Intelligent fault diagnosis of multi-source cross-machine bearings based on center-weighted optimal transport and class-level alignment domain adaptation DOI
Zhiwu Shang, Changchao Wu, Fei Liu

и другие.

Measurement Science and Technology, Год журнала: 2024, Номер 35(11), С. 116206 - 116206

Опубликована: Авг. 7, 2024

Abstract Most of the current domain adaptation research primarily focuses on single-source or multi-source transfer constructed under different working conditions same machine. However, when faced with cross-machine tasks significant discrepancies, forcing direct feature alignment between source and target samples may lead to negative transfer, thereby reducing model’s diagnostic performance. To overcome above limitations, this paper proposes a deep model based center-weighted optimal transport (CWOT) class-level adaptation. Firstly, enhance representation capability features, multi-structure network is enrich information capacity embedded within achieving better capabilities. Then, local maximum mean discrepancy introduced fully exploit fine-grained discriminative features among domains, minimizing distribution differences domains greatest extent, thus capturing reliable highly generalized invariant features. On basis, CWOT strategy designed, which comprehensively considers cost intra-domain uncertainty inter-domain correlation samples, establishing more effective alleviating problem sample improving Finally, instance studies are conducted through multiple tasks, demonstrating that proposed method outperforms existing methods in terms accuracy fault capability. This provides diagnosis for detecting health status rotating machinery equipment, promoting application technology practical industry.

Язык: Английский

Impact time domain decomposition: An adaptive decomposition method for multi-source impact signals based on envelope energy gradient characteristics DOI
Yuyang Chen, Jinjie Zhang, Nanyang Zhao

и другие.

Mechanical Systems and Signal Processing, Год журнала: 2024, Номер 219, С. 111637 - 111637

Опубликована: Июнь 15, 2024

Язык: Английский

Процитировано

3
A Hybrid Dynamic Principal Component Analysis Feature Extraction Method to Identify Piston Pin Wear for Binary Classifier Modeling DOI Creative Commons
Hao Yang, Yidi Zhai, M.M. Zheng

и другие.

Machines, Год журнала: 2025, Номер 13(1), С. 68 - 68

Опубликована: Янв. 18, 2025

The wear condition of a piston pin is main factor in determining the operational continuity and life cycle diesel engine; identifying its vibration feature paramount importance carrying out necessary maintenance early stage. As dynamic features are susceptible to environmental disturbance during operation, an effective signal processing method improve accuracy fineness extracted features, which essential build reliable precise binary classifier model identify based on features. Aiming at extraction requirements anti-noise, effectiveness, this paper proposes algorithm principal component analysis (DPCA) combined with variational mode decomposition (VMD) singular value (SVD). An orthogonal sensor layout applied collect under normal worn conditions, proved reducing disturbance. DPCA utilized extract dynamical by introducing time lag. Then, matrix further decomposed VMD obtain intrinsic functions (IMFs) as finer finally SVD compress thus improving classification efficiency To validate significance proposed method, support vector machine (SVM) employed classifiers evaluate performance trained different A modeling dataset containing 80 samples (40 40 samples) employed, five-round cross-validation adopted. For each round, two models empirical (EMD)–auto regressive (AR) spectrum fast Fourier transform (FFT) continuous wavelet (CWT), respectively; precision, recall ratio, F1 ratio obtained testing set contrasting overall performances cross-validation, be more noise reduction significant extraction, able for identification.

Язык: Английский

Процитировано

0
Enhancing the safety of hydroelectric power generation systems: an intelligent identification of axis orbits based on a nonlinear dynamics method DOI
Fei Chen, Jie Liu, Xiaoxi Hu

и другие.

Energy, Год журнала: 2025, Номер unknown, С. 135864 - 135864

Опубликована: Апрель 1, 2025

Язык: Английский

Процитировано

0
Method for determining the function of transfer of vibrational energy to the engine structure from the ignition distribution mechanism DOI
E. Kalinin,

Daria Lemishko

AIP conference proceedings, Год журнала: 2025, Номер 3306, С. 020001 - 020001

Опубликована: Янв. 1, 2025

Язык: Английский

Процитировано

0
Exploring evolutionary-tuned autoencoder-based architectures for fault diagnosis in a wind turbine gearbox DOI Creative Commons
Samuel M. Gbashi, Obafemi O. Olatunji, Paul A. Adedeji

и другие.

Smart Science, Год журнала: 2024, Номер unknown, С. 1 - 21

Опубликована: Июнь 11, 2024

Vibration-based fault diagnosis from rotary machinery requires prior feature extraction, selection, or dimensionality reduction. Feature extraction is tedious, and computationally expensive. selection presents unique challenges intrinsic to the method adopted. Nonlinear reduction may be achieved through kernel transformations, however there often a trade-off in information achieve this. Given above, this study proposes novel autoencoder (AE) pre-processing framework for vibration-based wind turbine (WT) gearboxes. In study, AEs are used learn features of WT gearbox vibration data while simultaneously compressing data, obviating need costly engineering The effectiveness proposed was evaluated by training genetically optimized linear discriminant analysis (LDA), multilayer perceptron (MLP), random forest (RF) models, with AE's latent space features. models were using known classification metrics. results showed that performance depends on size space. As increased, quality extracted improved until plateau observed at dimension 10. AE pre-processed RF, MLP, LDA designated AE-Pre-GO-RF, AE-Pre-GO-MLP, AE-Pre-GO-LDA, accuracy, sensitivity, specificity seven (7) conditions. AE-Pre-GO-RF model outperformed its counterparts, scoring 100% all metrics, though longest time (239.50 sec). Comparable found comparing similar investigations involving traditional processing techniques. More so, it established effective can manifold learning without expensive engineering.

Язык: Английский

Процитировано

2
Adaptive time-domain impact extraction method for multi-source impact vibration signal of diesel engine DOI
Nanyang Zhao, Chao Liu, Dongxiang Jiang

и другие.

Structural Health Monitoring, Год журнала: 2024, Номер unknown

Опубликована: Июль 25, 2024

Diesel engines are widely used in fields such as ships, vehicles, and nuclear power. The vibration signals of a diesel engine’s casing exhibit characteristic intermittent distribution multi-source impacts. In response to the challenges faced by existing feature extraction methods identifying localizing impact signals, this paper proposes adaptive time-domain (ATDIE) method, which is based on characteristics exhibiting local high-energy time domain rapid amplitude decay. purpose ATDIE method extract various components from signals. constructs solution model with goal minimizing signal’s multi-order central moments. By using residual iterative number extracted adaptively determined. Then, an window optimization function established enhance adaptability. Finally, test results both simulation engine demonstrate that possesses good capabilities for computational efficiency.

Язык: Английский

Процитировано

2
DSTF-Net: A Novel Framework for Intelligent Diagnosis of Insulated Bearings in Wind Turbines with Multi-Source Data and Its Interpretability DOI
Tongguang Yang, Ming Xu,

Chun‐Lung Chen

и другие.

Renewable Energy, Год журнала: 2024, Номер unknown, С. 121965 - 121965

Опубликована: Ноя. 1, 2024

Язык: Английский

Процитировано

2
Intelligent fault diagnosis of multi-source cross-machine bearings based on center-weighted optimal transport and class-level alignment domain adaptation DOI
Zhiwu Shang, Changchao Wu, Fei Liu

и другие.

Measurement Science and Technology, Год журнала: 2024, Номер 35(11), С. 116206 - 116206

Опубликована: Авг. 7, 2024

Abstract Most of the current domain adaptation research primarily focuses on single-source or multi-source transfer constructed under different working conditions same machine. However, when faced with cross-machine tasks significant discrepancies, forcing direct feature alignment between source and target samples may lead to negative transfer, thereby reducing model’s diagnostic performance. To overcome above limitations, this paper proposes a deep model based center-weighted optimal transport (CWOT) class-level adaptation. Firstly, enhance representation capability features, multi-structure network is enrich information capacity embedded within achieving better capabilities. Then, local maximum mean discrepancy introduced fully exploit fine-grained discriminative features among domains, minimizing distribution differences domains greatest extent, thus capturing reliable highly generalized invariant features. On basis, CWOT strategy designed, which comprehensively considers cost intra-domain uncertainty inter-domain correlation samples, establishing more effective alleviating problem sample improving Finally, instance studies are conducted through multiple tasks, demonstrating that proposed method outperforms existing methods in terms accuracy fault capability. This provides diagnosis for detecting health status rotating machinery equipment, promoting application technology practical industry.

Язык: Английский

Процитировано

0