Research On Distributed Fault Diagnosis For Flight Control System

With the continuous development of modern science and technology,the aircraft has been given more and more complex air roles in military and civil fields,such as battlefield search,geographical survey,and wildfire monitoring.In the face of large-scale complex information environment,the distributed design of the flight control system has become a development trend from the view of structure and mission.For this reason,to ensure the security and reliability of the distributed control system is very important.However,the traditional centralized fault diagnosis system is difficult to maintain,expand and the communication load is too large.It is not enough to adapt to the needs of distributed fault diagnosis.Therefore,distributed fault diagnosis based on the multi-agent technology is paid more and more attention by scholars.Fault estimation is one of the most important aspects of fault diagnosis.Compared with fault detection and isolation,the design of fault estimation is much more difficult,so the research of distributed fault estimation technology is also more challenging.In this paper,the fault estimation technique of the distributed flight control system under the actuator and sensor fault is studied based on the observer method,and the performance optimization measures are given.The main contents are as follows:Firstly,the communication topology of the distributed flight control system is described by means of algebraic graph theory,and the mathematical representation of the information sharing between agents is realized.By introducing the concept of relative output estimation error and the symbol of Kronecker product,the global fault estimation observer under a distributed communication topology is well designed.Secondly,for a class of distributed flight control system under an undirected communication topology,a distributed fault estimation algorithm based on adaptive observer is proposed,and the stability of the global error dynamic system based on Lyapunov function is proved.The calculation and simulation of the fault estimation algorithm for UAV formation are carried out in consideration of the occurrence of the constant or time-varying actuator fault.It is pointed out that the scheme also applies to the case of a directed communication topology.Furthermore,a distributed fault estimation method which can eliminate the external disturbance completely is studied in terms of the disturbance isolation characteristic of unknown input observer.Taking into account the actuator and sensor fault simultaneously,and through the unified design based on the concept of augmented state,it is access to obtain a distributed global fault augmented model,and give the proof of the existence of the distributed fault estimation observer.The disturbance elimination and unified design make a greater extent to improve the applicability of the design based on robust unknown input observer.The distributed fault estimation algorithm in Chapter 4 presupposes some strong hypotheses.However,it is obvious that not all systems can meet such preset conditions.In order to further improve the theoretical completeness,an advanced fault estimation method based on robust unknown input observer with relaxed hypothesis is studied.In order to further improve the performance of fault estimation,the fault estimation method with a finite time convergence constraint instead of ordinary disk region constraint is studied,and a rigorous theoretical proof and a solution method based on linear matrix inequality are given.Finally,in order to verify the proposed theoretical method and reflect the horizontal comparison of the research results,the studied fault estimation algorithms are applied to a class of distributed leader-follower UAV formation flight control system.The simulation results show that the proposed theoretical methods have a good ability of fault estimation for the actuator or sensor fault.These research results have certain theoretical academic value and practical application significance to enhance the reliability and security of the distributed flight control system.