Abstract

The main objective of this thesis is to develop a dynamic neural network-based fault detection and isolation (FDI) scheme for the Pulsed Plasma Thrusters (PPTs) that are used in the Attitude Control Subsystem (ACS) of satellites that are tasked to perform a formation flying mission. In order to accomplish these objectives three fault detection and isolation (FDI) approaches based on dynamic neural networks (DNN) are developed: (i) a "Low Level" FDI scheme, (ii) a "High Level" FDI scheme, and (iii) an "Integrated" FDI scheme. Based on data that is collected from the electrical circuit of the PPTs, our proposed "Low Level" FDI scheme can detect and isolate faults in the PPT actuators. Using a Confusion Matrix evaluation system we demonstrate that can achieve a high level of accuracy but the precision level is below expectations and the misclassification rate is expressed as False Healthy and False Faulty parameters is significant. Our proposed "High Level" FDI scheme utilizes data collected from the relative attitudes of the formation flying satellites. According to the simulation results, our proposed FDI scheme can detect the pair of thrusters which is faulty. It represents a promising detection capability, however its isolation capabilities are not adequate. Finally, our proposed "Integrated" FDI scheme takes advantage of the strengths of each of the previous schemes and at same time reduces their individual weaknesses. To demonstrate its capabilities, various fault scenarios were simulated. The results demonstrate a high level of accuracy (99.79%) and precision (99.94%) with a misclassification rate that is quite negligible. Furthermore, our proposed "Integrated" FDI scheme provides additional and interesting information related to the effects of faults in the thrust production levels that would not be available from simply the low and high levels separately.