A hybrid deep learning model towards fault diagnosis of drilling pump

Applied Energy, Год журнала: 2024, Номер 372, С. 123773 - 123773

Опубликована: Июнь 26, 2024

This paper proposes a novel method namely WaveletKernelNet-Convolutional Block Attention Module-BiLSTM for intelligent fault diagnosis of drilling pumps. Initially, the random forest is applied to determine target signals that can reflect characteristics Accordingly, Module Net constructed noise reduction and feature extraction based on signals. The Convolutional embedded in WaveletKernelNet-CBAM adjusts weight enhances representation channel spatial dimension. Finally, Bidirectional Long-Short Term Memory concept introduced enhance ability model process time series data. Upon constructing network, Bayesian optimization algorithm utilized ascertain fine-tune ideal hyperparameters, thereby ensuring network reaches its optimal performance level. With hybrid deep learning presented, an accurate real five-cylinder pump carried out results confirmed applicability reliability. Two sets comparative experiments validated superiority proposed method. Additionally, generalizability verified through domain adaptation experiments. contributes safe production oil gas sector by providing robust industrial equipment.

Язык: Английский

---

An Integrated Deep Learning Model for Intelligent Recognition of Long-distance Natural Gas Pipeline Features

Reliability Engineering & System Safety, Год журнала: 2024, Номер unknown, С. 110664 - 110664

Опубликована: Ноя. 1, 2024

Язык: Английский

---

A hybrid fault diagnosis scheme for milling tools using MWN-CBAM-PatchTST network with acoustic emission signals

Nondestructive Testing And Evaluation, Год журнала: 2025, Номер unknown, С. 1 - 29

Опубликована: Янв. 8, 2025

Milling tools are critical to machining and manufacturing processes. Accurate diagnosis identification of faults occurring in milling during their operation utmost importance for maintaining the reliability availability these tools, minimise machine downtime overall costs. This paper presents a fault network model based on acoustic emission signals. The integrates multilayer wavelet CNN (MWN) consisting discrete transform (DWT) convolutional neural (CNN), block attention module (CBAM), PatchTST module. MWN uses transformation withdraw multi-scale features from signals, thus improving sensitivity small variations emission. CBAM improves feature representation by focusing channels regions, while self-attention mechanism optimise processing long-range dependencies. synergy mechanisms results superior performance, outperforming traditional diagnostic methods. Bayesian optimisation is used select hyperparameters, eliminating subjective bias associated with manual range setting. Validation experiments using dataset, including ablation studies comparative tests, demonstrated that achieves an accuracy over 98%, validating its generalisation capability effectiveness diagnosing tool

Язык: Английский

---

An adaptive mixture prior in Bayesian convolutional autoencoder for early detecting anomalous degradation behaviors in lithium-ion batteries

Reliability Engineering & System Safety, Год журнала: 2025, Номер unknown, С. 110926 - 110926

Опубликована: Фев. 1, 2025

Язык: Английский

---

DCSIAN: A novel deep cross-scale interactive attention network for fault diagnosis of aviation hydraulic pumps and generalizable applications

Reliability Engineering & System Safety, Год журнала: 2024, Номер 249, С. 110246 - 110246

Опубликована: Май 27, 2024

Язык: Английский

---

A dual‐channel transferable RUL prediction method integrated with Bayesian deep learning and domain adaptation for rolling bearings

Quality and Reliability Engineering International, Год журнала: 2024, Номер 40(5), С. 2348 - 2366

Опубликована: Март 23, 2024

Abstract Many deep learning methods typically assume that the marginal probability distribution between training and testing bearing data is similar or same. However, of rolling bearings may deviate significantly under diverse working conditions. To address above limitations, a novel transferable remaining useful life (RUL) prediction method integrated with Bayesian unsupervised domain adaptation (DA) proposed. First, signal alignment executed on after first time to maintain same granularity scale across both source target domains. Second, multi‐domain features are extracted sent into dual‐channel Transformer network (DCTN) incorporating convolutional block attention module (CBAM) adequately exploit abundant degradation information. Then, DA incorporated model mitigate discrepancies high‐level merged Finally, by applying variational inference method, DCTN‐CBAM extended neural network, RUL its corresponding confidence intervals can be conveniently derived. In addition, generalization capability effectiveness validated through six bidirectional transfer tasks two datasets. The experimental results demonstrate it could provide more reliable efficiently account for uncertainty.

Язык: Английский

---

CCZM‐based fatigue analysis and reliability assessment for wind turbine blade adhesive joints considering parameter uncertainties

Quality and Reliability Engineering International, Год журнала: 2024, Номер 40(6), С. 3037 - 3054

Опубликована: Май 19, 2024

Abstract Wind turbine blades are complex structures composed of multiple bonded components. The fatigue performance these adhesive joints is crucial for ensuring operational safety over the blade's lifespan. Traditional structural analysis methods inadequate evaluating properties due to unique characteristics materials. Variations in material and dimensional parameters, as well fluctuating loads, further complicate wind blades. To tackle this issue, study introduces a reliability assessment method blades, employing Cyclic Cohesive Zone Model (CCZM) accounting parameter uncertainties. Specifically, novel methodology based on CCZM presented. programmatically implemented obtain life dataset through simulations, considering uncertainties dimensions, loads. Subsequently, model formulated evaluate under different conditions, sensitivity each investigated. findings offer valuable insights improving

Язык: Английский

---

Numerical analysis of the effect of hydrogen doping ratio on gas transmission in low-pressure pipeline network

International Journal of Hydrogen Energy, Год журнала: 2024, Номер 73, С. 868 - 884

Опубликована: Июнь 14, 2024

Язык: Английский

---

Semi-analytical investigation on hydrodynamic efficiency and loading of perforated breakwater-integrated OWCs

Ocean Engineering, Год журнала: 2024, Номер 309, С. 118460 - 118460

Опубликована: Июнь 18, 2024

Язык: Английский

---

A variable‐speed‐condition fault diagnosis method for crankshaft bearing in the RV reducer with WSO‐VMD and ResNet‐SWIN

Quality and Reliability Engineering International, Год журнала: 2024, Номер 40(5), С. 2321 - 2347

Опубликована: Март 20, 2024

Abstract Due to the noise interference and weak characterization ability of fault vibration signal rotation vector (RV) reducer crankshaft bearing, it is difficult obtain satisfactory results for available diagnosis methods. For that, this paper proposes a variable‐speed‐condition method with WSO‐VMD ResNet‐SWIN. A reconstruction was carried out, Firstly, performance VMD algorithm improved by using war strategy optimization select parameters adaptively. Then reconstructed considering characteristic frequency, so as realize reduction signal. By residual network module attention mechanism replace first stage original SWIN model, novel ResNet‐SWIN model established enhance feature extraction The experiments constant‐operating‐condition variable‐operating‐condition are out verify effectiveness proposed method. show whether at variable‐speed or constant‐speed conditions, WSO has been proven be fastest convergence speed compared WOA, SSA, NGO algorithms, WSO‐VMD, variance evaluation indicator 36%, 21%, 46%, 40%, respectively. achieved optimal accuracy SWIN, VIT, CNN‐SVM models in both conditions.

Язык: Английский

---