Unifying local and nonlocal corrosion frameworks: A convergent nonlocal extension of the KKS phase-field model

Christian J. Cyron 1,2 , Marvin Fritz 3 , Alexander Hermann 1,4,∗ , Tobias Köppl 5 , Arman Shojaei 1 , Stewart Silling 6 · 2025

Read Unifying local and nonlocal corrosion frameworks: A convergent nonlocal extension of the KKS phase-field model by Christian J. Cyron 1,2 , Marvin Fritz 3 , Alexander Hermann 1,4,∗ , Tobias Köppl 5 , Arman Shojaei 1 , Stewart Silling 6 (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-2509-23503.txt.

Introduction

Corrosion models come in local and nonlocal families that rarely speak to each other. Christian J. Cyron and 5 co-authors extend the Kim-Kim-Suzuki phase-field corrosion model so that it bridges the two. Gradient terms give way to integral terms over a finite reach, which adds a built-in length scale like the one in peridynamics. A nonlocal free energy combines the usual bulk term with this interaction term.

From it the team derives nonlocal versions of the Allen-Cahn and Cahn-Hilliard equations. They prove the equations are well-posed and that the model settles to the classic local one as the reach shrinks to zero. Numerical tests confirm the predicted convergence and show what the nonlocal term changes.

Catalog record

CREATORChristian J. Cyron 1,2 , Marvin Fritz 3 , Alexander Hermann 1,4,∗ , Tobias Köppl 5 , Arman Shojaei 1 , Stewart Silling 6
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE82 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2509-23503

Browse the full library · About the Tower of Babel

Full text

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

Unifying local and nonlocal corrosion frameworks: A convergent nonlocal extension of the KKS phase-field model Christian J. Cyron 1,2 , Marvin Fritz 3 , Alexander Hermann 1,4,∗ , Tobias Köppl 5 , Arman Shojaei 1 , Stewart Silling 6 Abstract We introduce a nonlocal extension of the Kim-Kim-Suzuki (KKS) phase-field corrosion model aimed at bridging local and nonlocal corrosion modeling approaches, such as phase-field and peridynamic frameworks. In this formulation, classical gradient operators are replaced with integral operators defined over a finite interaction horizon, naturally embedding an intrinsic length scale that aligns with nonlocal theories like peridynamics. Under precise assumptions on function spaces and kernel functions, we define a nonlocal free energy that integrates a standard bulk free energy density with a nonlocal interaction term. Through differentiation in an appropriate Hilbert space, we derive evolution equations, yielding a nonlocal Allen–Cahn equation for the phase-field and a nonlocal Cahn–Hilliard-type equation for the concentration. The latter is expressed as a gradient flow in a metric induced by the inverse of the nonlocal operator, mirroring the classical H − 1 H^{-1} metric for conserved dynamics. We establish the well-posedness of these equations using Galerkin approximations, uniform energy estimates, and compactness arguments. Furthermore, we prove the convergence of the nonlocal model to its local KKS counterpart as the interaction horizon approaches zero, effectively unifying local and nonlocal perspectives. Numerical experiments, implemented via finite difference spatial discretization and explicit time-stepping, demonstrate the effects of nonlocality and confirm the theoretical convergence, reinforcing the connection between the two modeling paradigms. keywords: phase-field, nonlocal, corrosion, well-posedness, nonlocal-to-local convergence, explicit scheme, peridynamics, Allen–Cahn/Cahn–Hilliard system, Kim–Kim–Suzuki phase-field model 1 Introduction Nonlocal phase-field models have attracted significant attention in scientific computing because they naturally incorporate long-range interactions and introduce an intrinsic length scale into the governing equations [ 39 ] . Unlike classical local formulations that employ differential operators—which capture only infinitesimal interactions—nonlocal formulations replace these operators with integral operators that account for interactions over a finite horizon. This replacement not only provides a natural regularization mechanism for evolving interfaces but also establishes a direct connection with peridynamic theories, where the interaction horizon is a fundamental parameter [ 39 ] . In many applications, including phase transitions [ 11 ] and corrosion phenomena (e.g., in biodegradable magnesium implants [ 25 , 29 ] ), classical local models fail to capture important multiscale effects due to the absence of an intrinsic length scale. For instance, in the two-phase metal–electrolyte setting of corrosion, the classical KKS (Kim–Kim–Suzuki) phase-field model [ 27 ] —with a free energy comprising a bulk term and a gradient term—does not explicitly incorporate a finite spatial interaction range. Recent studies have demonstrated that introducing nonlocal interactions via integral operators can significantly alter interface dynamics, sometimes even producing phenomena that are not observed in local theory [ 10 , 36 , 20 ] . Peridynamic models, initially proposed by Silling [ 39 ] , are reformulations of classical local continuum mechanics in the form of nonlocal integral differential equations. With the spatial derivatives of the stress tensor replaced by integrals of force density functions, a nonlocal model allows the natural treatment of balance laws on and off material discontinuities. We refer to the overview articles [ 40 , 41 ] and the books [ 4 , 28 ] on perdiynamic modeling. Motivated by these developments and rigorous convergence results in nonlocal-to-local settings of Cahn–Hilliard and Allen–Cahn models [ 1 , 12 , 17 , 31 ] , we present in this work a rigorous nonlocal formulation of the KKS phase-field corrosion model. Our approach is based on defining a free energy functional that couples a standard bulk free energy density with a nonlocal interaction term characterized by a finite cut-off radius, akin to those employed in peridynamic models [ 39 ] and in tumor growth models with long-range interaction radius [ 19 , 20 ] . We rigorously derive the associated evolution equations via differentiation in a suitable Hilbert space framework. Our derivation yields a nonlocal Allen–Cahn equation for the phase-field and a nonlocal Cahn–Hilliard-type equation for the concentration, the latter formulated as a gradient flow in a metric induced by the inverse of the nonlocal operator, analogous to the classical H^{-1} -metric that naturally enforces mass conservation. While our analysis, based on Galerkin approximations, uniform energy estimates, and compactness arguments, establishes the well-posedness of both equations and demonstrates that, under standard smoothness assumptions and proper scaling, the nonlocal model converges to its classical local counterpart, our primary focus is on clarifying the impact of the finite interaction horizon on the solution’s behavior. Furthermore, we highlight a structural equivalence between the nonlocal KKS model and peridynamic corrosion models under specific parameter regimes, thereby bridging phase-field and peridynamic approaches to corrosion modeling. The remainder of the paper is organized as follows: In Section 2 we introduce the mathematical framework and state our assumptions. Section 3 reviews the classical local KKS phase-field corrosion model. Furthermore, we introduce the nonlocal free energy functional and derive the corresponding evolution equations via differentiation. In Section 4 , we prove the existence and uniqueness of the decoupled and fully coupled nonlocal models. Section 5 discusses the nonlocal-to-local convergence of the derived operators, while Section 6 presents our finite difference spatial discretization and explicit time-stepping scheme and presents numerical experiments on a two-dimensional pitting corrosion benchmark, and Section 7 concludes with a discussion and outlook for future work. 2 Mathematical framework and standing assumptions In the following, we state the notational conventions and assumptions that are used throughout the paper. In our work, \Omega\subset\mathbb{R}^{d} ( d\in\{1,2,3\} ) is a bounded Lipschitz domain and T>0 is a fixed final time. The mean-free space of square-integrable functions is denoted by \mathring{L}^{2}(\Omega) . Moreover, we write \lesssim instead of \leq C when omitting a generic constant C , which may also change from line to line. Assumption 1 (Kernel properties) . The kernel J_{\delta}:\mathbb{R}^{d}\to\mathbb{R} , \delta>0 , satisfies: (J1) Nonnegativity : J_{\delta}(z)\geq 0 for all z\in\mathbb{R}^{d}. (J2) Radial symmetry : J_{\delta}(z)=j_{\delta}(\lvert z\rvert) for some j_{\delta}:[0,\infty)\to\mathbb{R}. (J3) Compact support : \textup{supp}(J_{\delta})\subseteq B_{\delta}(0). (J4) Embeddings : V_{\delta}\subset L^{4}(\Omega) and V_{\delta}\Subset L^{2}(\Omega) where V_{\delta}=\Bigl\{u\in L^{2}(\Omega):|u|_{{V_{\delta}}}^{2}:=\frac{1}{2}\iint_{\Omega\times\Omega}J_{\delta}(x-y)[u(x)-u(y)]^{2}\,\textup{d}(x,y)<\infty\Bigr\}. Before mentioning relevant examples of kernels satisfying these stated properties, we state one more assumption on the kernel function J_{\delta} that becomes relevant when considering the nonlocal-to-local convergence \delta\to 0 in Section 5 . Assumption 2 (Additional kernel property) . The kernel J_{\delta}:\mathbb{R}^{d}\to\mathbb{R} , \delta>0 , satisfies: (J5) Second-moment normalization : \int_{\mathbb{R}^{d}}J_{\delta}(z)\lvert z\rvert^{2}\,\textup{d}z=2d . Example 1 . The stated properties ( J1 )–( J5 ) are satisfied by: • the top-hat kernel J_{\delta}(z)=c_{\delta}\chi_{B_{\delta}(0)}(z)\quad\forall z\in\mathbb{R}^{d},\quad c_{\delta}=\frac{2d(d+2)}{\delta^{d+2}|S^{d-1}|}, (1) where S^{d-1} denotes the surface of the unit sphere in \mathbb{R}^{d} ; • the truncated fractional kernel [ 15 , Eq. (2.7)] J_{\delta}(z)=c_{\delta}\lvert z\rvert^{-d-2s}\chi_{B_{\delta}(0)}(z)\quad\forall z\in\mathbb{R}^{d},\quad d/2<s<2. where c_{\delta} is chosen as in [ 15 , Eq. (2.20)–(2.22)] , depending on the size of \delta , such that J_{\delta}\to-\Delta as \delta\to 0 and J_{\delta}\to(-\Delta)^{s} as \delta\to\infty . The first three properties of Assumption 1 are easy to check for the top-hat and truncated fractional kernels. Regarding the compact embedding ( J4 ), we refer to [ 37 , Thm. 1.9] and [ 13 , Cor 3.12] . It is well-known for these examples

Read the complete text