Optimal Fault Estimation

Abstract In this paper, whenever a single fault occurs at a time, we present conditions under which we can construct residual generators that enable us (1) to detect the fault either exactly or almostly (in either H2 or H∞ norm sense), and (2) to estimate the extent of fault (i.e. the fault signal) as best as possible, i.e. the estimation error is as small as possible in the sense of H2 or H∞ norm.