bearing fault detection
Recently Published Documents


TOTAL DOCUMENTS

390
(FIVE YEARS 124)

H-INDEX

40
(FIVE YEARS 7)

2021 ◽  
Vol 11 (24) ◽  
pp. 11663
Author(s):  
Eugenio Brusa ◽  
Cristiana Delprete ◽  
Luigi Gianpio Di Maggio

Today’s deep learning strategies require ever-increasing computational efforts and demand for very large amounts of labelled data. Providing such expensive resources for machine diagnosis is highly challenging. Transfer learning recently emerged as a valuable approach to address these issues. Thus, the knowledge learned by deep architectures in different scenarios can be reused for the purpose of machine diagnosis, minimizing data collecting efforts. Existing research provides evidence that networks pre-trained for image recognition can classify machine vibrations in the time-frequency domain by means of transfer learning. So far, however, there has been little discussion about the potentials included in networks pre-trained for sound recognition, which are inherently suited for time-frequency tasks. This work argues that deep architectures trained for music recognition and sound detection can perform machine diagnosis. The YAMNet convolutional network was designed to serve extremely efficient mobile applications for sound detection, and it was originally trained on millions of data extracted from YouTube clips. That framework is employed to detect bearing faults for the CWRU dataset. It is shown that transferring knowledge from sound and music recognition to bearing fault detection is successful. The maximum accuracy is achieved using a few hundred data for fine-tuning the fault diagnosis model.


2021 ◽  
Vol 11 (23) ◽  
pp. 11429
Author(s):  
Jurgen van den Hoogen ◽  
Stefan Bloemheuvel ◽  
Martin Atzmueller

With the developments in improved computation power and the vast amount of (automatic) data collection, industry has become more data-driven. These data-driven approaches for monitoring processes and machinery require different modeling methods focusing on automated learning and deployment. In this context, deep learning provides possibilities for industrial diagnostics to achieve improved performance and efficiency. These deep learning applications can be used to automatically extract features during training, eliminating time-consuming feature engineering and prior understanding of sophisticated (signal) processing techniques. This paper extends on previous work, introducing one-dimensional (1D) CNN architectures that utilize an adaptive wide-kernel layer to improve classification of multivariate signals, e.g., time series classification in fault detection and condition monitoring context. We used multiple prominent benchmark datasets for rolling bearing fault detection to determine the performance of the proposed wide-kernel CNN architectures in different settings. For example, distinctive experimental conditions were tested with deviating amounts of training data. We shed light on the performance of these models compared to traditional machine learning applications and explain different approaches to handle multivariate signals with deep learning. Our proposed models show promising results for classifying different fault conditions of rolling bearing elements and their respective machine condition, while using a fairly straightforward 1D CNN architecture with minimal data preprocessing. Thus, using a 1D CNN with an adaptive wide-kernel layer seems well-suited for fault detection and condition monitoring. In addition, this paper clearly indicates the high potential performance of deep learning compared to traditional machine learning, particularly in complex multivariate and multi-class classification tasks.


Measurement ◽  
2021 ◽  
pp. 110635
Author(s):  
Yuanbo Xu ◽  
Yongbo Li ◽  
Youming Wang ◽  
Junsuo Qu ◽  
Yu Wei ◽  
...  

2021 ◽  
Vol 13 (1) ◽  
Author(s):  
Graeme Garner

Although bearing condition monitoring and fault diagnosis is a widely studied and mature field, applications to automotive wheel bearings have received little attention. This is likely due to the lack of business case, as the vehicle’s four wheel bearings are typically designed to last the vehicle life with low failure rates. Rapid advancements in battery technology are expected to open the door for EVs with million-mile lifespans, exceeding the reliable life of existing low-cost wheel bearing designs. Vehicle designers and fleet owners must choose between paying a higher price for bearings with a longer life or replacing wheel bearings periodically throughout the vehicle life. The latter strategy can be implemented most effectively with the implementation of a low-cost fault detection system on the vehicle.   To develop such a system, data from systems with healthy and faulty wheel bearings is needed. This paper discusses the options for generating this data, such as simulation, bench tests, and vehicle-level tests. The challenges and limitations of each are explored, and the specific challenges of developing an approach for a wheel bearing fault detection system are discussed in detail. A method for injecting Brinell Dent failures is developed, and the results of injecting a total of 40 faulty wheel bearings are presented. Metrics of measuring and summarizing the ground-truth health of a wheel bearing using vibration signals recorded on a test bench are explored. These wheel bearings are used to collect preliminary vehicle data, and some initial analysis is shared highlighting the differences between healthy and faulty wheel bearings, setting the stage for future work to develop a low-cost wheel bearing fault detection system.


2021 ◽  
Vol 10 (1) ◽  
pp. 40
Author(s):  
Guilherme Beraldi Lucas ◽  
Bruno Albuquerque de Castro ◽  
Paulo José Amaral Serni ◽  
Rudolf Ribeiro Riehl ◽  
André Luiz Andreoli

Three-Phase Induction Motors (TIMs) are widely applied in industries. Therefore, there is a need to reduce operational and maintenance costs since their stoppages can impair production lines and lead to financial losses. Among all the TIM components, bearings are crucial in the machine operation once they couple rotor to the motor frame. Furthermore, they are constantly subjected to friction and mechanical wearing. Consequently, they represent around 41% of the motor fault, according to IEEE. In this context, several studies have sought to develop monitoring systems based on different types of sensors. Therefore, considering the high demand, this article aims to present the state of the art of the past five years concerning the sensing techniques based on current, vibration, and infra-red analysis, which are characterized as promising tools to perform bearing fault detection. The current and vibration analysis are powerful tools to assess damages in the inner race, outer race, cages, and rolling elements of the bearings. These sensing techniques use current sensors like hall effect-based, Rogowski coils, and current transformers, or vibration sensors such as accelerometers. The effectiveness of these techniques is due to the previously developed models, which relate the current and vibration frequencies to the origin of the fault. Therefore, this article also presents the bearing fault mathematical modeling for these techniques. The infra-red technique is based on heat emission, and several image processing techniques were developed to optimize bearing fault detection, which is presented in this review. Finally, this work is a contribution to pushing the frontiers of the bearing fault diagnosis area.


Sign in / Sign up

Export Citation Format

Share Document