This thesis takes a certain model of Fighters as the plant. The purpose is to study the technology of fault-tolerant control and fault diagnosis based on Neural network. In this paper, we designed a fault-tolerant controller and a way of fault diagnosis for the plant, with establishing by C++ program.First, a great deal of Aerodynamics and moment parameters of fault aircraft are processed by interpolation and a class of fault aircraft flight simulation model are established.Second, based on the T-S fuzzy model of aircraft, this paper consideres the actuator fault into the course of designing, chooses the mixed area which combines a vertical strip and a clipped-sector, and searches for the robust fault-tolerant controller with the algebraic riccati equations(AREs) iterative algorithm. In the course of searching a more robust degree controller, this thesis combines the AREs iterative algorithm with the niche genetic algorithm to own the better controller. To weaken the effect of modeling error, ARBFN who includes compensation item is used in this paper.Third, a method is proposed in this paper which combines an adaptive radial basis function(RBF) neural network and a genetic algorithm(GA) to obtain a more accurate class separability for the aircraft. The adaptive RBFN is used to add new hidden layer neurons or to determine the certain class which the input vector belongs to. The GA is used to search for the best value for the parameter of RBFN by estimating the input vectors. Our method which includes ARBFN and GA can select a parsimonious network architecture. Compared with other methods, the result shows that our method can achieve a more precise class separability.Finally, a great deal of simulation with MATLAB is used to find the area of fault-tolerant. Compared with the traditional robust fault-tolerant controller, the method in this paper can improve the dynamic performance greatly. In this paper, the fault-tolerant control law module are established to obtain the real-time simulation. |