Advances on finite element discretization of fluid-structure interaction problems
Najwa Alshehri, Daniele Boffi, Fabio Credali, Lucia Gastaldi · 2025
Read Advances on finite element discretization of fluid-structure interaction problems by Najwa Alshehri, Daniele Boffi, Fabio Credali, Lucia Gastaldi (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-2505-02594.txt.
Introduction
Najwa Alshehri, Daniele Boffi, Fabio Credali, and Lucia Gastaldi review in 2025 a numerical method for problems where a fluid meets a moving structure, from heart valves to parachutes. Simulating both sides at once is hard because the boundary between them moves and deforms.
Their method uses unfitted finite elements: the mesh ignores the interface, and a distributed Lagrange multiplier glues fluid and structure together in the spirit of fictitious domain methods. The authors recap what is proven about the scheme's stability and accuracy, plus error estimates for simpler interface problems. The review doubles as a guide to the method's remaining computational challenges.
Catalog record
| CREATOR | Najwa Alshehri, Daniele Boffi, Fabio Credali, Lucia Gastaldi |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 84 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-02594
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Advances on finite element discretization of fluid-structure interaction problems Najwa Alshehri, Daniele Boffi, Fabio Credali, Lucia Gastaldi Abstract . We review the main features of an unfitted finite element method for interface and fluid-structure interaction problems based on a distributed Lagrange multiplier in the spirit of the fictitious domain approach. We recall our theoretical findings concerning well-posedness, stability, and convergence of the numerical schemes, and discuss the related computational challenges. In the case of elliptic interface problems, we also present a posteriori error estimates. mathx"17 Introduction The mathematical description of many physical and engineering problems requires partial differential equations involving different operators or discontinuous coefficients so that the computational domain is partitioned into several (possibly time dependent) regions. The coupling between different models is enforced at the interface via suitable transmission conditions. Among the applications, we mention, for instance, thermo-mechanical problems [ 1 , 2 ] , cardiac simulations [ 3 , 4 ] , computational geo-science [ 5 , 6 ] , and many others. When designing effective numerical methods for the mentioned problems, an accurate representation of the underlying geometry is needed. Fitted approaches are based on a mesh which conforms to the interface and provide accurate solutions [ 7 , 8 , 9 , 10 , 11 ] . However, when applied to evolution problems, computational issues may arise. For instance, the Arbitrary Lagrangian–Eulerian approach [ 12 , 13 , 14 ] , appealing for fluid-structure interaction problems, needs an update of the mesh at each time step. In case of large displacements or deformations, the update may break shape regularity, thus requiring the generation of a new mesh, a costly operation in general. Moreover, when fluid and solid have similar or equal densities, the added mass effect may affect the performance of the method [ 15 ] . Such phenomenon may lead to the unconditional instability of the scheme regardless of the physical parameters. The added mass effect can be reduced by an appropriate treatment of transmission conditions, as described in [ 16 , 17 ] . Due to the critical aspects of fitted approaches, several unfitted methods have been introduced during the past decades. In this case, meshes are generated without taking care of the interface, which is allowed to cross elements. We mention, for instance, the Immersed Boundary Method [ 3 , 18 ] , where the interface position is tracked by Dirac delta functions, and the level set method [ 19 , 20 ] , where the interface corresponds to the zero level set of a certain function. Other techniques such as Nitsche-XFEM [ 21 ] , Cut-FEM [ 22 , 23 , 24 ] , Finite Cell methods [ 25 , 26 , 27 ] , and Ghost-FEM [ 28 ] are based on the use of penalty terms. The fictitious domain approach has been introduced to model the particulate flow as described in [ 29 , 30 ] , while the application to more general cases can be found in [ 31 , 32 ] . Another paradigm for tackling moving boundaries is given by diffuse interface approaches [ 33 , 34 , 35 ] , where quantities localized on the interface are distributed throughout a thick interfacial region. In [ 36 , 37 ] , we introduced a new fictitious domain formulation with distributed Lagrange multiplier (FD-DLM) for the discretization of elliptic interface and fluid-structure interaction problems. Indeed, these problems share the common feature that the domain is divided into two regions separated by an interface. In the case of elliptic interface problems this is due to the presence of discontinuous coefficients, while fluid-structure interactions involve different types of equations, corresponding to constitutive models for fluid and solid. This survey paper aims to collect, in a unified framework, the main properties of this approach, including the well-posedness and stability of the solution, convergence of the discrete scheme and computational aspects. The main idea of our approach is to extend one domain into the other. The two domains are then discretized by fixed independent meshes. In particular, in case of fluid-structure interactions, the fluid dynamics is described in Eulerian framework on the extended domain, while the deformation of the solid body is represented in Lagrangian setting by considering a reference domain which is mapped, at each time instant, into the actual configuration. In order to impose that the solution in the fictitious region coincides with the solution in the immersed domain, we consider a Lagrange multiplier and a coupling term is added to the model. We emphasize that all computations are done on fixed domains, so that meshes are generated only once. The paper is organized into two parts. In Section 1 , we focus on elliptic interface problems and discuss in some details the main features of our approach, which will be adapted later on for fluid-structure interactions. After presenting the continuous and discrete problems, we describe how to deal with the coupling term [ 38 , 39 ] . Indeed, the construction of its finite element counterpart requires integration over non-matching grids, which is a challenging aspect of several immersed methods [ 40 , 41 , 42 , 43 , 44 ] . Our formulation turns out to be quite robust from this point of view since, in some situations, optimal results can be achieved even if the coupling term is computed in approximate way. Moreover, the presence of small cut cells does not affect the stability and conditioning. Next, we discuss the design of effective block preconditioners, whose action is performed thorough direct inversion [ 45 ] or by employing multigrid algorithms [ 10 ] . Another issue is related to the fact that the solution of the interface problem can present singularities along the interface. A way to achieve optimal convergence properties is to employ adaptive mesh refinement based on suitable error indicators. This crucial aspect has been previously analyzed in e.g. [ 46 , 47 , 48 , 49 ] for other numerical schemes. Thus, we report the a posteriori error analysis presented in [ 50 ] for our FD-DLM formulation. The second section of the present paper extends the FD-DLM formulation and the related techniques to fluid-structure interaction problems, where an incompressible structure made of viscous hyper-elastic material is immersed in an incompressible Newtonian fluid (see [ 37 ] and references therein). In this case, the regions occupied by the fluid and the solid evolve in time and their configuration is itself an unknown of the problem. The actual position of the solid is obtained by mapping the reference domain through the deformation unknown: this gives further difficulties for the computation of the so-called coupling terms. Viscous hyper-elastic materials are useful when dealing with several applications: we mention, for instance, blood flow in heart [ 3 , 51 ] and vessels [ 52 , 53 , 4 ] , and flapping flexible filaments in soap films [ 54 , 55 ] . This choice of solid constitutive law is relevant for our formulation since it allows to separate the viscous and elastic parts of the Cauchy stress tensor so that we can treat each of them in the appropriate Eulerian and Lagrangian frameworks, respectively. Both