Research On Modeling And Control Of Morphing Flight Vehicles

The morphing flight vehicles become to the new research focus in last decades, for the reasons that morphing could resolve the design contradiction when flight vehicles flying in large range of space and speed, and playing multi-roles. However, the large scale morphing aerodynamic configuration and severely changeable dynamic parameters, bring great challenge in the the modeling and control of morphing vehicles. Based on this background, this dissertation focuses on modeling and control problem of morphing aircraft with variable sweep and variable span, and conducts deep research in the area of aerodynamic optimization, multi-rigid-body dynamics, longitudinal and lateral control. The author tries to give an overall description of the characteristics and provides a reference for the practical use of morphing vehicles.To solve the multidisciplinary coupled overall design and aerodynamic optimization problems of morphing aircraft with variable sweep and variable span, the 5 key design variables and 3 key design objectives are refined. Then an optimization platform based on Kriging surrogate models, CAD parametric design and CFD automated design is proposed. According to this platform, the Pareto set of optimized aerodynamic configuration is generated. On contrast with the original, the three design objectives are all significantly improved. The further CFD analysis of the optimized morphing aircraft shows that the aircraft could fly in low speed with the highest L/D ratio when in state of small sweep angle and large span and fly in high speed with the lowest drag when in state of large sweep angle and small span. The result also indicates the necessity of variable sweep and variable span.To solve the difficult dynamic modeling problem of morphing UAVs with variable sweep and variable span aroused by great time varying and strong coupling characteristics, a multi-rigid-body dynamic modeling method is proposed. Firstly, the body coordinate fixedly connected to the fuselage center of mass is introduced. The movement of the fuselage is equivalent to the movement of the aircraft. Then based on Newton-Euler method, independent analysis of each rigid body is conducted. Finally the complete multi-body dynamics equations are derived, and the additional force and additional moment compared with traditional aircraft are generated. The additional moment consists of two parts: the morphing-gravitational moment and the morphing-dynamic moment. Since the establishment of the dynamic model focuses on the fuselage, thus the derivative inertia items are avoid, which making the analysis of dynamics greatly simplified. The dynamic responses in different morphing patterns and speeds are also analysis.To solve the difficult control problem which aroused by the severe change of state parameters of morphing aircraft with variable sweep and variable span, the longitudinal LPV model is generated based on Jacobian method and a LPV-gain-scheduling state feedback controller based on LMI and parameter dependent Lyapunov function is proposed. When considering the speed changes when morphing to achieve the synchronization of configuration status and flight speed status, firstly a parameter dependent dimensionality reduction state observer is designed, secondly a gain scheduling controller based on polytope theorem and optimal theory is proposed. Nonlinear simulations show that the designed controller enables the aircraft fly in stable height and has good speed tracking performance when morphing.To solve the low roll efficiency problem in lateral control of morphing aircraft with variable sweep and variable span, the mechanism of asymmetric morphing to achieve roll control is studied. Then the concepts of span-rudder and asymmetirc-morphing-differential- action-equivalent method are proposed. The lateral LPV model is then established. In order to detect the failure of span-rudder, a mathematic model of possible fault is given and a fault observer based on augmented state and Gopinath method is proposed. The simulation results show that, in the case of disturbances, the fault observer is still able to accurately estimate the system failure and estimate the sideslip angle. Finally, based on LMI method, a LPV-H?-GS output feedback controller is proposed and the simulation results show that: in the absence of fault, the system could maintain good performance globally. For the case of possible stuck of span-rudder, a fault tolerant controller based on fault compensation is proposed. The controller is reconstructed in the original LPV-H?-GS controller and avoids the online solving of LMIs. The simulation results show that after the sudden failure of the actuator, the controller can still meet the control requirements.