Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
126 tokens/sec
GPT-4o
11 tokens/sec
Gemini 2.5 Pro Pro
47 tokens/sec
o3 Pro
5 tokens/sec
GPT-4.1 Pro
3 tokens/sec
DeepSeek R1 via Azure Pro
33 tokens/sec
2000 character limit reached

A critical analysis of some popular methods for the discretisation of the gradient operator in finite volume methods (1606.05556v6)

Published 17 Jun 2016 in cs.NA and cs.CE

Abstract: Finite volume methods (FVMs) constitute a popular class of methods for the numerical simulation of fluid flows. Among the various components of these methods, the discretisation of the gradient operator has received less attention despite its fundamental importance with regards to the accuracy of the FVM. The most popular gradient schemes are the divergence theorem (DT) (or Green-Gauss) scheme, and the least-squares (LS) scheme. Both are widely believed to be second-order accurate, but the present study shows that in fact the common variant of the DT gradient is second-order accurate only on structured meshes whereas it is zeroth-order accurate on general unstructured meshes, and the LS gradient is second-order and first-order accurate, respectively. This is explained through a theoretical analysis and is confirmed by numerical tests. The schemes are then used within a FVM to solve a simple diffusion equation on unstructured grids generated by several methods; the results reveal that the zeroth-order accuracy of the DT gradient is inherited by the FVM as a whole, and the discretisation error does not decrease with grid refinement. On the other hand, use of the LS gradient leads to second-order accurate results, as does the use of alternative, consistent, DT gradient schemes, including a new iterative scheme that makes the common DT gradient consistent at almost no extra cost. The numerical tests are performed using both an in-house code and the popular public domain PDE solver OpenFOAM.

Citations (53)

Summary

  • The paper shows that popular gradient discretization schemes in FVMs do not maintain consistent accuracy, which varies significantly with grid structure.
  • Specifically, the Divergence Theorem scheme can be zeroth-order on unstructured meshes, highlighting issues with mesh skewness, while Least-Squares provides more consistent first-order accuracy on arbitrary grids.
  • The study provides practical guidance for CFD engineers, warning about the limitations of DT on unstructured meshes and suggesting LS or iterative methods for improved accuracy in real-world applications.

Finite Volume Methods (FVMs) are integral to Computational Fluid Dynamics (CFD) and simulate fluid flow by managing the solution of partial differential equations (PDEs) on discretized spaces. However, discretizing the gradient operator, a crucial step in FVMs, particularly needs attention due to its impact on the method's accuracy. Two prevalent schemes, the Divergence Theorem (DT) and Least-Squares (LS), have traditionally been perceived as second-order accurate across varied grids. Alexandros Syrakos et al.'s paper challenges this notion, demonstrating through theoretical analysis and numerical simulations that these schemes offer different orders of accuracy contingent on the mesh structure.

Gradient Schemes and Orders of Accuracy

The divergence theorem method, also known as Green-Gauss, can be zeroth-order accurate on unstructured meshes due to skewness and unevenness inherent in these grids. On structured grids, its accuracy improves to second-order when error cancellation occurs between opposite faces. The paper proposes using corrector iterative schemes which achieve better accuracy than traditional DT, albeit at increased computational cost. The LS scheme, on the other hand, is shown to maintain first-order accuracy on arbitrary grids and achieve second-order accuracy on structured grids due to its ability to minimize the sum of the squares of directional derivative errors weighted by distances. Notably, adopting a specific weight exponent of 3/2-3/2 can enhance accuracy in certain grid configurations.

Implications and Practical Applications

The algorithms dissected in this paper have ramifications for structured and unstructured grid contexts used in industrial applications. Engineers should be wary of using DT without correcting for skewness, especially in automatic unstructured mesh generation processes prevalent in contemporary CFD practices. This work suggests employing LS schemes and highlights novel iterative methods to increase DT's computational viability without compromise in accuracy. Computational performance metrics, convergence behavior, and grid refinement impacts highlighted by this paper serve as practical guides for optimizing FVMs in CFD simulation contexts.

Future Directions

The exploration reveals that while DT and LS maintain theoretical second-order accuracies under assumptions, practical impediments arise from grid irregularity. Future research should focus on enhanced methodologies and alternative grid arrangements for high-precision gradient calculations in turbulent and non-Newtonian flow models, pushing the boundaries of current CFD capabilities. Moreover, exploring parallel algorithms and adaptive techniques could complement the iterative capacity advocated herein, potentially revolutionizing FVM processes.

In conclusion, Syrakos et al.’s paper serves as a pivotal resource, reframing understandings of gradient discretization in FVMs and preparing the research community for crafting improved CFD models that navigate the complex terrain of real-world engineering applications with greater reliability.

Youtube Logo Streamline Icon: https://streamlinehq.com