Towards high-fidelity aerospace design in the age of extreme scale supercomputing (Invited)

Abstract

Aerospace engineers have always been leading in the field of design using computational simulations. We design vehicles that go fast, fly high, and operate in extreme conditions. Because computation can help reduce expensive tests that physically simulate these extreme conditions, we benefit from computational simulation and design. This is why successful aerospace companies and agencies, such as NASA, have always been leading and investing in research and development of computational tools for simulation and design. Aerospace engineers have been leading contributors to computational tools and technology. NASA, for example, led the ICASE program which produced significant advance Computational Fluid Dynamics (CFD) methods. Advance in CFD revolutionized many aspects of aerospace engineering, including how commercial jets are designed. It helped reduce the cost of testing wing designs in wind tunnels when designing new airplanes. For example, Boeing had to test 77 wing designs when designing the 757 in late 1970s. By mid 1990s, it only needed to test 11 wing designs when designing the 737-NG. Our investment in computational simulation and design, including CFD, has paid back. Despite decades of progress, computational simulation and design, including CFD, is far from reaching its limit. High fidelity CFD simulations, in particular, may change the world of aerospace engineering in the near future. In 2014, NASA completed the CFD Vision 2030 Study: A Path to Revolutionary Computational Aerosciences. The Vision 2030 study highlighted the enormous potential of simulations and design based on high-fidelity CFD. High fidelity CFD includes Large Eddy Simulations (LES), wall-modeled LES, and hybrid RANS-LES. These unsteady simulations can capture more aerospace-relevant flow physics, work for complex geometries, and are becoming increasingly reasonable in terms of computational cost. Some of those leading the CFD Vision 2030 study predict that aerospace engineers will be using these high-fidelity simulations in the design, almost in real time. Computational simulations would be most useful if they are reliable enough to be trusted, and fast enough to be applied in real-time design. If this is achieved, it might become as easy to design and test aerospace components virtually as to design and test as bicycle components. They would not only accelerate aerospace design and development cycles, reduce the cost, but also may spawn a culture of unprecedented innovation in aerospace engineering.