On the development of an efficient numerical ice tank for the simulation of fluid-ship-rigid-ice interactions on graphics processing units

•A numerical ice tank for the simulation of fluid-ship-ice interactions is presented.•The model considers hydrodynamics, ice-floe-collisions and friction effects.•An efficient GPU acceleration allows for simulations within just a couple of hours. This paper reports on the adaptation of a Lattice Bol...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Computers & fluids 2017-09, Vol.155, p.22-32
Hauptverfasser: Janßen, Christian F., Mierke, Dennis, Rung, Thomas
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:•A numerical ice tank for the simulation of fluid-ship-ice interactions is presented.•The model considers hydrodynamics, ice-floe-collisions and friction effects.•An efficient GPU acceleration allows for simulations within just a couple of hours. This paper reports on the adaptation of a Lattice Boltzmann based free surface flow solver to the simulation of complex fluid-ship-ice interactions in marine engineering. The analysis is restricted to the interaction of already broken ice floes and the ship hull, aiming at the optimization of a ship hull’s capability to clear the ice and keep it away from the propulsion device. The ice floes and the ship hull are treated as rigid bodies. In order to model the dynamics of the colliding rigid multi-body systems, a coupling of the flow solver to the Open Dynamics Engine (ODE) is established. The basic methodology and initial validation of the fluid-structure coupling is presented. Then, basic validations of the employed collision and friction models are given, particularly focusing on interacting surface triangle meshes that later serve to describe the ice floes. Finally, a three-dimensional validation case shows that the ship-fluid-rigid-ice interaction forces agree well with available reference data. Apart from the numerical coupling, performance has to be addressed. The employed flow solver elbe uses graphics processing units (GPUs) to accelerate the numerical calculations. In order to make the GPU performance accessible to colliding multi-body systems, a careful and tailor-made implementation is presented in the paper. The resulting optimized elbe-ODE solver allows for the investigation of three-dimensional fluid-ship-ice interactions in a very competitive computational time, on off-the-shelf desktop hardware.
ISSN:0045-7930
1879-0747
DOI:10.1016/j.compfluid.2017.05.006