Abstract:
Based on radial basis function (RBF) neural networks and the
reference operating point, a new approach to fault detection and
diagnosis of highly nonlinear system is presented. It studies
nonlinear system directly. The effects caused by occurrence of
fault can be differed from the ones caused by variance of
operating points precisely. Finally the effectiveness of this
method is verified in an aeroplane actuating system.
Key words: Fault
detection and diagnosis Nonlinear system Neural network
Manuscript received on
June 7, 1996; revised manuscript September 18, 1996
|