Author(s): Teresa Alvarez, Solange Valente, Silvia Amaral, Maria Teresa Viseu, Teresa Alvarez, Rui M. L. Ferreira
Keywords: Embankment dams; Failure by overtopping; Pilot channel; Breaching process; Physical modelling;
Abstract: The present work consists in the comparison of two homogeneous dam breach experimental tests, performed under identical conditions of soil composition (same grain-size distribution curve), embankment material (compaction degree and water content), reservoir dimensions and inflow. The pilot channel (initial notch to start the overtopping) was the varied parameter throughout the experiment.
In the tests was possible to fully reproduce the main overtopping breaching mechanisms for cohesive embankments: headcut erosion, associated with mass detachment episodes induced by underscouring.
Results analysis was based on the comparison of the breach hydrograph and the breaching process evolution, through evolution of the breach width in the crest and 3D reconstructions of the breach. A depth sensing device (Kinect sensor) was used for this purpose. The 3D reconstruction of the breach required to stop the breaching process by cutting the inflow and emptying out the reservoir, since KINECT sensor is no able to acquire under the breach outflow.
In both tests were observed identical breach mechanisms and temporal evolutions. A small time delay is noticeable when comparing the time stamps of the two tests. This time delay results from the time required for the smaller pilot channel (1.5 cm deep) reaches a section equivalent to the larger pilot channel (5 cm deep).
Hence, it is apparent that the influence of the pilot channel is negligible in dam breach experiments.