Font Size: a A A

Fault Location Method Of Spacecraft Attitude Control System Based On Neural Network And Observer

Posted on:2022-09-03Degree:MasterType:Thesis
Country:ChinaCandidate:W K HuFull Text:PDF
GTID:2492306572955909Subject:Aeronautical and Astronautical Science and Technology
Abstract/Summary:PDF Full Text Request
The fault location of the spacecraft attitude control system is an important research content of fault diagnosis,and it is also a key guarantee for the smooth progress of aerospace missions.When the spacecraft is in orbit,the probability of failure of the attitude control system is high,and timely location of the source of the failure is the prerequisite for fault diagnosis.The actuators,gyroscopes and star sensors in the spacecraft attitude control system will all have various common failures,and will cause abnormal attitude angles and attitude angular speeds.Due to the lack of sufficient redundant information,it is often difficult to accurately locate the fault that has occurred.In this context,based on the idea of Elman neural network and observer,this paper focuses on the research of subsystem-level and component-level fault location methods for the spacecraft attitude control system from the perspective of data and models.In response to the need of the Elman neural network prediction model offline training data set,the attitude control system model is established and the sliding mode controller considering external interference is designed.First,design the sliding mode surface by taking the desired attitude angular velocity and the desired attitude quaternion of the spacecraft as the control objectives.Then,design the control law and conduct stability analysis based on Lyapunov theory.Finally,simulate the designed controller,analyze its performance and sample the relevant data of the simulation process to establish a health data set.Based on the prediction model and kinematics characteristics of Elman neural network,a fault location method for the attitude control subsystem of the spacecraft is proposed.First,perform offline training and testing of the Elman neural network prediction model,and perform online application simulation of the prediction model.Then,combine the prediction model and kinematics characteristics to propose a fault location method for the attitude control subsystem of the spacecraft.Simulation verification is carried out when there is no fault and when a fault occurs.Using the observer method,the actuator and sensor(gyroscope,star sensor)are respectively used for component-level accurate fault location.Based on the idea of dedicated observer,first design the actuator fault separation observer,then conduct stability analysis,and finally verify the effectiveness of the proposed method through simulation;for the problem of sensor fault location,a location based on adaptive observer is proposed The method is followed by stability analysis,and finally the method is simulated when there is no fault and when a fault occurs to verify the feasibility of the proposed method.
Keywords/Search Tags:spacecraft attitude control system, fault location, Elman neural network, observer, sliding mode control
PDF Full Text Request
Related items