Data validation using interval algebra

An application of interval algebra in the diagnosis framework is proposed. After a brief recall concerning interval algebra, the problem of data reconciliation is considered. A formulation of state estimation with measurements represented by intervals is proposed. The problem of fault detection is then tackled. Two model-based procedures of fault detection and isolation using an interval representation of the measurements are presented. The first one uses the parity space residuals and the second one uses the estimation residuals, which are obtained by comparing each measurement and its estimation.