arXiv subjects

Ernesto Castillo

Publications and source records attributed to Ernesto Castillo.

10 recordsLinked to original sources

arXiv2026

Stress-divergence, Laplacian, and rotational forms of the incompressible Navier--Stokes equations with variable viscosity

In the Navier--Stokes equations, incompressibility allows rewriting the viscous term in various forms leading to distinct numerical properties and flow descriptions. Furthermore, models accounting for non-Newtonian, thermal or turbulent effects often break the constant-viscosity assumption, thereby producing additional consistency terms. In this context, the present work compares the classical symmetric-gradient diffusion term with more recent variable-viscosity generalizations of the Laplacian and rotational forms. We discuss, analyze and test their differences with respect to implementation, efficiency, numerical stability and outflow boundary conditions. With a focus on time-dependent flows, we consider second-order implicit-explicit (IMEX) temporal discretizations aimed at improving efficiency and numerical stability. Through a rigorous stability analysis, we show how selected explicit treatments can bypass algorithmic nonlinearities without inducing CFL conditions. Our numerical results highlight important differences between the three viscous formulations---especially in the presence of outflow boundaries, for which the generalized Laplacian form proves more suitable in diffusion-dominated regimes. %(as widely known for constant viscosity).

math.NA↗

arXiv2026

Dynamic Reduced-Order Data Assimilation from Sparse Velocity Measurements

We present a novel reduced-order data assimilation framework, termed Reduced-Order Dynamical Assimilation (RODAS), for reconstructing high-resolution, time-resolved flow fields from sparse velocity measurements. The method combines low-dimensional experimental observations with a physics-based parametric reduced-order model, enabling both spatial extrapolation beyond the measurement region and temporal super-resolution. The approach first identifies the dominant dynamics from sparse measurements through Dynamic Mode Decomposition (DMD), and subsequently reconstructs the corresponding full-order flow evolution by projecting the identified dynamics onto a parametric Proper Orthogonal Decomposition (POD) manifold generated from high-fidelity numerical simulations. Unlike conventional reduced-order data assimilation methods that estimate independent snapshots or treat time as an additional parameter, RODAS reconstructs an entire dynamical trajectory in a single inference step while naturally incorporating parametric variability. We assess the proposed methodology on vortex shedding behind a circular cylinder for both Newtonian and non-Newtonian (Carreau-Yasuda) fluids. Numerical experiments demonstrate accurate reconstruction of high-resolution velocity fields from localized, low-resolution measurements, achieving sub-percent reconstruction errors with sufficiently rich reduced bases, robust performance under severe temporal undersampling, and accurate prediction of engineering quantities of interest such as the drag coefficient. These results demonstrate that RODAS provides an efficient framework for real-time, physics-informed reconstruction of unsteady flows from sparse experimental data.

physics.flu-dyn↗

arXiv2026

A Stress-Based Estimator for Pressure and Stress Recovery from Velocity Measurements

Non-invasive pressure field estimation from velocity measurements is a longstanding engineering problem. We propose, analyze, and test a pressure-recovery method that computes a full stress field from velocity measurements, and leaves the pressure estimation as a cheap post-processing step. The method relies on a stress-velocity first order formulation of the Navier-Stokes equations, and we show that the formulation accounts for deviations from incompressibility in the measured velocity field by construction. In addition, we theoretically establish the convergence of the finite element (FE) approximation scheme, the stability of the stress recovery with respect to finite-resolution velocity measurements, and then validate this theory numerically. Our results show that the proposed estimator is robust in convective flow regimes and remains accurate at reduced spatial resolution, improving upon state-of-the-art pressure-recovery strategies.

math.NA↗

arXiv2026

A term-by-term variational multiscale method with dynamic subscales for incompressible turbulent aerodynamics

Variational multiscale (VMS) methods offer a robust framework for handling under-resolved flow scales without resorting to problem-specific turbulence models. Here, we propose and assess a dynamic, term-by-term VMS stabilized formulation for simulating incompressible flows from laminar to turbulent regimes. The method is embedded in an incremental pressure-correction fractional-step framework and employs a minimal set of stabilization terms, yielding a unified discretization that (i) allows equal-order velocity--pressure interpolation and (ii) provides robust control of convection-dominated dynamics in complex three-dimensional settings. Orthogonal projections are a key ingredient and ensure that the non-residual, term-by-term structure induces dissipation through dynamic subscales suitable for turbulent simulations. The methodology is validated on large-scale external-aerodynamics configurations, including the Ahmed body at Re $ = 7.68\times 10^{5}$ for multiple slant angles, using unstructured tetrahedral meshes ranging from 3 to 40 million elements. Applicability is further demonstrated on a realistic Formula~1 configuration at $U_\infty=56~\mathrm{m/s}$ (201.6~km/h), corresponding to Re $ \approx 10^{6}$. The results show that the proposed stabilized pressure-segregated formulation remains robust at scale and captures key separated-flow features and coherent wake organization. Pointwise velocity and pressure spectra provide an a posteriori consistency indicator, exhibiting finite frequency ranges compatible with inertial-subrange reference slopes in the resolved band and supporting dissipation control in under-resolved regimes within a unified stabilized finite element framework.

physics.flu-dyn↗

arXiv2025

Improving performance estimation of a PCM-integrated solar chimney through reduced-order based data assimilation

This study evaluates a data assimilation framework based on reduced-order modeling (ROM-DA), complemented by a hybrid data-filling strategy, to reconstruct dynamic temperature fields in a phase-change-material (PCM) integrated solar chimney from limited temperature measurements. The goal is to enhance the estimation accuracy of the outlet airflow velocity. A regularized least-squares formulation is employed to estimate temperature distributions within an inclined solar chimney using RT-42 as the PCM. The methodology combines (i) a reduced-order model derived from high-fidelity finite-volume simulations of unsteady conjugate heat transfer with liquid-solid phase change and surface radiation, and (ii) three experimental datasets with 22, 135, and 203 measurement points. Missing data are reconstructed using a hybrid filling scheme based on boundary-layer and bicubic interpolations. The assimilated temperature fields are integrated into the thermally coupled forward solver to improve velocity predictions. Results show that the ROM-DA framework reconstructs the transient temperature fields in both the air and PCM domains with relative errors below 10 percent for sparse data and below 3 percent for expanded datasets. When applied to experimental measurements, the approach enhances the fidelity of temperature and velocity fields compared with the baseline model, reducing the outlet velocity RMS error by 20 percent. This represents the first application of a ROM-DA framework to a coupled multiphysics solar chimney with PCM integration, demonstrating its potential for near-real-time thermal state estimation and digital-twin development.

math.NA↗

arXiv2025

Estimation of Hemodynamic Parameters via Physics Informed Neural Networks including Hematocrit Dependent Rheology

Physics-Informed Neural Networks (PINNs) show significant potential for solving inverse problems, especially when observations are limited and sparse, provided that the relevant physical equations are known. We use PINNs to estimate smooth velocity and pressure fields from synthetic 4D flow Magnetic Resonance Imaging (MRI) data. We analyze five non-Newtonian dynamic 3D blood flow cases within a realistic aortic model, covering a range of hematocrit values from anemic to polycythemic conditions. To enhance state estimation results, we consider various design and training techniques for PINNs, including adaptive loss balancing, curriculum training, and a realistic measurement operator. Regarding blood rheology, the PINN approach accurately estimates viscosity globally and locally under peak systolic conditions. It also provides a clear pattern recognition for diastolic stages. Regarding mass conservation, PINN estimations effectively reproduce the bifurcation of flow through the different branches of the aorta, demonstrate an excellent representation of the non-slip conditions at the walls, and accurately estimate pressure drops with relative errors below the 5% in the whole pressure field. We test our pressure drop estimations against the state of the art Virtual Work Energy Relative Pressure (vWERP) estimator, and we observe how our results outperform vWERP in terms of both accuracy and time resolution. Additionally, we find that the best results are achieved by computing the velocity field using the PINN, which is then integrated into the vWERP framework, leading to time super-sampled and high-order approximations, with a clinically admissible accuracy.

math.NA↗

arXiv2024

A fast food-freezing temperature estimation framework using optimally located sensors

This article presents and assesses a framework for estimating temperature fields in real time for food-freezing applications, significantly reducing computational load while ensuring accurate temperature monitoring, which represents a promising technological tool for optimizing and controlling food engineering processes. The strategy is based on (i) a mathematical model of a convection-dominated problem coupling thermal convection and turbulence, and (ii) a least-squares approach for solving the inverse data assimilation problem, regularized by projecting the governing dynamics onto a reduced-order model (ROM). The unsteady freezing process considers a salmon slice in a freezer cabinet, modeled with temperature-dependent thermophysical properties. The forward problem is approximated using a third-order WENO finite volume solver, including an optimized second-order backward scheme for time discretization. We employ our data assimilation framework to reconstruct the temperature field based on a limited number of sensors and to estimate temperature distributions within frozen food. Sensor placement is optimized using a novel greedy algorithm, which maximizes the observability of the reduced-order dynamics for a fixed set of sensors. The proposed approach allows efficient extrapolation from external sensor measurements to the internal temperature of the food under realistic turbulent flow conditions, which is crucial for maintaining food quality.

math.NA↗

arXiv2024

Unconditionally stable, linearised IMEX schemes for incompressible flows with variable density

For the incompressible Navier--Stokes system with variable density and viscosity, we propose and analyse an IMEX framework treating the convective and diffusive terms semi-implicitly. This extends to variable density and second order in time some methods previously analysed for variable viscosity and constant density. We present three new schemes, both monolithic and fractional-step. All of them share the methodological novelty that the viscous term is treated in an implicit-explicit (IMEX) fashion, which allows decoupling the velocity components. Unconditional temporal stability is proved for all three variants. Furthermore, the system to solve at each time step is linear, thus avoiding the costly solution of nonlinear problems even if the viscosity follows a non-Newtonian rheological law. Our presentation is restricted to the semi-discrete case, only considering the time discretisation. In this way, the results herein can be applied to any spatial discretisation. We validate our theory through numerical experiments considering finite element methods in space. The tests range from simple manufactured solutions to complex two-phase viscoplastic flows.

math.NA↗

arXiv2024

Enhancing Hemodynamic Parameter Estimations: Nonlinear Blood Behavior in 4D Flow MRI

Hemodynamic parameters are often estimated assuming a constant Newtonian viscosity, even though blood exhibits shear-thinning behavior. This article investigates the influence of blood rheology and hematocrit (Hct) percentage on the estimation of Wall Shear Stress (WSS), rate of viscous Energy Loss ($\dot{E}_L$) at different points in the cardiac cycle, and the Oscillatory Shear Index (OSI). We focus on a hematocrit-dependent power-law non-Newtonian model, considering a wide range of Hct values at physiological temperature, with rheological parameters obtained from previously reported experimental data. In all cases, we systematically compared WSS, $\dot{E}_L$, and OSI using both Newtonian and power-law models, underscoring the crucial role of blood rheology in accurately assessing cardiovascular diseases. Our results show that, in in-silico experiments, differences in WSS and $\dot{E}_L$ across a wide range of Hct values can reach as high as 190\% and 113\% at systole, and as low as -72\% and -74\% at diastole, respectively. In in-vivo data, differences in WSS and $\dot{E}_L$ can reach up to -45\% and -60\% at systole, and range from -69\% to 73\% at diastole. This study enhances our understanding of the impact of blood rheology on hemodynamic parameter estimations using both in-silico and in-vivo aortic 4D Flow MRI data.

eess.SP↗

arXiv2023

Implicit-explicit schemes for incompressible flow problems with variable viscosity

In this work we study different Implicit-Explicit (IMEX) schemes for incompressible flow problems with variable viscosity. Unlike most previous work on IMEX schemes, which focuses on the convective part, we here focus on treating parts of the diffusive term explicitly to reduce the coupling between the velocity components. We present different, both monolithic and fractional-step, IMEX alternatives for the variable-viscosity Navier--Stokes system, analysing their theoretical and algorithmic properties. Stability results are proven for all the methods presented, with all these results being unconditional, except for one of the discretisations using a fractional-step scheme, where a CFL condition (in terms of the problem data) is required for showing stability. Our analysis is supported by a series of numerical experiments.

math.NA↗