Simulation Framework |
Verification & Validation
Integrated Multicode Simulation Framework
CHIMPS, Coupler for High-performance Integrated Multi-Physics Simulations
CITS overarching problem is the simulation of the complete aero-thermodynamic flow path through a jet engine. CITS simulation environment is based on Python and built around the Center's flagship codes, CDP (Large Eddy Simulation approach) and TFLO (based on Reynolds-averaged Navier-Stokes equations). It also includes a set of general-purpose interpolation and communication libraries (CHIMPS, Coupler for High-performance Integrated Multi-Physics Simulations).
Basic concept: a parallel, scalable, distributed module that automatically takes care of mesh / solution interpolations to exchange information between CFD solvers, Structral dynamics codes, etc. The code-interfacing is designed around a distributed "interpolation server". Participating codes "register" the required data and their locations and "provide" their local mesh and solutions to the other components in the integrated environment. Information exchange (volume/overlap, surface/minimum distance and failsafe) happens transparently to the user through straightforward API.
The advantages of this approach are:
- User can manage multiple codes (or multiple instances of several codes) in truly INTEGRATED simulations without the need to worry about the complicated details of the coupling itself
- All the necessary searches and interpolations (both non-conservative and conservative) can be done with the most advanced, parallel methodologies with very small memory overhead
- Once the coupling problem has been defined (at the level of the master Python script) the search/interpolation procedures can be repeated many times (as the meshes move or move relative to each other)
- Coupler is being constructed to treat a large number of potential interface problems (fluid-fluid, solid-fluid, fluid-acoustics)