This SBIR Phase I project proposes to use artificial intelligence for knowledge extraction from large amounts of CFD solution data to make engineering decisions in real time. With rapidly growing computing power, the scales of problems solved by CFD and thereby the resulting CFD solution data become larger and larger. It becomes more and more challenging to manage those ever-growing CFD solution data. Furthermore, the current computing power still does not allow the direct CFD computations in the real-time simulation, design and optimization environments. The proposed effort aims to address this issue by applying artificial intelligence technique to train a surrogate model out from a large amount of CFD solution data. The obtained surrogate model can then be used in the real-time simulation, design and optimization environments to capture all the fidelity of CFD results.
Various programs and projects of NASA missions use CFD for advanced aircraft concepts, launch vehicle design, and planetary entry vehicles. The developed technology will facilitate the utilization of those simulation data in the real-time simulation, design and optimization environments and significantly improve the accuracy of those designs, optimizations and simulations by Aeronautics Research Mission Directorate (ARMD) and Human Exploration Operations Mission Directorate (HEOMD).
Design engineers in various industries can use the developed technology to train their required surrogate models out from large amounts of simulation data and use the obtained surrogate models to improve their designs, optimizations and simulations. Business can further benefit from the developed technology through reduction of design cycle.