Computational fluid dynamics (CFD) has remained a vital tool for aircraft design assessment ever since its inception. Its use has already brought a paradigm shift in the aircraft design process.
However, since past decade, the advancements in this technology has reached a stagnation point as RANS methods have become a high-fidelity method of choice as it permits the use of large meshes, complex geometries and more number of runs. As such, there has been little contribution in the past decade to improve the algorithms.
The CFD technology at present has matured enough as hardware costs have reduced to a great extent. However, the present matured algorithms are found to be poorly positioned to sync with the rapidly changing HPC architectures.
In a recent report on CFD Vision 2030, NASA states that the present CFD capabilities must undergo a revolutionary change, in order to meet future aerodynamic goals. The problem with present CFD techniques that simulate the airflow around the aircraft is that it is designed to solve traditional tube and wing configurations. Even then, these simulations based on RANS methods are confined to small region of the flight envelope, limiting CFD effectiveness through all phases of the flight. Modern aircrafts on the contrary, may look entirely different; designed to achieve most prevailing aero science goals such as reduced emissions, fuel consumption and noise.
Modern aircrafts may have longer and skinnier wings held up using trusses; hulls may be broad and flat with pointed nose. Jet engines may be mounted differently or the wing and body might be blended into a seamless contour. As such, it is important to understand the physics of these new variables affecting airflow during all phases of flight, requiring a strong emphasis on physics based predictive modeling.
Adequate Prediction of Turbulent Flows:
One of the important simulation capabilities required for future aerodynamic analysis will be the ability to adequately predict turbulent flows with boundary layer transition and flow separation. There is already a good amount of research conducted on post-separation physics with the use of hybrid RANS-LES models.
However, pre-separation physics is still being provided by RANS models, which haven’t seen any development from past two decades mainly due to the changing nature of the interface between RANS and LES regions. Thus, a seamless, automatic RANS-to-LES transition in the boundary layer is highly required for future aircraft simulations. As per the survey conducted by NASA, majority of the scientific community believes that the expanded use of RANS-LES methods will be the norm in 2030.
Utilization of High Performance Computing (HPC):
The simulation capabilities will also rely heavily on the effective utilization of High-Performance Computing (HPC). Supercomputers having the capability of 30 exaflops (floating-point operations per second) as envisioned by NASA, will have enough power to simulate turbulent flow around a complete aircraft, in order to perform multidisciplinary design optimizations of future aircraft configurations.
However, the radical technologies such as quantum computing, low-power memory and parallel molecular computing required to develop powerful supercomputers of tomorrow will require different algorithms and software infrastructures. Thus, an advanced programming environment will be required for future HPC ecosystem, requiring the development of highly scalable, error-free algorithms, software architectures and programming environments.
Robust Multi-Physics Simulations:
Future aerospace engineering problems will also require seamless interface between CFD and other high-fidelity analyses such as vibrations, structures, reacting flows, radiation and dynamics & control. At present, multi-physics analyses are still in its embryonic stage, limiting the accuracy and stability of the simulations.
However, the need in the future to simulate complete aerospace systems, multi-physics is bound to become a norm rather than exception. Multi-physics capabilities will be required to be robust enough with rapid turnaround and the ability to provide uncertainty information along with the capability to leverage future HPC resources.
Source: http://www10.mcadcafe.com/nbc/articles/1/1326780/Need-Revolutionize-CFD-Capabilities-Future-Aircraft-Simulations