Fault Severity Assessment Research Articles

Optimal filtering of angle-time cyclostationary signals: Application to vibrations recorded under nonstationary regimes

Mechanical signals are often a mixture of multiple components produced by different sources. The separation of these components is beneficial for differential diagnosis, fault severity assessment and prognosis. An optimal way to accomplish this task is to apply a linear periodically time-variant filter, also known as the cyclic Wiener filter (CWF), assuming the signal to be cyclostationary and assuming that the source periodicity is a priori known. For the first assumption to be valid, the machine must operate under a stationary regime which is generally a restrictive condition. This paper addresses this issue by proposing a formal extension of the CWF to the nonstationary regime case within the angle-time cyclostationary framework. This framework was specifically designed to describe and process machine signals recorded under variable speed conditions. In addition to the theoretical formalisation of the so-called angle-time CWF, a simple and fast algorithm based on the Welch estimator is proposed. The efficiency of the new filter is demonstrated, and compared to the classical one, on synthetic and real vibration signals.

Knowledge extraction from deep convolutional neural networks applied to cyclo-stationary time-series classification

Modelling complex processes from raw time series increases the necessity to build Deep Learning (DL) architectures that can manage this type of data structure. However, as DL models become deeper, larger and more diverse datasets are necessary and knowledge extraction will become more difficult. In an attempt to sidestep these issues, in this paper a methodology based on two main steps is presented, the first being to increase size and diversity of time-series datasets for training, and the second to retrieve knowledge from the obtained model. This methodology is compared with other approaches reported in the literature and is tested under two configuration setups of Condition-Based Maintenance problems: fault diagnosis of bearing, and fault severity assessment of a helical gearbox, obtaining not only a performance improvement in comparison, but also in retrieving knowledge about how the signals are being classified.

An improved singular value decomposition-based method for gear tooth crack detection and severity assessment

Gear tooth crack fault detection and severity assessment using vibration analysis rely on the extraction of fault induced periodic impulses. Singular value decomposition (SVD)-based methods have been used by researchers for periodic impulses extraction. Most such methods extract high-energy signal components (SCs) but ignore weak-energy SCs. A newly reported reweighted SVD (RSVD) method uses periodic modulation intensity (PMI) criterion to measure how influential the periodic impulses are in each SC. This method effectively selects the correct SCs regardless of their energy. However, RSVD suffers from two issues: 1) it does not consider interference from non-fault related vibration components on the PMI, which leads to a high miss and/or false alarm rate; and 2) it directly conduct reweighted summation of all SCs with PMI values that exceed the threshold for signal reconstruction, which undermines periodic impulses extraction. To overcome these issues, this paper proposes an improved SVD-based method. First, the improved method adopts an auto-regression model-based baseline removal approach. SVD is applied to decompose the residual signal rather than the raw signal. Interference from non-fault related vibration components on the PMI is therefore eliminated. Second, the improved method selects the SCs such that the PMI of the reconstructed signal is maximized. Periodic impulses extraction is therefore strengthened. An experimental study was conducted. Results show that the improved method outperforms RSVD in terms of fault detection and severity assessment without creating a considerable computational burden.

A Multi-Stage Hybrid Fault Diagnosis Approach for Rolling Element Bearing Under Various Working Conditions

To timely detect bearing operating condition, and accurately identify bearing fault type and fault severity, a novel multi-stage hybrid fault diagnosis strategy for rolling bearing is proposed in this paper, which mainly consists of three stages (i.e. fault initial detection, fault type recognition and fault severity assessment). Firstly, the procedure of permutation entropy (PE)-based fault initial detection is performed to estimate bearing operating condition. If the bearing fault exists, the next two stages are conducted for fault type recognition and fault severity assessment. Specifically, in the second and third stages, for each dataset under different fault conditions, hybrid-domain features including time-domain, frequency-domain and time-frequency domain are firstly extracted to establish high-dimensional feature space based on statistical analysis and variational mode decomposition (VMD). Then, locality preserving projection (LPP) is introduced to compress high-dimensional dataset into low-dimensional feature space which can reflect preferably intrinsic information of the raw signal and remove information redundancy embedded in hybrid-domain features. Finally, the obtained low-dimensional dataset is imported into Fuzzy C-means (FCM) clustering for recognizing bearing fault type and fault severity. The efficacy of the proposed approach is verified by experimental bearing data under different working conditions. The results indicate that our proposed method can both assess effectively bearing health status and recognize accurately bearing fault type and fault severity. In addition, our proposed approach has higher diagnosis precision than traditional single-stage diagnosis method mentioned in this paper.

Gear fault diagnosis using the general linear chirplet transform with vibration and acoustic measurements

This work is aimed to develop a parameterized time–frequency analysis method combined with vibration and acoustic measurements for gear fault diagnosis. To achieve this aim, the work introduces the combined use of the residual method and general linear chirplet transform using acoustic and vibration measurements from a single stage spur gearbox. Experimental works were undertaken on a developed gearbox test rig. It was found from experiments that despite acoustic measurements were heavily corrupted by measurement noise, the use of the combined general linear chirplet transform method provided more accurate fault severity assessment compared to other commonly used diagnostic methods: continuous wavelet transform and pseudo Wigner–Ville distribution methods. The combined general linear chirplet transform method allows an accurate determination of the angular location of gear fault and a better representation of sidebands associated with the severity level of gear fault. The results demonstrate the potential of using non-contact acoustic measurement using the combined general linear chirplet transform method as an alternative sensing method for gear condition monitoring applications.

A review on data-driven fault severity assessment in rolling bearings

Abstract Health condition monitoring of rotating machinery is a crucial task to guarantee reliability in industrial processes. In particular, bearings are mechanical components used in most rotating devices and they represent the main source of faults in such equipments; reason for which research activities on detecting and diagnosing their faults have increased. Fault detection aims at identifying whether the device is or not in a fault condition, and diagnosis is commonly oriented towards identifying the fault mode of the device, after detection. An important step after fault detection and diagnosis is the analysis of the magnitude or the degradation level of the fault, because this represents a support to the decision-making process in condition based-maintenance. However, no extensive works are devoted to analyse this problem, or some works tackle it from the fault diagnosis point of view. In a rough manner, fault severity is associated with the magnitude of the fault. In bearings, fault severity can be related to the physical size of fault or a general degradation of the component. Due to literature regarding the severity assessment of bearing damages is limited, this paper aims at discussing the recent methods and techniques used to achieve the fault severity evaluation in the main components of the rolling bearings, such as inner race, outer race, and ball. The review is mainly focused on data-driven approaches such as signal processing for extracting the proper fault signatures associated with the damage degradation, and learning approaches that are used to identify degradation patterns with regards to health conditions. Finally, new challenges are highlighted in order to develop new contributions in this field.

Automatic feature extraction of time-series applied to fault severity assessment of helical gearbox in stationary and non-stationary speed operation

Signals captured in rotating machines to obtain the status of their components can be considered as a source of massive information. In current methods based on artificial intelligence to fault severity assessment, features are first generated by advanced signal processing techniques. Then feature selection takes place, often requiring human expertise. This approach, besides time-consuming, is highly dependent on the machinery configuration as in general the results obtained for a mechanical system cannot be reused by other systems. Moreover, the information about time events is often lost along the process, preventing the discovery of faulty state patterns in machines operating under time-varying conditions. In this paper a novel method for automatic feature extraction and estimation of fault severity is proposed to overcome the drawbacks of classical techniques. The proposed method employs a Deep Convolutional Neural Network pre-trained by a Stacked Convolutional Autoencoder. The robustness and accuracy of this new method are validated using a dataset with different severity conditions on failure mode in a helical gearbox, working in both constant and variable speed of operation. The results show that the proposed unsupervised feature extraction method is effective for the estimation of fault severity in helical gearbox, and it has a consistently better performance in comparison with other reported feature extraction methods.

A fuzzy transition based approach for fault severity prediction in helical gearboxes

Rotating machinery is an important device supporting manufacturing processes, and a wide research works are devoted to detecting and diagnosing faults in such machinery. Recently, prognosis and health management in rotating machinery have received high attention as a research area, and some advances in this field are focused on fault severity assessment and its prediction. This paper applies a fuzzy transition based model for predicting fault severity conditions in helical gears. The approach combines Mamdani models and hierarchical clustering to estimate the membership degrees to fault severity levels of samples extracted from historical vibration signals. These membership degrees are used to estimate the weighted fuzzy transitions for modelling the evolution along the fault severity states over time, according to certain degradation path. The obtained fuzzy model is able of predicting the one step-ahead membership degrees to the severity levels of the failure mode under study, by using the current and the previous membership degrees to the severity levels of two available successive input samples. This fuzzy predictive model was validated by using real data obtained from a test bed with different damages of tooth breaking in the helical gears. Results show adequate predictions for two scenarios of fault degradation paths.

Fault identification and severity assessment of rolling element bearings based on EMD and fast kurtogram

Faults in rolling element bearings often cause the breakdown of rotating machinery. Not only the fault type identification but also the fault severity assessment is important. So this paper emphasizes the fault severity assessment. The method proposed in this paper contains two steps: first, identify the fault type based on the combination of empirical mode decomposition (EMD) and fast kurtogram; Second, assess the fault severity. In the first step, the original signal is firstly decomposed into some intrinsic mode functions (IMFs) and the representative IMFs are selected based on correlation analysis, and then the reconstruction signal (RS) is generated; Secondly, the fast kurtogram method is applied to the RS, and the optimum band width and center frequency is obtained. The fault type can be identified based on the fault characteristic frequency marked in the envelope demodulation spectrum. In the second step, the energy percentage of the most fault-related IMF is chosen as an indicator of the fault severity assessment. Experimental data of rolling element bearings inner raceway fault (IRF) with three severities at four running speeds were analyzed. The results show that the IRF identification and fault severity assessment is realized. The breakthrough attempt provides the great potential in the application of condition monitoring of bearings.

Modeling and Dynamic Behavior Analysis of Magnet Defect Signatures in Permanent Magnet Synchronous Motors

In this paper, a motor specific fault severity assessment method is proposed to calculate the amplitude of magnet defect fault signatures in the stator current and back-electromotive force (EMF) through machine and controller parameters of permanent magnet synchronous motor (PMSM) drive. A detailed mathematical analysis is developed based on the linear model of PMSM to predict the behavior of fault signatures in motor variables at various operating points. In order to understand and decouple the effects of motor controllers and operating points, the derivations are further extended to clarify the effects of current loop gains. Under the light of findings, the fault severity impact on the current and back-EMF fault signatures is investigated exhaustively. For this purpose, the fault intensity imposed to a motor is increased in four steps by removing 25% of one magnet at each step. Throughout each fault level, the changes in the corresponding spectrums are analyzed, which provides an essential information to design accurate threshold to minimize false alarms. The theoretical derivations are validated through comparative finite element analysis (FEA) simulations and experiments.

Vibration response mechanism of faulty outer race rolling element bearings for quantitative analysis

For the quantitative fault diagnosis of rolling element bearings, a nonlinear vibration model for fault severity assessment of rolling element bearings is established in this study. The outer race defect size parameter is introduced into the dynamic model, and vibration response signals of rolling element bearings under different fault sizes are simulated. The signals are analyzed quantitatively to observe the relationship between vibration responses and fault sizes. The impact points when the ball rolls onto and away from the defect are identified from the vibration response signals. Next, the impact characteristic that reflects the fault severity in rolling element bearings is obtained from the time interval between two impact points. When the width of the bearing fault is small, the signals are presented as clear single impact. The signals gradually become double impacts with increasing size of defects. The vibration signals of a rolling element bearings test rig are measured for different outer race fault sizes. The experimental results agree well with the results from simulations. These results are useful for understanding the vibration response mechanism of rolling element bearings under various degrees of fault severity.

Acoustic emission-based early fault detection in tapered roller bearings

This paper proposes an acoustic emissions-based method for monitoring tapered roller bearings. This method monitors tracking time-based fault indicators (i.e. RMS, peak value, ring-down count and kurtosis) obtained using the AE signal. Experiments were carried out on a dedicated test bench for different levels of tapered roller bearing outer and inner race defect severity. Although the fault indicators studied could not discriminate between outer and inner raceway defects, the experimental results highlighted the proposed indicators' tapered roller bearing fault detection and fault severity assessment ability (i.e. peak value, RMS and ring-down count).

Gear tooth crack detection using dynamic response analysis

Efficient non-destructive fault detection and severity assessment can be performed using vibration analysis. This paper studies gear tooth crack detection through investigating the natural frequenc ...

Fault severity assessment for rolling element bearings using the Lempel–Ziv complexity and continuous wavelet transform

This paper proposes a new version of the Lempel–Ziv complexity as a bearing fault (single point) severity measure based on the continuous wavelet transform (CWT) results, and attempts to address the issues present in the current version of the Lempel–Ziv complexity measure. To establish the relationship between the Lempel–Ziv complexity and bearing fault severity, an analytical model for a single-point defective bearing is adopted and the factors contributing to the complexity value are explained. To avoid the ambiguity between fault and noise, the Lempel–Ziv complexity is jointly applied with the CWT. The CWT is used to identify the best scale where the fault resides and eliminate the interferences of noise and irrelevant signal components as much as possible. Then, the Lempel–Ziv complexity values are calculated for both the envelope and high-frequency carrier signal obtained from wavelet coefficients at the best scale level. As the noise and other un-related signal components have been largely removed, the Lempel–Ziv complexity value will be mostly contributed by the bearing system and hence can be reliably used as a bearing fault measure. The applications to the bearing inner- and outer-race fault signals have demonstrated that the revised Lempel–Ziv complexity can effectively measure the severity of both inner- and outer-race faults. Since the complexity values are not dependent on the magnitude of the measured signal, the proposed method is less sensitive to the data sets measured under different data acquisition conditions. In addition, as the normalized complexity values are bounded between zero and one, it is convenient to observe the fault growing trend by examining the Lempel–Ziv complexity.

Gear tooth stiffness reduction measurement using modal analysis and its use in wear fault severity assessment of spur gears

Due to excessive service load, inappropriate operating conditions or simply end of life fatigue, damage can occur in gears. When a fault, either distributed or localised, is incurred by gears, the stiffness and consequently vibration characteristics of the damaged tooth will change. A possible non-destructive technique for damage detection and severity assessment can be derived from vibration analysis. This paper presents the use of vibration analysis in the detection, quantification, and advancement monitoring of damage incurred by spur gear teeth. The stiffness of a single spur gear tooth is analysed theoretically, and due to the difficulties in measuring the gear tooth stiffness, an experimental procedure based on the modal analysis is developed to assess the severity of the gear tooth damage. A pair of spur gears was tested under accelerated wear conditions, and conventional time and frequency domain techniques are applied to the gear vibrations to indicate the presence and progression of the wear. The developed modal stiffness assessment technique is then used to quantify the resulting wear damage to the spur gear teeth.

Recurrent Neural Networks for Fault Diagnosis and Severity Assessment of a Screw Compressor

This paper describes a method to diagnose the most frequent faults of a screw compressor and assess magnitude of these faults by tracking changes in compressor’s dynamics. To determine the condition of the compressor, a feedforward neural network model is first employed to identify the dynamics of the compressor. A recurrent neural network is then used to classify the model into one of the three conditions including baseline, gaterotor wear and excessive friction. Finally, another recurrent neural network estimates the magnitude of a fault from the model. The method’s ability to generalize was evaluated. Experimental validation of the method was also performed. The results show significant improvement over the previous method which used only feedforward neural networks.

Feedforward neural networks for fault diagnosis and severity assessment of a screw compressor

Using feedforward neural networks (FNNs), a fault diagnosis and severity assessment (FDIA) scheme for a screw compressor has been established. This FDIA method consists of non-linear model identification of the compressor and pattern classification of parameters of identified models corresponding to the various faulty conditions. First, a non-linear input/output model is identified in the form FNNs. Then an FNN classifies the FNN model into one of the possible faults or the baseline conditions. If the model is classified as a faulty one, another FNN is used to assess the severity of the fault. A fully automatic structure and weight learning algorithm for FNNs is utilised to identify an FNN non-linear model from operating data of the compressor. To establish training data for the classifiers, measurements of motor current and shaft speed were made under baseline conditions and faulty conditions such as various extents of gaterotor wear and increased rolling friction. Experimental results show that the scheme is capable of not only diagnosing the faults but also assessing the magnitude of faults.