To reduce the economic losses caused by bearing failures and prevent safety accidents, it is necessary to develop an effective method to predict the remaining useful life (RUL) of the rolling bearing. However, the degradation inside the bearing is difficult to monitor in real-time. Meanwhile, external uncertainties significantly impact bearing degradation. Therefore, this paper proposes a new bearing RUL prediction method based on long-short term memory (LSTM) with uncertainty quantification. First, a fusion metric related to runtime (or degradation) is proposed to reflect the latent degradation process. Then, an improved dropout method based on nonparametric kernel density is developed to improve estimation accuracy of RUL. The PHM2012 dataset is adopted to verify the proposed method, and comparison results illustrate that the proposed prediction model can accurately obtain the point estimation and probability distribution of the bearing RUL.
The effects of oxidative aging on the static and dynamic properties of nitrile rubber at the molecular scale were investigated by molecular dynamics simulation. The aged nitrile rubber models were constructed by introducing hydroxyl groups and carbonyl groups into rubber molecular chains to mimic oxidative aging. The static and dynamic properties of the unaged and aged nitrile rubber under different conditions were evaluated by mean square displacement, self-diffusion coefficients, hydrogen bond, fractional free volume, radial distribution function, cohesive energy density and solubility parameter. The results show that the elevated temperature intensified significantly the mobility of rubber molecular chains and fractional free volume, while the compressive strain displayed the opposite effect resulting in packing and rearrangement of rubber chains. The introduction of hydroxyl groups and carbonyl groups enhanced the polarity, intermolecular interactions, the volume and rigidity of molecular chains, implying weaker mobility of molecular chains as compared to unaged models. The compressive strain and oxidative aging both decreased the fractional free volume, which inhibited gaseous and liquid diffusion into the rubber materials, and slowed down the oxidative aging rate. This study provides insights to better understand the effect of molecular changes due to oxidative aging on the structural and dynamic properties of rubber materials at the molecular level.
Early detection of defects inside a rail is of great significance to ensure the safety of rail transit. This work investigated the ability of ultrasonic guided waves (UGWs) to detect internal defects in a rail head. First, the model of UGW propagation in rail, which has an irregular cross-section, was constructed based on the semi-analytical finite element (SAFE) method. Fundamental characteristics, such as wavenumber, phase or group velocity, and wave structure inside the rail, were then calculated. Following modal and vibration energy distribution analysis, a guided wave mode that is sensitive to transverse fissure (TF) defects was selected, and its excitation method was proposed. The effectiveness of the excitation method was confirmed by simulations performed in the ABAQUS software. According to the simulation data, the dispersion curve calculated by using the two-dimensional Fourier fast transform (2D-FFT) coincided well with that of the SAFE method. After that, the sensitivity of the selected mode to internal rail defects was validated and its ability to locate defects was also demonstrated. Finally, the effects of excitation frequency, defect size, and vertical and horizontal defect depth on the reflection waveforms were investigated.
Early detection of defects inside the rail is of great significance to ensure the safety of rail transit. This work investigates the ability of ultrasonic guided waves (UGWs) to detect internal defects of the rail head. First, the model of UGWs propagation in rail, which has an irregular cross-section, is constructed based on semi-analytical finite element (SAFE). Fundamental characteristics such as wavenumber, phase or group velocity, and wave structure inside the rail are then calculated. Following modal and vibration energy distribution analysis, a guided wave mode sensitive to transverse fissure (TF) defects is selected and its excitation method is proposed. The effectiveness of the excitation method is confirmed by simulations performed in the ABAQUS software. According to the simulation data, the dispersion curve calculated by using the two-dimensional Fourier Fast Transform (2D-FFT) coincides well with that of SAFE. After that, the sensitivity of the selected mode to rail internal defect have been validated and its ability to locate defects has also been demonstrated. Finally, the effect of excitation frequency, defect size, defect vertical and horizontal depth on the reflection waveforms is investigated.
The ultrasonic guide wave (UGW) has good application prospects in steel rail damage diagnosis, but the features of the rail damage implied in the UGW are complex. Deep learning enables an end-to-end approach to fault diagnosis. Nevertheless, a large amount of diversity data is needed for training, whereas the ultrasonic wave guide signals of simulation and repeated experiments lack diversity. Therefore, in this paper, a diagnostic framework based on simulation and transfer learning for rail damage is developed to tackle the problems mentioned above. The proposed framework is based on deep learning with a simulation pretraining strategy to build convolutional neural network (CNN) models through parameter fine-tuning for damage diagnosis. Specifically, for the problem that the simulation data lacks diversity, a damage mechanism-based data diversity augmentation method is proposed; this obtains the diagnostic high-value simulation data including supporting features, and expanded the diversity of the simulation data. Adopting the proposed method of data augmentation and transfer learning (TL), a diagnostic model for rail damage utilizing augmented UGW signals is constructed. The finite element simulation data of UGW with damages at different locations and depths of rails are augmented to achieve the pretraining of CNN models, and the model transfer is performed with the experimental data of rails. Ultimately, through comparative studies it can be concluded that (1) The TL diagnostic framework makes full use of the finite element simulation data to realize the model pretraining. (2) The proposed data augmentation method realizes the diversity expansion of simulation data containing supporting features and ensures the efficient application of simulation data in model pretraining.
Convenient and fast fault diagnosis is the key to improving the service safety and maintenance efficiency of gearboxes. However, the environment and working conditions under complex service conditions are variable, and there is a lack of fault samples in engineering applications. These factors lead to difficulties in intelligent diagnosis methods based on machine learning, while traditional mechanism-based fault diagnosis requires high expertise and long time periods for the manual analysis of data. For the requirements of diagnostic convenience, an automatic fault diagnosis method for gearboxes is proposed in this paper. The method achieves accurate acquisition of rotational speed by constructing a rotational frequency search algorithm. The self-referencing characteristic frequency identification method is proposed to avoid manual signal analysis. On this basis, a framework of anti-interference automatic diagnosis is constructed to realize automatic diagnosis of gear faults. Finally, a gear fault experiment is carried out based on a high-fidelity experimental bench of bogie to verify the effectiveness of the proposed method. The proposed automatic diagnosis method does not rely on a large number of fault samples and avoids the need for diagnosis through professional knowledge, thus saving time for data analysis and promoting the application of fault diagnosis methods.
Bolt connections are subjected to severe service conditions, such as cyclic loading and mechanical shock, leading to loosening failure. Commonly, the degradation of the bolt pretightening state is a multistage process, consisting of the tight contact stage (TCS) and significant loosening stage. Therefore, utilizing a single model to monitor the pretightening state in the full degradation stage is difficult. Here, a method based on nonlinear Lamb waves to identify the TCS of bolts and quantitatively monitor the pretightening state to bolt loosening is proposed. In the proposed method, phase reversal technology is first adopted to enhance the sensitivity and reduce the calculation errors of nonlinear damage indexes for bolt loosening in the TCS, and then the phase reversal relative nonlinear coefficient (PRC) is constructed. This indicator overcomes the disadvantage that linear indicators are insensitive to early loosening and realizes the identification of critical points between the TCS and the significant loosening stage, which provides a prerequisite for constructing a staged loosening monitoring model. After the TCS is determined, a quantitative monitoring model for loosening, which fuses seven nonlinear damage indexes, is established based on canonical correlation forests to evaluate the pretightening state. To verify the effectiveness of the method, an experimental study of bolts is carried out, the lamb signals under different loosening states are measured, and the monitoring effects of different indicators are compared and analyzed. The comparison results show that the proposed method has higher accuracy than conventional approaches.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.