Analysis of MHD Flow With Convective Boundary Conditions Over a Permeable Stretching Surface Using a Physics‐Informed Neural Network

Heat Transfer, Год журнала: 2025, Номер unknown

Опубликована: Янв. 3, 2025

ABSTRACT In this study, we examine the impact of heat and mass transfer magnetohydrodynamic (MHD) flow through a stretching permeable surface while considering chemical reaction convective boundary conditions. A physics‐informed neural network (PINN) approach is employed to obtain precise solutions, representing key novelty work. The governing partial differential equations were transformed into nonlinear ordinary by applying similarity transformations. These are integrated PINN's loss function enforce initial conditions, enabling model learn effectively during training. We analyze various parameters related velocity, thermal, concentration distributions present results graphically. findings indicate that injecting fluid leads reduction in velocity gradient as moves away from surface, whereas suction has opposite effect, increasing gradient. parameter significantly reduces layer thickness, an effect further enhanced magnetic parameter. thermal layers primarily affected Schmidt Prandtl numbers. Additionally, slows near sheet, increases temperature at plate's surface. Our proposed method shows significant agreement with previous studies, validating its effectiveness solving complex MHD problems. provide deeper insights dynamics flows have implications for applications involving transfer, such reactors, cooling systems, material processing, environmental management.

Язык: Английский

---

PF-PINNs: Physics-informed neural networks for solving coupled Allen-Cahn and Cahn-Hilliard phase field equations

Journal of Computational Physics, Год журнала: 2025, Номер unknown, С. 113843 - 113843

Опубликована: Фев. 1, 2025

Язык: Английский

---

Modeling water flow in unsaturated soils through physics-informed neural network with principled loss function

Computers and Geotechnics, Год журнала: 2023, Номер 161, С. 105546 - 105546

Опубликована: Июнь 1, 2023

Язык: Английский

---

Physics-informed neural networks for incompressible flows with moving boundaries

Physics of Fluids, Год журнала: 2024, Номер 36(1)

Опубликована: Янв. 1, 2024

Physics-informed neural networks (PINNs) employed in fluid mechanics deal primarily with stationary boundaries. This hinders the capability to address a wide range of flow problems involving moving bodies. To this end, we propose novel extension, which enables PINNs solve incompressible flows time-dependent More specifically, impose Dirichlet constraints velocity at interfaces and define new loss functions for corresponding training points. Moreover, refine points around boundaries accuracy. effectively enforces no-slip condition With an initial condition, extended unsteady still have flexibility leverage partial data reconstruct entire field. Therefore, version inherits amalgamation both physics from original PINNs. series typical problems, demonstrate effectiveness accuracy The proposed concept allows solving inverse as well, calls further investigations.

Язык: Английский

---

Mixtures Recomposition by Neural Nets: A Multidisciplinary Overview

Journal of Chemical Information and Modeling, Год журнала: 2024, Номер 64(3), С. 597 - 620

Опубликована: Янв. 29, 2024

Artificial Neural Networks (ANNs) are transforming how we understand chemical mixtures, providing an expressive view of the space and multiscale processes. Their hybridization with physical knowledge can bridge gap between predictivity understanding underlying This overview explores recent progress in ANNs, particularly their potential 'recomposition' mixtures. Graph-based representations reveal patterns among mixture components, deep learning models excel capturing complexity symmetries when compared to traditional Quantitative Structure–Property Relationship models. Key such as Hamiltonian networks convolution operations, play a central role representing The integration ANNs Chemical Reaction Physics-Informed for inverse kinetic problems is also examined. combination sensors shows promise optical biomimetic applications. A common ground identified context statistical physics, where ANN-based methods iteratively adapt by blending initial states training data. concept recomposition unveils reciprocal inspiration reactive highlighting behaviors influenced environment.

Язык: Английский

---

Fluidized Bed Scale-Up for Sustainability Challenges. 1. Tomorrow’s Tools

Industrial & Engineering Chemistry Research, Год журнала: 2024, Номер 63(6), С. 2519 - 2533

Опубликована: Фев. 5, 2024

The scaling up of fluidized beds has been purposefully pursued for more than 100 years. Yet, over that time, scale-up tools have not significantly changed. Data analysis is typically a standard variances statistical exercise, perhaps reinforced with design experimental procedure. Flowsheeting and equipment are based on institutional knowledge, albeit graphical user interface-based process flow models make job manageable. Advanced such as computational fluid dynamics used but often supplement primary driver. As result, the bed can take 10 Fluidized remain at forefront present time-critical sustainability challenges, e.g., carbon capture by particulate sorbents, methane-to-hydrogen, plastic-to-chemicals, etc. In view exigency toward net zero, today's efforts need to be accelerated, leveraging advanced new become readily available. problem neglected, inadequately implemented, ineffectively resourced, and/or poorly understood. This motivated current effort, which targeted reviewing how evolved years promising tools, addressing some barriers these in beds, well contemplating what done circumvent barriers. follow up, companion part 2 (Cocco, R. A.; Chew, J. W. Ind. Eng. Chem. Res., submitted publication) proposes path achieve timely implementation green processes.

Язык: Английский

---

Grid adaptive reduced-order model of fluid flow based on graph convolutional neural network

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

Опубликована: Авг. 1, 2022

In the interdisciplinary field of data-driven models and computational fluid mechanics, reduced-order model for flow prediction is mainly constructed by a convolutional neural network (CNN) in recent years. However, standard CNN only applicable to data with Euclidean spatial structure, while non-Euclidean properties can be convolved after pixelization, which usually leads decreased accuracy. this work, novel framework based on graph convolution (GCN) proposed allow operator predict dynamics non-uniform structured or unstructured mesh data. This achieved fact that inherit characteristics message passing mechanism GCN. The conversion method from form operation GCN are clarified. Moreover, additional relevance features weight loss function dataset also investigated improve performance. learns an end-to-end mapping between physical field. Through our studies various cases internal flow, it shown GCN-based offers excellent adaptability non-uniformly distributed data, achieving high accuracy three-order speedup compared numerical simulation. Our generalizes opens door further extending most existing architectures future.

Язык: Английский

---

Physics-informed neural network based on a new adaptive gradient descent algorithm for solving partial differential equations of flow problems

Physics of Fluids, Год журнала: 2023, Номер 35(6)

Опубликована: Июнь 1, 2023

Physics-informed neural network (PINN) is an emerging technique for solving partial differential equations (PDEs) of flow problems. Due to the advantage low computational cost, gradient descent algorithms coupled with weighted objectives method are usually used optimize loss functions in PINN training. However, interaction mechanisms between gradients not fully clarified, leading poor performances optimization. For this, adaptive algorithm (AGDA) proposed based on analyses and then validated by analytical PDEs First, training traditional Adam optimizer analyzed. The main factors responsible identified. Then, a new AGDA developed two modifications: (1) balancing magnitude difference (2) eliminating directions conflict. Finally, three types (elliptic, hyperbolic, parabolic) four viscous incompressible problems selected validate algorithm. It found that reach specified accuracy, required time about 16%–90% 41%–64% PCGrad optimizer, demanded number iterations 10%–68% 38%–77% optimizer. Therefore, more efficient robust

Язык: Английский

---

Physics-informed neural network frameworks for crack simulation based on minimized peridynamic potential energy

Computer Methods in Applied Mechanics and Engineering, Год журнала: 2023, Номер 417, С. 116430 - 116430

Опубликована: Сен. 22, 2023

Язык: Английский

---

Exploring hidden flow structures from sparse data through deep-learning-strengthened proper orthogonal decomposition

Physics of Fluids, Год журнала: 2023, Номер 35(3)

Опубликована: Фев. 25, 2023

Proper orthogonal decomposition (POD) enables complex flow fields to be decomposed into linear modes according their energy, allowing the key features of extracted. However, traditional POD requires high-quality inputs, namely, high-resolution spatiotemporal data. To alleviate dependence on quality and quantity data, this paper presents a method that is strengthened by physics-informed neural network (PINN) with an overlapping domain strategy. The loss function convergence are considered simultaneously determine PINN-POD model. proposed framework applied past two-dimensional circular cylinder at Reynolds numbers ranging from 100 10 000 achieves accurate robust extraction structures spatially sparse observation spatial dominant frequency can also extracted under high-level noise. These results demonstrate reliable tool for extracting data fields, potentially shedding light data-driven discovery hidden fluid dynamics.

Язык: Английский

---