Robust observer based FDI and its application to the monitoring of a distillation column

Parametric design for observer gain matrices is utilised to formulate a fault detection and diagnosis algorithm for a class of dynamic systems. The systems considered are subject to both un-modelled dynamic uncertainties and unknown inputs. By making use of the free parameters available in the parametric design approach, a diagnostic vector is constructed which is shown to be capable of isolating faults. The remaining free parameters are then used to generate an optimally robust residual signal for fault detection. It is shown that this residual signal is insensitive to the model uncertainty and unknown inputs, but sensitive to system faults. The proposed FDI approach is demonstrated through its application to the monitoring of a simulated distillation column.