Searcharxiv⌕ Search

arXiv subjects

Gianluigi Rozza

Publications and source records attributed to Gianluigi Rozza.

At least 19 recordsLinked to original sources

Graph Neural Network based non-local models: Toward application to RANS turbulence modeling

Most machine-learning closures for the Reynolds-averaged Navier-Stokes (RANS) equations are local, predicting the closure quantities at each point solely from local flow features, despite the non-local nature of turbulent interactions. We propose a non-local closure framework based on Graph Convolutional Networks (GCNs) that exploits the mesh connectivity. Each finite-volume cell becomes a graph node, and cells sharing a face are connected by an edge. Through message passing, each layer aggregates information from neighboring nodes, so that after $k$ layers the prediction at a cell depends on its $k$-hop neighborhood, which naturally accounts for non-local effects. To make training feasible when each simulation yields a single large graph, we also introduce a breadth-first-search $k$-hop subgraph sampling strategy, which acts as data augmentation and cuts the cost per processed graph during training. We compare GCNs with feed-forward networks (FNNs) of equal size on two periodic-hills benchmarks: a laminar advection-diffusion transport problem that mimics the non-local character of RANS closures, and the turbulent eddy-viscosity prediction of the Closure Challenge dataset at $Re = 5600$. In the laminar case, FNNs are accurate only within the training regime and with sufficiently informative inputs, whereas GCNs generalize better to unseen geometries, including out-of-distribution flows, and partly compensate for missing input features. In the turbulent proof-of-concept case, the graph-based models achieve modest but consistent improvements compared to FNNs, supporting GCNs as a promising route to non-local RANS closures.

physics.flu-dyn↗

Parametric spatio-temporal reduced-order modelling via Spectral Proper Orthogonal Decomposition and Gaussian Process Regression

We present SPOD-GPR, a non-intrusive, parametric, spatio-temporal reduced-order model (ROM) that combines Proper Orthogonal Decomposition (POD), Spectral POD (SPOD), and Gaussian Process Regression (GPR). A parametric POD first projects the high-dimensional state onto a low-dimensional subspace. The reduced trajectories are then divided into windowed sections, transformed into the frequency domain, and truncated to the most energetic frequencies. SPOD is then performed independently at each retained frequency, yielding temporally coherent modes and coefficients. The parametric dependence of the POD and SPOD bases is modelled by GPR on a local chart of the Grassmann manifold after Procrustes alignment. The SPOD modal coefficients are learned both as explicit functions of the parameters and through a parameter-conditioned autoregressive model that allows for temporal extrapolation beyond the training window. Gaussian processes are sampled jointly and pathwise via Matheron conditioning, ensuring that each predictive realization carries uncertainty quantification consistent with the posterior distribution. The method is tested on four parametric examples where the parameter values and the prediction horizon lie outside the training data: the two-dimensional anisotropic Kuramoto--Sivashinsky equation in a time-periodic and a travelling-wave regime, each parametrised by one of its two coefficients; flow past a cylinder parametrised by the Reynolds number; and flow past an elliptic obstacle parametrised by its semi-axes. In all cases, SPOD-GPR reduces the normalised mean absolute error compared to a POD-GPR baseline, and improves the calibration of the predictive uncertainty, especially in the fluid flow cases, where POD-GPR is severely overconfident. This opens the way to spectral, uncertainty-aware reduced-order models for many-query and real-time applications, such as digital twins.

physics.flu-dyn↗

Meshless Latent Dynamics Models for Model Order Reduction of time-dependent Partial Differential Equations

Latent dynamics learning has emerged as a prominent approach for Model Order Reduction (MOR) of time-dependent, parameterized Partial Differential Equations (PDEs). Traditional autoencoder-based approaches are intrinsically tied to fixed spatial discretizations during both training and inference, motivating the development of meshless methodologies. In this context, conditioning provides a natural mechanism for representing temporal evolution through spatially continuous functions, enabling mesh-agnostic Latent Dynamics Models (LDMs). In this work, we introduce a unified mathematical framework for mesh-agnostic LDMs based on conditioning in Neural Fields (NeFs). Leveraging a taxonomy of conditioning mechanisms (weak, strong, and pure), we formulate meshless LDMs as NeFs conditioned on a latent state. We establish a Universal Approximation Theorem (UAT) for this class of models and analyze their stability by deriving Lipschitz bounds for different conditioning mechanisms. Our analysis reveals an expressivity--stability trade-off, providing insight into why more expressive conditioning, despite its higher representational capacity, can potentially degrade performance during temporal extrapolation. Finally, we provide a comprehensive comparison of meshless and mesh-based architectures in the continuous limit of the spatial discretization, showing that meshless models guarantee grid-independent Lipschitz bounds, unlike mesh-based counterparts. Our experiments across Burgers' equation, Navier--Stokes flow past a cylinder, and linear advection show that meshless LDMs can achieve predictive accuracy comparable to or better than their mesh-based counterparts while using only 2% of the trainable parameters and exhibiting nearly constant error across spatial discretizations.

math.NA↗

Reservoir computing based predictive reduced order model for steel grade intermixing in an industrial continuous casting tundish

Continuous casting is a widely adopted process in the steel industry, where maintaining high steel quality is paramount. Efficient prediction of grade intermixing during ladle changeover operations is critical for maintaining steel quality and minimizing material losses in the continuous casting process. Among various factors influencing grade intermixing, operating parameters play a significant role, in addition to tundish geometry and flow control devices. In this study, three-dimensional, transient, two-phase turbulent flow simulations are conducted to investigate the ladle changeover operation. During this process, the molten steel level in the tundish typically varies over time, significantly affecting the grade intermixing phenomena. The influence of ladle change time on intermixing time has been presented. However, high-fidelity full-order simulations of such complex transient phenomena are computationally expensive and are impractical for real-time monitoring or design-space exploration in industrial-scale applications. To address this issue, a reduced order modelling approach based on proper orthogonal decomposition (POD) and reservoir computing (RC) is employed to efficiently predict intermixing time. The proposed reduced order model (ROM) demonstrates excellent predictive accuracy using limited training data while requiring significantly less computational resources and training time. The results demonstrate the potential of the proposed methodology as a fast, reliable tool for real-time process monitoring and optimization in industrial continuous casting operations.

physics.flu-dyn↗

Data-driven reduced order model for residence time distribution analysis of an industrial-scale continuous casting tundish

The continuous casting tundish plays a critical role as a metallurgical reactor in the continuous casting process, with its flow characteristics serving as a key parameter in the production of high-quality steel. These characteristics are typically assessed through residence time distribution (RTD) curves. This study examines the flow behaviour in a single-strand continuous casting tundish through a combination of numerical simulations and experimental validation. Steady-state full order model (FOM) simulations are performed under both isothermal and non-isothermal conditions to evaluate the influence of thermal buoyancy on the velocity field, which is found to be negligible. The resulting flow fields are used to initialize transient tracer transport simulations for determining the RTD and flow volume partitioning. Subsequently, a data-driven reduced order model (ROM) is developed to predict the RTD response. Comparison of RTD curves obtained from experiments, FOM, and ROM shows excellent agreement, with the ROM accurately capturing the key flow characteristics at a fraction of the computational cost. These results highlight the potential of ROM techniques for efficient real-time analysis, design, and optimization of tundish operations in metallurgical processes.

physics.flu-dyn↗

Cluster Attention Neural Operators for Solving Parametric Partial Differential Equations

Traditional simulations of parametric partial differential equations (PDEs) rely on repetitive computations for each parameter, which makes high-fidelity design impractical. Neural operators address this issue by learning solution operators, accelerating parameter-space mapping by orders of magnitude. Recent Transformer-based neural operators attempt to capture global dependencies, but often at the cost of quadratic attention complexity. Transolver resolves this problem by projecting physical states into a reduced slice space for attention computation. Although fast, this projection sacrifices fine spatial information. Moreover, by operating in this reduced space with shared weights across attention heads, it may constrain the model's flexibility, thereby limiting its capacity to capture complex phenomena. To address these issues, we propose the Cluster Attention Neural Operator (CANO), which reformulates attention via a novel cross-attention mechanism that dynamically clusters queries while preserving full-resolution keys and values. This avoids slice compression loss and removes weight-sharing limits. At the same time, the model remains fast without losing global interactions. Empirically, CANO achieves state-of-the-art performance across canonical PDE benchmarks, covering fluid and solid dynamics (e.g., Navier-Stokes, Airfoil, Plasticity), irregular unstructured geometries (e.g., Pipe Turbulence, Composites), and long-term temporal rollouts. Across solid deformation and turbulent flow benchmarks, CANO achieves lower errors than baselines and exhibits strong geometric adaptability and temporal consistency.

math-ph↗

A Nonlinear Physics-based Reduced Order Model with Convolutional-based Operator Compression

Reduced-order models (ROMs) are widely used to accelerate the solution of parametrized partial differential equations, with projection-based methods relying on a linear representation of the governing equations in a low-dimensional space. Their efficiency, however, is fundamentally limited by the Kolmogorov N-width of the solution manifold: when the N-width decays slowly, accurate approximations require a large number of linear modes. This limitation has motivated nonlinear ROMs, which typically replace the linear representation of the solution field with nonlinear latent representations learned through autoencoder architectures. In this work, we propose a different perspective by introducing nonlinearity directly into the representation of the governing operators. Rather than compressing only the solution field, we learn nonlinear low-dimensional representations of the full-order operators while retaining an explicit reduced system of equations. Convolutional architectures are employed to exploit the locality and spatial structure of differential operators, and the resulting compressed operators are coupled with a nonlinear representation of the solution through a training strategy that accounts for both reconstruction accuracy and the error of the reduced-order solution. To address large sparse operators arising from fine and unstructured discretizations, continuous convolutions are used to operate directly on their non-zero entries. Finally, radial basis function interpolation enables the prediction of the compressed operators at unseen parameter configurations, providing an efficient online stage without assembling the corresponding full-order system. Numerical experiments on three parametrized problems demonstrate the potential of nonlinear operator compression as an alternative strategy for equation-based ROMs.

math.NA↗

A Comparative Study of Finite-Volume-based Coupled and Segregated Reduced-Order Models for Incompressible Flows in Parametrized Domains

This work presents a comparative analysis of Reduced-Order Models (ROMs) applied to incompressible fluid dynamics within geometrically parametrized domains. Two distinct reduced-order solution strategies are investigated and compared: a monolithic coupled solver and a segregated SIMPLE-based algorithm. Their performance is assessed on two steady, two-dimensional benchmark cases: a lid-driven cavity flow and a flow past a cylindrical obstacle. The two algorithms are compared in terms of fields evaluation and aerodynamic coefficients prediction. The computational results highlight a fundamental trade-off between accuracy and numerical efficiency. On the one hand, after an opportune supremizers enrichment, the coupled approach guarantees a faster convergence, despite the need of a larger number of degrees of freedom. On the other hand, the segregated SIMPLE algorithm yields superior reconstruction accuracy, particularly for lower-dimensional reduced spaces, at the cost of a slower convergence rate.

physics.flu-dyn↗

A Dynamical Approximation Scheme on the Stiefel manifold for Wasserstein Gradient Flows

We propose a meshless Lagrangian dynamical method for approximating Wasserstein gradient flows (WGFs). The evolving measure is represented as the pushforward of the initial measure $μ_0$ through a transport map in the weighted Hilbert space $L^2_{μ_0}$. We approximate this map in time-dependent linear subspaces of $L^2_{μ_0}$, whose orthonormal frames are evolved by a Dirac--Frenkel dynamical principle on a Stiefel manifold constrained to a finite-dimensional background space, adaptively constructed via local approximations of the WGF velocity field. We prove that the resulting transport map induces an absolutely continuous curve of probability measures in Wasserstein space, whose velocity is obtained by projecting the exact WGF velocity onto the background space, and we show that the approximation preserves the energy dissipation structure up to the projection error of the velocity. Moreover, for geodesically convex energies, we derive an a posteriori estimate controlling such projection error through the adaptive construction of the background space, yielding as well a bound on the approximation error of the pushforward measure in the Wasserstein metric. Numerical experiments on linear and nonlinear Fokker--Planck equations, porous-medium diffusion, and interaction energies demonstrate the accuracy of the method, its energy-dissipation properties, and the advantages of the adaptive construction.

math.NA↗

Trainable Spline Representations for Physics-Informed Learning

This work introduces Physics-Informed Splines (PI-Splines), a structured spline-based architecture for physics-informed learning. Instead of representing the solution of a differential equation with a neural network, PI-Splines directly parametrize the unknown field through a tensor-product B-spline expansion with trainable control coefficients. This formulation preserves the residual-based training paradigm of Physics-Informed Neural Networks while providing compact support, explicit smoothness control, analytical derivatives, and a direct geometric interpretation of the trainable parameters. When compatible with the spline representation, boundary conditions can be imposed strongly by fixing suitable boundary control coefficients. The proposed method is evaluated on several benchmark problems of increasing difficulty and compared with standard physics-informed frameworks under matched governing equations, collocation sets, loss terms, and optimization procedures, so as to isolate the effect of the approximation architecture. Numerical experiments show that PI-Splines provide a competitive and stable alternative to neural physics-informed architectures, particularly in settings where structured representations, locality, and parameter efficiency are desirable.

cs.LG↗

A Generative Model-Free Form Deformation Approach for the Generation of Mesh Motions with Applications to PDE

We introduce a topology-agnostic framework for matching deformations of three-dimensional shapes with non-isomorphic mesh graphs by modelling the deformation as the flow of an Ordinary Differential Equation (ODE). The velocity field is parameterised by a time-dependent Free Form Deformation (FFD), expressed through displacements of a coarse control lattice, yielding a smooth and low-dimensional representation that decouples the deformation model from the discretisation of the source and target surfaces. Under mild regularity assumptions, we prove that the induced ODE map is a universal approximator (in the sup norm) for mappings between genus-0 surfaces, providing a theoretical expressivity guarantee. To further compress the representation and enable probabilistic inference, we couple the ODE--FFD model with a flow-based generative approach in the TarFlow framework, learning a compact latent parametrisation over time series of FFD maps. The resulting method supports efficient sampling and optimisation of plausible deformation trajectories while preserving mesh quality, and it enables scalable reduced-order modelling. Experiments on deforming-body flow benchmarks demonstrate improved accuracy and computational efficiency of reduced-order models constructed from the learned latent dynamics.

math.NA↗

Bifurcation curve detection with deflation for multiparametric PDEs

This work presents a comprehensive framework for capturing bifurcating phenomena and detecting bifurcation curves in nonlinear multiparametric partial differential equations, where the system exhibits multiple coexisting solutions for given values of the parameters. Traditional continuation methods for one-dimensional parameterizations employ the previously computed solution as the initial guess for the next parameter value. These are usually very inefficient, since small step sizes increase computational cost, while larger steps could jeopardize the method convergence jumping to a different solution branch or missing the bifurcation point. To address these challenges, we propose a novel framework that combines: (i) arclength continuation, adaptively selecting new parameter values in higher dimension, and (ii) the deflation technique, discovering multiple branches to construct complete bifurcation diagrams without requiring a costly spectral analysis of the system. In particular, the arclength continuation method is designed to handle multiparametric scenarios, where the parameter vector $λ\in \mathbb{R}^p$ traces a curve $g(λ)$ within a $p$-dimensional parameter space. In addition, we introduce a zigzag path-following strategy to robustly track the bifurcation curves and surfaces, respectively, for two- and three-dimensional parametric spaces. Finally, we demonstrate its performance on three benchmark problems of increasing complexity: from the 1D/2D Bratu and Allen--Cahn equations to the 2D/3D Rayleigh--Benard convection problem.

math.NA↗

An isotropic recovery-based error estimator algorithm for mesh adaptation in a finite volume environment with application to atmospheric flows

We develop an Isotropic Recovery-based Error Estimator (IREE) to drive mesh adaptation within a finite volume framework. Recovery-based error estimators are widely used in practice thanks to their simplicity, relying solely on the available discrete solution and a suitable post-processing step. While recovery-based estimators are well established in the finite element framework, their application to finite volume methods, despite the widespread use of the latter in industrial and commercial codes, remains largely unexplored and motivates the present study. We assess the performance of the proposed IREE-driven mesh adaptation procedure through well-known benchmarks for dry atmospheric flows modeling and compare it against a widely-used plain mesh adaptation algorithm. Both qualitative and quantitative results show that the IREE algorithm is more effective at suppressing numerical instabilities compared with the plain approach and, for most of the simulation time, provides a superior accuracy, though at a moderately higher computational cost. In addition, both the adaptive approaches offer substantial computational savings (between 35% and 94%) relative to simulations based on fixed uniformly fine meshes. These findings demonstrate that IREE-based mesh adaptation is a promising and effective strategy for atmospheric flow simulations and has the potential to substantially reduce the computational effort of generating reanalysis-quality data.

math.NA↗

A Quadratic Order Reduction -- Gaussian Process Ordinary Differential Equation framework for the inference of Large Continuous Dynamical Systems

Forecasting the evolution of complex dynamical systems remains a fundamentally challenging task, primarily due to pronounced nonlinear interactions, high-dimensional state spaces, and the concomitant requirement for rigorous and reliable uncertainty quantification. Contemporary reduced-order modelling (ROM) frameworks frequently exhibit inherent trade-offs among predictive accuracy, numerical stability, and interpretability, and thus often fail to achieve an optimal balance among these competing objectives. To address these limitations, we propose a framework for forecasting complex dynamical systems via a kernel autonomous ordinary differential equation approach based on Gaussian Processes and Quadratic Order Model Reduction. Our base method, the Gaussian Process Ordinary Differential Equations model, allows accurate short-term forecasting with uncertainty quantification, and it provably converges to the real autonomous equation in the smooth case. We integrate it with quadratic order reduced-order modelling and sphere projection for learning the latent dynamics efficiently while preserving stability. Numerical experiments demonstrate that our full model outperforms ROM forecasting methods such as Extended Dynamic Mode Decomposition, Bagging Optimised Dynamic Mode Decomposition and Linear and Nonlinear Disambiguation Optimisation in terms of accuracy or computational costs. These results demonstrate the potential of the framework as a robust and stable tool for forecasting complex dynamical systems with rigorous uncertainty quantification.

math.NA↗

A Structured Review of Reduced Order Modeling for Domain Decomposition Problems: State of the Art and Perspectives

Reduced Order Models (ROMs) have been regarded as an efficient alternative to conventional high-fidelity Computational Fluid Dynamics (CFD) for accelerating the design and optimization processes in engineering applications. Many industrial geometries feature repeating subdomains or contain sub-regions governed by distinct physical phenomena, making them well-suited to Domain Decomposition (DD) techniques. The integration of ROM and DD is promising to further reduce computational costs by constructing local ROMs and assembling them into global solutions. Due to the complexity and necessity of coupling ROMs, many approaches have been proposed in recent years. This review provides a concise overview of existing methodologies combining ROM and DD. We categorize existing methods into intrusive (projection-based) and non-intrusive (data-driven) frameworks. Various strategies for generating local reduced bases and coupling them across subdomains are illustrated. Particular emphasis is placed on intrusive techniques, including equations, numerical algorithms, and practical implementations. The non-intrusive framework is also discussed, highlighting its general procedures, basic formulations, and underlying principles. Finally, we summarise the state of the literature, identify open challenges, and present perspectives on future implementation from an engineering viewpoint.

math.NA↗

Constraint-driven Optimization and Parametrization of Industrial NURBS Geometries via Neural Deformation Field

This work presents a differentiable framework for the parametrization and shape optimization of industrial CAD geometries represented by multi-patch NURBS surfaces. The method enables the deformation of complex CAD models through a physics-informed geometric parametrization, allowing direct morphing driven by physical constraints without the need to prescribe a predefined deformation strategy. A neural displacement field, implemented as a multi-layer perceptron acting on the NURBS control points, provides a compact parametrization of the admissible design space while preserving patch connectivity. Global geometric quantities relevant to hydrostatic design, including displaced volume, wetted surface area and buoyancy centroid, are formulated as differentiable integral operators evaluated on the parametric domain. These quantities are computed through Gauss-Legendre quadrature combined with analytical B-spline derivatives for surface metric evaluation, allowing gradient propagation to the deformation parameters while limiting the computational overhead of automatic differentiation. The proposed framework operates directly on CAD representations without intermediate mesh generation. Numerical experiments on a modified KVLCC2 hull demonstrate the capability of the method to satisfy competing hydrostatic constraints while producing smooth CAD-compatible geometries and showing stable convergence across multiple random initializations.

math.NA↗

Surrogate normal-forms for the numerical bifurcation and stability analysis of navier-stokes flows via machine learning

Inspired by the Equation-Free paradigm, we propose an ``embed-learn-lift'' framework for constructing minimal-dimensional surrogate ROMs for the numerical analysis of high-fidelity Navier-Stokes simulations, even in the presence of symmetries that standard machine-learning surrogates often fail to preserve. The framework consists of four main stages. First, manifold learning (here both POD and Diffusion Maps) is used to uncover the intrinsic geometry and dimensionality of the latent space underlying the high-dimensional spatio-temporal Navier-Stokes dynamics across the parameter space. Second, we construct ROMs (here, via Gaussian Process regression (GPR)) of minimal dimension -- by learning the evolution equations directly on the identified latent space. Third, we exploit the toolkit of numerical bifurcation analysis to construct bifurcation diagrams and perform systematic stability analysis directly in the latent coordinates. This enables, for example, the efficient continuation of branches of limit cycles emerging from Andronov-Hopf and Neimark-Sacker bifurcations, together with the computation of limit-cycles periods and stability properties via Floquet multipliers. Such analysis is effectively intractable for the full Navier-Stokes equations. Finally, by solving the pre-image problem in manifold learning, we reconstruct the bifurcating steady and time-periodic states in the original high-dimensional physical space, thus closing the ``lift'' step of the pipeline. We show that DMs-based ROMs allow for a computationally efficient and accurate numerical bifurcation and stability analysis, thus outperforming the widely used POD-ROMs by providing a geometrically consistent parametrization and correctly identifying the intrinsic dimension even in the presence of secondary instabilities, highlighting the need for nonlinear manifold learning methods in CFD.

physics.flu-dyn↗

End-to-end optimization of subgrid scale models for discontinuous spectral element schemes based on the discrete adjoint method

In computational fluid dynamics, Large Eddy Simulation (LES) offers a compelling balance between accuracy and computational cost by resolving large-scale flow structures while modeling unresolved subgrid scales. However, its predictive capacity is critically dependent on the choice and calibration of subgrid-scale (SGS) models, which often involve problem-dependent parameters and exhibit intricate interactions with the numerical discretization. In this work, we propose a discrete-adjoint framework to optimize SGS model parameters in the loop, leveraging automatic differentiation within a high-order Spectral Difference (SD) solver. Coarse-grained simulations of Forced Homogeneous Isotropic Turbulence (FHIT), together with filtered Direct Numerical Simulation (DNS) data, are used to optimize a limited set of parameters for classical SGS models, including the Smagorinsky model and non-linear tensor-basis formulations. For chaotic systems such as LES, the choice of objective function plays a crucial role in the stability and accuracy of the optimization. Here, we consider the spatio-temporally averaged decay of the Legendre modal coefficients as the quantity of interest for the SD scheme. The optimization is performed across different grid resolutions and polynomial orders, highlighting the impact of numerical discretization on model performance. The methodology is applied to both one-dimensional Burgers turbulence and fully three-dimensional turbulence. The trained models are subsequently assessed on out-of-sample configurations, including Decaying Homogeneous Isotropic Turbulence (DHIT) and the Taylor-Green vortex. Variations in polynomial order, grid resolution, and Reynolds number are considered to evaluate robustness and generalization. In all test cases, the optimized models demonstrate significant improvements over baseline SGS closures.

physics.flu-dyn↗