VPCA-based fault diagnosis of spacecraft reaction wheels

Fault detection and isolation methods based on the Principal Component Analysis (PCA) have been widely used for monitoring complex industrial processes with multiple variables and diagnosing process and sensor faults. This approach has been mainly developed for the analysis of single valued variables without considering any uncertainty in the systems. The main objective of this paper is then to develop a diagnosis process, of the chaser reaction wheels used during the rendezvous phase of the Mars Sample Return (MSR) mission, based on the Vertices Principal Component Analysis (VPCA) as an extension of the classical PCA for interval valued data. An industrial “high-fidelity” simulator of the MSR has provided a set of interval valued data describing the rotation rates of the spacecraft reaction wheels. The results have proven the efficiency of the proposed FDI approach in the diagnosing process.