Fuzzy-model-based parity equations for fault isolation

Abstract In this paper, a new approach to model-based fault detection and isolation (FDI) for nonlinear processes is presented. A local linear fuzzy model of the process is used for the generation of structured parity equations. The model is run both in parallel and in series–parallel to the process, which leads to residuals with different sensitivities. The sensitivities of the parallel and series–parallel residuals are compared, and the most sensitive residuals are selected for FDI. The practical applicability is illustrated on an industrial-scale thermal plant. Here, five different sensor faults can be detected and isolated continuously, over all ranges of operation.