Title :
Using LS-SVM inverse model on actuator fault diagnosis and accommodation for nonlinear systems
Author :
Ma Yongwei ; Xingke, Gou ; Xianjun, Du ; Jie, Cao
Author_Institution :
Coll. of Electr. & Inf. Eng., Lanzhou Univ. of Technol., Lanzhou, China
Abstract :
For general nonlinear reversible multiple-input multiple-output (MIMO) nonlinear systems, a new approach is used in actuator fault estimation and accommodation. The new approach primary uses inverse system method to estimate the value of actuator fault and accommodates the inverse model of plant according to the estimation value. The series system via inverse model and plant has to remain the same pseudo-linear system. It´s difficult to describe precisely the system in engineering actual. LS-SVM is used to identify the inverse model of nonlinear system, and it´s sure that the existence of error when using identification tool to establish the inverse model of the plant, the controller with good robust is used in system.
Keywords :
MIMO systems; actuators; fault diagnosis; inverse problems; least squares approximations; nonlinear control systems; robust control; support vector machines; LS-SVM inverse model; MIMO nonlinear system; actuator fault accommodation; actuator fault diagnosis; actuator fault estimation; identification tool; nonlinear reversible multiple-input multiple-output system; pseudo-linear system; robust controller; series system; Actuators; Estimation; Fault tolerance; Inverse problems; Mathematical model; Nonlinear systems; Support vector machines; diagnosis; inverse model; nonlinear;
Conference_Titel :
Natural Computation (ICNC), 2010 Sixth International Conference on
Conference_Location :
Yantai, Shandong
Print_ISBN :
978-1-4244-5958-2
DOI :
10.1109/ICNC.2010.5583512