Experimental Study of Evolution of Breach Resulting from Piping at Upper Part of Earth-Fill Dam

Piping and overtopping are the most important causes of earth-fill dam failure. Such dams may erode under seepage, causing a reduction in the structural strength. The aim of this study was to investigate the temporal evolution of the breach and flow rate from the breach resulting from the piping in earth-fill dams. The experiments were carried out at Hydraulics Laboratory of Civil Engineering Department within İzmir University of Economics. The dam was constructed by using a mixture consisting of 85 % sand and 15 % fine (low plasticity clay). In the first scenario a circular tunnel with a diameter of 2 cm was created along the centreline at 6 cm below the dam crest whereas in the second one it was located at the upper edge. Six cameras at different locations recorded the evolution of the progress of the breach formation. The pump flow rate was measured by magnetic flow meter, and the continuity equation was used to calculate the flow rate values from the breach. The time-varied values of the total breach areas were determined using the Gauss Area formula. The image processing method was also applied in the determination of the breach areas. The time-dependent changes of water depth in the channel were also recorded. The obtained experimental findings are presented and commented, together with the universal dimensionless curves. The failure of the dams occurred mainly because of the headcut erosion developed from downstream to upstream. When breaching started, the orifice flow was converted to open channel flow where breach bottom behaved like a broad crested weir. In the second scenario, the rigid lateral side considerably influenced the flow rate and the development of the breach. The peak flow rate corresponding to the first scenario was found approximately 2.3 times greater than that of the second one. The maximum values of all the breach parameters were reached earlier in the case of the seepage along the centerline. The ratios between the values corresponding to the first and the second scenarios were found as 3.25 and 1.75 for maximum breach areas at downstream and at upstream sides, respectively. These ratios were 2.44 and 1.37 for the average breach widths at downstream and upstream sides, respectively. A very good agreement was found between the area values obtained from Gauss area method and image processing technique, in both scenarios. This fact demonstrated that either of these two approaches can be used to determine the time-dependent breach areas. These experimental findings provide the opportunities for the calibration and validation of the numerical models used in the relevant numerical investigations. This study also offers guidance for the strategies concerning emergency action plans related to the failure of homogeneous earth-fill dams when the piping starts at upper part of the homogeneous earth-fill dams.