Fault Diagnosis Research And Its Application To Spacecraft

Spacecraft has become an indispensable part of national defense construction and economic development.How to improve spacecraft reliability,security and reduce the risk of failure events has become the focus of attention at home and abroad among the people who are in related fields.Spacecraft fault diagnosis,prognosis and health management methods are studied in this paper,and their research contents are outlined as follows:For sudden failure diagnosis,Adaptive observers based on kalman filter is designed,which is capable of joint estimating system states and parameters accurately.Whether a failure occurs can be determined according to the threshold detection for system states.Compared with the traditional adaptive observer,this method is simple,practical,and it can estimate system states and fault parameters accurately.For degradation failure diagnosis,a remaining useful life estimation method based on the degradation process is applied to failure time prediction.Considering the performance degradation models are non-linear in nature,a non-monotonic stochastic degradation process is used for system modeling.The remaining useful life of the device is derived through the distribution of first hitting time of the degradation process.Finally,monitoring data of gyroscope of inertial navigation platform are applied to this method to verify its rationality and feasibility.In order to apply diagnosis,prognosis and health management method to engineering application,we develop a "spacecraft failure analysis and decision support system" software based on graphic interpretation of fault tree analysis method.The software includes data processing and fault tree analysis section,which can output reliability report according to analysis result of system so as to provide an effective technical support for the system health management.The failure of spacecraft oil system is analyzed by the software,which shows the usefulness of the software.