|
A NEURAL NETWORK OBSERVER APPROACH FOR ACTUATOR FAULT DETECTION AND DIAGNOSIS IN NONLINEAR SYSTEMS
Liling Ma, Yinghua Yang, Fuli Wang, Ningyun Lu
P.O. Box 131
The School of Information Science and Engineering,
Northeastern University, Shenyang, 110004, P.R. China
E-mail: maliling1974@yahoo.com.cn Fax:(86-24) 23890912
A novel approach for the fault detection and diagnosis of actuators in nonlinear systems using a neural network observer technique has been studied. This kind of nonlinear systems do not satisfy the matching condition and have unmeasured states. A neural network is used to approximate unknown nonlinearities, and an adaptive diagnostic algorithm is developed to diagnose the fault. The theoretic analysis guarantees the convergence of the observer. Simulation results have shown the feasibility and effectiveness of the method.
Keywords: fault detection, fault diagnosis, neural network, observer, nonlinear system
Session slot T-We-M10: Fault Diagnosis of Actuator Systems/Area code 7e : Fault Detection, Supervision and Safety of Technical Processes
 | 