Effect of leaks in dams and trials to detect leakages by geophysikal means

Abstract

A research project was carried out over a period of 3.5 years, dealing with the hydraulics of leaks in dam seals and with geophysical methods to detect and to localize these leaks. The main object of the investigations was a large model dam (H = 3.5 m, L = 20 m, V = 600 m-) in a huge open pit which was sealed with HDPE plates to form a watertight basin and an "impervious base" of the dam. The dam body was made of sand and had an upstream sealing face which was constructed with a number of artifical leaks to be operated under full reservoir conditions. The instrumentation of the dam body consisted of a number of piezometer and temperature gauges for observing the seepage processes and the related temperature changes. A net of self-potential electrodes was installed in the downstream face of the dam which were observed along with the numerous tests performed under varying conditions. A thermo-camera was also installed and was used to observe the infrared thermo-reflection of the downstream face of the dam during the leakage tests. The tests showed that even small leaks in sealing faces can lead to considerable water losses and extensive percolations of embankment dams. Local infiltrations through leaks in an upstream sealing spread out widely into the dam body and do not result in local leakages in the downstream faces of homogeneous embankment dams. Additional safety elements like toe drains are urgently to be recommended. The geophysical measurements show that the self-potentials react immediately to changes in the seepage conditions, at least in certain regions; but a generalization of these results is not yet possible. Also the temperature measurements show significant reactions, which are substantially influenced by seasonal and day/night fluctuations. The investigations show that the geophysical methods applied can be useful for the detection of seepages through dams, but only under certain favourable conditions. More research work and systematic investigations seem to be required.