Optimal design of robust analytical redundancy for uncertain systems

In this paper a criterion is constructed for determining a set of parity vectors which result in different robustness of fault detection and identification (FDI). The solution of the criterion is obtained by using the theory of generalized eigenstructure. A jet engine system with a set of model variants corresponding to different system operation points is given for the illustration of FDI design using the proposed approach. Simulation results show that robust FDI can be achieved. >