Fault diagnosis of nonlinear system based on fuzzy dynamic model

Presents a dynamic fuzzy model (DFM) based fault detection and isolation (FDI) scheme for the nonlinear system. The dynamic behavior of a nonlinear process is represented by a fuzzy aggregation of a set of local linear models. The parameter variations of the DFM identified in online and the validity of the local linear models are used to generate a residual vector. The neural network classifier, which learned the relationship between the residual vector and fault type, detects and isolates the process faults. We apply the proposed FDI scheme to the design of an FDI system of a two-tank system and show the usefulness of the proposed scheme.