Intelligent Aerodynamics (IA), a combination of artificial intelligence (AI) and aerodynamics that integrates the fourth research paradigm (Big Data), is growing to be a distinct cross-discipline. This paper first introduces IA's concept and connotation and sketches the main research methods. Secondly, the research progress and the work carried out in China Aerodynamics Research and Development Center (CARDC) are reviewed regarding AI-enhanced flow prediction, transition/turbulence modeling, multi-source data fusion, aerodynamic/aerothermal modeling, and extraction of critical flow-field information. Finally, the future research work and development trends of IA are prospected.

Modulation of dispersed particles on the carrying fluid is a common phenomenon in aeronautic and aerospace engineering, including the ablation particle-laden boundary layer, the particles of solid fuel combustion in the rocket engine and the droplets of liquid fuel. Although theoretical, numerical and experimental studies on the particle laden incompressible flow have been well established, the particle laden compressible flow is distinctly different. In view of this, it is necessary to conduct a theoretical study on the dynamics and thermodynamics of particle laden compressible flow. We established the theoretical equations, and discovered the unique particle modulation parameters which are different from those of the incompressible counterpart. By considering the extreme mechanical environmental condition, the particle dynamic and thermodynamic equations have been derived and coupled to the governing equations of the carrying phase. It is found that, even with complicate multi-physics effects, the pathway of particle modulation on the compressible flow is deterministic, i.e., the carrying fluid momentum is modulated by the drag force of particles, while the energy is exchanged due to the work done by the drag force as well as the thermal convection. The present study lays a theoretical foundation for developing the particle solver for compressible flow.

A direct numerical simulation of hypersonic Shock wave and Turbulent Boundary Layer Interaction(STBLI) at Mach 6.0 on a sharp 7° half-angle circular cone/flare configuration at zero angle of attack is performed. The flare angle is 34° and the momentum thickness Reynolds number based on the incoming turbulent boundary layer on the sharp circular cone is Re_θ = 2506. It is found that the mean flow is separated and the separation bubble occurring near the corner exhibits unsteadiness. The Reynolds analogy factor changes dramatically across the interaction, and varies between 1.06 and 1.27 in the downstream region, while the QP85 scaling factor has a nearly constant value of 0.5 across the interaction. The evolution of the reattached boundary layer is characterized in terms of the mean profiles, the Reynolds stress components, the anisotropy tensor and the turbulence kinetic energy. It is argued that the recovery is incomplete and the near-wall asymptotic behavior does not occur for the hypersonic interaction. In addition, mean skin friction decomposition in an axisymmetric turbulent boundary layer is carried out for the first time. Downstream of the interaction, the contributions of transverse curvature and body divergence are negligible, whereas the positive contribution associated with the turbulence kinetic energy production and the negative spatial-growth contribution are dominant. Based on scale decomposition, the positive contribution is further divided into terms with different spanwise length scales. The negative contribution is analyzed by comparing the convective term, the streamwise-heterogeneity term and the pressure gradient term.

When flying at extremely high Mach numbers at low and medium altitudes, the new hypersonic vehicles will encounter a complicated flow environment, i.e. the coupling between turbulence and chemical non-equilibrium. However, there are only very few related studies so far. In this work, direct numerical simulations (DNS) were performed for the high-temperature and non-equilibrium turbulent boundary layer after the leading shock of a cone using two different gas models, i.e. the calorically perfect gas model and the chemical non-equilibrium gas model. The Walz equation, the strong Reynolds analogy (SRA) relation, the turbulent kinetic energy production-dissipation mechanism and the compressibility effects were compared and analyzed. The results show that the weak compressibility hypothesis still holds in the chemical non-equilibrium turbulent boundary layer. When using the relation between the non-dimensional recovery enthalpy and the velocity, effects of the Mach number, the chemical reaction, etc. can be eliminated, and the prediction agrees well with the DNS result. GHSRA (Generalized Huang’s SRA) is appropriate to describe the relation between the temperature fluctuation and the velocity fluctuation. The ‘semi-local’ scaling can collapse the distribution profiles of the turbulent kinetic energy budgets under different flow conditions. The compressibility effect caused by the chemical non-equilibrium and the high Mach number is limited.