Dynamic Temporal Denoise Neural Network with Multi-Head Attention for Fault Diagnosis Under Noise Background DOI Creative Commons
Zhongzhi Li, Rong Fan, Jinyi Ma

и другие.

Sensors, Год журнала: 2024, Номер 24(21), С. 6813 - 6813

Опубликована: Окт. 23, 2024

Fault diagnosis plays a crucial role in maintaining the operational safety of mechanical systems. As intelligent data-driven approaches evolve, deep learning (DL) has emerged as pivotal technique fault research. However, collected vibrational signals from systems are usually corrupted by unrelated noises due to complicated transfer path modulations and component coupling. To solve above problems, this paper proposed dynamic temporal denoise neural network with multi-head attention (DTDNet). Firstly, model transforms one-dimensional into two-dimensional tensors based on periodic self-similarity signals, employing multi-scale convolution kernels extract signal features both within across periods. Secondly, for problem lacking denoising structure traditional convolutional networks, variable (TVD) module nonlinear processing is filter noises. Lastly, fusion (MAF) used weight denoted different Evaluation two datasets, Case Western Reserve University bearing dataset (single sensor) Real aircraft sensor (multiple sensors), demonstrates that DTDNet can reduce useless achieve remarkable improvement classification performance compared state-of-the-art method. provides high-performance solution potential noise may occur actual tasks, which important application value.

Язык: Английский

Transformer fault diagnosis method based on SMOTE and NGO-GBDT DOI Creative Commons
Lizhong Wang,

Jianfei Chi,

Ye-qiang Ding

и другие.

Scientific Reports, Год журнала: 2024, Номер 14(1)

Опубликована: Март 26, 2024

Abstract In order to improve the accuracy of transformer fault diagnosis and influence unbalanced samples on low model identification caused by insufficient training, this paper proposes a method based SMOTE NGO-GBDT. Firstly, Synthetic Minority Over-sampling Technique (SMOTE) was used expand minority samples. Secondly, non-coding ratio construct multi-dimensional feature parameters, Light Gradient Boosting Machine (LightGBM) optimization strategy introduced screen optimal subset. Finally, Northern Goshawk Optimization (NGO) algorithm optimize parameters Decision Tree (GBDT), then realized. The results show that proposed can reduce misjudgment Compared with other integrated models, has high accuracy, rate stable performance.

Язык: Английский

Процитировано

12
Systematic Review on Fault Diagnosis on Rolling-Element Bearing DOI

M. Pandiyan,

T. Narendiranath Babu

Journal of Vibration Engineering & Technologies, Год журнала: 2024, Номер unknown

Опубликована: Апрель 10, 2024

Язык: Английский

Процитировано

11
Review of research on signal decomposition and fault diagnosis of rolling bearing based on vibration signal DOI
Junning Li, Luo Wen-guang,

Mengsha Bai

и другие.

Measurement Science and Technology, Год журнала: 2024, Номер 35(9), С. 092001 - 092001

Опубликована: Май 22, 2024

Abstract Rolling bearings are critical components that prone to faults in the operation of rotating equipment. Therefore, it is utmost importance accurately diagnose state rolling bearings. This review comprehensively discusses classical algorithms for fault diagnosis based on vibration signal, focusing three key aspects: data preprocessing, feature extraction, and identification. The main principles, features, application difficulties, suitable occasions various thoroughly examined. Additionally, different methods reviewed compared using Case Western Reserve University bearing dataset. Based current research status diagnosis, future development directions also anticipated. It expected this will serve as a valuable reference researchers aiming enhance their understanding improve technology diagnosis.

Язык: Английский

Процитировано

11
Deep learning based approaches for intelligent industrial machinery health management and fault diagnosis in resource-constrained environments DOI Creative Commons

Ali Saeed,

Muazzam A. Khan, Usman Akram

и другие.

Scientific Reports, Год журнала: 2025, Номер 15(1)

Опубликована: Янв. 7, 2025

Industry 4.0 represents the fourth industrial revolution, which is characterized by incorporation of digital technologies, Internet Things (IoT), artificial intelligence, big data, and other advanced technologies into processes. Industrial Machinery Health Management (IMHM) a crucial element, based on (IIoT), focuses monitoring health condition machinery. The academic community has focused various aspects IMHM, such as prognostic maintenance, monitoring, estimation remaining useful life (RUL), intelligent fault diagnosis (IFD), architectures edge computing. Each these categories holds its own significance in context In this survey, we specifically examine research RUL prediction, edge-based architectures, diagnosis, with primary focus domain diagnosis. importance IFD methods ensuring smooth execution processes become increasingly evident. However, most are formulated under assumption complete, balanced, abundant often does not align real-world engineering scenarios. difficulties linked to classifications IMHM have received noteworthy attention from community, leading substantial number published papers topic. While there existing comprehensive reviews that address major challenges limitations field, still gap thoroughly investigating perspectives across complete To fill gap, undertake survey discusses achievements domain, focusing IFD. Initially, classify three distinct perspectives: method processing aims optimize inputs for model mitigate training sample set; constructing model, involves designing structure features enhance resilience challenges; optimizing training, refining process models emphasizes ideal data process. Subsequently, covers techniques related prediction edge-cloud resource-constrained environments. Finally, consolidates outlook relevant issues explores potential solutions, offers practical recommendations further consideration.

Язык: Английский

Процитировано

1
Biologically inspired compound defect detection using a spiking neural network with continuous time–frequency gradients DOI
Zisheng Wang, Shaochen Li, Jianping Xuan

и другие.

Advanced Engineering Informatics, Год журнала: 2025, Номер 65, С. 103132 - 103132

Опубликована: Янв. 24, 2025

Язык: Английский

Процитировано

1
Investigation of ship energy consumption based on neural network DOI
Yaqing Shu, Benshuang yu, Wei Liu

и другие.

Ocean & Coastal Management, Год журнала: 2024, Номер 254, С. 107167 - 107167

Опубликована: Май 10, 2024

Язык: Английский

Процитировано

7
Cross-modal zero-sample diagnosis framework utilizing non-contact sensing data fusion DOI
Sheng Li, Ke Feng, Yadong Xu

и другие.

Information Fusion, Год журнала: 2024, Номер 110, С. 102453 - 102453

Опубликована: Май 1, 2024

Язык: Английский

Процитировано

5
Gearbox faults severity classification using Poincaré plots of acoustic emission signals DOI Creative Commons
Rubén Medina, René–Vinicio Sánchez, Diego Cabrera

и другие.

Applied Acoustics, Год журнала: 2024, Номер 219, С. 109918 - 109918

Опубликована: Фев. 22, 2024

Classification of fault severity in gearboxes using Acoustic Emission (AE) signals is challenging because such represent a highly non-linear and possibly chaotic system. Due to the common assumption linearity, statistical features extracted from these systems are suboptimal for classification severity. Hence, this paper uses Poincaré plot (PP) extract useful classify type gearboxes. For development, four types were applied over different gears then tested on an experimental condition monitoring bench: broken tooth, pitting, scuffing, cracks, each with nine levels. Then, feature set was conventional 2-D PP, composed shape-related known as complex correlation measurements (CCM). The performed frequency bands. Low band-pass filtered obtained highest accuracy Random Forest (RF): classified 99.69%, depending corresponding pitting 98.76%, cracks 98.71%, tooth 98.96%, scuffing 98.51%. PP has low computational cost even large datasets representing AE signals, which can benefit practical possibility implementation high levels gearbox.

Язык: Английский

Процитировано

4
A novel deep ensemble reinforcement learning based control method for strip flatness in cold rolling steel industry DOI
Wen Peng,

Jiawei Lei,

Chengyan Ding

и другие.

Engineering Applications of Artificial Intelligence, Год журнала: 2024, Номер 134, С. 108695 - 108695

Опубликована: Май 31, 2024

Язык: Английский

Процитировано

4
Tri‐Plane Dynamic Neural Radiance Fields for High‐Fidelity Talking Portrait Synthesis DOI Creative Commons
Xueping Wang,

Xueni Guo,

Jun Xu

и другие.

IET Image Processing, Год журнала: 2025, Номер 19(1)

Опубликована: Янв. 1, 2025

ABSTRACT Neural radiation field (NeRF) has been widely used in the of talking portrait synthesis. However, inadequate utilisation audio information and spatial position leads to inability generate images with high audio‐lip consistency realism. This paper proposes a novel tri‐plane dynamic neural (Tri‐NeRF) that employs an implicit study impacts on facial movements. Specifically, Tri‐NeRF propose offset network (TPO‐Net) positions three 2D planes guided by audio. allows for sufficient learning features from image low dimensional state more accurate lip In order better preserve texture details, we innovatively new gated attention fusion module (GAF) dynamically fuse based strong weak correlation cross‐modal features. Extensive experiments have demonstrated can portraits

Язык: Английский

Процитировано

0