Project A1 (finished)
Project A1 - Armin Iske, Stephan Olbrich and Thomas Rung: Adaptive Kernel-Based Particle Simulation Approaches for Large-Scale High-Performance Computations
Scientific Background and Motivation
Contact problems between different materials which are subjected to large deformations are particularly challenging for simulation approaches. An example refers to multi-physics fluid-structure-soil-interaction, which is of significance for many geotechnical, aeronautical and marine engineering problems. Relevant applications are related to e.g. harbour flows, mudslides & granular landslides, off-shore engineering, aircraft crash simulations, tsunamis or volcano eruptions. A characteristic feature off such problems is, that experimental investigations are hardly possible or suffer from severe scale effects. Thus, the analysis is either based on empirical data or simulations. Due to the large deformations of inherently different interacting materials, meshless Lagrangian kernel-based simulation approaches, e.g. Smoothed-Particle-Hydrodynamics (SPH), which discretize the various involved continua by means of finite particles rather then meshing the domain, are an ideal candidate to investigate such problems. The simulations usually imply large physical domains, complex geometries and the necessity to resolve local details. This typically yields particle numbers of the order of many 10 Mio. or many 100 Mio. particles and hinders the application of particle methods for large-scale problems. Accordingly, a pressing need for advanced strategies to increase the computational performance exists.
Aims and Objectives
The projects aims to introduce high-performance computing aspects as well as refined numerical approximation and modeling techniques to SPH in a joint effort between engineers from TUHH and mathematicians and computer scientists from UniHH. Emphasis is put on geotechnical and harbour flow simulations in conjunction with the parallel SPH model GADGETH2O. The research will focus on four major methodical aspects to advance the numerical capabilities of particle methods based upon harmonised contributions from experts in the area of high-performance computing (UniHH/TUHH), kernel-based numerical approximation (UniHH) and particle-based computational engineering (TUHH).
PhD student: Adeleke Bankole
Adeleke Bankole successfully defended his PhD on 1 February 2017.
Contact: Prof. Dr. Armin Iske, Department Mathematik, email: armin.iske(armin.iske"AT"uni-hamburg.de)
Bankole, A.O., Dumbser, M., Iske, A., Rung, T.: A meshfree semi-implicit Smoothed-Particle-Hydrodynamics method for free surface flow. Meshfree Methods for Partial Differential Equations VIII, M. Griebel and M.A. Schweitzer (eds.), Springer LNCSE, 115, 35-52. 2017.