Abstract
In this paper, we propose a cell-centered finite volume method to numerically solve the G-equation on polyhedral meshes in three-dimensional space, that is, a general type of the level-set equation including advective, normal, and mean curvature flow motions. The main contribution is to design a numerical algorithm for the regularized mean curvature flow equation that can be consistently combined regarding the size of the time step with previous algorithms for the advective and normal flows on polyhedral meshes. For a spatial discretization, we use a flux-balanced approximation with an orthogonal splitting of displacement vector from a center of the cell to a center of the face. For a temporal discretization, we use a nonlinear Crank–Nicolson method with a deferred correction method which gives us, firstly, second-order accuracy in space and time similarly to the algorithms for the advective and normal flow equations, and, secondly, a possibility of straightforward domain decomposition for efficient parallel computation. Numerical experiments quantitatively show that the size of time step proportional to an average size of computational cells is enough to obtain the second-order convergence in space and time for smooth solutions of the general level set equation. A qualitative comparison is presented for a nontrivial example to compare numerical results obtained with hexahedral and polyhedral meshes. Finally, an example of solving numerically the G-equation used in combustion literature is given.
Similar content being viewed by others
Data Availability
The data that support the findings of this study are available on request from the corresponding author.
References
Williams, F.A.: Turbulent combustion. In: Buckmaster, J.D. (ed.) The Mathematics in Combustion, pp. 97–131. SIAM, Philadelphia (1985)
Peters, N.: The turbulent burning velocity for large-scale and small-scale turbulence. J. Fluid Mech. 384, 107–132 (1999)
Peters, N.: Turbulent Combustion. Cambridge Monographs on Mechanics. Cambridge University Press, Cambridge (2000)
Kolář, M., Kobayashi, S., Uegata, Y., Yazaki, S., Beneš, M.: Analysis of Kuramoto-Sivashinsky model of flame/smoldering front by means of curvature driven flow. In: Vermolen, F.J., Vuik, C. (eds.) Numerical Mathematics and Advanced Applications ENUMATH 2019. Lecture Notes in Computational Science and Engineering, vol. 139, pp. 615–624. Springer, Cham (2021)
Goto, M., Kuwana, K., Kushida, G., Yazaki, S.: Experimental and theoretical study on near-floor flame spread along a thin solid. Proc. Combust. Inst. 37, 3783–3791 (2019)
Evans, L.C., Spruck, J.: Motion of level sets by mean curvature. I. J. Differ. Geom. 33, 635–681 (1991)
Osher, S., Fedkiw, R.: Level Set Methods and Dynamic Implicit Surfaces. Springer, Berlin (2000)
Sethian, J.A.: Level Set Methods and Fast Marching Methods, Evolving Interfaces in Computational Geometry, Fluid Mechanics, Computer Vision, and Materical Science. Cambridge University Press, New York (1999)
Gibou, F., Fedkiw, R., Osher, S.: A review of level-set methods and some recent applications. J. Comput. Phys. 353, 82–109 (2018)
Perič, M.: Flow simulation using control volumes of arbitrary polyhedral shape. In: ERCOFTAC Bulletin, vol. 62, pp. 25–29 (2004)
Eymard, R., Gallouët, T., Herbin, R.: Finite volume methods. Handb. Numer. Anal. 7, 713–1018 (2000)
Lipnikov, K., Manzini, G., Shashkov, M.: Mimetic finite difference method. J. Comput. Phys. 257, 1163–1227 (2014)
da Veiga, L.B., Breszzi, F., Cangiani, A., Manzini, G., Marini, L.D., Russo, A.: Basic principles of virtual element methods. Math. Models Methods Appl. Sci. 23, 199–214 (2013)
Coudière, Y., Manzini, G.: The discrete duality finite volume method for convection-diffusion problems. SIAM J. Numer. Anal. 47, 4163–4192 (2010)
Barth, T.J., Sethian, J.: Numerical schemes for the Hamilton–Jacobi and level set equations on triangulated domains. J. Comput. Phys. 145(1), 1–40 (1998)
Kees, C.E., Akkerman, I., Farthing, M.W., Bazilevs, Y.: A conservative level set method suitable for variable-order approximations and unstructured meshes. J. Comput. Phys. 230(12), 4536–4558 (2011)
Lipnikov, K., Morgan, N.: A high-order discontinuous Galerkin method for level set problems on polygonal meshes. J. Comput. Phys. 397, 108834 (2019)
Frolkovič, P., Mikula, K.: High-resolution flux-based level set method. SIAM J. Sci. Comput. 29, 579–597 (2007)
Hahn, J., Mikula, K., Frolkovič, P., Basara, B.: Inflow-based gradient finite volume method for a propagation in a normal direction in a polyhedron mesh. J. Sci. Comput. 72, 442–465 (2017)
Mikula, K., Ohlberger, M.: A new level set method for motion in normal direction based on a semi-implicit forward–backward diffusion approach. SIAM J. Sci. Comput. 32, 1527–1544 (2010)
Mikula, K., Ohlberger, M.: Inflow-implicit/outflow-explicit scheme for solving advection equations. In: Fort, J. (ed.) Finite Volumes for Complex Applications VI—Problems and Perspectives. Springer Proceedings in Mathematics 4, vol. 1, pp. 683–691. Springer, Berlin (2011)
Mikula, K., Ohlberger, M., Urbán, J.: Inflow-implicit/outflow-explicit finite volume methods for solving advection equations. Appl. Numer. Math. 85, 16–37 (2014)
Frolkovič, P., Mikula, K., Urbán, J.: Semi-implicit finite volume level set method for advective motion of interfaces in normal direction. Appl. Numer. Math. 95, 214–228 (2015)
Hahn, J., Mikula, K., Frolkovič, P., Basara, B.: Semi-implicit level set method with inflow-based gradient in a polyhedron mesh. In: Cancès, C., Omnes, P. (eds.) Finite Volumes for Complex Applications VIII - Hyperbolic, Elliptic and Parabolic Problems, pp. 81–89 (2017). Springer International Publishing
Hahn, J., Mikula, K., Frolkovič, P., Medl’a, M., Basara, B.: Iterative inflow-implicit outflow-explicit finite volume scheme for level-set equations on polyhedron meshes. Comput. Math. Appl. 77, 1639–1654 (2019)
Deckelnick, K., Dziuk, G., Elliott, C.M.: Computation of geometric partial differential equations and mean curvature flow. Acta Numer. 14, 139–232 (2005)
Walkington, N.J.: Algorithms for computing motion by mean curvature. SIAM J. Numer. Anal. 33, 2215–2238 (1996)
Handlovičová, A., Mikula, K., Sgallari, F.: Semi-implicit complementary volume scheme for solving level set like equations in image processing and curve evolution. Numer. Math. 93, 675–695 (2003)
Mikula, K., Sarti, A., Sgallari, F.: Co-volume level set method in subjective surface based medical image segmentation, Chap. 11. In: Suri, J.S., Wilson, D.L., Laxminarayan, S. (eds.) Handbook of Biomedical Image Analysis. International Topics in Biomedical Engineering. Springer, Boston, pp. 583–626 (2005). https://doi.org/10.1007/0-306-48551-6_11
Mikula, K., Sarti, A., Sgallari, F.: Co-volume method for Riemannian mean curvature flow in subjective surfaces multiscale segmentation. Comput. Vis. Sci. 9, 23–31 (2006)
Corsaro, S., Mikula, K., Sarti, A., Sgallari, F.: Semi-implicit covolume method in 3d image segmentation. SIAM J. Sci. Comput. 28, 2248–2265 (2006)
Handlovičová, A., Mikula, K.: Stability and consistency of the semi-implicit co-volume scheme for regularized mean curvature flow equation in level set formulation. Appl. Math. 53, 105–129 (2008)
Eymard, R., Handlovičová, A., Mikula, K.: Study of a finite volume scheme for the regularized mean curvature flow level set equation. IMA J. Numer. Anal. 31, 813–846 (2011)
Eymard, R., Handlovičová, A., Mikula, K.: Approximation of nonlinear parabolic equations using a family of conformal and non-conformal schemes. Commun. Pure Appl. Anal. 11, 147–172 (2012)
Frolkovič, P., Mikula, K., Hahn, J., Martin, D., Basara, B.: Flux balanced approximation with least-squares gradient for diffusion equation on polyhedral mesh. Discrete Contin. Dyn. Syst. S (2020). https://doi.org/10.3934/dcdss.2020350
Hahn, J., Mikula, K., Frolkovič, P., Balažovjech, M., Basara, B.: Cell-centered finite volume method for regularized mean curvature flow on polyhedral meshes. In: Klöfkorn, R., Keilegavlen, E., Radu, F.A., Fuhrmann, J. (eds.) Finite Volumes for Complex Applications IX—Methods, Theoretical Aspects, Examples. Springer International Publishing, pp. 755–763 (2020). https://doi.org/10.1007/978-3-030-43651-3_72
Demirdžić, I.: On the discretization of the diffusion term in finite-volume continuum mechanics. Numer. Heat Transfer B Fundam. 68, 1–10 (2015)
Niceno, B.: A three dimensional finite volume method for incompressible Navier–Stokes equations on unstructured staggered grids. In: European Conference on Computational Fluid Dynamics ECCOMAS CFD 2006, Egmond Aan Zee, The Netherlands, Paper 196 (2006)
Balažovjech, M., Mikula, K.: A higher order scheme for a tangentially stabilized plane curve shortening flow with a driving force. SIAM J. Sci. Comput. 33(5), 2277–2294 (2011)
Balažovjech, M., Frolkovič, P., Frolkovič, R., Mikula, K.: Semi-implicit second order accurate finite volume method for advection-diffusion level set equation. In: Finite Volumes for Complex Applications VII-Elliptic, Parabolic and Hyperbolic Problems. Springer, Berlin, pp. 479–487 (2014)
Böhmer, K., Hemker, P.W., Stetter, H.J.: The defect correction approach. In: Defect Correction Methods. Springer, pp. 1–32 (1984)
Pitsch, H.: A consistent level set formulation for large-eddy simulation of premixed turbulent combustion. Combust. Flame 143, 587–598 (2005)
Acknowledgements
The work was supported by grants VEGA 1/0314/23, VEGA 1/0436/20, and APVV-19-0460. This project No. 2140/01/01 has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No. 945478.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix A: Iterative IIOE method
Appendix A: Iterative IIOE method
In order to complete the finite volume method to solve (2), we briefly review the iterative inflow-implicit outflow-explicit method (IIOE) [24, 25] of discretizing the advective and normal flow equations which appear in the second and third term of (2):
The following two forms of the velocity function \({\textbf{u}}\) are considered:
that describes the advective and normal flow, respectively, for an evolution of surface described by a zero level set of \(\phi \). Without a loss of generality, we choose the spatial variable \(\delta = \delta ({\textbf{x}})\) as the constant \(\delta = 1\) in (4) for the rest of derivation of the method.
Firstly, we introduce a spatial discretization, and, secondly, we discuss a temporal discretization. Let us denote the set of indices \(\bar{{\mathcal {F}}}_p\) and \(\bar{{\mathcal {B}}}_p\) of triangles obtained by a tessellation of internal and boundary faces \(e_f\), \(f \in {\mathcal {F}}_p\) and \(e_b\), \(b \in {\mathcal {B}}_p\), respectively; see more details how to construct the tessellation in [25]. For a triangle index \(i \in \bar{{\mathcal {F}}}_p\), \(\bar{e}_i\) is the triangle as a subset of some face \(e_f\), \(f \in {\mathcal {F}}_p\). Furthermore, \(\bar{{\textbf{n}}}_{pi}\) is the outward normal vector to the triangle whose length is the area of \(\bar{e}_i\), and \(\bar{{\textbf{x}}}_i\) is the center of the triangle. Using the Gauss divergence theorem on (A1), the spatial discretization is obtained:
where \(\phi _{pi}\) is an approximated value defined at the center of the triangle \(\bar{{\textbf{x}}}_i\). Note that the flux \(a_{pi}\) is computed using Gaussian quadrature of degree 1 which insures the exactness of integration as polynomials of degree 1.
A crucial point in (A3) is how to approximate the value \(\phi _{pi}\). In order to obtain a second-order accurate upwind scheme, we prepare some values at vertices and gradients at centers of tessellated triangles in the following procedure. A value \(\phi ({\textbf{x}}_v)\) at an internal vertex is computed by an inverse distance weighted average from adjacent cells. For the fixed vertex \({\textbf{x}}_v\), let us denote a set of indices of cells containing \({\textbf{x}}_v\) as a vertex, \({\mathcal {N}}_v \equiv \left\{ p \in {\mathcal {I}}: {\textbf{x}}_v \in \partial \Omega _p \right\} .\) Then the value at internal vertex is obtained by the inverse distance weighted average of the first order Taylor polynomial:
Note that a value at boundary vertex is directly assigned by Dirichlet boundary condition. From the obtained vertex values, a gradient at a center of tessellated triangle is computed. For an internal triangle \(\bar{e}_i \subset e_f\), \(f\in {\mathcal {F}}_p\), there exists \(q \in {\mathcal {I}}\) such that \(e_f \subset \partial \Omega _p \cap \partial \Omega _q\). Then, we consider two tetrahedrons whose apices are \({\textbf{x}}_p\) and \({\textbf{x}}_q\) and they have the common base \(\bar{e}_i\). We denote the set of all vertices of the tetrahedrons as \({\mathcal {P}}_i\). Then, the gradient \(\bar{\varvec{\beta }}_i\) at the center of \(\bar{e}_i\) is computed by the weighted least-squares minimization:
where \(\bar{{\textbf{x}}}_i\) is the center of \(\bar{e}_i\) the weight function is defined by \(w_i({\textbf{y}}) = |{\textbf{y}} - \bar{{\textbf{x}}}_i|^{-2}\). The formula (A4) is a generalization of the diamond-cell method described for a regular structured hexahedron cell in [20]. Now, we define a so-called average-based gradient [19] as the inverse distance weighted average of gradients:
Finally, we compute the value \(\phi _{pi}\) at an internal triangle \(\bar{e}_i \subset e_f\), \(f\in {\mathcal {F}}_p\), in (A3) using the average-based gradient and the upwind principle:
where the neighbor cell index \(q \in {\mathcal {I}}\) is such that \(e_f \subset \partial \Omega _p \cap \partial \Omega _q\). The value \(\phi _{pi}\) at a boundary triangle \(\bar{e}_i \subset e_b\), \(b\in {\mathcal {B}}_p\), in (A3) is computed by
where \(\phi _{bi} \equiv \phi _b(\bar{{\textbf{x}}}_i)\) from Dirichlet boundary condition. We finally obtain the spatial discretization from (A6) and (A7):
where \(\bar{{\mathcal {F}}}^{-}_{p}\) and \(\bar{{\mathcal {B}}}^{-}_{p}\) are the subsets of \(\bar{{\mathcal {F}}}_p\) and \(\bar{{\mathcal {B}}}_p\) with \(a_{pi} < 0\), respectively, and \(\bar{{\mathcal {F}}}^{+}_{p} \equiv \bar{{\mathcal {F}}}_p {\setminus } \bar{{\mathcal {F}}}^{-}_{p}\), \(\bar{{\mathcal {B}}}^{+}_{p} \equiv \bar{{\mathcal {B}}}_p {\setminus } \bar{{\mathcal {B}}}^{-}_{p}\).
Concerning the time discretization, we apply the deferred correction method because of the same reasons as discussed in Sect. 3. Following the analogous notations as in previous sections, the IIOE method inspired by [20,21,22,23] is defined by using an implicit and explicit time discretization of terms in (A8) on an inflow and outflow triangle, respectively:
where \(\phi ^{n,0} = \phi ^{n-1}\). Note that the average-based gradient \({\mathcal {D}}_p \phi ^{n,k-1}\) is computed by values at centers of cells from \(\phi ^{n,k-1}\) and values at centers of boundary faces from the nth time level.
About this article
Cite this article
Hahn, J., Mikula, K., Frolkovič, P. et al. Second-order accurate finite volume method for G-equation on polyhedral meshes. Japan J. Indust. Appl. Math. 40, 1053–1082 (2023). https://doi.org/10.1007/s13160-023-00574-x
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s13160-023-00574-x
Keywords
- G-equation
- Level-set equations
- Mean curvature flow
- Polyhedral meshes
- Nonlinear Crank–Nicolson method
- Flux-balanced approximation
- Second-order experimental order of convergence