H infinity bounded fault detection filter

A key issue in practical fault detection filter applications is sensitivity to system disturbances and sensor noise, In this paper, a stabilizing faul t detection filter gain is found that bounds the H-infinity norm of the tra nsfer matrix from system disturbances and sensor noise to the residual. For multidimensional faults, a residual direction is identified that enhances the fault signal-to-noise ratio while maintaining the H-infinity norm bound , By retaining the fault detection filter structure, the practical applicab ility of the proposed approach is the same as for fault detection filter de signs found by conventional eigenvector assignment. In contrast to an unkno wn input observer approach wherein noise is explicitly decoupled from the f ault signal, a noise bounding approach does not impose additional geometric constraints on the filter structure. This allows a given filter to isolate more faults, an important feature because in practical applications low-or der filter dynamics are common. Further, the possibility of ill-conditioned filter eigenvectors, occasionally imposed by geometric constraints, may be reduced.