Rapid Performance Analysis Method for Geometrically Adjustable Air Intakes Based on Surrogate Models

To address the unclear coordinated control mechanism of the dual wedge angles in a two-dimensional mixed-compression hypersonic variable-geometry inlet, as well as the high computational cost of numerical simulations, this study investigates rapid aerodynamic performance prediction and parameter-control mechanisms for the inlet. The freestream Mach number Ma = 6 is selected as the design point. A parametric inlet configuration is established using a three-shock external compression system combined with a Bézier-curve internal compression section. An automated simulation workflow is then developed to construct a dual-parameter dataset based on the deflection angles δ₁ and δ₂ of the first two wedge surfaces. Under the design condition, the inlet achieves a total pressure recovery coefficient of 48.44% and a mass flow coefficient of 98.97%, verifying the basic aerodynamic performance of the designed configuration.On this basis, the off-design condition of Ma = 5.5 is selected as a typical slice. An indirect prediction model for the outlet flow field and a direct prediction model for aerodynamic performance are constructed, and RBF, Kriging, and two-dimensional linear interpolation models are introduced for comparison. The results show that the direct prediction model can reduce error accumulation caused by secondary conversion of outlet flow-field parameters, and it performs better in predicting the flow distortion index and static pressure ratio. The RBF model shows certain advantages in predicting the total pressure recovery coefficient. The dual-wedge-angle control results indicate that δ₂ has a more significant influence on outlet flow distortion, while δ₁ mainly affects shock intensity and mass flow capture capability. The high-value regions of the total pressure recovery coefficient and static pressure ratio are mainly concentrated within δ₁∈4.3°, 8.3° and δ₂∈6.33°, 9.33°.The results demonstrate that the analysis method based on parametric CFD data and surrogate models can achieve rapid performance prediction for variable-geometry inlets at a fixed Mach number. It also provides a data basis and methodological reference for the subsequent development of wide-speed-range multi-parameter surrogate models.