Fault diagnosis of gas turbine engines: A symbolic multiple model approach

This work is mainly concerned with the fault detection and isolation (FDI) problem of nonlinear systems by using a symbolic-based linear multiple model approach. To accomplish this goal, first a novel symbolic computational method for the linearization of a nonlinear system is developed such that the obtained linear models are subjected to the symbolic fault variables. By obtaining different linear models (healthy and faulty), a modified linear multiple model approach is then developed for the FDI problem. In order to demonstrate and illustrate the capability of our proposed FDI method a single spool gas turbine application is considered where extensive simulation studies are conducted.