A Shifted Boundary Method for Thermal Flows

Cheng-Hau Yang, Guglielmo Scovazzi, Adarsh Krishnamurthy, Baskar Ganapathysubramanian · 2025

Read A Shifted Boundary Method for Thermal Flows by Cheng-Hau Yang, Guglielmo Scovazzi, Adarsh Krishnamurthy, Baskar Ganapathysubramanian (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-00143.txt.

Introduction

Cheng-Hau Yang and three co-authors brought a shifted edge method to heat flows. Their paper, from twenty twenty-five, deals with joined flow and heat runs on meshes that do not fit the body's shape. Their build uses an octree mesh, part filled, with the shifted boundary method. A half implicit form of the thermal flow laws keeps runs fast yet true. The method can pin both field and slope terms on cut cells.

That gives true flux reads near hard shapes. Slope terms, the hard case, held exact on their meshes. Runs span flow regimes from slow to swift. Readers in flow sims get a way to keep edge truth with no body fit mesh grind.

Catalog record

CREATORCheng-Hau Yang, Guglielmo Scovazzi, Adarsh Krishnamurthy, Baskar Ganapathysubramanian
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE92 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-00143

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

A Shifted Boundary Method for Thermal Flows Cheng-Hau Yang, Guglielmo Scovazzi, Adarsh Krishnamurthy, Baskar Ganapathysubramanian Abstract This paper presents an incomplete Octree mesh implementation of the Shifted Boundary Method (Octree-SBM) for multiphysics simulations of coupled flow and heat transfer. Specifically, a semi-implicit formulation of the thermal Navier-Stokes equations is used to accelerate the simulations while maintaining accuracy. The SBM enables precise enforcement of field and derivative boundary conditions on cut (intercepted) elements, allowing for accurate flux calculations near complex geometries, when using non-boundary fitted meshes. Both Dirichlet and Neumann boundary conditions are implemented within the SBM framework, with results demonstrating that the SBM ensures precise enforcement of Neumann boundary conditions on Octree-based meshes. We illustrate this approach by simulating flows across different regimes, spanning several orders of magnitude in both the Rayleigh number ( R ​ a ∼ 10 3 Ra\sim 10^{3} – 10 9 10^{9} ) and the Reynolds number ( R ​ e ∼ 10 0 Re\sim 10^{0} – 10 4 10^{4} ), and covering the laminar, transitional, and turbulent flow regimes. Coupled thermal-flow phenomena and their statistics across all these regimes are accurately captured without any additional numerical treatments, beyond a Residual-based Variational Multiscale formulation (RB-VMS). This approach offers a reliable and efficient solution for complex geometries, boundary conditions and flow regimes in computational multiphysics simulations. 1 Introduction Natural and forced convection are fundamental mechanisms in heat transfer, influencing a wide range of engineering applications, such as the optimization of the design of thermal exchangers [ Bhutta et al., 2012 , Abeykoon, 2020 ] . Beyond industrial contexts, convection plays a key role in addressing climate-related challenges. For example, the urban heat island effect has become increasingly prominent due to rising global temperatures and urbanization, making the understanding and management of convection in such environmental flows more critical than ever [ Priyadarsini et al., 2008 , Allegrini and Carmeliet, 2018 ] . In sustainable building design, natural ventilation has emerged as a promising strategy for improving energy efficiency by harnessing wind and thermal energy [ Chenari et al., 2016 , Zhang et al., 2021 ] . Furthermore, accurately modeling the interaction between airflow and temperature in built environments is essential for ensuring indoor comfort and safety. This is particularly relevant in the context of public health, where effective airflow control is crucial in preventing the spread of infectious diseases [ Bhattacharyya et al., 2020 , Li et al., 2021 , Foster and Kinzel, 2021 , Saurabh et al., 2021 , Tan et al., 2023 , Rayegan et al., 2023 ] . Simulating these processes, however, often involves geometrically complex domains, such as urban layouts in heat island studies, human anatomy in aerosolized virus transmission modeling, or intricate configurations in thermal exchangers. Generating boundary-fitted meshes for such geometries is time-consuming and labor-intensive. Moreover, during simulations, engineers frequently discover that some regions require finer resolution, while others can afford coarser discretization. Adjusting these resolutions typically necessitates revisiting meshing procedures, producing impractical situations in workflows demanding rapid iterations. Another example of such situations is the increased interest in creating large computational datasets of flow past complex geometries for training AI-driven algorithms [ Tali et al., 2024 ] . Here again, the bottleneck is the creation of good body-fitted meshes and accurate imposition of boundary conditions. The immersed boundary method (IBM) [ Peskin, 1972 , Mittal and Iaccarino, 2005 , Colonius and Taira, 2008 , Wang and Liu, 2004 , Zhang et al., 2004 , Borazjani et al., 2008 , Zhao and Yan, 2022 ] offers an alternative, as it avoids boundary-fitted meshes, significantly simplifying the meshing process. This approach decouples the computational grid from geometric shapes, allowing greater efficiency in testing and simulation. However, traditional IBM implementations, such as the Finite Cell Method (FCM) [ Parvizian et al., 2007 , Düster et al., 2008 , Schillinger et al., 2013 , Stavrev et al., 2016 , de Prenter et al., 2017 , Jomo et al., 2021 ] and immersogeometric analysis (IMGA) [ Kamensky et al., 2015 , Xu et al., 2016 , Wang et al., 2017 , Hoang et al., 2019 , de Prenter et al., 2019 , Zhu et al., 2019 , Xu et al., 2021 , Xu et al., 2019 , Kamensky, 2021 , Jaiswal et al., 2024 ] , face inherent challenges. Issues such as the small-cut cell problem and load balancing inefficiencies arise because elements intersected by the geometry often require a disproportionately large number of integration points, leading to uneven computational loads across processors. This remains an area of active development. To address these limitations, a recent development in IBMs, the Shifted Boundary Method (SBM) [ Main and Scovazzi, 2018a , Main and Scovazzi, 2018b , Karatzas et al., 2020 , Atallah et al., 2020 , Atallah et al., 2021a , Atallah et al., 2021b , Colomés et al., 2021 , Saurabh et al., 2021 , Atallah et al., 2022 , Zeng et al., 2022 , Heisler et al., 2023 , Yang et al., 2024 ] proposes to integrate the variational forms over a surrogate domain instead of directly working on the cut (or Intercepted ) elements. The SBM has demonstrated its versatility in various applications, including fluid dynamics [ Main and Scovazzi, 2018 ] , structural simulations [ Atallah et al., 2021a ] , free surface flows [ Colomés et al., 2021 ] , and one-way coupled fluid-structure interaction (FSI) [ Xu et al., 2024 ] . By eliminating the need for boundary-fitted meshing, the SBM significantly reduces preprocessing time while maintaining high accuracy. However, efficiently and automatically generating non-boundary-fitted meshes for the SBM or IBM simulations remains challenging. To overcome this, there have been recent efforts to employ Octree meshes, which enable faster and parallelized generation of non-boundary-fitted meshes. Octree meshes stand out due to their favorable aspect ratios, intrinsic hierarchical structure, and compatibility with parallel computing frameworks [ Popinet, 2003 , Losasso et al., 2004 , Chen et al., 2009 , Theillard et al., 2013 , Papac et al., 2013 , Guittet et al., 2015 , Sousa et al., 2019 , Egan et al., 2021 , Saurabh et al., 2021 , Bayat et al., 2022 , van Hooft and Popinet, 2022 , Yu et al., 2022 , Kim et al., 2023 , Blomquist et al., 2024 ] . These properties minimize inter-processor communication overhead by localizing the required neighborhood element information, making them highly efficient for large-scale, distributed simulations. The Octree-SBM framework, which combines Octree meshes with the SBM, has been applied to a variety of PDEs [ Yang et al., 2024 , Yang et al., 2024 ] . Together, they provide robust capabilities for handling geometrically complex domains while maintaining computational efficiency and accuracy, making them a promising approach for modern simulation challenges. Modeling thermal flows using the SBM is, therefore, a promising avenue for a variety of applications, but has been largely unexplored. Existing research has been limited to a single example involving a one-dimensional convection-diffusion equation [ Main and Scovazzi, 2018a ] . Motivated by this gap, this paper investigates the efficacy of the SBM for thermal incompressible flow simulations, with a focus on a coupled solver framework for the Navier-Stokes and energy (convection-diffusion) equations. The simulation framework developed in

Read the complete text