面向特情改出的民机多模型切换控制技术研究

发布时间：2018-06-18 03:49

本文选题：容错飞行控制 + 多模型自适应切换　；参考：《南京航空航天大学》2017年硕士论文

【摘要】：飞行失控严重威胁飞行品质与安全,民机飞行若产生失控且未被及时改出,易造成重大飞行事故。其中,飞行故障是飞行失控主要诱因之一。因而,对飞行故障预测及失控后的特情改出保护研究可提高民机飞行可靠性。针对上述研究意义,论文主要开展具体工作如下:(1)民机复杂动力学模型线化方法。采用线性变参数(Linear Parameter Varying,LPV)建模理论对复杂强耦合非线性系统线性化。考虑复杂系统凸胞型变参数实现过程困难,采用依赖网格划分LPV理论中函数替换方法,并采用启发式搜索算法中的遗传算法对LPV建模优化问题求解,对民机正常飞行状态、单发停车以及风场LPV模型进行设计。通过时域响应、可达集分析、周期模态分析等仿真分析验证遗传算法求解的LPV模型对原模型具有更高逼真度;通过对飞机着陆过程穿越风场仿真实验,验证含扰动风LPV模型有效性;另外,采用全发失效仿真验证发动机失效LPV模型可逼近原非线性模型。(2)多模型自适应预测(Multi Model Adaptive Estimator,MMAE)方法下飞行故障诊断与隔离(Fault Detection and Isolation,FDI)系统搭建。研究包括有色噪声卡尔曼滤波器设计、FDI残差信号获取以及故障概率计算。通过增加主动监督机制和流程标识位方法对原FDI改进。实验结果表明改进后FDI系统可一定程度提高系统故障检测性能以及系统稳定性。(3)特情控制改出问题研究。考虑主动FDI、被动FDI特性以及FDI系统模型表现形式,选用两类改出控制器进行研究仿真,即基于r性能的鲁棒综合控制器及带控制约束、状态约束、控制增量约束的模型预测控制器(Model Predictive Control,MPC)仿真。与线性二次型最优调节器(Linear Quadratic Regular,LQR)信号跟踪结果对比验证两类控制器改出结果有效。(4)容错飞行控制(Fault Tolerant Control,FTC)系统综合及扰动风场参数辨识。以鲁棒综合控制器作为被动FTC控制,以MPC作为主动FTC控制,构建针对风场扰动下的单发停车和升降舵卡阻两类常用故障模式的FTC系统。与基于LQR的FTC系统改出结果证明:鲁棒综合控制可抑制扰动风场影响;升降舵卡阻改出策略选择水平安定面为控制舵面,采用鲁棒综合控制器的FTC可实现升降舵卡阻改出控制目标;针对单发停车故障改出,采用MPC对剩余发动机控制分配问题求解,MPC的FTC系统同样可保证飞机平稳飞行。
[Abstract]:Flight runaway is a serious threat to flight quality and safety. If civil aircraft flight is out of control and is not corrected in time, it is easy to cause serious flight accidents. Among them, the flight failure is one of the main causes of flight out of control. Therefore, the prediction of flight failure and the study of special protection after runaway can improve the flight reliability of civil aircraft. In view of the above research significance, the main work of this paper is as follows: 1) linearization of complex dynamic model of civil aircraft. Linear variable Parameter Linear Parameter (LPV) modeling theory is used to linearize complex strongly coupled nonlinear systems. Considering the difficulty of realizing the variable parameters of convex cell type of complex system, the function replacement method in LPV theory is adopted, and the genetic algorithm in heuristic search algorithm is used to solve the optimization problem of LPV modeling, and the normal flight state of civil aircraft is obtained. Single stop and wind field LPV model are designed. The simulation results of time domain response, reachable set analysis and periodic modal analysis show that the LPV model solved by genetic algorithm has a higher fidelity to the original model. The validity of the LPV model with disturbance wind is verified, in addition, the engine failure LPV model can approach the original nonlinear model by full failure simulation, and the flight fault diagnosis and isolation fault detection and isolation (FDI) system is built under the method of Multi Model Adaptive Estimator (MMAE). The research includes the design of colored noise Kalman filter and the acquisition of FDI residual signal and the calculation of fault probability. The original FDI is improved by adding active supervision mechanism and process marking bit method. The experimental results show that the improved FDI system can improve the system fault detection performance and system stability to some extent. Considering the characteristics of active FDI, passive FDI and FDI system model, two kinds of modified output-out controllers are selected for simulation, that is, robust synthesis controller based on r performance, robust synthesis controller with control constraints and state constraints. Model Predictive Control (MPC) Simulation of Model Predictive Controller for controlling incremental constraints. Compared with linear quadratic optimal regulator linear Quadric regular LQR signal tracking results, it is proved that the two kinds of controllers are effective. 4) Fault tolerant flight control fault Tolerant control FTC) system synthesis and disturbance wind field parameter identification. Taking robust integrated controller as passive FTC control and MPC as active FTC control, the FTC system with two common fault modes of single stop and hoist resistance under wind field disturbance is constructed. The results show that robust integrated control can restrain the influence of disturbance wind field, and the horizontal stabilizer surface is chosen as the control rudder surface. The FTC with robust integrated controller can realize the control target of hoist stop and change out, and the FTC system using MPC to solve the residual engine control assignment problem can also ensure the smooth flight of the aircraft.
【学位授予单位】：南京航空航天大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：V328;V249.1

【参考文献】