Augmentation of piezoelectric thin-film flapping velocimetry turbulence strength detection via machine learning DOI Creative Commons

Ted Sian Lee,

Ean Hin Ooi, Wei Sea Chang

и другие.

Physics of Fluids, Год журнала: 2025, Номер 37(1)

Опубликована: Янв. 1, 2025

Qualitatively evaluating the fundamental mechanical characteristics of square-fractal-grid (SFG)-generated turbulent flow using piezoelectric thin-film flapping velocimetry (PTFV) is rather time-consuming. More importantly, its sensitivity in detecting high-frequency, fine-scale fluctuations constrained by high-speed camera specifications. To reduce dependency on imaging future PTFV implementations, regression models are trained with supervised machine learning to determine correlation between piezoelectric-generated voltage V and corresponding local equivalent velocity fluctuation. Using tip deflection δ data as predictors responses, respectively, Trilayered Neural Network (TNN) emerges best-performing model compared linear regression, trees, support vector machines, Gaussian process ensembles trees. TNN from (i) lower quarter, (ii) bottom left corner, (iii) central opening SFG-grid provide accurate predictions insert-induced centerline streamwise cross-sectional lateral turbulence intensity root mean square-δ, average errors not exceeding 5%. The output predicted response, which considers small-scale across entire surface, better expresses integral length scale (38% smaller) forcing (270% greater), particularly at corner SFG where eddies significant. Furthermore, effectively captures occasional extensive excitation forces large-scale eddies, resulting a more balanced force distribution. In short, this study paves path for comprehensive expedited dynamics characterization detection via PTFV, potential deployment high Reynolds number flows generated various grid configurations.

Язык: Английский

Three-dimensional wake transition of rectangular cylinders and temporal prediction of flow patterns based on a machine learning algorithm DOI
A. Mashhadi, A. Sohankar, M. M. Moradmand

и другие.

Physics of Fluids, Год журнала: 2024, Номер 36(9)

Опубликована: Сен. 1, 2024

This study investigates the three-dimensional (3D) wake transition in unconfined flows over rectangular cylinders using direct numerical simulation (DNS). Two different cross-sectional aspect ratios (AR) and Reynolds numbers (Re) are scrutinized: AR = 0.5 at Re 200 3 600. The investigation focuses on characterizing flow patterns forecasting their temporal evolution utilizing proper orthogonal decomposition (POD) technique coupled with a long short-term memory (LSTM) network. DNS results reveal emergence of an ordered mode A for 3, attributed to stabilizing effect elongated AR. On other hand, case smaller (= 0.5) exhibits mode-swapping regime characterized by modes B's distinct simultaneous manifestation. spanwise wavelengths B approximately 4.7 1.2 D 0.5, while wavelength is 3.5 3. POD serves as dimensionality reduction technique, LSTM facilitates prediction. algorithm demonstrates satisfactory performance predicting patterns, including instabilities B, across both transverse directions. employed adeptly predicts pressure time series surrounding cylinders. duration training only about 0.5% required computations. research, first time, effectiveness POD–LSTM complex 3D instantaneous past

Язык: Английский

Процитировано

7
XLB: A differentiable massively parallel lattice Boltzmann library in Python DOI
Mohammadmehdi Ataei, Hesam Salehipour

Computer Physics Communications, Год журнала: 2024, Номер 300, С. 109187 - 109187

Опубликована: Март 27, 2024

Язык: Английский

Процитировано

6
Artificial intelligence in fluid mechanics DOI
Weiwei Zhang, Bernd R. Noack

Acta Mechanica Sinica, Год журнала: 2021, Номер 37(12), С. 1715 - 1717

Опубликована: Дек. 1, 2021

Язык: Английский

Процитировано

33
Deep neural network based reduced-order model for fluid–structure interaction system DOI

Renkun Han,

Yixing Wang, Weiqi Qian

и другие.

Physics of Fluids, Год журнала: 2022, Номер 34(7)

Опубликована: Июль 1, 2022

Fluid–structure interaction analysis has high computing costs when using computational fluid dynamics. These become prohibitive optimizing the fluid–structure system because of huge sample space structural parameters. To overcome this realistic challenge, a deep neural network-based reduced-order model for is developed to quickly and accurately predict flow field in system. This network can at next time step based on current motion conditions. A be constructed by combining with dynamic solver. Through learning structure evolution different systems, trained systems parameters only initial Within learned range parameters, prediction accuracy good agreement numerical simulation results, which meet engineering needs. The speed increased more than 20 times, helpful rapid optimal design systems.

Язык: Английский

Процитировано

25
Data-driven quantification of model-form uncertainty in Reynolds-averaged simulations of wind farms DOI
Ali Eidi, Navid Zehtabiyan-Rezaie, Reza Ghiassi

и другие.

Physics of Fluids, Год журнала: 2022, Номер 34(8)

Опубликована: Авг. 1, 2022

Computational fluid dynamics using the Reynolds-averaged Navier-Stokes (RANS) remains most cost-effective approach to study wake flows and power losses in wind farms. The underlying assumptions associated with turbulence closures are one of biggest sources errors uncertainties model predictions. This work aims quantify model-form RANS simulations farms at high Reynolds numbers under neutrally stratified conditions by perturbing stress tensor through a data-driven machine-learning technique. To this end, two-step feature-selection method is applied determine key features model. Then, extreme gradient boosting algorithm validated employed predict perturbation amount direction modeled toward limiting states on barycentric map. procedure leads more accurate representation anisotropy. trained high-fidelity data obtained from large-eddy simulation specific farm, it tested two other (unseen) distinct layouts analyze its performance cases different turbine spacing partial wake. results indicate that, unlike data-free which uniform constant entire computational domain, proposed framework yields an optimal estimation uncertainty bounds for RANS-predicted quantities interest, including velocity, intensity,

Язык: Английский

Процитировано

25
Explainability analysis of neural network-based turbulence modeling for transonic axial compressor rotor flows DOI

Chutian Wu,

Shizhao Wang, Xinlei Zhang

и другие.

Aerospace Science and Technology, Год журнала: 2023, Номер 141, С. 108542 - 108542

Опубликована: Авг. 2, 2023

Язык: Английский

Процитировано

13
Policy Implementation Roadmap, Diverse Perspectives, Challenges, Solutions Towards Low-Carbon Hydrogen Economy DOI Creative Commons
Harshit Mittal, Omkar Singh Kushwaha

Green and Low-Carbon Economy, Год журнала: 2024, Номер unknown

Опубликована: Март 11, 2024

Hydrogen is a nearly emission-free energy carrier with many enticing qualities, including wide availability, environmental friendliness, and high calorific value. There have constantly been lot of challenges to establish an entire fledge low carbon hydrogen economy in the past century. This study aims critically analyse economic, environmental, technological, policy implementation division low-carbon find novel solutions, bridging gaps giving perspective approach study. Differentiation various (LCH) components, green blue hydrogen, was also proposed based on life cycle assessment emissions (LCAE). Current perspectives Promised Pledged Perspectives are considered project demand 2030. A thorough economic analysis system technologies conducted from both production storage by comparing systems. Policies towards LCH were viewed policymakers, consumers, R & D perspectives, through which several challenges, gaps, keynote necessities stated.

Язык: Английский

Процитировано

5
Physics-Informed Machine Learning of Argon Gas-Driven Melt Pool Dynamics DOI
Rahul Sharma, Y. B. Guo, Maziar Raissi

и другие.

Journal of Manufacturing Science and Engineering, Год журнала: 2024, Номер 146(8)

Опубликована: Май 3, 2024

Abstract Melt pool dynamics in metal additive manufacturing (AM) is critical to process stability, microstructure formation, and final properties of the printed materials. Physics-based simulation, including computational fluid (CFD), dominant approach predict melt dynamics. However, physics-based simulation approaches suffer from inherent issue very high cost. This paper provides a physics-informed machine learning method by integrating conventional neural networks with governing physical laws dynamics, such as temperature, velocity, pressure, without using any training data on velocity pressure. avoids solving nonlinear Navier–Stokes equation numerically, which significantly reduces cost (if generation). The difficult-to-determine parameters' values equations can also be inferred through data-driven discovery. In addition, network (PINN) architecture has been optimized for efficient model training. data-efficient PINN attributed extra penalty incorporating PDEs, initial conditions, boundary conditions model.

Язык: Английский

Процитировано

5
On the prediction of the turbulent flow behind cylinder arrays via Echo State Networks DOI Creative Commons
Mohammad Sharifi Ghazijahani, Christian Cierpka

Machine Learning Science and Technology, Год журнала: 2024, Номер 5(3), С. 035005 - 035005

Опубликована: Июнь 4, 2024

Abstract This study aims at the prediction of turbulent flow behind cylinder arrays by application Echo State Networks (ESN). Three different arrangements seven cylinders are chosen for current study. These represent regimes: single bluff body flow, transient and co-shedding flow. allows investigation flows that fundamentally originate from wake yet exhibit highly diverse dynamics. The data is reduced Proper Orthogonal Decomposition (POD) which optimal in terms kinetic energy. Time Coefficients POD Modes (TCPM) predicted ESN. network architecture optimized with respect to its three main hyperparameters, Input Scaling (INS), Spectral Radius (SR), Leaking Rate (LR), order produce best predictions Weighted Prediction Score (WPS), a metric leveling statistic deterministic prediction. In general, ESN capable imitating complex dynamics even longer periods several vortex shedding cycles. Furthermore, mutual interdependencies TCPM well preserved. However, hyperparameters depend strongly on characteristics. Generally, as become faster more intermittent, larger LR INS values result better predictions, whereas less clear trends SR observable.

Язык: Английский

Процитировано

5
ViscoelasticNet: A physics informed neural network framework for stress discovery and model selection DOI
Sukirt Thakur, Maziar Raissi, Arezoo M. Ardekani

и другие.

Journal of Non-Newtonian Fluid Mechanics, Год журнала: 2024, Номер 330, С. 105265 - 105265

Опубликована: Июнь 5, 2024

Язык: Английский

Процитировано

5