Failure diagnosis of actuator using the EKF and NP theorem

• Published in: International journal of control, automation, and systems Vol. 11; no. 3; pp. 450 - 459
• Main Authors: Kim, Donghoon, Choi, Daegyun, Oh, Hwa-Suk
• Format: Journal Article
• Language: English
• Published: Heidelberg Institute of Control, Robotics and Systems and The Korean Institute of Electrical Engineers 01.06.2013; Springer Nature B.V; 제어·로봇·시스템학회
• Subjects: Aerospace engineering; Control; Control systems; Engineering; Eulers equations; Failure analysis; Kalman filters; Kinematics; Mechatronics; Methods; Parameter estimation; Regular Paper; Robotics; Spacecraft; Studies; 제어계측공학; failure diagnosis; extended Kalman filter; Actuator; Neyman-Pearson theorem; probability density function
• ISSN: 1598-6446; 2005-4092
• DOI: 10.1007/s12555-012-9222-1

There are many classical methods available for generating maneuver commands for spacecraft, where the nominal control system design consists of three torque inputs. A standard problem for the spacecraft feedback control consists of developing attitude stabilization methods that send the angular velocity to zero, if it exists. If a malfunction occurs on-orbit that eliminates or degrades the capability of torque inputs, reconfiguration of the spacecraft system is required for performing missions with the current condition. In this paper, the model-based failure diagnosis method for actuators of spacecraft, such as reaction wheels, is proposed. After generating residuals of attitude through the extended Kalman filter, failure detection of the actuator is performed. Moreover, the probability of actuator failure is conducted using the Neyman-Pearson theorem in order to recognize a degree of failure. Partial and complete failure cases are considered and successful results of failure diagnosis are presented.