Optimal entry flight of hypersonic vehicles requires achieving specific mission objectives under complex nonlinear flight dynamics constraints. The challenge lies in rapid generation of optimal or near-optimal flight trajectories with significant changes in the initial flight conditions during entry. Deep Neural Networks (DNNs) have shown the capability to capture the inherent nonlinear mapping between states and optimal actions in complex control problems. This paper focused on comprehensive investigation and evaluation of a DNN-based method for three-dimensional hypersonic entry flight trajectory generation. The network is designed using cross-validation to ensure its performance, enabling it to learn the mapping between flight states and optimal actions. Since the time-consuming training process is conducted offline, the trained neural network can generate a single optimal control command in about 0.5 milliseconds on a PC, facilitating onboard applications. With the advantages in mapping capability and calculating speed of DNNs, this method can rapidly generate control action commands based on real-time flight state information from the DNN model. Simulation results demonstrate that the proposed method maintains a high level of accuracy even in scenarios where the initial flight conditions (including altitude, velocity, and flight path angle) deviate from their nominal values, and it has certain generalization ability.
15 April 2019 in Journal of Mechanical Science and Technology
The highly nonlinear and coupling characteristics of a flexible air-breathing hypersonic vehicle create great challenges to its flight control design. A unique parameter adaptive nonsingular terminal sliding mode method is proposed for longitudinal control law design of a flexible coupling air-breathing hypersonic vehicle. This method uses adaptive reaching law gain instead of the additional adaptive compensation term to handle the uncertainty to improve robustness. The stability of the close loop system is proved via a Lyapunov way. The longitudinal tracking control law for velocity and angle of attack is designed based on a rigid dynamic model of a flexible air-breathing hypersonic vehicle. A strong coupling model of the same vehicle, considering aerodynamic-scramjet engine-flight dynamic-elastic couplings, is established as the verification platform of the designed control law. The remarkable differences of flight dynamic characteristics between this strong coupling model and the rigid body model can be seen, which mean the controller needs to endure very great uncertainty, unmodeled dynamics, and other types of internal disturbance. Simulation results based on the coupling model demonstrate that the designed control law has good performance and acceptable robustness.
The aim of this study was to solve the frequently occurring rotor-stator rub-impact fault in aero-engines without causing a significant reduction in efficiency. We proposed a fault mitigation scheme, using shape memory alloy (SMA) wire, whereby the tip clearance between the rotor and the stator is adjusted. In this scheme, an acoustic emission (AE) sensor is utilized to monitor the rub-impact fault. An active control actuator is designed with pre-strained two-way SMA wires, driven by an electric current via an Arduino control board, to mitigate the rub-impact fault once it occurs. In order to investigate the feasibility of the proposed scheme, a series of tests on the material properties of NiTi wires, including heating response rate, ultimate strain, free recovery rate, and restoring force, were carried out. A prototype of the actuator was designed, manufactured, and tested under various conditions. The experimental result verifies that the proposed scheme has the potential to mitigate or eliminate the rotor-stator rub-impact fault in aero-engines.
2 April 2020 in Proceedings of the Institution of Mechanical Engineers, Part G: Journal of Aerospace Engineering
This paper studies the dynamic retrieval process of a towed target system under perturbation by simulating the reel-in operation under the towing airplane’s wake and atmospheric turbulence. To settle the computational problem of cable tension, the constant-length method is proposed to transform this variable-mass problem into a constant-length problem based on a mass-spring model that discretizes the cable. A three-dimensional atmospheric turbulence field is built using a recursive function to model the complex perturbation field along the cable. A horseshoe vortex model is adopted to simulate the towing airplane’s wake. Simulation analysis shows that both the reel-in speed and turbulence significantly affect the retrieval deviation, whereas the airplane’s wake has little influence. The reel-in speed should be chosen in concert with the turbulence intensity to achieve a high retrieval success rate. Moreover, a higher reel-in speed is preferred to maintain cable strain and smoothen the process when the retrieval tolerance is satisfied.