Parallel Processing Strategy for Solving the Thermal-Mechanical Coupled Problem Applied to a 4D System using the Finite Element Method
Abstract
We propose a high performance computing
strategy (HPC) to simulate the deformation of a solid
body through time as a consequence of the internal
forces provoked by its temperature change, using the
Finite Element Method (FEM). The program finds a
solution of a multi-physics problem, solving the heat
diffusion problem and the linear strain problem for
homogeneous solids at each time step, exchanging
information between both solutions to simulate the
material distortion. The HPC strategy approach
parallelizes vector and matrix operations as well as
system equation solvers. The tests were realized over
a model simulating a car braking system (a rotating disk
velocity decreased by friction). Then we performed a
quantitative analysis of stress, strain and temperature in
some points of the geometry, and a qualitative analysis
to show some visualizations of the simulation.
strategy (HPC) to simulate the deformation of a solid
body through time as a consequence of the internal
forces provoked by its temperature change, using the
Finite Element Method (FEM). The program finds a
solution of a multi-physics problem, solving the heat
diffusion problem and the linear strain problem for
homogeneous solids at each time step, exchanging
information between both solutions to simulate the
material distortion. The HPC strategy approach
parallelizes vector and matrix operations as well as
system equation solvers. The tests were realized over
a model simulating a car braking system (a rotating disk
velocity decreased by friction). Then we performed a
quantitative analysis of stress, strain and temperature in
some points of the geometry, and a qualitative analysis
to show some visualizations of the simulation.
Keywords
Parallel computing, HPC, simulation, FEM, finite element, thermal-mechanical coupled problem, dynamic analysis, heat distortion.