Fluid–structure interaction analysis has high computing costs when using computational fluid dynamics. These costs become prohibitive when optimizing the fluid–structure interaction system because of the huge sample space of structural parameters. To overcome this realistic challenge, a deep neural network-based reduced-order model for the fluid–structure interaction system is developed to quickly and accurately predict the flow field in the fluid–structure interaction system. This deep neural network can predict the flow field at the next time step based on the current flow field and the structural motion conditions. A fluid–structure interaction model can be constructed by combining the deep neural network with a structural dynamic solver. Through learning the structure motion and fluid evolution in different fluid–structure interaction systems, the trained model can predict the fluid–structure interaction systems with different structural parameters only with initial flow field and structural motion conditions. Within the learned range of the parameters, the prediction accuracy of the fluid–structure interaction model is in good agreement with the numerical simulation results, which can meet the engineering needs. The simulation speed is increased by more than 20 times, which is helpful for the rapid analysis and optimal design of fluid–structure interaction systems.

The incorporation of nature-inspired techniques to control or reduce boundary layer separation, to bring about performance enhancements on air/water vehicles, has been an active research area for many years. In this paper, a baseline NACA0012 airfoil is modified using a short flap on its upper surface at a Reynolds number of Re = 1000. The impact of the flap configuration—described by length, attachment position, deployment angle, and material properties, on the aerodynamic performance of the airfoil—quantified by mean and fluctuating forces, is investigated, and the flow field is analyzed. Inspired by the observation of pop-up feathers on a bird’s wing, the flap is first set to be rigid for a range of location, size, and inclination angles. After the optimal location of a rigid flap has been established, the flap is then allowed to be flexible, its motion is coupled to the encircling flow field, and it is tested for a range of mass ratios and bending stiffness values. The fluid motion is obtained by solving the lattice Boltzmann equation, while the dynamics of the flexible flap are calculated using the finite element method and the coupling between the flow and flap handled by the immersed boundary method. For the flexible flap, two flapping patterns are observed and the mechanism of separation control via rigid/flexible flap is explained. Compared to the flapless NACA0012 airfoil case, in the case with a flap of optimal configuration, the mean lift coefficient is improved by 13.51%, the mean drag coefficient is decreased by 3.67%, the mean lift-drag ratio is improved by 17.84%, the maximum lift fluctuation is decreased by 40.90%, and the maximum drag fluctuation is decreased by 56.90%.

In nature, creatures such as birds, insects, and fish have excellent flight and mobility capabilities. The prominent flight performance of many creatures employing flapping wings has attracted researchers to study the aerodynamics of bionic flapping wings, which has potential application in designing micro air vehicles and autonomous underwater vehicles. Bionic movements usually have to adapt to the low Reynolds number environment. It is noteworthy that the flow field of a flapping wing at low Reynolds numbers flow state is closely related to the complex non-linear shedding and viscous phenomenon, especially in a three-dimensional (3D) flapping wing. In order to observe the influence of the viscous phenomenon on flapping wing propulsive performance at low Reynolds numbers, the flow field characteristics of the 3D flapping wing under different Reynolds numbers are discussed using the immersed boundary-lattice Boltzmann method with the Chinese supercomputer TianHe-II in this paper. The influence of kinematic parameters on the flow characteristics at low Reynolds number is particularly emphasized, considering that the biological movement involves many kinematic parameters, the unsteady flow field and vortex structure around the flapping wing are analyzed in detail. This study reports that the law of the flapping wing propulsive performance strongly depends on kinematic parameters that affect the vortex changes. The underlying flow mechanism behind flapping wing performance at low Reynolds numbers has been explored, which will make it possible to apply superior kinematic parameters to improve the propulsive performance of a flapping-like new airplane.

After billions of years of natural selection, flying animals with flapping wings have superior flight and mobility capabilities. The aerodynamic characteristics and the propulsion mechanism of bionic wings have attracted a large number of researchers because they will be beneficial to novel bio-inspired micro air or underwater vehicle design. Except the single activities, for fish, birds, and insects, there is a very popular and interesting biological clustering phenomenon known as schooling. Considering the real biological movements in schooling under low Reynolds number, the study of the flow mechanisms and thrust performance of bionic multiflapping wings in different schooling configurations could be applied to the design of future bionic flapping wing aircraft formation. The unsteady flow mechanisms and the thrust performance of three-dimensional multiflapping wings in three different schooling configurations are numerically investigated using the immersed boundary-lattice Boltzmann method with the Chinese TianHe-II supercomputer. The influences of different schooling configurations and individual distances on the thrust performance of multiflapping wings are thoroughly investigated. Numerical results indicate that the individual horizontal distance has great effects on the thrust performance of multiflapping wings in schooling, and the average thrust coefficient of each flapping wing in different schooling configurations at a specific individual horizontal distance is larger than that of the single flapping wing. There is an optimum distance for different schooling configurations, where the individual interaction lead to best propulsion performance. Different from the simple tandem schooling, the closer the individual distance, the better the overall thrust performance obtained for triangle and diamond schooling.

In this paper, we expolore Multi-Agent Reinforcement Learning (MARL) methods for unmanned aerial vehicle (UAV) cluster. Considering that the current UAV cluster is still in the program control stage, the fully autonomous and intelligent cooperative combat has not been realised. In order to realise the autonomous planning of the UAV cluster according to the changing environment and cooperate with each other to complete the combat goal, we propose a new MARL framework. It adopts the policy of centralised training with decentralised execution, and uses Actor-Critic network to select the execution action and then to make the corresponding evaluation. The new algorithm makes three key improvements on the basis of Multi-Agent Deep Deterministic Policy Gradient (MADDPG) algorithm. The first is to improve learning framework; it makes the calculated Q value more accurate. The second is to add collision avoidance setting, which can increase the operational safety factor. And the third is to adjust reward mechanism; it can effectively improve the cluster’s cooperative ability. Then the improved MADDPG algorithm is tested by performing two conventional combat missions. The simulation results show that the learning efficiency is obviously improved, and the operational safety factor is further increased compared with the previous algorithm.