نوع مقاله : مقاله پژوهشی
1 دانشجوی کارشناسی ارشد / دانشگاه آزاد اسلامی قزوین
2 پژوهشیار / مجتمع دانشگاهی هوافضا، دانشگاه صنعتی مالک اشتر
عنوان مقاله [English]
Improvement of reliability and accuracy of satellite ADCS with fault detection approach is considered. Rapid fault detection avoids propagation of unwanted effects. With this approach, probability of system failure is decreased and reliability of system is increased. In this paper, a new proposed approach based on neural network data is presented .In this approach, investigation of Dynamical behavior of reaction wheel is presented. Then a multilayer prespetron neural network with linear delay learns the Dynamical behavior of reaction wheel. Achieved neural model has more accuracy and sensitivity of it with respect to unusual behavior of reaction wheel via residual signal is high. Investigation of this approach is done with software model of reaction wheel In Matlab software.