Nonlinear unsteady aerodynamic modeling at high angles of attack is critical for high-precision control law design of modern aircraft. Current modeling approaches primarily fall into two categories: expert’s experience-informed models and data-driven models. The accuracy of expert’s experience-informed models is limited by the a priori expression terms. Data-driven model has a strong nonlinear mapping ability, but its performance depends on sample size and has insufficient generalization ability in small samples. To address these limitations, this paper proposes a physics-informed data-driven modeling framework, in which a Long Short-Term Memory (LSTM) neural network is trained to reconstruct the a priori expression terms in the differential equation model. While retaining the physical mechanism of the expert’s experience-informed model, the data-driven method is utilized to enhance the prediction accuracy of the model. To validate the model, this paper conducts missile single-degree-of-freedom pitching and fighter two-degree-of-freedom aerodynamics modeling at high angles of attack. Results show that, compared to a traditional differential equation model, a standalone LSTM network, and a hybrid multi-fidelity neural network, the proposed method achieves superior accuracy and generalizability in both cases, providing an effective solution for modeling complex nonlinear unsteady aerodynamic behaviors.

Low drag and low sonic-boom design technologies both play significant roles in the re-introduction and sustainable commercial operations of next-generation supersonic civil aircraft. Although the potential of natural laminar flow wing to reduce the drag of supersonic civil aircraft has been verified, further research is still needed to achieve laminar flow wing design under the constraint of low sonic-boom intensity. Considering that the laminar flow region of the wing is determined by the wing pressure distribution, the influence of wing pressure distribution on the sonic-boom characteristics of the aircraft is first investigated. A method for supersonic natural laminar flow wing design considering low sonic-boom intensity requirement of the aircraft is then proposed. Firstly, the influence of wing pressure distribution on the equivalent areas and sonic-boom waveform of the aircraft is analyzed. Results indicate that on the low-boom configuration, changing the wing pressure distribution mainly leads to changes in the distribution of axial lift of the aircraft. In addition, the original configuration's shock-expansion waves that feature low-sonic boom characteristics are also affected, leading to increase of intensity of ground sonic boom. Based on these discoveries, a three-step method for supersonic natural laminar flow wing design is proposed considering the sonic-boom constraint of the aircraft. The first step is to conduct low boom design to determine the layout of the configuration and fuselage shape. The second step is to conduct multiple rounds of pressure gradient inverse design on the wing to obtain the pressure distribution required by natural laminar flow. The third step is to carry out multiple rounds of low sonic-boom design on the fuselage, horizontal tail and vertical tail to compensate for changes in sonic boom characteristics caused by changes in wing pressure distribution. The proposed method is applied to achieve natural laminar flow on a wing of a 30 t low boom supersonic civil aircraft configuration. The designed configuration not only maintains low sonic-boom characteristics (sonic boom intensity equals 81.7 PLdB), but also achieves 33% natural laminar flow on the upper surface of the wing and a 5.2% reduction in friction drag compared to the baseline configuration. The results verify the effectiveness of the proposed method.

To shorten the response time of air-to-air missile and realize rapid attack on backward target in close combat, a rapid backward-launching method of short-range air-to-air missile is developed, which takes advantage of the static instability of missile at large/very large angle of attack, auxiliary jet direct force control and active rudder deflection control. To simulate the influence of preset rudder deflection, jet direct force and active rudder deflection control strategy in the new launching method, a high-precision computational fluid dynamics/flight dynamics numerical simulation method coupling with active rudder deflection control strategy is developed based on the dynamic Chimera grid. Numerical simulation results show that the proposed method can achieve rapid reorientation of the missile in 1.53 s, and maintain a stable and controllable attitude at the end of the flipping phase. The new method can reduce the response time of over-the-shoulder launch to about 1/10 of the traditional method.

Efficient mesh deformation methods can significantly improve computational efficiency in fluid structure interaction numerical simulation and aerodynamic shape optimization based on the high-fidelity CFD method. The mesh deformation method based on original Radial Basis Functions (RBF) can result in a significant increase in computational cost when the number of deformation control points increases. The method of reducing RBF modeling calculation time by reducing deformation control points can result in a loss of fitting accuracy. Since the existing deformation methods based on RBF cannot guarantee the computational efficiency and fitting accuracy simultaneously, this paper proposes a Spatially-Nested Radial Basis Function (SN-RBF) model and develops an efficient mesh deformation method based on SN-RBF. The proposed mesh deformation method maintains the accuracy of mesh deformation, while significantly improving the efficiency of mesh deformation. The spatially-nested radial basis function model utilizes multiple spatially overlapping sub models to replace the radial basis function model which has large number of deformation control points, greatly reducing the modeling time of RBF method in mesh deformation. Considering the good parallelism of the proposed method, the strategies for parallel modeling and parallel mesh deformation based on this method have been adopted, further improving the efficiency of mesh deformation. The test cases show that the more modeling sample points, the more significant the improvement in grid deformation efficiency. For the CRM wing-body configuration case, the maximum improvement in modeling efficiency is 16 947 times, and the maximum improvement in mesh deformation efficiency is 5 218 times.