Nonlinear actuator fault diagnosis: An application of model reference adaptive technique (Q2744511)

The author considers a problem of nonlinear actuator diagnosis using a model reference adaptive approach. That is, she supposes that the nonlinear actuator is characterized by a parameter vector and thinks that the parameter vector will be changed in unexpected mode when faults happen. Thus, she proposes designing two observers: one is a fixed observer for detecting whether the faults happen, and the other one is a nonlinear adaptive observer for estimating the actual parameter vector. A simulation example is presented to illustrate the effectiveness of her algorithm.