Traditional aircraft configuration design faces significant efficiency challenges. This study addresses the critical bottleneck of slow sampling in point cloud diffusion models for generating 3D aerodynamic configurations under multidisciplinary constraints. We introduce the denoising diffusion implicit model (DDIM) acceleration strategy to substantially reduce the required sampling iterations, leveraging its non-Markovian skip-step mechanism without model retraining. Specifically, reducing the sampling steps from 1000 to 50 cuts the generation time by 57.5% (from 30.32 s to 12.89 s), while the average aerodynamic performance relative error increases only from 3.62% to 8.52%. To further optimize the balance between speed, accuracy, and diversity, we propose a novel “deterministic-stochastic” hybrid sampling strategy. This approach dynamically identifies critical timesteps by analyzing the temporal evolution of latent point cloud feature gradients and employs a trained classifier to adaptively modulate the noise strength parameter (η) across regions of varying criticality. Experimental validation demonstrates that the hybrid strategy operating at 50 steps delivers generation time below 15 s, achieves a 76.6% satisfaction rate for Coverage (COV, chamfer distance) below 10%, and attains a 73.3% satisfaction rate for aerodynamic performance error below 10%, outperforming static noise sampling. This work successfully integrates DDIM acceleration with dynamic noise regulation into a point cloud diffusion framework for aircraft configuration generation, effectively overcoming the sampling efficiency hurdle and enabling the rapid production of diverse, constraint-satisfying designs. Future efforts will focus on automating the optimization of the classifier and noise control parameters.

Shape stealth is the primary factor determining the stealth performance of an aircraft. In the shape design of modern military aircraft, both aerodynamic performance and stealth performance need to be comprehensively considered. Aerodynamic stealth optimization is a complex multimodal problem. Given the high computational cost of global optimization and the tendency of gradient optimization to fall into local optima, an optimal design method combining global multimodal optimization for airfoils and gradient optimization for layouts is proposed. To address the insufficient local search capability of the classical multimodal particle swarm optimization algorithm, the fitness-euclidean distance ratio ring topology local-best particle swarm optimization (FER-R3PSO) algorithm is introduced by combining the ring topology particle swarm optimization algorithm with the fitness-euclidean distance particle swarm optimization algorithm, which enhances the local search capability of the classical multimodal particle swarm optimization algorithm. To combine the multimodal search algorithm with surrogate models and reduce the computational burden of using the multimodal algorithm in engineering applications, a surrogate model point addition strategy and a peak extraction method suitable for the multimodal search algorithm are proposed. Function tests, airfoil stealth optimization, and standard aerodynamic examples are used to verify the effectiveness of the multimodal algorithm. The global/gradient coupling design for the flying wing layout is proposed, and aerodynamic stealth optimization based on the adjoint method is carried out using the three-dimensional shape obtained through global multimodal algorithm optimization of the airfoil assembly as the initial stage. The optimization results show that, compared to gradient optimization, the proposed method can achieve a shape with superior aerodynamic and stealth characteristics at a lower computational cost.

The strong and unsteady wind imposes severe challenges to the safe flight and aerodynamic prediction of the fixed-wing aircraft. Traditional aerodynamic models established in the wind-oriented coordinate system have a clear physical meaning but cannot be readily applied to unsteady windy environments. This paper proposes an innovative "neural"-fly aerodynamic modeling method based on deep meta-learning to accurately predict the aerodynamic forces and moments online for fixed-wing aircraft subjected to strong and unsteady wind. Based on variables in a coordinate system relative to the ground, this method decomposes the aerodynamic forces and moments into the sum of polynomial multiplication and constructs the common aerodynamic base functions by a three-step deep meta-learning algorithm using the Generative Adversarial Network. The application of the method for the fixed-wing aircraft F-18 demonstrates that the method can accurately predict the aerodynamic forces and moments under unknown wind conditions, laying a good foundation for real-time aerodynamic modeling.

The aerial refueling hose-drogue exhibits unstable motion because of atmospheric turbulence and other disturbances, increasing the uncertainty and risk of the docking process. An actively stabilized drogue with spinning momentum rings is designed, integrating a pair of bias momentum rings at the drogue cup to achieve pendulous suppressionib and avoid collision between rudders and the refueling pod when recovering the drogue. A dynamic model of the hose-drogue considering rotation of momentum rings is established based on the Lagrange equation where the hose is modeled by a series of pendulums-connected rigid links. Based on this model, the gust and atmospheric turbulence interference on the hose-drogue is analyzed. A pendulous motion suppression strategy based on momentum rings control is proposed to change the angular momentum of the system by adjusting the rotational rate of the momentum rings. The simulation results show that the regulating time is significantly shortened under gust interference, and the pendulous amplitude of the drogue is effectively reduced under turbulence interference, attenuating the high-frequency chattering of the drogue motion and creating conditions for precise docking control.