Fluid-structure interactions (FSI) are invited by the coupling between a fluid flow and the movement or deformation of a structure. This type of interactions can be found in a great amount of industrial applications, such as the design of aircrafts, the study of aerolastic instabilities of bridges and buildings, the optimisation of wind turbines or in bio-fluid mechanics research like heart valve development, among others. In those cases, the understanding of the fluid flow behaviour or the awareness of the dynamic action of the structure can be essential to improve engineering designs and optimisations. Therefore, accompanied by a great enhance in computational resources, in the last years the numerical simulation has become an important tool to predict FSI applications, even in turbulent regimes. As a consequence, a partitioned FSI solution coupled algorithm focused on the study of the flow phenomena in these type of problems has been implemented.

To attempt the numerical simulation of these phenomena, a robust, accurate and fast method which redistributes the computational mesh in accordance with the movement of the domain is needed. In general terms, the case is formulated as a three-field problem: the fluid domain, the structure domain and the dynamic mesh. Thus, each subdomain is solved separatly and a coupling criterion manages the transfer of information between domains. This staggered procedure allows to use the desired or most case-suitable solver for the fluid, the structure and the moving mesh, independently of the other domains.

In the fluid domain, the incompressible Navier-Stokes equations concerning dynamic mesh are solved by means of a three-dimensional explicit finite volume fractional-step algorithm formulated in a second-order, conservative and collocated unstructured grid arrangement. To compute the grid velocity considering the conservation principle and avoiding the loss of mass and momentum, the space conservation law is applied. Large-eddy simulations (LES) are performed for turbulent flows.

Referring to the dynamic mesh technique, a radial basis function (RBF) method is involved. The RBF approximation is becoming increasingly popular as a method the interpolate scattered data. In our context, the basic idea is to solve an interpolation problem in order to transfer the known displacements of a set of control points to the rest of grid points.

Different strategies have been carried out in the solid domain; for instance, a simplified model based on a steady state deformation law or a routine implying modal analysis of the solid dynamic action. Moreover, a finite volume computational structural dynamics (CSD) solver has been elaborated to solve the transient and dynamic elastic deformation of a St. Venant-Kirchhof material.