A study of troubled-cell indicators applied to finite volume methods using a novel monotonicity parameter

2025

Read A study of troubled-cell indicators applied to finite volume methods using a novel monotonicity parameter (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-2504-11056.txt.

Introduction

Rao and Ramakrishna adapt in twenty twenty-five a shock-spotting tool from one simulation method to another. Solvers for gas flow laws must add smoothing near shocks, but smoothing everywhere slows convergence and blurs results. Their tool flags only the troubled cells, using data from each cell and its direct neighbours, so smoothing applies just where needed.

They also define a new score for how clean the solution looks near a shock. Tests on two-dimensional gas flow cases with a standard solver backed the approach. Limiting only in flagged cells converged better than limiting everywhere, with no loss of accuracy. The authors present the score as a fair way to judge such tools.

Catalog record

YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE44 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-11056

Browse the full library · About the Tower of Babel

Full text

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

A study of troubled-cell indicators applied to finite volume methods using a novel monotonicity parameter Abstract We adapt a troubled-cell indicator developed for discontinuous Galerkin (DG) methods [ 1 ] to the finite volume method (FVM) framework for solving hyperbolic conservation laws. This indicator depends solely on the cell-average data of the target cell and its immediate neighbours. Once the troubled-cells are identified, we apply the limiter only in these cells instead of applying in all computational cells. We introduce a novel technique to quantify the quality of the solution in the neighbourhood of the shock by defining a monotonicity parameter μ \mu . Numerical results from various two-dimensional simulations on the hyperbolic systems of Euler equations using a finite volume solver employing MUSCL reconstruction validate the performance of the troubled-cell indicator and the approach of limiting only in the troubled-cells. These results show that limiting only in the troubled-cells is preferable to limiting everywhere as it improves convergence without compromising on the solution accuracy. Keywords: Euler equations, Finite Volume Method, MUSCL, Limiter, Troubled-cell indicator, Monotonicity parameter 1 Introduction In this paper, we adapt a troubled-cell indicator originally developed for discontinuous Galerkin (DG) methods to finite volume methods (FVM). We introduce a novel technique to quantify the quality of the solution in the neighbourhood of the shock. We show that limiting only in troubled-cells is preferable to limiting everywhere. The Finite Volume Method (FVM) is a widely employed numerical technique for solving hyperbolic conservation laws arising in various scientific and engineering disciplines, particularly compressible fluid dynamics [ 2 ] due to its robustness and adaptability across diverse applications. The Monotone Upstream-centered Schemes for Conservation Laws (MUSCL) [ 3 ] reconstruction is widely used within the finite volume framework to attain high-order accuracy by reconstructing cell interface values from neighboring cell-average data. However, like other high-order numerical schemes, MUSCL face challenges in accurately capturing shocks and discontinuities without introducing non-physical oscillations, which can potentially lead to numerical instability. Limiting techniques are often incorporated into the MUSCL reconstruction to prevent such oscillations. These limiters control spurious oscillations in the vicinity of discontinuities, thereby preserving monotonicity while retaining high-order accuracy of the numerical scheme in regions of smooth solution [ 3 , 4 ] . In conventional implementations of the MUSCL reconstruction, the limiter function is typically applied to all computational cells. However, this function remains inactive in most of the cells, particularly, in regions of smooth solution and it often identifies regions near smooth extrema as requiring limiting [ 5 ] . This approach guarantees monotonicity but it can lead to unnecessary computational effort and may adversely affect the convergence of the solution to a steady state [ 6 ] . In the subsequent sections, we refer to this as the “ limiting everywhere approach ”. Krivodonova et al. [ 7 ] proposed a discontinuity detector for Discontinuous Galerkin (DG) methods. By applying limiters exclusively in the computational cells suggested by discontinuity detector, they demonstrated a significant improvement over the conventional approach of applying limiters in all computational cells. Qiu and Shu [ 8 ] proposed a weighted essentially non-oscillatory (WENO) finite volume methodology as limiters for DG methods. Their approach involves detecting “troubled-cells” that potentially contain discontinuities and may require limiting. Within these troubled cells, the original high-order DG solution polynomials are replaced with WENO reconstructed polynomials to retain the original high-order accuracy of the DG method. In the present paper, we investigate this approach of limiting only in the regions near discontinuities for an FVM solver employing MUSCL reconstruction. To detect such regions near discontinuities, where limiting is needed, we adapt the troubled-cell indicator developed by Fu and Shu [ 1 ] for DG methods to our FVM framework. To validate the effectiveness of this limiting approach, we employ a series of well-defined two-dimensional simulations on the hyperbolic systems of Euler equations involving discontinuities, particularly shocks. Limiting in troubled-cells alone shows better convergence without compromising the solution accuracy as compared to limiting in all computational cells. In the subsequent sections, we refer to the method of applying the limiter function only in troubled-cells as the “ limiting restricted region approach ”. We also present a new technique to quantify the quality of the solution in the neighbourhood of the shock by defining a monotonicity parameter \mu . Using the monotonicity parameter, we compare the solutions obtained from the limiting restricted region approach for different threshold constants of the troubled-cell indicator, as well as the solution obtained from the limiting everywhere approach for the given test problem. The outline of the rest of the paper is as follows. In section 2 , we briefly review the two-dimensional compressible Euler equations governing inviscid flow, along with the FVM employing the MUSCL reconstruction scheme to solve these equations. This section also presents the description of various two-dimensional test cases employed in this study. Section 3 introduces the troubled-cell indicator, adapted from DG methods to the FVM framework, and demonstrates its effectiveness across various test cases. Section 4 presents numerical results for the test problems outlined in Section 2 , by solving them using both limiting approaches and comparing the solutions through various metrics, including the monotonicity parameter. Finally, we conclude in section 5 . 2 Numerical Methodology In this section, we provide a concise overview of the governing equations and the base solver employed to solve them. We describe the various two-dimensional test cases, both steady-state and unsteady problems, employed in this study. The base solver utilizes a finite volume approach with MUSCL reconstruction and the Koren limiter [ 9 ] . No other scheme is used to improve the quality of the solution or enhance convergence to steady-state so as to isolate the effect of the limiting strategy. As a consequence, there are instances where the base solver stalls and convergence to steady-state is not achieved. Governing Equations and Base Solver: The governing equations for compressible inviscid flows can be written in integral form over a control volume with volume \mathcal{V} and surface area S as \frac{d}{dt}\int_{\mathcal{V}}\textbf{Q}\,d\mathcal{V}+\int_{S}\textbf{H}(\textbf{Q})\cdot\,\hat{\textbf{n}}\,dS=0. (1) In two dimensions, the vector of conservative variables \textbf{Q}=[\rho,\rho u,\rho v,\rho E]^{T} , the convective flux vector \textbf{H}=(\textbf{F},\textbf{G}) , \textbf{F}=[\rho u,\rho u^{2}+p,\rho uv,\rho uH]^{T} , \textbf{G}=[\rho v,\rho uv,\rho v^{2}+p,\rho vH]^{T} , pressure p=(\gamma-1)\rho\left[E-\frac{1}{2}(u^{2}+v^{2})\right] , enthalpy H=E+p/\rho and \gamma=1.4 . The region of interest in the flow field is discretized and replaced by an array of non-overlapping cells (or volumes). For a given cell, using the finite volume method, equation ( 1 ) can be discretized in space to obtain the semi-discretized form \Omega\,\frac{d\bar{\textbf{Q}}}{dt}+\textbf{R}(\bar{\textbf{Q}})=0 (2) where, \bar{\textbf{Q}} is the cell average of Q and \Omega is the volume of the cell. The residue \textbf{R}(\bar{\textbf{Q}}) is given as \textbf{R}(\bar{\textbf{Q}})=\sum_{f}\tilde{\textbf{H}}_{f}(\bar{\textbf{Q}})\cdot\,\hat{\textbf{n}}_{f}\,s_{f} (3) where, f is an index over the faces of the cell, \tilde{\textbf{H}}_{f} is the numerical inviscid flux vector at the face, \hat{\textbf{n}}_{f} is the unit normal vector at the face and s_{f} is the area of the cell face. In this investigation, we use the Advection Upstream Splitting Method (AUSM) family scheme AUSM+ [ 10 ] to evaluate the numerical inviscid flux \tilde{\textbf{H}}_{f} . This requires the state to the left and right of the face where \tilde{\textbf{H}}_{f} is being evaluated. We use the higher order

Read the complete text