Inland navigation networks are mainly used for transport with economic and environmental benefits. In a climate change context which leads to the scarcity of the water resource, the control of navigation levels and the supervision of these networks become crucial. Thus, this paper is focused on the sensors Fault Detection and Isolation of inland navigation reaches. A modeling method based on the identification technique is proposed. Then, based on residuals, the dynamic classification algorithm AUDyC leads to the detection and diagnosis of sensors faults. Setting errors and slow drifts are considered. The proposed methods are applied on the Cuinchy- Fontinettes reach benchmark.

Peer Reviewed