Fault identification for Lipschitz nonlinear systems

Fault identification for a class of nonlinear systems is considered, where the nonlinear term in the system satisfies local Lipschitz conditions. Under some fault detectability conditions are satisfied, the system is first decomposed into two subsystems, where one is coupled with and another is decoupled from faults. One asymptotic state observer is then designed for the fault-coupled subsystem, in which a fictitious input is introduced to assure the state observer is asymptotic sable. It is proven that the actuator faults would be reconstructed from the fictitious input under steady state. Both stability analysis of the state observer and fault identification scheme design is given in a constructive manner and finally such a procedure is equivalent to solving a set of linear matrix inequalities.