region estimation for the rolling bearing operational status was performed by two-class LSSVM. The states of normal bearing, ball fault, inner race fault, and outer race fault were identified by the multi-class LSSVM. The results show that the estimation accuracy for both the safety region and the states identification reached 95%, and that the validity of the proposed method was verified.

The idea of safety region was introduced into the rolling bearing condition monitoring. The safety region estimation and the state identification of the rolling bearing operational were performed by the comprehensive utilization of Empirical Mode Decomposition (EMD), Principal Component Analysis (PCA), and the Least Square Support Vector Machine (LSSVM). The collected vibration data was segmented according to a certain time interval, and then the Intrinsic Mode Functions (IMFs) of each piece of the data were obtained by EMD. The control limits of two statistical variables extracted by PCA were presented as state characteristics. The safety region estimation for the rolling bearing operational status was performed by two-class LSSVM. The states of normal bearing, ball fault, inner race fault, and outer race fault were identified by the multi-class LSSVM. The results show that the estimation accuracy for both the safety region and the states identification reached 95%, and that the validity of the proposed method was verified.

In order to discover the fault with roller bearing in time, a new fault diagnosis method based on Empirical mode decomposition (EMD) and BP neural network is put forward in the paper. First, we get the fault signal through experiments. Then we use EMD to decompose the vibration signal into a series of single signals. We can extract main fault information from the single signals. The kurtosis coefficient of the single signals forms a feature vector which is used as the input data of the BP neural network. The trained BP neural network can be used for fault identification. Through analyzing, BP neural network can distinguish the fault into normal state, inner race fault, outer race fault. The results show that this method can gain very stable classification performance and good computational efficiency.

As one of the most important parts on the passenger ropeway, the rolling bearing is difficult to be disassembled for in-service inspection. Online monitoring of rolling bearing on passenger ropeway could be carried out by acoustic emission technology. This chapter presents two tests of acoustic emission detection of rolling bearing on a ropeway. The AE characteristics of the normal rolling bearing are investigated using comparative analysis of AE statistical parameters, AE parameter courses, waveform, and spectrum. The results prove that the AE technology has great potential for the condition monitoring and fault diagnosis of rolling bearings on passenger ropeway.

The failure of a rolling bearing may cause the shutdown of mechanical equipment and even induce catastrophic accidents, resulting in tremendous economic losses and a severely negative impact on society. Fault diagnosis of rolling bearings becomes an important topic with much attention from researchers and industrial pioneers. There are an increasing number of publications on this topic. However, there is a lack of a comprehensive survey of existing works from the perspectives of fault detection and fault type recognition in rolling bearings using vibration signals. Therefore, this paper reviews recent fault detection and fault type recognition methods using vibration signals. First, it provides an overview of fault diagnosis of rolling bearings and typical fault types. Then, existing fault diagnosis methods are categorized into fault detection methods and fault type recognition methods, which are separately revised and discussed. Finally, a summary of existing datasets, limitations/challenges of existing methods, and future directions are presented to provide more guidance for researchers who are interested in this field. Overall, this survey paper conducts a review and analysis of the methods used to diagnose rolling bearing faults and provide comprehensive guidance for researchers in this field.

Rolling bearing is a key component of rotary machinery, and its working state tends to affect the reliability and lifetime of the equipment. The traditional way to ensure productivity is to replace rolling bearings regularly which is a waste of resources and could increase maintenance costs, because of the large discreteness of rolling bearing's lifetime. So changing the fail-and-fix practices into a predict-and-prevent methodology is the most effective method. At present, the most common way used for rolling bearing fault diagnosis is vibration signal analysis. This paper explores the Local Mean Decomposition (LMD) algorithm to self-adaptively decompose the vibration signal into several PF components, and further calculates the power spectrum of each PF component, and extracts the sum of power value at the characteristic frequency band of inner ring fault/outer ring fault/ rolling element fault. The sum of power value is used to train the BP neural network. And the experimental results show that the BP neural network trained in this way has high classification ability.

Rolling element bearings are very important components in electrical machines. Almost 50% of the faults that occur in the electrical machines occur in the bearings. This makes bearings as one of the most critical components in electrical machinery. Bearing fault diagnosis has drawn the attention of many researchers. Generally, vibration signals from the machine’s accelerometer are used for the diagnosis of bearing faults. In literature, application of Deep Learning algorithms on these vibration signals has resulted in the fault detection accuracy that is close to 100%. Although, fault detection using vibration signals from the machine is ideal but measurement of vibration signals requires an additional sensor, which is absent in many machines, especially low voltage machines as it significantly adds to its cost. Alternatively, bearing fault diagnosis with the help of the stator current or Motor Current Signal (MCS) is also gaining popularity. This paper uses MCS for the diagnosis of bearing inner raceway and outer raceway fault. Diagnosis using MCS is difficult as the fault signatures are buried beneath the noise in the current signal. Hence, signal-processing techniques are employed for the extraction of the fault features. The paper uses the Paderborn University damaged bearing dataset, which contains stator current data from healthy, real damaged inner raceway, and real damaged outer raceway bearings with different fault severity. Fault features are extracted from MCS by first filtering out the redundant frequencies from the signal and then extracting eight features from the filtered signal, which include three features from time domain and five features from time–frequency domain by using the Wavelet Packet Decomposition (WPD). After the extraction of these eight features, the well-known Deep Learning algorithm Long Short-Term Memory (LSTM) is used for bearing fault classification. The Deep Learning LSTM algorithm is mostly used in speech recognition due to its time coherence, but in this paper, the ability of LSTM is also demonstrated with the fault classification accuracy of 97%. A comparison of the proposed algorithm is done with the traditional Machine Learning techniques, and it is shown that the proposed methodology outperforms all the traditional algorithms which are used for the classification of bearing faults using MCS. The method developed is independent of the speed and the loading conditions.

The defect characteristics of rolling bearing are difficult to excavate at the incipient injury phase; in order to effectively solve this issue, an original strategy fusing recursive singular spectrum decomposition (RSSD) with optimized cyclostationary blind deconvolution (OCYCBD) is put forward to achieve fault characteristic enhanced detection. In this diagnosis strategy, the data-driven RSSD method without predetermined component number is proposed. In addition, a new morphological difference operation entropy (MDOE) indicator, which takes advantage of morphological transformation and Shannon entropy, is developed for confirming the influencing parameters of cyclostationary blind deconvolution (CYCBD). During the process of fault detection, RSSD is firstly adopted to preprocess the original signal, and the most sensitive singular spectrum component (SSC) is selected by the envelope spectrum peak (ESP) indicator. Then, the grid search algorithm is adopted to precisely confirm the optimal parameters and OCYCBD is further performed as a postprocessing technology on the most sensitive component to suppress the residual interferences and amplify the fault signatures. Finally, the enhanced fault detection of rolling bearing is able to achieve by analyzing the envelope spectrum of deconvolution signal. The feasibility of the proposed strategy is verified by the simulated and the measured signals, respectively, and its superiority is also demonstrated through several comparison methods. The results manifest this novel strategy has praisable advantages on weak characteristic extraction and intensification.