Informationen für

Direkt zu


 

Coupling Smoothed Particle Hydrodynamics and Finite Volume methods for incompressible flow simulations

Coupling Smoothed Particle Hydrodynamics and Finite Volume methods for incompressible flow simulations

Datum: 27. April 2017 16:00

Ort: Allmandring 3, Seminarraum 1.079

Veranstalter: ICP

Coupling Smoothed Particle Hydrodynamics and Finite Volume methods for incompressible flow simulations

Enrico Napoli, Università di Palermo Viale delle Scienze

Finite Volumes (FV) and Smoothed Particle Hydrodynamics (SPH) are well-established numerical methods for the simulation of compressible and incompressible fluid flows, widely used in several different engineering and scientific applications.

Although grid-based methods like FV have achieved large robustness and efficiency in the Computational Fluid Dynamics, several difficulties still remain when dealing with highly complex geometries, solid-fluid moving interfaces or rapidly evolving free-surface flows (e.g. wave breaking). In the last decade thus a growing interest has developed in the use of mesh-less methods, among which SPH is probably one of the most used.

Since SPH is still less computationally efficient than FV, a coupled approach can be used in order to make use of the specific advantages of both methods.

The approach presented in the talk is based on the partitioning of the computational domain into a portion discretized with a structured grid of hexahedral elements (the FV-domain) and a portion filled with Lagrangian particles (the SPH-domain), separated by an interface made of triangular elements. A smooth transition between the solutions in the FV and SPH regions is guaranteed by the introduction of a layer of grid cells in the SPH-domain and of a band of virtual particles in the FV one, on which the hydrodynamic variables are obtained through suitable interpolation procedures from the local solutions. Several test cases are used in order to test the efficiency and accuracy of the coupled approach, showing that a significant reduction in the computational efforts can be achieved with respect to the standard SPH method.

<< Zurück