Search arXivSearch

arXiv subjects

Stephan Simonis

Publications and source records attributed to Stephan Simonis.

At least 19 recordsLinked to original sources

Computing statistical Euler limits of the Navier--Stokes equations in three dimensions

We develop a Monte Carlo lattice Boltzmann method to compute statistical solutions to the three-dimensional incompressible Navier--Stokes and Euler equations. Entropic space-time adaptive relaxation of the higher order kinetic moments yields stable numerical solutions with decreasing viscosity. We provide a convergence analysis that is conditional on four explicitly stated assumptions regarding the discrete dynamics. Under diffusive scaling, the laws of the discrete ensemble converge along a subsequence to a limit satisfying the Foias--Temam Liouville formulation of the Navier--Stokes equations. Consequently, provided the structure function scaling holds uniformly, the vanishing viscosity limit of these measures satisfies the multi-point statistical Euler hierarchy of Fjordholm, Mishra, and Weber. The limit measures inherit the known weak-strong uniqueness principle on the interval of existence of a strong Euler solution. Under explicit scaling assumptions, a Kuznetsov-type argument yields a fractional 1-Wasserstein convergence rate. We present three-dimensional computations of time-dependent statistical solutions along the inviscid limit of the incompressible Navier--Stokes equations together with convergence measurements in the Wasserstein metric. Numerical experiments on a randomized Taylor--Green vortex with 24-dimensional initial uncertainty recover Kolmogorov's K41 scaling for energy spectra and structure functions, exhibit the failure of pathwise strong convergence, and yield Wasserstein convergence rates of about $0.5$ at the onset of turbulence. Finally, error measurements with respect to spectral hyperviscosity computations indicate that the computed limit measure is independent of the numerical regularization.

math.NA

Computing statistical solutions of a Mach 2000 astrophysical jet

The multi-dimensional compressible Euler equations admit non-unique entropy solutions in turbulent regimes, and extreme-Mach astrophysical flows are a natural setting in which this breakdown of deterministic well-posedness becomes computationally visible. We compute statistical solutions of a Mach~2000 astrophysical jet, defined as the pushforward of an initial probability measure through a vectorial lattice Boltzmann method, by Monte Carlo sampling with $M=1000$ realizations on grids of up to $12.8$ million cells. Under mesh refinement the individual realizations diverge pathwise, while the statistical solution converges: Wasserstein distances of the one- and two-point marginals, the ensemble mean, and the ensemble standard deviation all exhibit stable positive convergence rates. A spatially resolved analysis along the jet axis traces this dichotomy to the structure of the one-point laws, which are numerically Dirac in the undisturbed core, skewed in the sheared turbulent regions, and intermittent two-state mixtures at the random leading front. We conclude that the computed statistical solution is non-Dirac and remains stable in the extreme compressible regime, in which no strong solution is expected to exist.

math.NA

Efficient Wall-Modelled Large Eddy Simulation of Rotors using Homogenized Lattice Boltzmann Methods

Accurately capturing the dynamic forces acting on rotors as well as their wake effects presents a significant challenge for computational fluid dynamics (CFD) due to high Reynolds numbers and a large range of spatio-temporal scales. The present work proposes a novel blade-resolved wall-modeled large eddy simulation (WMLES) approach based on the lattice Boltzmann method (LBM). A homogenized hybrid regularized recursive collision scheme targeting the filtered Brinkman--Navier--Stokes equations is combined with a novel wall-model. This is implemented in the context of a platform-transparent framework for fluid-structure interaction in the open source LBM framework OpenLB. Convergence order and accuracy are validated against both experimental and numerical data for a model wind turbine, demonstrating excellent agreement for integral forces and wake velocity profiles. Computational efficiency and parallel scalability was investigated by roofline analysis and weak scaling studies for up to 384 rotors resolved by 41 billion lattice cells on the Karolina supercomputer. The proposed framework enables efficient blade-resolved WMLES of entire wind farms and offers a new methodology for other complex wall-modeled fluid-structure interaction applications.

physics.flu-dyn

Uncertain data assimilation for urban wind flow simulations with OpenLB-UQ

Accurate prediction of urban wind flow is essential for urban planning, pedestrian safety, and environmental management. Yet, it remains challenging due to uncertain boundary conditions and the high cost of conventional CFD simulations. This paper presents the use of the modular and efficient uncertainty quantification (UQ) framework OpenLB-UQ for urban wind flow simulations. We specifically use the lattice Boltzmann method (LBM) coupled with a stochastic collocation (SC) approach based on generalized polynomial chaos (gPC). The framework introduces a relative-error noise model for inflow wind speeds based on real measurements. The model is propagated through a non-intrusive SC LBM pipeline using sparse-grid quadrature. Key quantities of interest, including mean flow fields, standard deviations, and vertical profiles with confidence intervals, are efficiently computed without altering the underlying deterministic solver. We demonstrate this on a real urban scenario, highlighting how uncertainty localizes in complex flow regions such as wakes and shear layers. The results show that the SC LBM approach provides accurate, uncertainty-aware predictions with significant computational efficiency, making OpenLB-UQ a practical tool for real-time urban wind analysis.

physics.flu-dyn

OpenLB-UQ: An Uncertainty Quantification Framework for Incompressible Fluid Flow Simulations

Uncertainty quantification (UQ) is crucial in computational fluid dynamics to assess the reliability and robustness of simulations, given the uncertainties in input parameters. OpenLB is an open-source lattice Boltzmann method library designed for efficient and extensible simulations of complex fluid dynamics on high-performance computers. In this work, we leverage the efficiency of OpenLB for large-scale flow sampling with a dedicated and integrated UQ module. To this end, we focus on non-intrusive stochastic collocation methods based on generalized polynomial chaos and Monte Carlo sampling. The OpenLB-UQ framework is extensively validated in convergence tests with respect to statistical metrics and sample efficiency using selected benchmark cases, including two-dimensional Taylor--Green vortex flows with up to four-dimensional uncertainty and a flow past a cylinder. Our results confirm the expected convergence rates and show promising scalability, demonstrating robust statistical accuracy as well as computational efficiency. OpenLB-UQ enhances the capability of the OpenLB library, offering researchers a scalable framework for UQ in incompressible fluid flow simulations and beyond.

physics.flu-dyn

Analytical solution for dynamic evaporation of liquid in isothermal condition

An analytical solution based on a diffuse interface model is presented for an isothermal evaporation problem under sub-saturation pressure. The macroscopic equations are derived from the free-energy method, widely recognized in the lattice Boltzmann literature, distinguishing our approach from conventional evaporation models that rely on jump conditions or pure kinetic theory. The interface behavior is fully described by differential equations, eliminating the need for assumptions such as local equilibrium at the interface. We derive an exact analytical solution for the inviscid case and propose an approximate solution when viscosity effects are considered. Our model unveils a novel relationship between evaporation rate and viscosity, providing new insights that have not been thoroughly explored in the literature. The analytical results are validated through numerical simulations using the open-source parallel library OpenLB, demonstrating excellent agreement in predicting the physical behavior of the evaporation phenomena within the framework of diffuse interface methods.

physics.flu-dyn

On the consistency of pseudo-potential lattice Boltzmann methods

We derive the partial differential equation (PDE) to which the pseudo-potential lattice Boltzmann method (P-LBM) converges under diffusive scaling, providing a rigorous basis for its consistency analysis. By establishing a direct link between the method's parameters and physical properties-such as phase densities, interface thickness, and surface tension-we develop a framework that enables users to specify fluid properties directly in SI units, eliminating the need for empirical parameter tuning. This allows the simulation of problems with predefined physical properties, ensuring a direct and physically meaningful parametrization. The proposed approach is implemented in OpenLB, featuring a dedicated unit converter for multiphase problems. To validate the method, we perform benchmark tests-including planar interface, static droplet, Galilean invariance, and two-phase flow between parallel plates-using R134a as the working fluid, with all properties specified in physical units. The results demonstrate that the method achieves second-order convergence to the identified PDE, confirming its numerical consistency. These findings highlight the robustness and practicality of the P-LBM, paving the way for accurate and user-friendly simulations of complex multiphase systems with well-defined physical properties.

physics.flu-dyn

On the range of validity of parabolic models for fluid flow through isotropic homogeneous porous media

Lattice-Boltzmann methods are established mesoscopic numerical schemes for fluid flow, that recover the evolution of macroscopic quantities (viz., velocity and pressure fields) evolving under macroscopic target equations. The approximated target equations for fluid flows are typically parabolic and include a (weak) compressibility term. A number of Lattice-Boltzmann models targeting, or making use of, flow through porous media in the representative elementary volume, have been successfully developed. However, apart from two exceptions, the target equations are not reported, or the assumptions for and approximations of these equations are not fully clarified. Within this work, the underlying assumption underpinning parabolic equations for porous flow in the representative elementary volume, are discussed, clarified and listed. It is shown that the commonly-adopted assumption of negligible hydraulic dispersion is not justifiable by clear argument - and in fact, that by not adopting it, one can provide a qualitative and quantitative expression for the effective viscosity in the Brinkman correction of Darcy law. Finally, it is shown that, under certain conditions, it is possible to interpret porous models as Euler-Euler multiphase models wherein one phase is the solid matrix.

physics.flu-dyn

Predicting Filter Medium Performances in Chamber Filter Presses with Digital Twins Using Neural Network Technologies

Efficient solid-liquid separation is crucial in industries like mining, but traditional chamber filter presses depend heavily on manual monitoring, leading to inefficiencies, downtime, and resource wastage. This paper introduces a machine learning-powered digital twin framework to improve operational flexibility and predictive control. A key challenge addressed is the degradation of the filter medium due to repeated cycles and clogging, which reduces filtration efficiency. To solve this, a neural network-based predictive model was developed to forecast operational parameters, such as pressure and flow rates, under various conditions. This predictive capability allows for optimized filtration cycles, reduced downtime, and improved process efficiency. Additionally, the model predicts the filter mediums lifespan, aiding in maintenance planning and resource sustainability. The digital twin framework enables seamless data exchange between filter press sensors and the predictive model, ensuring continuous updates to the training data and enhancing accuracy over time. Two neural network architectures, feedforward and recurrent, were evaluated. The recurrent neural network outperformed the feedforward model, demonstrating superior generalization. It achieved a relative $L^2$-norm error of $5\%$ for pressure and $9.3\%$ for flow rate prediction on partially known data. For completely unknown data, the relative errors were $18.4\%$ and $15.4\%$, respectively. Qualitative analysis showed strong alignment between predicted and measured data, with deviations within a confidence band of $8.2\%$ for pressure and $4.8\%$ for flow rate predictions. This work contributes an accurate predictive model, a new approach to predicting filter medium cycle impacts, and a real-time interface for model updates, ensuring adaptability to changing operational conditions.

cs.LG

A Digital Urban Twin Enabling Interactive Pollution Predictions and Enhanced Planning

Digital twin (DT) technology is increasingly used in urban planning, leveraging real-time data integration for environmental monitoring. This paper presents an urban-focused DT that combines computational fluid dynamics simulations with live meteorological data to analyze pollution dispersion. Addressing the health impacts of pollutants like particulate matter and nitrogen dioxide, the DT provides real-time updates on air quality, wind speed, and direction. Using OpenStreetMaps XML-based data, the model distinguishes between porous elements like trees and solid structures, enhancing urban flow analysis. The framework employs the lattice Boltzmann method (LBM) within the open-source software OpenLB to simulate pollution transport. Nitrogen dioxide and particulate matter concentrations are estimated based on traffic and building emissions, enabling hot-spot identification. The DT was used from November 7 to 23, 2024, with hourly updates, capturing pollution trends influenced by wind patterns. Results show that alternating east-west winds during this period create a dynamic pollution distribution, identifying critical residential exposure areas. This work contributes a novel DT framework that integrates real-time meteorological data, OpenStreetMap-based geometry, and high-fidelity LBM simulations for urban wind and pollution modeling. Unlike existing DTs, which focus on structural monitoring or large-scale environmental factors, this approach enables fine-grained, dynamic analyses of urban airflow and pollution dispersion. By allowing interactive modifications to urban geometry and continuous data updates, the DT serves as a powerful tool for adaptive urban planning, supporting evidence-based policy making to improve air quality and public health.

physics.soc-ph

Nonuniqueness of lattice Boltzmann schemes derived from finite difference methods

Recently, the construction of finite difference schemes from lattice Boltzmann schemes has been rigorously analyzed [Bellotti et al. (2022), Numer. Math. 152, pp. 1-40]. It is thus known that any lattice Boltzmann scheme can be expressed in terms of a corresponding multi-step finite difference scheme on the conserved variables. In the present work, we provide counterexamples for the conjecture that any multi-step finite difference scheme has a unique lattice Boltzmann formulation. Based on that, we indicate the existence of equivalence classes for discretized relaxation systems.

math.NA

Generative AI for fast and accurate statistical computation of fluids

We present a generative AI algorithm for addressing the pressing task of fast, accurate, and robust statistical computation of three-dimensional turbulent fluid flows. Our algorithm, termed as GenCFD, is based on an end-to-end conditional score-based diffusion model. Through extensive numerical experimentation with a set of challenging fluid flows, we demonstrate that GenCFD provides an accurate approximation of relevant statistical quantities of interest while also efficiently generating high-quality realistic samples of turbulent fluid flows and ensuring excellent spectral resolution. In contrast, ensembles of deterministic ML algorithms, trained to minimize mean square errors, regress to the mean flow. We present rigorous theoretical results uncovering the surprising mechanisms through which diffusion models accurately generate fluid flows. These mechanisms are illustrated with solvable toy models that exhibit the mathematically relevant features of turbulent fluid flows while being amenable to explicit analytical formulae. Our codes are publicly available at https://github.com/camlab-ethz/GenCFD.

cs.LG

Homogenized lattice Boltzmann methods for fluid flow through porous media -- part I: kinetic model derivation

In this series of studies, we establish homogenized lattice Boltzmann methods (HLBM) for simulating fluid flow through porous media. Our contributions in part I are twofold. First, we assemble the targeted partial differential equation system by formally unifying the governing equations for nonstationary fluid flow in porous media. A matrix of regularly arranged, equally sized obstacles is placed into the domain to model fluid flow through porous structures governed by the incompressible nonstationary Navier--Stokes equations (NSE). Depending on the ratio of geometric parameters in the matrix arrangement, several homogenized equations are obtained. We review existing methods for homogenizing the nonstationary NSE for specific porosities and discuss the applicability of the resulting model equations. Consequently, the homogenized NSE are expressed as targeted partial differential equations that jointly incorporate the derived aspects. Second, we propose a kinetic model, the homogenized Bhatnagar--Gross--Krook Boltzmann equation, which approximates the homogenized nonstationary NSE. We formally prove that the zeroth and first order moments of the kinetic model provide solutions to the mass and momentum balance variables of the macrocopic model up to specific orders in the scaling parameter. Based on the present contributions, in the sequel (part II), the homogenized NSE are consistently approximated by deriving a limit consistent HLBM discretization of the homogenized Bhatnagar--Gross--Krook Boltzmann equation.

math.NA

OpenLB User Guide: Associated with Release 1.6 of the Code

OpenLB is an object-oriented implementation of LBM. It is the first implementation of a generic platform for LBM programming, which is shared with the open source community (GPLv2). Since the first release in 2007, the code has been continuously improved and extended which is documented by thirteen releases as well as the corresponding release notes which are available on the OpenLB website (https://www.openlb.net). The OpenLB code is written in C++ and is used by application programmers as well as developers, with the ability to implement custom models OpenLB supports complex data structures that allow simulations in complex geometries and parallel execution using MPI, OpenMP and CUDA on high-performance computers. The source code uses the concepts of interfaces and templates, so that efficient, direct and intuitive implementations of the LBM become possible. The efficiency and scalability has been checked and proved by code reviews. This user manual and a source code documentation by DoxyGen are available on the OpenLB project website.

cs.MS

Binary mixture flow with free energy lattice Boltzmann methods

We use free energy lattice Boltzmann methods (FRE LBM) to simulate shear and extensional flow of a binary mixture in two and three dimensions. To this end, two classical configurations are digitally twinned, namely a parallel-band device for binary shear flow and a four-roller apparatus for binary extensional flow. The FRE LBM and the test cases are implemented in the open-source C++ framework OpenLB and evaluated for several non-dimensional numbers. Characteristic deformations are captured, where breakup mechanisms occur for critical capillary regimes. Though the known mass leakage for small droplet-domain ratios is observed, suitable mesh sizes show good agreement to analytical predictions and reference results.

math.NA

Constructing relaxation systems for lattice Boltzmann methods

We present the first top-down ansatz for constructing lattice Boltzmann methods (LBM) in d dimensions. In particular, we construct a relaxation system (RS) for a given scalar, linear, d-dimensional advection-diffusion equation. Subsequently, the RS is linked to a d-dimensional discrete velocity Boltzmann model (DVBM) on the zeroth and first energy shell. Algebraic characterizations of the equilibrium, the moment space, and the collision operator are carried out. Further, a closed equation form of the RS expresses the added relaxation terms as prefactored higher order derivatives of the conserved quantity. Here, a generalized (2d+1)x(2d+1) RS is linked to a DdQ(2d+1) DVBM which, upon complete discretization, yields an LBM with second order accuracy in space and time. A rigorous convergence result for arbitrary scaling of the RS, the DVBM and conclusively also for the final LBM is proven. The top-down constructed LBM is numerically tested on multiple GPUs with smooth and non-smooth initial data in d=3 dimensions for several grid-normalized non-dimensional numbers.

math.NA

Consistent lattice Boltzmann methods for the volume averaged Navier-Stokes equations

We derive a novel lattice Boltzmann scheme, which uses a pressure correction forcing term for approximating the volume averaged Navier-Stokes equations (VANSE) in up to three dimensions. With a new definition of the zeroth moment of the Lattice Boltzmann equation, spatially and temporally varying local volume fractions are taken into account. A Chapman-Enskog analysis, respecting the variations in local volume, formally proves the consistency towards the VANSE limit up to higher order terms. The numerical validation of the scheme via steady state and non-stationary examples approves the second order convergence with respect to velocity and pressure. The here proposed lattice Boltzmann method is the first to correctly recover the pressure with second order for space-time varying volume fractions.

math.NA

Limit Consistency of Lattice Boltzmann Equations

We establish the notion of limit consistency as a modular part in proving the consistency of lattice Boltzmann equations (LBEs) with respect to a given partial differential equation (PDE) system. The incompressible Navier--Stokes equations (NSE) are used as paragon. Based upon the hydrodynamic limit of the Bhatnagar--Gross--Krook (BGK) Boltzmann equation towards the NSE, we provide a successive discretization by nesting conventional Taylor expansions and finite differences. We track the discretization state of the domain for the particle distribution functions and measure truncation errors at all levels within the derivation procedure. Via parametrizing equations and proving the limit consistency of the respective families of equations, we retain the path towards the targeted PDE at each step of discretization, i.e. for the discrete velocity BGK Boltzmann equations and the space-time discretized LBEs. As a direct result, we unfold the discretization technique of lattice Boltzmann methods as chaining finite differences and provide a generic top-down derivation of the numerical scheme which upholds the continuous limit.

math.NA