In this study, an adaptive impact-time-control cooperative guidance law based on deep reinforcement learning considering field-of-view (FOV) constraints is proposed for high-speed UAVs with time-varying velocity. Firstly, a reinforcement learning framework for the high-speed UAVs’ guidance problem is established. The optimization objective is to maximize the impact velocity; and the constraints for impact time, dive attacking, and FOV are considered simultaneously. The time-to-go estimation method is improved so that it can be applied to high-speed UAVs with time-varying velocity. Then, in order to improve the applicability and robustness of the agent, environmental uncertainties, including aerodynamic parameter errors, observation noise, and target random maneuvers, are incorporated into the training process. Furthermore, inspired by the ${\mathrm{RL}}^2$ algorithm, the recurrent layer is introduced into both the policy and value network. In this way, the agent can automatically adapt to different mission scenarios by updating the hidden states of the recurrent layer. In addition, a compound reward function is designed to train the agent to satisfy the requirements of impact-time control and dive attack simultaneously. Finally, the effectiveness and robustness of the proposed guidance law are validated through numerical simulations conducted across a wide range of scenarios.

In this research, we studied the interaction between the ultra-intense laser and multiple copper layers covered with multiple hydrogen layers. The research conditions are based on the symmetric and asymmetric structure of multilayer copper and hydrogen. It was found that the acceleration obtained from the first copper and hydrogen layer plasma was higher and occurred earlier than the second copper and hydrogen layer plasma. We investigated the spatial distribution and phase-space distribution of copper electrons, copper ions, hydrogen electrons, and hydrogen protons with different widths of the front hydrogen layer and the front copper layer, respectively. Theoretical simulations show that when the ultra-intense laser was irradiated in multiple copper layers coated with multiple hydrogen layers targets, some plasma phase-space distribution varied clearly in the different thicknesses of the first hydrogen layer or first copper layer, while some plasma were not influenced by the thickness of these two layers.

The blades of some airborne equipment are prone to icing under supercooled cloud conditions. In this paper, we propose an anti-deicing spray method to prevent blades from icing at low temperatures. Using computational fluid dynamics modeling and orthogonal experimental methods, we investigated the effects of the blade angle of attack, inlet wind speed, and nozzle mass flow rate on the thickness and coverage of the liquid layer of spray material and examined the use of deflectors in this study. We found the magnitude and change rule of the influence of the previously mentioned parameters on the liquid film thickness and coverage of sprayed material to be the nozzle mass flow rate is greater than the blade angle of attack and greater than the inlet wind speed. Under the optimal combination of conditions of α = 30°, u₀ = 6 m/s, and Q = 0.003 kg/s, the liquid film thickness was maximized, and the liquid film thickness was 0.037 mm; under the optimal combination of conditions of α = 60°, u₀ = 6 m/s, and Q = 0.003 kg/s, the liquid film coverage was maximized, and the liquid film coverage was 99.81%. The anti-deicer spraying method proposed herein for use on blades is effective when considered from a number of perspectives. It provides an innovative and feasible solution to the wind turbine blade freezing problem. However, the method must be explored and modified to maximize its chances of general application, and other factors must also be considered to fully optimize the sprinkler de-icing technique to improve the performance and reliability of blades.

With increasing launch frequency in China, rocket debris impact threats have become significant.Traditional trajectory methods often neglect nonlinear stochastic factors and spatial heterogeneity of population activities, leading to risk assessment deviations.This study aims to develop a comprehensive threat assessment framework to enhance statistical confidence in impact prediction.Employing Monte Carlo simulations incorporating mass deviations, engine parameters, and stochastic wind fields, we modeled first-stage debris impact points.A bivariate normal distribution characterization constrained by a 99%confidence ellipse is established to describe impact point dispersion.The proposed data-driven framework integrates stochastic trajectory simulation and fine-grained population data, providing a quantitative approach for ground risk evaluation.Utilizing high-resolution WorldPop data in China and a"severity"index, we optimized ground avoidance zone polygons via boundary merging algorithms.Results demonstrate that impact points exhibit a distinct two-dimensional normal distribution under random couplings.Risk assessment reveals that cumulative risk values increase significantly when intersecting high-density areas.The proposed data-driven framework effectively integrates Monte Carlo simulation with multi-source stochastic disturbances and population data, providing a quantitative method for risk evaluation.The comprehensive risk value(R)offers a single, actionable metric to optimize launch operations and mitigate casualties.