FlowVision-HPC combined with Abaqus offers unique capabilities for solving heavily coupled fluid-structure-interaction problems. The implementation is based on the SGGR (Sub-Grid-Geometry-resolution) method used in FlowVision for grid generation. The SGGR method provides natural link between CFD grid and FEA mesh. The process is fully automatic and can involve geometry of arbitrary complexity. The MPM (Multi-Physics-Manager) controls both FlowVision and ABAQUS during the simulation process. The data is exchanged using direct coupling providing optimal data transfer rates and accurate data exchange. Direct coupling allows using parallel processing both on fluid (FlowVision) as structure (Abaqus) side resulting in realistic simulations times. Most complex FSI problems can be solved over the night (e.g. hydroplaning of automotive tire). Simple set-up with and possibility to use not matched meshes (fluid-structure) allows solving the most complex fluid structure interaction problems involving highly deformable and moving structures. |  | Sub-Grid Geometry Resolution (SGGR) method FlowVision-HPC adaptive grid generator is fully automatic and operates directly on 3D CAD data. A CAD model, defined by 3D surfaces is imported to FlowVision-HPC and surrounded by a (rectangular) Cartesian mesh. The computation domain is defined as result of the Boolean subtraction of the CAD model exterior boundaries and the Cartesian CFD mesh. On the boundary the original hexahedral CFD cells are trimmed by the initial CAD geometry in such a way that the exact geometry shape is maintained. The flow equations in the boundary cells are specially adapted to approximate the boundary conditions and take into account the cell shape. | 
| Original Cartesian hexahedral cell trimmed by the CAD geometry forms a boundary cell with arbitrary form. The boundary cell reflects exactly the CAD model shape without any simplification. SGGR method is suitable for catching fluid characteristics with small amount of cells. |
Natural link between CFD and FEA systems
When the boundary of the computational domain is represented by the FE mesh, the link between the CFD grid and FE mesh is automatically established. FE mesh is matched with fluid cells allowing precise load transfer between fluid and structure. In this way heavily coupled multi-physic problems as the fluid-structure interaction can be efficiently and accurately solved. |  | FE model represented by 3D mesh submerged into CFD computational domain. Natural link between FE elements and CFD cells is established automatically including not matching meshes. |
Multi-Physics Manager (MPM) for Fluid-Structure Interaction (FSI)
FSI simulation takes into account large structural deformations and moving bodies caused by the hydrodynamic loads imposed on the boundaries of the computation domain are (e.g., a heart valve operation).
The MPM module control and synchronize the operation of both FlowVision-HPC and ABAQUS (explicit or implicit).
- Starts and stops FlowVision-HPC and ABAQUS solvers,
- Controls bi-directional data transfer (using sockets) at pre-defined time step interval
- Enables control of different computers platforms (Windows, Linux )
- Monitors selected set of parameters through graphical window.
The FSI calculation proceeds in a fully automatic regime, monitoring the crucial characteristics through the MPM graphical window, enabling the user to analyze the solution in the course of the on-going calculations. |  | Direct 2-way coupling interface between Abaqus and FlowVision, explicit time stepping procedure (at FSI exchange time interval), notice different time steps on FlowVision and Abaqus side.
|
Moving bodies and free surfaces
The FSI simulation supports presents of moving bodies (user defined rotation and translation) and trucking of free surfaces analysis (e.g., car tire aquaplaning). The grid generation process is fully automatic, with rebuilding of the computational grid at each time step the data is exchanged with Abaqus calculating the structural deformations. |  |  | | Moving body | Moving body and free surface
|
Parallel calculations
FlowVision-HPC parallel solver supports multi-core, multi-processor and mixed configurations. It operates on Microsoft Windows or Linux operating systems. The parallel solver is optimized to work on parallel computers with shared, distributed, and combined memory (the latter implies Non Uniform Memory Access architecture). The solver implementation is highly scalable and offers possibilities to speed up the most complex simulations: typical simulation requiring a week to complete on a single CPU runs one day on a multi-processor cluster. |  | Aerodynamic influence on F1 cars following each other |  |  | | Simulation 12.000.000 cells | Automatic domain decomposition on 16-processors |  Scalability |
Integrated Application
The FlowVision HPC is an integrated application consisting of pre/post processor and a parallel solver. The solver can be configured to operate on Microsoft Windows or Linux clusters (32/64 bit). The pre/postprocessor runs only under Microsoft Windows offering extensive OpenGL based graphics capability for post processing and on-line monitoring of the results during on-going simulation. The system can be configured for multiple users accessing one common solver, multiple solvers (utilizing optimally large cluster configuration) and remote access (users accessing solver from remote location). |
|
Products