Computers & Fluids最新文献

英文中文

Dual scale Residual-Network for turbulent flow sub grid scale resolving: A prior analysis

IF 2.5 3区工程技术 Q3 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers & Fluids

Pub Date : 2025-03-18 DOI: 10.1016/j.compfluid.2025.106592

Omar Sallam, Mirjam Fürth

This paper introduces generative Residual Networks (ResNet) as a surrogate Machine Learning (ML) tool for Large Eddy Simulation (LES) Sub Grid Scale (SGS) resolving. The study investigates the impact of incorporating Dual Scale Residual Blocks (DS-RB) within the ResNet architecture. Two LES SGS resolving models are proposed and tested for prior analysis test cases: a super-resolution model (SR-ResNet) and a SGS stress tensor inference model (SGS-ResNet). The SR-ResNet model task is to upscale LES solutions from coarse to finer grids by inferring unresolved SGS velocity fluctuations, exhibiting success in preserving high-frequency velocity fluctuation information, and aligning with higher-resolution LES solutions’ energy spectrum. Furthermore, employing DS-RB enhances prediction accuracy and precision of high-frequency velocity fields compared to Single Scale Residual Blocks (SS-RB), evident in both spatial and spectral domains. The SR-ResNet model is tested and trained on filtered/downsampled 2-D LES planar jet injection problems at two Reynolds numbers, two jet configurations, and two upscale ratios. In the case of SGS stress tensor inference, both SS-RB and DS-RB exhibit higher prediction accuracy compared to other explicit closure models such as the Smagorinsky model or the Approximate Deconvolution Model (ADM) with reference to the true DNS SGS stress tensor, with DS-RB-based SGS-ResNet showing stronger statistical alignment with DNS data. The SGS-ResNet model is tested on a filtered/downsampled 2-D DNS isotropic homogeneous decay turbulence problem. The adoption of DS-RB incurs notable increases in network size, training time, and forward inference time, with the network size expanding by over tenfold, and training and forward inference times increasing by approximately 0.5 and 3 times, respectively.

{"title":"Dual scale Residual-Network for turbulent flow sub grid scale resolving: A prior analysis","authors":"Omar Sallam, Mirjam Fürth","doi":"10.1016/j.compfluid.2025.106592","DOIUrl":"10.1016/j.compfluid.2025.106592","url":null,"abstract":"<div><div>This paper introduces generative Residual Networks (ResNet) as a surrogate Machine Learning (ML) tool for Large Eddy Simulation (LES) Sub Grid Scale (SGS) resolving. The study investigates the impact of incorporating Dual Scale Residual Blocks (DS-RB) within the ResNet architecture. Two LES SGS resolving models are proposed and tested for prior analysis test cases: a super-resolution model (SR-ResNet) and a SGS stress tensor inference model (SGS-ResNet). The SR-ResNet model task is to upscale LES solutions from coarse to finer grids by inferring unresolved SGS velocity fluctuations, exhibiting success in preserving high-frequency velocity fluctuation information, and aligning with higher-resolution LES solutions’ energy spectrum. Furthermore, employing DS-RB enhances prediction accuracy and precision of high-frequency velocity fields compared to Single Scale Residual Blocks (SS-RB), evident in both spatial and spectral domains. The SR-ResNet model is tested and trained on filtered/downsampled 2-D LES planar jet injection problems at two Reynolds numbers, two jet configurations, and two upscale ratios. In the case of SGS stress tensor inference, both SS-RB and DS-RB exhibit higher prediction accuracy compared to other explicit closure models such as the Smagorinsky model or the Approximate Deconvolution Model (ADM) with reference to the true DNS SGS stress tensor, with DS-RB-based SGS-ResNet showing stronger statistical alignment with DNS data. The SGS-ResNet model is tested on a filtered/downsampled 2-D DNS isotropic homogeneous decay turbulence problem. The adoption of DS-RB incurs notable increases in network size, training time, and forward inference time, with the network size expanding by over tenfold, and training and forward inference times increasing by approximately 0.5 and 3 times, respectively.</div></div>","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"292 ","pages":"Article 106592"},"PeriodicalIF":2.5,"publicationDate":"2025-03-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143643968","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A lattice Boltzmann flux solver for numerical simulation of flows with the viscoelastic fluid

IF 2.5 3区工程技术 Q3 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers & Fluids

Pub Date : 2025-03-13 DOI: 10.1016/j.compfluid.2025.106593

Hua Zhang , Chang Shu , Lian-Ping Wang , Yaguang Liu , Lailai Zhu

In this paper, a viscoelastic lattice Boltzmann flux solver (VLBFS) is developed to simulate incompressible flows of viscoelastic fluids with linear and non-linear constitutive models. In this method, the macroscopic equations are solved by the finite volume method, where the fluxes at the cell interface are evaluated by local reconstruction of the solutions of lattice Boltzmann equations (LBE). Two sets of distribution functions are introduced to reconstruct the cell-interface fluxes, one used for mass and momentum fluxes and the other for the conformation tensor flux in the polymer constitutive equation. The elastic-viscous stress splitting (EVSS) and the solvent-polymer stress splitting (SPSS) techniques are incorporated into the present LBFS to improve the numerical stability. The standard lattice Boltzmann method (LBM) for solving the polymer constitutive equation contains redundant diffusion terms, but this problem is resolved in the current LBFS by setting the relaxation time corresponding to the true diffusion-free limit thus the correct polymer constitutive equation can be recovered. Furthermore, VLBFS eliminates other disadvantages of the standard LBM, such as the LBM on-grid advection coupling the time interval with grid spacing, complicated treatment of the mesoscopic boundary conditions, dependence on uniform grids, and the larger memory requirement due to solving the phase-space discrete distributions. Several flows of a viscoelastic fluid, namely, the two-dimensional plane Poiseuille flow, two-dimensional simplified four-roll mill flows, and three-dimensional Taylor–Green vortex flows, are considered to investigate the accuracy and stability of the present method. The results are found to be in good agreement with the analytical solutions and the previous numerical results. Numerical error analyses show that the present method owns a second-order accuracy in space. The developed VLBFS extends the application domain of LBFS and serves as a basis for simulating viscoelastic flows at high Weissenberg numbers.

{"title":"A lattice Boltzmann flux solver for numerical simulation of flows with the viscoelastic fluid","authors":"Hua Zhang , Chang Shu , Lian-Ping Wang , Yaguang Liu , Lailai Zhu","doi":"10.1016/j.compfluid.2025.106593","DOIUrl":"10.1016/j.compfluid.2025.106593","url":null,"abstract":"<div><div>In this paper, a viscoelastic lattice Boltzmann flux solver (VLBFS) is developed to simulate incompressible flows of viscoelastic fluids with linear and non-linear constitutive models. In this method, the macroscopic equations are solved by the finite volume method, where the fluxes at the cell interface are evaluated by local reconstruction of the solutions of lattice Boltzmann equations (LBE). Two sets of distribution functions are introduced to reconstruct the cell-interface fluxes, one used for mass and momentum fluxes and the other for the conformation tensor flux in the polymer constitutive equation. The elastic-viscous stress splitting (EVSS) and the solvent-polymer stress splitting (SPSS) techniques are incorporated into the present LBFS to improve the numerical stability. The standard lattice Boltzmann method (LBM) for solving the polymer constitutive equation contains redundant diffusion terms, but this problem is resolved in the current LBFS by setting the relaxation time corresponding to the true diffusion-free limit thus the correct polymer constitutive equation can be recovered. Furthermore, VLBFS eliminates other disadvantages of the standard LBM, such as the LBM on-grid advection coupling the time interval with grid spacing, complicated treatment of the mesoscopic boundary conditions, dependence on uniform grids, and the larger memory requirement due to solving the phase-space discrete distributions. Several flows of a viscoelastic fluid, namely, the two-dimensional plane Poiseuille flow, two-dimensional simplified four-roll mill flows, and three-dimensional Taylor–Green vortex flows, are considered to investigate the accuracy and stability of the present method. The results are found to be in good agreement with the analytical solutions and the previous numerical results. Numerical error analyses show that the present method owns a second-order accuracy in space. The developed VLBFS extends the application domain of LBFS and serves as a basis for simulating viscoelastic flows at high Weissenberg numbers.</div></div>","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"292 ","pages":"Article 106593"},"PeriodicalIF":2.5,"publicationDate":"2025-03-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143643967","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

3D mesh regularization within an ALE code using a weighted line sweeping method

IF 2.5 3区工程技术 Q3 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers & Fluids

Pub Date : 2025-03-06 DOI: 10.1016/j.compfluid.2025.106591

Jérôme Breil , Guillaume Damour , Sébastien Guisset , Arnaud Colaïtis

The Lagrangian formalism is widely used to simulate hydrodynamic responses in complex engineering applications, particularly those involving strong shock waves. However, as the mesh moves with the fluid, it can become highly distorted, requiring a regularization step. This involves constructing a new grid and remapping conservative quantities onto it to restore mesh quality. This work introduces a regularization method for block-structured meshes within a 3D ALE (Arbitrary Lagrangian–Eulerian) code. The proposed approach prevents mesh tangling while preserving the anisotropic features of the initial Lagrangian mesh. This regularization technique incorporates aspect ratio-based weights to control mesh smoothing. Unlike uniform rezoning techniques, this weighted approach maintains proximity to the Lagrangian mesh while improving mesh quality. The method effectively handles concave geometries by mitigating the grid attraction phenomenon, which typically leads to mesh concentration along concave edges. Numerical experiments demonstrate its efficiency in regularizing severely deformed meshes, and its integration within the ALE framework is validated on challenging hydrodynamic test cases, including the triple point problem.

{"title":"3D mesh regularization within an ALE code using a weighted line sweeping method","authors":"Jérôme Breil , Guillaume Damour , Sébastien Guisset , Arnaud Colaïtis","doi":"10.1016/j.compfluid.2025.106591","DOIUrl":"10.1016/j.compfluid.2025.106591","url":null,"abstract":"<div><div>The Lagrangian formalism is widely used to simulate hydrodynamic responses in complex engineering applications, particularly those involving strong shock waves. However, as the mesh moves with the fluid, it can become highly distorted, requiring a regularization step. This involves constructing a new grid and remapping conservative quantities onto it to restore mesh quality. This work introduces a regularization method for block-structured meshes within a 3D ALE (Arbitrary Lagrangian–Eulerian) code. The proposed approach prevents mesh tangling while preserving the anisotropic features of the initial Lagrangian mesh. This regularization technique incorporates aspect ratio-based weights to control mesh smoothing. Unlike uniform rezoning techniques, this weighted approach maintains proximity to the Lagrangian mesh while improving mesh quality. The method effectively handles concave geometries by mitigating the grid attraction phenomenon, which typically leads to mesh concentration along concave edges. Numerical experiments demonstrate its efficiency in regularizing severely deformed meshes, and its integration within the ALE framework is validated on challenging hydrodynamic test cases, including the triple point problem.</div></div>","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"292 ","pages":"Article 106591"},"PeriodicalIF":2.5,"publicationDate":"2025-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143591446","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Advanced numerical methods for conjugate heat transfer problems 共轭传热问题的高级数值方法

IF 2.5 3区工程技术 Q3 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers & Fluids

Pub Date : 2025-03-03 DOI: 10.1016/j.compfluid.2025.106594

Marc-Paul Errera

Conjugate heat transfer (CHT) analysis is a simulation process that addresses the thermal interaction between a solid body and a fluid. It is a crucial aspect in a wide range of engineering applications, especially in the aerospace industry. This paper focuses on implementing adaptive coupling coefficients to optimize CHT by improving stability and simplicity. A mathematical model based on a normal mode stability analysis is employed. This study highlights the importance of a new dimensionless number, the "numerical Biot number", and explores adaptive coupling coefficients in three distinct aerothermal situations: steady coupling, steady coupling with radiation, and unsteady coupling. The main results of these three cases are compared, illustrated, and analyzed. The results demonstrate the potential of the theoretical approach, particularly in understanding the impact of different phenomena on the stability process and the challenges of convergence in certain conditions.

引用次数: 0

Optimizing fluid mixing in channel flow using wall-mounted flexible structures

IF 2.5 3区工程技术 Q3 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers & Fluids

Pub Date : 2025-02-27 DOI: 10.1016/j.compfluid.2025.106590

Gaurav Singh , Arahata Senapati , Arnab Atta , Rajaram Lakkaraju

In channel mixers with two parallel streams of fluid, mixing is achieved by either laminar diffusion at low Reynolds numbers or from flow agitation due to geometric variations. Traditionally, rigid obstructions or textures are used in various spatial arrangements to improve fluid mixing. In our work, we have numerically investigated the mixing performance of a passive scalar in a two-dimensional channel flow accompanied by wall-mounted flexible plates as obstructions for a wide range of low Reynolds numbers (

R e

). The thin plates are arranged on opposite walls of the channel, and the distance between them is varied in the range

0 h

2 h

, where

h

is the channel lateral width. The different arrangements result in corresponding flow paths, thereby affecting fluid mixing and flow rate due to the pressure head losses. We assessed the mixing performance in the channel via the mixing index and the head loss. Our results show that the channel with the two plates when arranged exactly opposite the walls (without a separation gap), offers the highest mixing with significant pressure drop. In contrast, either a single plate or two plates widely separated result in nearly similar levels of mixing index with a lower head loss. We devised a performance index based on a cost-benefit analogy by comparing the flexible plate configurations with the plane channel (i.e, without any obstruction) so as to assess the mixing effectiveness and found that a single flexible plate in the channel in the flow conditions with

R e \approx 400

suit the best for the mixing of two fluid in the channel.

{"title":"Optimizing fluid mixing in channel flow using wall-mounted flexible structures","authors":"Gaurav Singh , Arahata Senapati , Arnab Atta , Rajaram Lakkaraju","doi":"10.1016/j.compfluid.2025.106590","DOIUrl":"10.1016/j.compfluid.2025.106590","url":null,"abstract":"<div><div>In channel mixers with two parallel streams of fluid, mixing is achieved by either laminar diffusion at low Reynolds numbers or from flow agitation due to geometric variations. Traditionally, rigid obstructions or textures are used in various spatial arrangements to improve fluid mixing. In our work, we have numerically investigated the mixing performance of a passive scalar in a two-dimensional channel flow accompanied by wall-mounted flexible plates as obstructions for a wide range of low Reynolds numbers (<span><math><mrow><mi>R</mi><mi>e</mi></mrow></math></span>). The thin plates are arranged on opposite walls of the channel, and the distance between them is varied in the range <span><math><mrow><mn>0</mn><mi>h</mi></mrow></math></span> to <span><math><mrow><mn>2</mn><mi>h</mi></mrow></math></span>, where <span><math><mi>h</mi></math></span> is the channel lateral width. The different arrangements result in corresponding flow paths, thereby affecting fluid mixing and flow rate due to the pressure head losses. We assessed the mixing performance in the channel via the mixing index and the head loss. Our results show that the channel with the two plates when arranged exactly opposite the walls (without a separation gap), offers the highest mixing with significant pressure drop. In contrast, either a single plate or two plates widely separated result in nearly similar levels of mixing index with a lower head loss. We devised a performance index based on a cost-benefit analogy by comparing the flexible plate configurations with the plane channel (i.e, without any obstruction) so as to assess the mixing effectiveness and found that a single flexible plate in the channel in the flow conditions with <span><math><mrow><mi>R</mi><mi>e</mi><mo>≈</mo><mn>400</mn></mrow></math></span> suit the best for the mixing of two fluid in the channel.</div></div>","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"291 ","pages":"Article 106590"},"PeriodicalIF":2.5,"publicationDate":"2025-02-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143519699","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

An adaptive solver for accurate simulation of multicomponent shock-interface problems for thermally perfect species

IF 2.5 3区工程技术 Q3 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers & Fluids

Pub Date : 2025-02-25 DOI: 10.1016/j.compfluid.2025.106587

Yuqi Wang , Ralf Deiterding , Jianhan Liang

A second-order-accurate finite volume method, hybridized by blending an extended double-flux algorithm and a traditionally conservative scheme, is developed. In this scheme, hybrid convective fluxes and hybrid interpolation techniques are designed to ensure stability and accuracy in the presence of both material interfaces and shocks. Two approaches, extended from the original double-flux model, are presented to eliminate the well-known ”pressure oscillation” phenomenon at material interfaces observed with the traditional conservative scheme. Numerous verification simulations confirm that the method can handle multi-dimensional shock-interface problems reliably and efficiently, even in the presence of viscous and reactive terms.

引用次数: 0

A differentiated geometry blade parameterization methodology for gas turbines

IF 2.5 3区工程技术 Q3 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers & Fluids

Pub Date : 2025-02-25 DOI: 10.1016/j.compfluid.2025.106588

Christian Voß , Martin Siggel , Jan Backhaus , Georgios Goinis , Andreas Pahs

The present paper is about the turbomachinery blade geometry parameterization, which has been developed at the DLR Institute of Propulsion Technology for the last two decades. It was implemented into the software BladeGen, which is used to design blades and vanes for axial and radial compressors as well as for fans, propellers and turbines. The parameterization uses typical geometric variables such as characteristic angles and lengths in combination with B-spline parameters like control points and knot vectors to generate a variable number of two-dimensional profile-sections, which are transformed by a conformal mapping into three-dimensional space and connected via NURBS surfaces to form blade surfaces and solids. This publication presents the application philosophy and some parameterization and implementation details. The intuitive handling and some software enhancements such as algorithmic differentiation capabilities are also demonstrated by application examples.

引用次数: 0

An efficient class of increasingly high-order ENO schemes with multi-resolution

IF 2.5 3区工程技术 Q3 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers & Fluids

Pub Date : 2025-02-24 DOI: 10.1016/j.compfluid.2025.106589

Hua Shen

We construct an efficient class of increasingly high-order (up to 17th-order) essentially non-oscillatory schemes with multi-resolution (ENO-MR) for solving hyperbolic conservation laws. The candidate stencils for constructing ENO-MR schemes range from the first-order one-point stencil increasingly up to the designed very high-order stencil. The proposed ENO-MR schemes adopt a simple and efficient strategy that only requires the computation of the highest-order derivatives of a part of candidate stencils. Theoretical analysis and numerical computations indicate that ENO-MR schemes achieve designed high-order convergence in smooth regions which may contain high-order critical points (local extrema) and retain ENO property for strong shocks. Moreover, the performance of ENO-MR schemes does not depend on the scale of the solutions.

引用次数: 0

Incremental singular value decomposition based model order reduction of scale resolving fluid dynamic simulations

IF 2.5 3区工程技术 Q3 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers & Fluids

Pub Date : 2025-02-21 DOI: 10.1016/j.compfluid.2025.106579

Niklas Kühl

<div><div>Scale-resolving flow simulations often feature several million [thousand] spatial [temporal] discrete degrees of freedom. When storing or re-using these data, e.g., to subsequently train some sort of data-based surrogate or compute consistent adjoint flow solutions, a brute-force storage approach is practically impossible. Therefore, – mandatory incremental – Reduced Order Modeling (ROM) approaches are an attractive alternative since only a specific time horizon is effectively stored, usually aligned with the amount of fast available, e.g., Random Access Memory (RAM). This bunched flow solution is then used to enhance the already computed ROM so that the allocated memory can be released and the procedure repeats.</div><div>This paper utilizes an incremental truncated Singular Value Decomposition (itSVD) procedure to compress flow data resulting from scale-resolving flow simulations. To this end, two scenarios are considered, referring to an academic Large Eddy Simulation (LES) around a circular cylinder at <span><math><mrow><msub><mrow><mi>Re</mi></mrow><mrow><mi>D</mi></mrow></msub><mo>=</mo><mn>1</mn><mo>.</mo><mn>4</mn><mi>⋅</mi><mn>1</mn><msup><mrow><mn>0</mn></mrow><mrow><mn>5</mn></mrow></msup></mrow></math></span> as well as an industrial case that employs a hybrid filtered/averaged Detached Eddy Simulation (DES) on the flow around the superstructure of a full-scale feeder ship at <span><math><mrow><msub><mrow><mi>Re</mi></mrow><mrow><mi>L</mi></mrow></msub><mo>=</mo><mn>5</mn><mo>.</mo><mn>0</mn><mi>⋅</mi><mn>1</mn><msup><mrow><mn>0</mn></mrow><mrow><mn>8</mn></mrow></msup></mrow></math></span>.</div><div>The paper’s central focus is on an aspect of severe practical relevance: how much information of the computed scale-resolving solution should be used by the ROM, i.e., how much redundancy occurs in the resolved turbulent fluctuations that favors ROM. In the course of the tSVD employed, this goes hand in hand with the question of ”how many singular values of the flow-solution-snapshot-matrix should be neglected (or considered)” – without (a) re-running the simulation several times in a try-and-error procedure and (b) still obtain compressed results below the model and discretization error. An adaptive strategy is utilized, which features two comparatively simple adjusting screws, for which appropriate decision support is provided. Next to a general feasibility study, reported results show the capability to obtain a fully adaptive data reduction of <span><math><mrow><mi>O</mi><mrow><mo>(</mo><mn>95</mn><mo>)</mo></mrow></mrow></math></span> percent via a computational overhead of <span><math><mrow><mi>O</mi><mrow><mo>(</mo><mn>10</mn><mo>)</mo></mrow></mrow></math></span> percent with a mean accuracy of reconstructed local and global flow data of <span><math><mrow><mi>O</mi><mrow><mo>(</mo><mn>1</mn><msup><mrow><mn>0</mn></mrow><mrow><mo>−</mo><mn>3</mn></mrow></msup><mo>−</mo><mn>1</mn><msup><mrow><mn>0</mn></mrow><mrow><mo>−</m

{"title":"Incremental singular value decomposition based model order reduction of scale resolving fluid dynamic simulations","authors":"Niklas Kühl","doi":"10.1016/j.compfluid.2025.106579","DOIUrl":"10.1016/j.compfluid.2025.106579","url":null,"abstract":"<div><div>Scale-resolving flow simulations often feature several million [thousand] spatial [temporal] discrete degrees of freedom. When storing or re-using these data, e.g., to subsequently train some sort of data-based surrogate or compute consistent adjoint flow solutions, a brute-force storage approach is practically impossible. Therefore, – mandatory incremental – Reduced Order Modeling (ROM) approaches are an attractive alternative since only a specific time horizon is effectively stored, usually aligned with the amount of fast available, e.g., Random Access Memory (RAM). This bunched flow solution is then used to enhance the already computed ROM so that the allocated memory can be released and the procedure repeats.</div><div>This paper utilizes an incremental truncated Singular Value Decomposition (itSVD) procedure to compress flow data resulting from scale-resolving flow simulations. To this end, two scenarios are considered, referring to an academic Large Eddy Simulation (LES) around a circular cylinder at <span><math><mrow><msub><mrow><mi>Re</mi></mrow><mrow><mi>D</mi></mrow></msub><mo>=</mo><mn>1</mn><mo>.</mo><mn>4</mn><mi>⋅</mi><mn>1</mn><msup><mrow><mn>0</mn></mrow><mrow><mn>5</mn></mrow></msup></mrow></math></span> as well as an industrial case that employs a hybrid filtered/averaged Detached Eddy Simulation (DES) on the flow around the superstructure of a full-scale feeder ship at <span><math><mrow><msub><mrow><mi>Re</mi></mrow><mrow><mi>L</mi></mrow></msub><mo>=</mo><mn>5</mn><mo>.</mo><mn>0</mn><mi>⋅</mi><mn>1</mn><msup><mrow><mn>0</mn></mrow><mrow><mn>8</mn></mrow></msup></mrow></math></span>.</div><div>The paper’s central focus is on an aspect of severe practical relevance: how much information of the computed scale-resolving solution should be used by the ROM, i.e., how much redundancy occurs in the resolved turbulent fluctuations that favors ROM. In the course of the tSVD employed, this goes hand in hand with the question of ”how many singular values of the flow-solution-snapshot-matrix should be neglected (or considered)” – without (a) re-running the simulation several times in a try-and-error procedure and (b) still obtain compressed results below the model and discretization error. An adaptive strategy is utilized, which features two comparatively simple adjusting screws, for which appropriate decision support is provided. Next to a general feasibility study, reported results show the capability to obtain a fully adaptive data reduction of <span><math><mrow><mi>O</mi><mrow><mo>(</mo><mn>95</mn><mo>)</mo></mrow></mrow></math></span> percent via a computational overhead of <span><math><mrow><mi>O</mi><mrow><mo>(</mo><mn>10</mn><mo>)</mo></mrow></mrow></math></span> percent with a mean accuracy of reconstructed local and global flow data of <span><math><mrow><mi>O</mi><mrow><mo>(</mo><mn>1</mn><msup><mrow><mn>0</mn></mrow><mrow><mo>−</mo><mn>3</mn></mrow></msup><mo>−</mo><mn>1</mn><msup><mrow><mn>0</mn></mrow><mrow><mo>−</m","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"291 ","pages":"Article 106579"},"PeriodicalIF":2.5,"publicationDate":"2025-02-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143471388","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Break-up of the Taylor bubble

IF 2.5 3区工程技术 Q3 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS

Computers & Fluids

Pub Date : 2025-02-20 DOI: 10.1016/j.compfluid.2025.106577

Evgenii L. Sharaborin , Oleg A. Rogozin , Aslan R. Kasimov

High-resolution direct numerical simulation is used to study the motion of a Taylor bubble in a cylindrical microtube under conditions that lead to the bubble break-up. It is observed that the initial bubble elongates and deforms such that its front part retains a bullet-like shape while its back part forms a skirt shape. Subsequently, the carrier fluid surrounded by the skirt penetrates into the bubble forming a finger that transitions into a bulb shape. The bulb then increases in size until it touches the near-wall liquid film and as a result splits the bubble into two comparable daughter bubbles. Various dynamical features of this break-up process are explored and described in detail.

引用次数: 0

下一页尾页

类型

全部化学•材料生命科学医学物理工程技术环境•农林材料科学地球科学法学管理学化学环境科学与生态学计算机科学教育学经济学农林科学人文科学生物学数学物理与天体物理心理学综合性期刊其他工业工程理学历史学农学文学信息工程

数据库

全部 ACS Publications Elsevier ieeexplore Springer The Royal Society of Chemistry Wiley

期刊

Computers & Fluids

全部 Acc. Chem. Res. ACS Applied Bio Materials ACS Appl. Electron. Mater. ACS Appl. Energy Mater. ACS Appl. Mater. Interfaces ACS Appl. Nano Mater. ACS Appl. Polym. Mater. ACS BIOMATER-SCI ENG ACS Catal. ACS Cent. Sci. ACS Chem. Biol. ACS Chemical Health & Safety ACS Chem. Neurosci. ACS Comb. Sci. ACS Earth Space Chem. ACS Energy Lett. ACS Infect. Dis. ACS Macro Lett. ACS Mater. Lett. ACS Med. Chem. Lett. ACS Nano ACS Omega ACS Photonics ACS Sens. ACS Sustainable Chem. Eng. ACS Synth. Biol. Anal. Chem. BIOCHEMISTRY-US Bioconjugate Chem. BIOMACROMOLECULES Chem. Res. Toxicol. Chem. Rev. Chem. Mater. CRYST GROWTH DES ENERG FUEL Environ. Sci. Technol. Environ. Sci. Technol. Lett. Eur. J. Inorg. Chem. IND ENG CHEM RES Inorg. Chem. J. Agric. Food. Chem. J. Chem. Eng. Data J. Chem. Educ. J. Chem. Inf. Model. J. Chem. Theory Comput. J. Med. Chem. J. Nat. Prod. J PROTEOME RES J. Am. Chem. Soc. LANGMUIR MACROMOLECULES Mol. Pharmaceutics Nano Lett. Org. Lett. ORG PROCESS RES DEV ORGANOMETALLICS J. Org. Chem. J. Phys. Chem. J. Phys. Chem. A J. Phys. Chem. B J. Phys. Chem. C J. Phys. Chem. Lett. Analyst Anal. Methods Biomater. Sci. Catal. Sci. Technol. Chem. Commun. Chem. Soc. Rev. CHEM EDUC RES PRACT CRYSTENGCOMM Dalton Trans. Energy Environ. Sci. ENVIRON SCI-NANO ENVIRON SCI-PROC IMP ENVIRON SCI-WAT RES Faraday Discuss. Food Funct. Green Chem. Inorg. Chem. Front. Integr. Biol. J. Anal. At. Spectrom. J. Mater. Chem. A J. Mater. Chem. B J. Mater. Chem. C Lab Chip Mater. Chem. Front. Mater. Horiz. MEDCHEMCOMM Metallomics Mol. Biosyst. Mol. Syst. Des. Eng. Nanoscale Nanoscale Horiz. Nat. Prod. Rep. New J. Chem. Org. Biomol. Chem. Org. Chem. Front. PHOTOCH PHOTOBIO SCI PCCP Polym. Chem.

﹀