Computers & Fluids最新文献

{"title":"A hybrid finite-volume reconstruction framework for efficient high-order shock-capturing on unstructured meshes","authors":"Yiren Tong,&nbsp;Panagiotis Tsoutsanis","doi":"10.1016/j.compfluid.2026.106988","DOIUrl":"10.1016/j.compfluid.2026.106988","url":null,"abstract":"<div><div>In this paper, we present a multi-dimensional, arbitrary-order hybrid reconstruction framework for compressible flows on unstructured meshes. The proposed method advances state-of-the-art high-resolution schemes by combining the efficiency of linear reconstruction with the robustness of high-order non-oscillatory formulations, activated only where necessary through a novel a priori detection strategy. This approach minimises the use of costly Compact Weighted Essentially Non-Oscillatory (CWENOZ) or Monotonic Upstream-centered Scheme for Conservation Laws (MUSCL) reconstructions, thereby substantially reducing computational overhead without compromising accuracy or stability. The framework integrates the strengths of CWENOZ formulations and the Multi-dimensional Optimal Order Detection (MOOD) paradigm, while introducing a redesigned Numerical Admissibility Detector (NAD) that classifies the local flow field in a single step into smooth, weakly non-smooth, and discontinuous regions. Each region is then reconstructed using an optimal method: a high-order linear scheme in smooth areas, CWENOZ in weakly non-smooth zones, and a second-order MUSCL scheme near discontinuities. This targeted, a priori allocation preserves high-order accuracy where possible and guarantees non-oscillatory, stable solutions near shocks and strong gradients. The proposed hybrid strategy is implemented within the open-source unstructured finite-volume solver UCNS3D and supports arbitrary-order reconstructions on mixed-element meshes. Comprehensive two- and three-dimensional benchmark tests demonstrate that the method maintains the designed order of accuracy in smooth regions while significantly enhancing robustness in shock-dominated flows. Owing to the reduced frequency of expensive nonlinear reconstructions, the framework achieves up to a 2.5 ×  speed-up compared to a CWENOZ scheme of the same order in 3D compressible turbulence simulations. Overall, this hybrid framework brings high-order accuracy closer to in industrial-scale CFD simulations through its combination of reduced computational cost, improved robustness, and reliability.</div></div>","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"308 ","pages":"Article 106988"},"PeriodicalIF":3.0,"publicationDate":"2026-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146076190","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":"An enhanced second-order discretization scheme for free surface and vicinity particles in MPS method aided by surface mesh","authors":"Gen Li ,&nbsp;Yunlong Liao ,&nbsp;Peitao Yao","doi":"10.1016/j.compfluid.2026.106996","DOIUrl":"10.1016/j.compfluid.2026.106996","url":null,"abstract":"<div><div>Moving Particle Semi-implicit (MPS) method is an emerging numerical method for free-surface flow involving complex deformation, fragmentation and coalescence of various fluid interfaces. However, higher-order discretization schemes for the MPS method remain imperfect. The non-uniform particle distribution near the free surface is highly prone to causing numerical divergence. The conventional virtual-particle-based second-order discretization scheme degrades the discretization scheme for the free surface and particles in its vicinity to a lower-order format. This treatment thus leads to error accumulation and propagation. To solve these problems, an improved second-order discretization scheme was developed with the aid of a surface mesh. A surface mesh constructed at the free surface provided the missing position information for neighboring particles required by the surface particles’ second-order discretization and compensated for particle number density deficiencies. A sensitivity analysis was conducted on surface mesh resolution and the particle-to-mesh size ratio was determined to balance computational efficiency and accuracy. Compared to the prior virtual-particle-based second-order method, the proposed approach enabled accurate discretization for free surface particles, preventing error accumulation caused by non-uniform particle distributions. Validations were conducted by simulating four benchmark cases of still water pool pressure, dam break flow, elliptical droplet evolution, and square droplet rotation. The results demonstrated that the proposed surface-mesh-based method exhibited superior performance in pressure calculation accuracy, free surface particle distribution uniformity, and surface consistency.</div></div>","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"308 ","pages":"Article 106996"},"PeriodicalIF":3.0,"publicationDate":"2026-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146076191","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":"Bayesian calibration of RANS model parameters based on hybrid surrogate modeling and adaptive sampling","authors":"Xiang Qiu ,&nbsp;Yujie Wen ,&nbsp;Xueqin Zhou ,&nbsp;Yulu Liu","doi":"10.1016/j.compfluid.2026.106982","DOIUrl":"10.1016/j.compfluid.2026.106982","url":null,"abstract":"<div><div>This study introduces a new Bayesian uncertainty quantification and calibration method, employing a hybrid surrogate model for parameter calibration and uncertainty analysis in turbulence modeling. The proposed method integrates three surrogate modeling techniques, including Gaussian Process Regression, Polynomial Chaos Expansion and BackPropagation neural networks, by employing a weighted averaging approach to construct a high-accuracy and robust hybrid surrogate model. To optimize model performance, an adaptive sampling method is employed, adjusting the distribution of samples dynamically based on output uncertainty and dispersion. This approach improves model fitting accuracy and predictive capability while reducing computational costs. Validation of the surrogate modeling method is carried out through mathematical function fitting, demonstrating its ability to improve accuracy and decrease sample requirements. Furthermore, the method is applied to two representative turbulence cases: periodic hill flow and backward-facing step flow. In the periodic hill case, model calibration performance is assessed using high-fidelity Direct Numerical Simulation data, showing that the calibrated model reduces prediction errors in the recirculation zone. In the backward-facing step case, the model’s applicability in separated turbulence is further verified through parameter calibration and posterior uncertainty quantification. The simulation reveals that the refined model effectively improves consistency with experimental data, reducing errors in predicting key flow characteristics, especially in regions with intense flow variations.</div></div>","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"308 ","pages":"Article 106982"},"PeriodicalIF":3.0,"publicationDate":"2026-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146076125","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":"Symmetry-preserving neural indicators for discontinuity detection in high-order discontinuous Galerkin solvers","authors":"G.G. Kokkinakis,&nbsp;A.I. Delis","doi":"10.1016/j.compfluid.2026.106984","DOIUrl":"10.1016/j.compfluid.2026.106984","url":null,"abstract":"<div><div>This paper presents a novel symmetry-aware troubled-cell indicator (TCI) for high-order ADER Discontinuous Galerkin (ADER-DG) schemes, designed to improve robustness and generalization in the presence of discontinuities on unstructured triangular meshes. The proposed method leverages a Siamese Convolutional Neural Network (SCNN-TCI) that explicitly encodes invariance to geometric transformations-rotations and reflections-by design. Six transformed versions of each input patch are processed through identical CNN branches with shared weights, and their features are fused using pixel-wise max-pooling to achieve transformation-invariant classification. This approach eliminates the need for extensive data augmentation and ensures consistent predictions across varying mesh orientations. Numerical experiments on analytical functions, as well as the two-dimensional Burgers and Euler equations, demonstrate that the SCNN-TCI accurately detects both sharp gradients and shock regions, while preserving rotational and reflectional symmetry. The architecture integrates seamlessly with the ADER-DG solver, triggering sub-cell limiting only when necessary, and represents a compact, interpretable, and computationally efficient alternative to existing ML-based TCIs.</div></div>","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"308 ","pages":"Article 106984"},"PeriodicalIF":3.0,"publicationDate":"2026-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146076123","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":"Neural physics: Using AI libraries to develop physics-based solvers for incompressible computational fluid dynamics","authors":"Boyang Chen ,&nbsp;Claire E. Heaney ,&nbsp;Christopher C. Pain","doi":"10.1016/j.compfluid.2026.106981","DOIUrl":"10.1016/j.compfluid.2026.106981","url":null,"abstract":"<div><div>Numerical discretisations of partial differential equations (PDEs) can be written as discrete convolutions, which, themselves, are a key tool in AI libraries and used in convolutional neural networks (CNNs). We therefore propose to implement numerical discretisations as convolutional layers of a neural network, where the weights or filters are determined analytically rather than by training. Furthermore, we demonstrate that these systems can be solved entirely by functions in AI libraries, either by using Jacobi iteration or multigrid methods, the latter realised through a U-Net architecture. Some advantages of the Neural Physics approach are that (1) the methods are platform agnostic; (2) the resulting solvers are fully differentiable, ideal for optimisation tasks; and (3) writing CFD solvers as (untrained) neural networks means that they can be seamlessly integrated with trained neural networks to form hybrid models. We demonstrate the proposed approach on a number of test cases of increasing complexity from advection-diffusion problems, the non-linear Burgers equation to the Navier-Stokes equations. We validate the approach by comparing our results with solutions obtained from traditionally written code and common benchmarks from the literature. We show that the proposed methodology can solve all these problems using repurposed AI libraries in an efficient way, without training, and presents a new avenue to explore in the development of methods to solve PDEs with implicit methods.</div></div>","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"308 ","pages":"Article 106981"},"PeriodicalIF":3.0,"publicationDate":"2026-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146076122","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":"Dynamic stall mitigation of a pitching aerofoil using a data-driven model","authors":"Luca Damiola ,&nbsp;Jan Decuyper ,&nbsp;Mark C. Runacres ,&nbsp;Tim De Troyer","doi":"10.1016/j.compfluid.2026.106986","DOIUrl":"10.1016/j.compfluid.2026.106986","url":null,"abstract":"<div><div>Dynamic stall is an unsteady aerodynamic phenomenon which temporarily enhances lift and delays flow separation on a lifting surface, but is also associated with large load fluctuations that may compromise the structural integrity of the system. The present work, based on transient computational fluid dynamics (CFD) simulations, proposes a methodology to mitigate the undesired effects of dynamic stall on a pitching NACA 0018 aerofoil undergoing a large-amplitude sinusoidal oscillation. The study aims to alleviate the post-stall load fluctuations by introducing small modifications to the pitching kinematics of the aerofoil. The approach relies on the construction of a nonlinear data-driven model of the system, which is capable of predicting the time-varying lift, drag, and moment coefficients from a given angle-of-attack time series. This fast and accurate nonlinear model, based on neural networks, is coupled with a multi-objective genetic algorithm designed to optimise two competing objectives: the negative peak pitching moment coefficient and the mean lift coefficient. The optimised pitching parameters are identified by modifying the original sinusoidal motion through the superposition of two higher harmonics, with their amplitudes and phases being the design variables. The optimised aerofoil motion proposed by the genetic algorithm is subsequently evaluated through CFD analysis to verify the accuracy of the model predictions. Results show good agreement between the predicted and the actual transient aerodynamic coefficients, demonstrating that small adjustments to the pitching trajectory can lead to substantial reduction of the peak loads during deep dynamic stall. The obtained results further underscore the usefulness of nonlinear data-driven models, which are particularly well-suited for integration into optimisation and control frameworks that require both accuracy and a fast evaluation time.</div></div>","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"308 ","pages":"Article 106986"},"PeriodicalIF":3.0,"publicationDate":"2026-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146076124","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":"Prediction of surface pressure distributions of non-parametric airfoils using geometric deep learning methods","authors":"Derrick Hines ,&nbsp;Philipp Bekemeyer","doi":"10.1016/j.compfluid.2026.106979","DOIUrl":"10.1016/j.compfluid.2026.106979","url":null,"abstract":"<div><div>Computational Fluid Dynamics (CFD) simulations are one of the cornerstones in providing aerodynamic data required for aircraft design and optimization. However, using these simulations extensively is limited by their high computational demands. Therefore, it is essential to create efficient data-driven surrogate models for CFD solvers. In many practical scenarios an explicit unifying parameterization of aircraft configurations is not available. This highlights the need for models that operate directly on raw geometric representations. Geometric deep learning has emerged as a class of deep learning techniques capable of operating on such data, enabling predictive modeling without the need of an explicit parameterization. In this paper, we extend and investigate two geometric deep learning approaches for the prediction of surface pressure distributions of non-parametric airfoils. These methods are Bi-Stride Multi-Scale Graph Neural Network and Implicit Neural Representation of the signed distance function coupled with a Multi-Layer Perceptron. To enhance both of these methods, we propose the use of area-weighted loss functions to better account for variations in node density in the meshes. Moreover, in the formulation of the graph neural network we introduce edge completion at the coarsest level to account for interactions between different connected components, such as flap, main element and slat in a 3-element high-lift airfoil. These methods are compared to the well-established method Proper Orthogonal Decomposition coupled with Interpolation, which is allowed to use an explicit parameterization and serves as a baseline. Two high-fidelity datasets with CFD simulations solving the Reynold-Averaged Navier Stokes equations are created. The first one features a varied set of single-element airfoils with simulations in the subsonic and transonic regime, while the second one features high-lift multi-element airfoils with a variable flap position with simulations in the subsonic regime. The results show that both geometric deep learning approaches outperform the established baseline across various data regimes. These approaches can capture shocks and flow separation with more accuracy. The use of an area-weighted loss function enhances area-weighted performance and leads to faster performance gains in the early training epochs. These findings support the potential of geometric deep learning methods as data-driven surrogates of CFD solvers for varying geometries.</div></div>","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"308 ","pages":"Article 106979"},"PeriodicalIF":3.0,"publicationDate":"2026-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146076192","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":"Burst buffer accelerated direct numerical simulation of turbulence generated by 3D sparse multiscale grids","authors":"Syed M. Usama ,&nbsp;Nadeem A. Malik ,&nbsp;Umair Umer ,&nbsp;Amjad Shaikh ,&nbsp;Zhigang Sun","doi":"10.1016/j.compfluid.2026.106985","DOIUrl":"10.1016/j.compfluid.2026.106985","url":null,"abstract":"<div><div>This study investigates the free-stream turbulence characteristics generated by a novel three-dimensional sparse multiscale grid (3DSG) using a burst buffer accelerated direct numerical simulation algorithm (BBDSA). The BBDSA achieves a fivefold reduction in computation time and an eightfold improvement in parallel efficiency by employing burst buffers to absorb large data volumes, thereby mitigating I/O bottlenecks in parallel file systems and enabling faster computation of turbulent flows. The algorithm was applied to examine turbulence modulation induced by three types of turbulence generating grids: two-dimensional classical, two-dimensional fractal, and 3DSG configurations. Simulations were conducted for a uniform inflow at a Reynolds number of 4000 within a conduit representative of a wind tunnel. Comparative analyses revealed that the 3DSG with a 24% blockage ratio produced turbulence intensities and Reynolds stresses comparable to those generated by classical and fractal grids with substantially higher blockage ratios. These findings advance the understanding of turbulent flow, highlighting the potential of sparse multiscale grids for efficient turbulence production and their applicability in the design of flow-sensitive engineering systems.</div></div>","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"308 ","pages":"Article 106985"},"PeriodicalIF":3.0,"publicationDate":"2026-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146076189","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 separation prediction in a simplified notchback car model by assimilating global luminescent oil film measurements","authors":"Takashi Misaka ,&nbsp;Takuji Nakashima ,&nbsp;Keigo Shimizu ,&nbsp;Masato Hijikuro ,&nbsp;Masayuki Anyoji ,&nbsp;Yoshiyuki Furukawa","doi":"10.1016/j.compfluid.2026.106983","DOIUrl":"10.1016/j.compfluid.2026.106983","url":null,"abstract":"<div><div>This study presents an enhancement of Reynolds-averaged Navier-Stokes (RANS) simulations for predicting the flow around a simplified notchback car model by incorporating experimental data from global luminescent oil film (GLOF) measurements. The simulations employ the SST <em>k</em>-<em>ω</em> turbulence model, with a particular focus on improving predictions of flow separation in the rear regions of the car model, including the rear window and rear side surfaces. To achieve this, the parameter <em>a</em>₁ of the SST <em>k</em>-<em>ω</em> turbulence model is zonally optimized using an ensemble Kalman filter (EnKF), ensuring that the predicted separation locations align locally with near-wall flow features captured by the GLOF measurements. The data assimilation framework is first validated through a numerical data assimilation experiment (twin experiment) that mimics GLOF measurements within the RANS simulation. Following this validation, the system is applied to actual GLOF measurements. The resulting GLOF-informed <span><math><msub><mi>a</mi><mn>1</mn></msub></math></span> distribution yields near-wall flow patterns that closely match those observed in the experiment, demonstrating the effectiveness of the proposed approach.</div></div>","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"308 ","pages":"Article 106983"},"PeriodicalIF":3.0,"publicationDate":"2026-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146076126","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":"Effect of temperature and curvature on surface tension and Tolman length in the multiphase lattice Boltzmann method","authors":"Fu Ling ,&nbsp;Yonggang Zhang ,&nbsp;Binghai Wen","doi":"10.1016/j.compfluid.2026.106980","DOIUrl":"10.1016/j.compfluid.2026.106980","url":null,"abstract":"<div><div>The nucleation behavior of nanobubbles and nanodroplets is highly sensitive to how the liquid-gas surface tension depends on temperature and curvature, and accurately modeling this dependence is crucial for understanding and predicting micro/nano-scale phase transition processes. We establish a dimensional transformation and use a chemical-potential multiphase lattice Boltzmann method to systematically study the effects of temperature and curvature on surface tension and Tolman length for two typical fluids: water and methane. The Tolman length is used to quantify the deviation of interfacial tension from the flat interface limit. The simulation results show that both water and methane exhibit exponential changes in surface tension with temperature at a flat interface. An equation for predicting surface tension is then derived by considering the effects of temperature and curvature. Further analysis reveals that as curvature increases, the surface tension of nanobubbles increases while the Tolman length decreases, whereas nanodroplets exhibit the opposite trends.</div></div>","PeriodicalId":287,"journal":{"name":"Computers & Fluids","volume":"308 ","pages":"Article 106980"},"PeriodicalIF":3.0,"publicationDate":"2026-01-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146076121","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}