peices / mto

Parallel solver for thermal-fluid-structural topology optimization on structured grids.

MTO——Multiphysics Topology Optimization

Description

The code presented is a parallel solver for multi-physics topology optimization on structured grids. Density-based method is adopted for solving 2D or 3D thermal-fluid-structural topology optimization problems based on OpenFOAM.

Installation

Before running this solver, following softwares are needed.
(a) OpenFOAM 5.0
(b) PETSc 3.12
(c) swak4Foam

Parallel version of MMA written by Aage is also needed for updating the design variables (https://github.com/topopt/TopOpt_in_PETSc). We have changed this code to apply it in our solver. For simplicity, the C++ script have been compiled into a dynamic library (libMMA_yu.so), users should put this file in FOAM_USER_LIBBIN after installing OpenFOAM.

Run the solver

After finishing above works, users should run wmake in the src folder, and then run blockMesh and decomposePar in the app folder. Finally, the multi-physics topology optimization solver can run in the app folder by mpirun –n 20 MTO_solid –parallel, where 20 is the default number of cores for parallel computing.

Postplot

After the optimization, users should run reconstructPar in the app folder and then run paraFoam to view the results with Paraview.

Solid displacement problem-2D

Solid displacement problem-3D

Thermal-fluid problem-2D

Thermal-fluid problem-3D