INTELLIGENT LIGHT™ has collaborated with a major engine manufacturer in Germany in a joint research project funded by the DLR to advance computational simulation capabilities for complete aero engines. Kombyne was called upon for two aspects of this effort, in transit post-processing within the company’s Hydra, an internally developed CFD solver used by Rolls Royce, and steering of the code during runtime.
Also Read: IBM Takes on the 5G Era with Acquisition of Sentaca
Rolls Royce Deutschland engineer, Marcus Meyer Ph.D., presented a landmark paper at AIAA SciTech 2022 (January, San Diego, CA). In it, Dr. Meyer described the use of “Computational Steering,” where the Rolls-Royce flow solver Hydra was instrumented with the Kombyne™ software to enable on-the-fly changes of the computational setup for creating compressor maps.
According to Dr. Meyer’s paper, compared to conventional CFD techniques, the steering approach “removes the need to submit separate jobs for each operating point and treats the CFD simulation like a virtual experiment where operating conditions can be changed while the virtual engine is running. File I/O and initialization happens only once, rather than repeated for different jobs.”
The steering of a simulation, applied to solvers running on high-performance computing (HPC) platforms, increases overall efficiency. According to Steve Legensky, President & CTO of Intelligent Light (and co-author of the AIAA paper), “Getting many results in less time saves money as well as the valuable time of engineers. When Kombyne™ is integrated into a solver code, the simulation can be treated more like a flight test or wind tunnel ensemble.”
Interested parties should download Kombyne™ Lite directly at kombyne.net. Installation is extremely easy for Kombyne™ using prebuilt modules for OVERFLOW2, Hydra, PyFR, OpenFOAM, and more every day. For other solvers or inhouse solvers, a small adapter is used to link Kombyne™ into the solver. Kombyne™ provides native APIs in FORTRAN, C/C++ and python, and is scalable via MPI. In batch operation, a simple parameter datafile sets up the functions to be performed during simulation.