Static and Dynamic Data Reconciliation for an Irrigation Canal

Abstract

This paper deals with the problem of fault detection and isolation in irrigation canals. We have developed a method which combines static and dynamic data reconciliation for the validation of measurements, detection, and isolation of sensors and actuator faults and reconstruction of missing data. Static data reconciliation uses static models at a regulation gate to validate measurements and detect sensor and actuator faults. It also enabled us to detect a drift in the stage discharge rating curve. The dynamic data reconciliation uses additional measurements and a dynamic model of the canal in order to validate measurements and detect faults and withdrawals. The combination of the two methods allowed us to distinguish between withdrawals and faults. Both methods are evaluated on measurements from a real irrigation canal located in the South of France.