As an important part of a satellite, the satellite attitude control system controls the attitude of the satellite, and its normal operation is the basic guarantee for the overall function of the satellite. The system has a broad application prospect for the exploration mission in the space. The severe challenges to the system propose a higher reliability demand. Under the research background of the satellite attitude control system, this paper focuses on the hierarchical fault diagnosis of the sensors and the actuators. The main tasks are as follows:A fault detection algorithm based on the unknown input observer is proposed. A simple unknown input observer is designed to generate the residuals, which are unaffected by the external disturbances. The features of this fault detection algorithm can be summarized as its simple design and disturbance decoupling.A fault isolation algorithm based on the functional fuzzy neural network is proposed, and the residuals, obtained from the fault detection observer, are used as the input vector of the functional fuzzy neural network. Moreover, the hybrid PSO-black stork foraging learning algorithm is adopted to tune the weights of the functional fuzzy neural network, and to enhance the learning ability of the neural network. The simulation results demonstrate that, compared to the traditional PSO algorithm, the convergence speed and the accuracy of the neural network are effectively improved by using the hybrid PSO-black stork foraging learning algorithm.The unknown input observer based on the radial basis function neural network is designed, which can estimation the fault of the satellite attitude control system. The unknown input observer in this paper is very easy to be implemented, and can eliminate the influence of external disturbances. In addition, the fault estimation of the satellite attitude control system is realized by using the radial basis function neural network. |