Computers & Fluids最新文献

{"title":"High order well-balanced Arbitrary-Lagrangian-Eulerian ADER discontinuous Galerkin schemes on general polygonal moving meshes","authors":"Elena Gaburro","doi":"10.1016/j.compfluid.2025.106764","DOIUrl":"10.1016/j.compfluid.2025.106764","url":null,"abstract":"<div><div>In this work, we present a novel family of high order accurate numerical schemes for the solution of hyperbolic partial differential equations (PDEs) which combines several geometrical and physical structure preserving properties. First, we settle our methods in the Lagrangian framework, where each element of the mesh evolves following as close as possible the local fluid flow, so to <em>reduce</em> the <em>numerical dissipation</em> at contact waves and moving interfaces and to satisfy the <em>Galilean and rotational invariance</em> properties of the studied PDEs system. In particular, we choose the direct Arbitrary-Lagrangian-Eulerian (ALE) approach which, in order to always guarantee the <em>high quality of the moving mesh</em>, allows to combine the Lagrangian motion with mesh optimization techniques. The employed polygonal tessellation is thus regenerated at each time step, the previous one is connected with the new one by spacetime control volumes, including <em>hole-like sliver</em> elements in correspondence of topology changes, over which we integrate a spacetime divergence form of the original PDEs through a <em>high order</em> accurate ADER discontinuous Galerkin (DG) scheme. Mass <em>conservation</em> and adherence to the <em>GCL condition</em> are guaranteed by construction thanks to the integration over closed control volumes, and <em>robustness</em> over shock discontinuities is ensured by the use of an <em>a posteriori</em> subcell finite volume (FV) limiting technique. On top of this effective moving mesh framework, we have also modified the full ADER-DG scheme with <em>a posteriori</em> subcell FV limiter to be, for the first time in literature, <em>well-balanced</em>. This is achieved by ensuring that any projection, reconstruction and integration procedures were always performed by summing up the exact value of a given equilibrium plus the high order accurate evolution of the fluctuations w.r.t. said equilibrium. The paper is closed by a wide set of numerical results, including simulations of Keplerian disks, which demonstrate all the claimed properties and the increased accuracy of our novel family of schemes, in particular for the evolution of small perturbations arising over moving equilibrium profiles.</div></div>","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"301 ","pages":"Article 106764"},"PeriodicalIF":3.0,"publicationDate":"2025-07-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144925216","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}

{"title":"Fake it till you make it: Synthetic turbulence to achieve swift converged turbulence statistics in a pressure-driven channel flow","authors":"Akshay Patil,&nbsp;Clara García-Sánchez","doi":"10.1016/j.compfluid.2025.106733","DOIUrl":"10.1016/j.compfluid.2025.106733","url":null,"abstract":"<div><div>In this study, we introduced a simple yet innovative application: the isotropic synthetic turbulence field generator (iSTFG), based on the synthetic turbulent inflow generator. The iSTFG leverages the homogeneity in the streamwise direction for channel flows and triggers turbulence to achieve statistically stationary flow conditions faster than standard community-used strategies. We compare this new method with two other well-established methods: linear and log-law profiles superposed with random noise and descending counter-rotating vortices. We find that iSTFG provides a computationally cheap and effective way to reduce simulation spin-up costs/time/emissions to achieve statistically stationary flow conditions when a precursor turbulent initial condition is unavailable. At a one-time cost between 1-10 Central Processing Unit (CPU) hour(s) to generate the synthetic turbulent initial condition based on the target friction Reynolds numbers (1 CPU hour - <span><math><mrow><mi>R</mi><msub><mrow><mi>e</mi></mrow><mrow><mi>τ</mi></mrow></msub><mo>=</mo><mn>500</mn></mrow></math></span>, 7 CPU hours - <span><math><mrow><mi>R</mi><msub><mrow><mi>e</mi></mrow><mrow><mi>τ</mi></mrow></msub><mo>=</mo><mn>2000</mn></mrow></math></span>), the flow achieves statistically stationary turbulent flow (SSTF) state within three eddy turnovers for all the parameters of interest in wall-bounded pressure-driven channel flow simulations when compared to other alternatives that can take more than ten eddy turnovers resulting in substantial savings in the computational cost. We also demonstrate that the transition and convergence to an SSTF state using conventional methods are sensitive to the computational domain size, while iSTFG is agnostic to the domain size. Furthermore, we explored the sensitivity of the iSTFG method to the non-dimensional integral length scale parameter and mismatch in reference and target input data to find iSTFG robust in such scenarios.</div></div>","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"301 ","pages":"Article 106733"},"PeriodicalIF":3.0,"publicationDate":"2025-07-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144749406","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}

{"title":"A structurally compatible forcing scheme for three-dimensional cumulant lattice Boltzmann method","authors":"Jie Kang,&nbsp;Dongyin Wu","doi":"10.1016/j.compfluid.2025.106762","DOIUrl":"10.1016/j.compfluid.2025.106762","url":null,"abstract":"<div><div>This study introduces a structurally compatible forcing scheme for the three-dimensional cumulant lattice Boltzmann method (LBM) based on the D3Q27 velocity set, and successfully integrates the pseudopotential model into the three-dimensional cumulant LBM framework. Asymptotic analysis confirms the scheme’s capability to recover the macroscopic Navier–Stokes equations, demonstrating its compatibility with the cumulant LBM structure. Numerical experiments reveal that under constant external force fields, the proposed scheme reduces the relative errors from 2.44<span><math><mrow><mo>×</mo><mn>1</mn><msup><mrow><mn>0</mn></mrow><mrow><mo>−</mo><mn>5</mn></mrow></msup></mrow></math></span> (the conventional scheme) to 4.09<span><math><mrow><mo>×</mo><mn>1</mn><msup><mrow><mn>0</mn></mrow><mrow><mo>−</mo><mn>8</mn></mrow></msup></mrow></math></span> by accounting for high-order cumulants in the force coupling. Notably, the integration of the pseudopotential model via this scheme effectively suppresses anisotropic distortions observed in the traditional scheme, reducing the maximum dimensionless radial deviation of the droplet from 7.21% to 0.11%. Furthermore, wettability simulations confirm the reliability of the scheme, showing a strong linear correlation (<span><math><mrow><msup><mrow><mi>R</mi></mrow><mrow><mn>2</mn></mrow></msup><mo>&gt;</mo><mn>0</mn><mo>.</mo><mn>99</mn></mrow></math></span>) between the contact angle and the adhesive parameter<!--> <span><math><msub><mrow><mi>G</mi></mrow><mrow><mi>w</mi></mrow></msub></math></span>. Additionally, droplet collision tests validate its robustness in dynamic scenarios. This work develops a structurally compatible forcing scheme for the three-dimensional cumulant LBM, which modifies the force effects on high-order cumulants. The proposed scheme maintains the cumulant-based LBM structure while enabling robust simulations across diverse flow regimes, from single-phase flow to dynamic droplet collision.</div></div>","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"300 ","pages":"Article 106762"},"PeriodicalIF":2.5,"publicationDate":"2025-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144712933","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}

{"title":"Three-dimensional SPH modeling of brittle fracture under hydrodynamic loading","authors":"Vishabjeet Singh ,&nbsp;Chong Peng ,&nbsp;Md Rushdie Ibne Islam","doi":"10.1016/j.compfluid.2025.106763","DOIUrl":"10.1016/j.compfluid.2025.106763","url":null,"abstract":"<div><div>A three-dimensional SPH computational framework is presented for modeling fluid–structure interactions with structural deformation and failure. We combine weakly compressible SPH with a pseudo-spring-based SPH solver to capture the fluid flow and deformable structures. A unified modeling approach captures the solid boundaries and fluid–structure interfaces without penalty-based contact force. The <span><math><mi>δ</mi></math></span>-SPH technique improves the pressure calculations in the fluid phase, while structural damage is modeled using a pseudo-spring approach, with particle interactions limited to its neighbors. The present framework can capture the three-dimensional crack surfaces in structures without any computationally intensive crack-tracking algorithm or visibility criteria. The framework has been proven effective against existing models and experimental data, demonstrating high accuracy and robustness in simulating detailed fracture patterns and offering insights into the impact of hydrodynamic events on structural integrity.</div></div>","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"301 ","pages":"Article 106763"},"PeriodicalIF":3.0,"publicationDate":"2025-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144739595","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}

{"title":"Non-intrusive model reduction of advection-dominated hyperbolic problems using neural network shift augmented manifold transformation","authors":"Harshith Gowrachari ,&nbsp;Nicola Demo ,&nbsp;Giovanni Stabile ,&nbsp;Gianluigi Rozza","doi":"10.1016/j.compfluid.2025.106758","DOIUrl":"10.1016/j.compfluid.2025.106758","url":null,"abstract":"<div><div>Advection-dominated problems are predominantly noticed in nature, engineering systems, and various industrial processes. Traditional linear compression methods, such as proper orthogonal decomposition (POD) and reduced basis (RB) methods are ill-suited for these problems, due to slow Kolmogorov <em>n</em>-width decay. This results in inefficient and inaccurate reduced order models (ROMs). There are few non-linear approaches to accelerate the Kolmogorov <em>n</em>-width decay. In this work, we use a neural network shift augmented transformation technique that employs automatic shift detection. This approach leverages a deep-learning framework to derive a parameter-dependent mapping between the original manifold <span><math><mi>M</mi></math></span> and the transformed manifold <span><math><mover><mrow><mi>M</mi></mrow><mrow><mo>̃</mo></mrow></mover></math></span>. We apply a linear compression method to obtain a low-dimensional linear approximation subspace of the transformed manifold <span><math><mover><mrow><mi>M</mi></mrow><mrow><mo>̃</mo></mrow></mover></math></span>. Furthermore, we construct non-intrusive reduced order models on the resulting transformed linear approximation subspace and employ automatic shift detection for predictions in the online stage. We propose a complete framework, the neural network shift-augmented proper orthogonal decomposition-based reduced order model (NNsPOD-ROM) algorithm, comprising both offline and online stages for model reduction of advection-dominated problems. We test our proposed methodology on numerous experiments to evaluate its performance on the 1D linear advection equation, a higher order method benchmark case - the 2D isentropic convective vortex, and 2D two-phase flow.</div></div>","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"300 ","pages":"Article 106758"},"PeriodicalIF":3.0,"publicationDate":"2025-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144739539","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}

{"title":"Flow around square, rounded, and round-convex cylinders at Reynolds numbers 20 to 22,000","authors":"Haotian Dong ,&nbsp;Shuting Fang ,&nbsp;Xiaoqing Du","doi":"10.1016/j.compfluid.2025.106771","DOIUrl":"10.1016/j.compfluid.2025.106771","url":null,"abstract":"<div><div>Flow around a square-like cylinder with round corners at a radius <em>R</em> = <em>B</em>/7 (<em>B</em>: nominal side length) and round sides at a radius <em>K</em> = <em>B</em> were numerically investigated at Reynolds numbers <em>Re</em> = 20 to 22,000. The <em>Re</em>-dependence of flow-induced loading and flow features of this “round-convex” cylinder was compared with square and rounded cylinders. Compared with the square and rounded cylinders, the round-convex cross-section has a smaller and declining drag, lower fluctuating loads, and a weaker vortex shedding from <em>Re</em> = 1000 to 22,000, showing distinct Reynolds number effects. A steady regime at <em>Re</em> = 20 and 45, a 2D regime at <em>Re</em> = 100 and 200, a 3D regime at <em>Re</em> = 300, 500, and 1000, and a turbulent regime at <em>Re</em> = 3900, 10,000, and 22,000 are identified for the round-convex and rounded cylinders, while a 3D flow is observed at <em>Re</em> = 300 for the square cylinder. The flow separation shifts from the rear corner to the front corner at <em>Re</em> = 200 and 300 for the square and rounded cylinders, respectively, while the separation point gradually moves to the middle of lateral sides for round-convex cylinders, which is similar to a circular cylinder.</div></div>","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"300 ","pages":"Article 106771"},"PeriodicalIF":2.5,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144714216","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}

{"title":"Benchmarking convolutional neural network and graph neural network based surrogate models on a real-world car external aerodynamics dataset","authors":"Sam Jacob Jacob ,&nbsp;Markus Mrosek ,&nbsp;Carsten Othmer ,&nbsp;Harald Köstler","doi":"10.1016/j.compfluid.2025.106760","DOIUrl":"10.1016/j.compfluid.2025.106760","url":null,"abstract":"<div><div>Aerodynamic optimization is crucial for developing eco-friendly, aerodynamic, and stylish cars, which requires close collaboration between aerodynamicists and stylists, a collaboration impaired by the time-consuming nature of aerodynamic simulations. Surrogate models offer a viable solution to reduce this overhead, but they are untested in real-world aerodynamic datasets. We present a comparative evaluation of two surrogate modeling approaches for predicting drag on a real-world dataset: a Convolutional Neural Network (CNN) model that uses a signed distance field as input and a commercial tool based on Graph Neural Networks (GNN) that directly processes a surface mesh. In contrast to previous studies based on datasets created from parameterized geometries, our dataset comprises 343 geometries derived from 32 baseline vehicle geometries across five distinct car projects, reflecting the diverse, free-form modifications encountered in the typical vehicle development process. Our results show that the CNN-based method achieves a mean absolute error of 2.3 drag counts, while the GNN-based method achieves 3.8. Both methods achieve approximately 77% accuracy in predicting the direction of drag change relative to the baseline geometry. While both methods effectively capture the broader trends between baseline groups (set of samples derived from a single baseline geometry), they struggle to varying extents in capturing the finer intra-baseline group variations. In summary, our findings suggest that aerodynamicists can effectively use both methods to predict drag in under two minutes, which is at least 600 times faster than performing a simulation. However, there remains room for improvement in capturing the finer details of the geometry.</div></div>","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"300 ","pages":"Article 106760"},"PeriodicalIF":2.5,"publicationDate":"2025-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144711726","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}

{"title":"Enhanced simulation techniques in predicting sonic boom loudness using CFD","authors":"S. Neuhoff,&nbsp;C. Ashby,&nbsp;J. Housman,&nbsp;J. Duensing","doi":"10.1016/j.compfluid.2025.106759","DOIUrl":"10.1016/j.compfluid.2025.106759","url":null,"abstract":"<div><div>This paper outlines advancements in predicting sonic boom loudness within the Launch, Ascent, and Vehicle Aerodynamics computational framework. Traditionally, a two step process consisting of a steady state computational fluid dynamics problem for near-field analysis and a far-field propagation solver for calculation of loudness metrics has been used. Improvements to this process made in this work include utilizing a high-order space marching method for mid-field computations, applying an output-based mesh adaptation method targeting error in near-field pressure signatures, and implementing a scripting system using curvilinear grids to increase robustness and simplify the process of running large databases of simulation cases. These advancements are detailed and applied to the simulation of the X-59 as a representative example, presenting comparative cost and timing analyses between the prior two step workflow and the current three step procedure. We achieve increased accuracy and robustness for loudness predictions with at least a 50% computational cost reduction.</div></div>","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"300 ","pages":"Article 106759"},"PeriodicalIF":2.5,"publicationDate":"2025-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144703705","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}

{"title":"Mixture-conservative temperature-based Baer–Nunziato solver for efficient full-disequilibrium simulations of real fluids","authors":"G. Sirianni ,&nbsp;B. Re ,&nbsp;R. Abgrall","doi":"10.1016/j.compfluid.2025.106761","DOIUrl":"10.1016/j.compfluid.2025.106761","url":null,"abstract":"<div><div>In this work, we present a primitive update scheme for the full-disequilibrium Baer–Nunziato equations that is conservative in the total energy of the mixture and valid for generic equations of state. The update scheme is derived for a generic thermodynamic variable and is independent of the chosen spatial discretization. We show results of various Riemann problems from the literature obtained by updating phasic temperatures through the proposed scheme and compare them to the standard approach and analytical solutions. The total energy imbalance of the mixture is assessed, and computational speed-ups using the Span–Wagner equation of state are briefly discussed. Finally, the scheme is tested in complex thermodynamic conditions on a two-phase non-ideal and a two-fluid non-classical Riemann problem, using the Span–Wagner equation of state with vanishing phases.</div></div>","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"300 ","pages":"Article 106761"},"PeriodicalIF":2.5,"publicationDate":"2025-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144695156","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}

{"title":"Localized implicit iterative shifting to improve particle distribution in Smoothed Particle Hydrodynamics","authors":"Mohamad Amin Ghazi ,&nbsp;Renato Vacondio ,&nbsp;Jean-Christophe Marongiu","doi":"10.1016/j.compfluid.2025.106753","DOIUrl":"10.1016/j.compfluid.2025.106753","url":null,"abstract":"<div><div>This study presents a novel Implicit Iterative Particle Shifting method designed to enhance the accuracy of Smoothed Particle Hydrodynamics (SPH) simulations by improving particle distribution through the solution of a linear system of equations. The proposed approach addresses key limitations of existing techniques, namely distribution instability and lack of control over maximum particle displacement by accounting for the contribution of neighboring particles on a given particle when constructing the linear system. The method also imposes a constraint on the shifting magnitude, allowing for better control of particle movement and improved distribution stability also in presence of free surface. To reduce computational cost, a localization procedure has been introduced that restricts particle shifting to regions where it is necessary. The effectiveness of the method is demonstrated through a series of benchmark tests, including the Taylor–Green vortex, a moving square box, an oscillating droplet, and a dam break flow. The results show substantial improvements in the uniformity of the particle distribution, computational efficiency, and simulation accuracy compared to established methods.</div></div>","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"300 ","pages":"Article 106753"},"PeriodicalIF":2.5,"publicationDate":"2025-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144695168","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}