Physical monotonicity is a pervasive phenomenon in the aerodynamic characteristics of aircraft, where the aerodynamic lift consistently increases with the angle of attack within the stalling range. Existing machine learning models for aerodynamic predictions often overlook this monotonicity, resulting in poor interpretability and credibility. To address this issue, we introduce a monotonic model, the Deep Lattice Network, which integrates the monotonicity constraint of the lift coefficient into machine learning based aerodynamic prediction framework. In this paper, we propose a novel deep learning model, Deep Lattice Cross Network, which aims to rapidly predict aerodynamic forces with high precision while ensuring monotonic constraints. Multi-Task Learning method is utilized to simultaneously predict both lift and drag coefficients, thereby enhancing the efficiency of the model. To optimize the training process and minimize costs, we adopt a unique two-phase deep network training strategy. Based on computational fluid dynamics simulation datasets of a morphing aircraft, the model is trained, and the efficacy of the model is tested by two interpolation and two extrapolation datasets. The results show a remarkable alignment with computational fluid dynamics outcomes across all test scenarios. Extended testing across a wider range of attack angles further highlights the superiority of the Deep Lattice Cross Network in upholding monotonicity. Incorporating monotonicity constraints not only improves predictive accuracy of the model but also greatly enhances its physical interpretability, which is crucial for advancing the development of more dependable aerodynamic prediction models.

Distributed electric propulsion is widely recognized as a disruptive technology in aviation, and the aerodynamic characteristics of a multi-propeller are critical to the design of such configurations. A reformulated vortex particle method is adopted to provide an in-depth analysis of the aerodynamic characteristics between the multiple propellers operating in close proximity. According to the symmetrical distribution features of the particle field, the entire flowfield is divided into a noninterference region and an interference region. The results show that the aerodynamic performance of the middle propeller fluctuates more than the other two adjacent propellers in the hover state, and the flowfield in the noninterference zone exhibits time-independent characteristics. As the advance ratio increases, the performance fluctuations decrease, the radial contraction of the tip vortex is gradually attenuated, and the initial vortex strength decreases. For the interference zone in hover, the aerodynamic load of each blade drops as the propellers approach the interference region and the wake geometry shows asymmetry. The slipstream deformation of the multi-propeller is more pronounced in hover. The downwash from the previous propeller is responsible for the sequential decrease in thrust coefficient for multi-propeller systems operating at different sideslip angles, and the middle propeller experiences the greatest thrust fluctuation.

A vortex ring is a flow phenomenon involving a complex toroidal vortex system. When a rotor is in descending flight, it may enter the vortex ring state (VRS), which will affect the rotor’s aerodynamic characteristics and even endanger it. In this paper, to clarify the aerodynamic mechanism by which the vortex ring exerts its effects on an axial descending rotor, an unsteady numerical simulation method for the rotor integrated with structured moving overset grids is proposed and validated using experimental data. This numerical simulation method is then applied to analyze the aerodynamic characteristics of the rotor in descending flight. The variations of the aerodynamic forces and the flow characteristics in the slipstream are analyzed to elucidate the physical mechanisms responsible for the relationships between the aerodynamic loads, flow field, and vortices when the rotor is in the VRS. The effects of the VRS cause a sharp drop in the average aerodynamic forces, which directly affects the safety and reliability of a rotor aircraft. In the slipstream of the descending rotor, a distinct vortex ring forms and moves upward as the velocity of descent increases. The most severe VRS occurs at a nondimensional velocity of descent of 1.13 when the center of the highly unsteady vortex ring is right at the blade tip, and this can be used as an indicator of the VRS. The physical mechanism by which the VRS exerts its effects can be attributed to the low pressure induced by strong vortices. The VRS decreases the pressure on the lower surface of the blade and increases the pressure on the upper surface, resulting in a reduction in the aerodynamic loads. In comparison with the hovering state, the VRS results in a much larger vortex strength and possesses a different vortex structure.

Flow over a radiating multi-filamentous structure with various opening angles has been investigated numerically at low and moderate Reynolds numbers. The opening angle, as the major geometric parameter for investigation, varies from 180° to 40° and the overall shape changes from disk-like to cone-like. The overall model drag decreases monotonously as the magnitude of the opening angle decreases. The opening angle effects on the drag-enhancement are then investigated in terms of the velocity distributions, the structures of the recirculation region, and the pressure distributions. As the Reynolds number increases, the wake structure loses its steadiness and the optimal opening angle is found to be approximately 60° when the critical Reynolds number for a steady recirculation structure becomes maximum. The opening angle effects on the wake steadiness in terms of the intensity of the shear layer are investigated as well. On the one hand, the exit velocity—a quantitative indication of the rate of the base bleed—decreases with the decreasing of the opening angle, which promotes wake unsteadiness. This phenomenon can be attributed to the effective gap width between adjacent filaments and the intensity of the virtual barrier from both macro- and micro-perspectives. On the other hand, a small opening angle decreases the thickness of shear layer and thus the unsteadiness is suppressed. These two effects compete and the balance is reached when the opening angle is near 60°.

A fast multiobjective optimization method for S-duct scoop inlets considering both inflow and outflow is developed and validated. To reduce computation consumption of optimization, a simplified efficient model is proposed, in which only inflow region is simulated. Inlet pressure boundary condition of the efficient model is specified by solving an integral model with both inflow and outflow. An automated optimization system integrating the computational fluid dynamics analysis, nonuniform rational B-spline geometric representation technique, and nondominated sorting genetic algorithm II is developed to minimize the total pressure loss and distortion at the exit of diffuser. Flow field is numerically simulated by solving the Reynolds-averaged Navier–Stokes equation coupled with k–ω shear stress transport turbulence model, and results are validated to agree well with previous experiment. S-duct centreline shape and cross-sectional area distribution are parameterized as the design variables. By analyzing the results of a suggested optimal inlet chosen from the obtained Pareto front, total pressure recovery has increased from 97% to 97.4%, and total pressure distortion DC60 has decreased by 0.0477 (21.7% of the origin) at designed Mach number 0.7. The simplified efficient model has been validated to be reliable, and by which the time cost for the optimization project has been reduced by 70%.