Fault Diagnosis of Nonlinear Systems Using a Hybrid Approach
Theincreasingcomplexityofspacevehiclessuchassatellites,andthecostreduction measures that have affected satellite operators are increasingly driving the need for more autonomy in satellite diagnostics and control systems. Current methods for detecting and correcting anomalies onboard the spacecraft as well as on the ground are primarily manual and labor intensive, and therefore, tend to be slow. Operators inspect telemetry data to determine the current satellite health. They use various statisticaltechniques andmodels,buttheanalysisandevaluation ofthelargevolume of data still require extensive human intervention and expertise that is prone to error. Furthermore, for spacecraft and most of these satellites, there can be potentially unduly long delays in round-trip communications between the ground station and the satellite. In this context, it is desirable to have onboard fault-diagnosis system that is capable of detecting, isolating, identifying or classifying faults in the system withouttheinvolvementandinterventionofoperators.Towardthisend,theprinciple goal here is to improve the ef?ciency, accuracy, and reliability of the trend analysis and diagnostics techniques through utilization of intelligent-based and hybrid-based methodologies.
Presents a novel integrated hybrid approach for fault diagnosis (FD) of nonlinear systems; taking advantage of both systems' mathematical model and the adaptive nonlinear approximation capability of computational intelligence techniquesSimultaneously accomplishes fault detection, isolation, and identification (FDII) within a unified diagnostic modulePresents fault detection, isolation, and identification (FDII) of reaction wheels of a 3-axis stabilized satellite in presence of disturbances and noise demonstrating effectiveness under both full and partial-state measurements