Q. Yang and G. Webb (Eds.): PRICAI 2006, LNAI 4099, pp. 308 – 318, 2006. © Springer-Verlag Berlin Heidelberg 2006 Dynamic Neural Network-Based Fault Diagnosis for Attitude Control Subsystem of a Satellite Z.Q. Li 1 , L. Ma 1,2 , and K. Khorasani 1 1 Department of Electrical and Computer Engineering Concordia University, Montreal, Quebec H3G 1M8 Canada 2 Department of Applied Computer Science Tokyo Polytechnic University Atsugi, Kanagawa, Japan kash@ece.concordia.ca Abstract. The objective of this paper is to develop a dynamic neural network scheme for fault detection and isolation (FDI) in the reaction wheels of a satellite. The goal is to decide whether a bus voltage fault, a current loss fault or a temperature fault has occurred in one of the three reaction wheels and further to localize which wheel is faulty. In order to achieve these objectives, three dynamic neural networks are introduced to model the dynamics of the wheels on all three axes independently. Due to the dynamic property of the wheel, the architecture utilized is the Elman recurrent network with backpropagation learning algorithm. The effectiveness of this neural network-based FDI scheme is investigated and a comparative study is conducted with the performance of a generalized observer-based scheme. The simulation results have demonstrated the advantages of the proposed neural network-based method. 1 Introduction The attitude control of spacecraft has been widely studied since the late 1950s [1-2]. The attitude control subsystem stabilizes the spacecraft and orients it in the desired set point position in short time despite the presence of external disturbance torques. The control torques could be formed from combination of momentum wheels, reaction wheels, control moment gyros, thrusters or magnetic torquers. Normally, there are four reaction wheels (3 active and 1 redundant) on the spacecraft to be controlled. Each active reaction wheel is aligned with one of the body axis of spacecraft, it can rotate in either direction and provide reaction torque for the related axis control. The redundant one will be engaged in case any of the three wheels fails. Any fault that occurs in the wheels should be detected and isolated as early as possible to avoid serious damage to the attitude control subsystem of the satellite. Normally, there are three types of faults in a wheel that deserve special attention. The first is the bus voltage fault. The bus voltage should be sufficiently high to avoid elimination of the voltage headroom. Low bus voltage will result in reduced torque capacity and will seriously cause instability of the attitude of the satellite. The same