Fig. 1. Location of Fontaine Gazelles embankment dam, Algeria. (374m) until (883m), which was carried out accordingly. The hydraulic conductivities used at the time of construction were calibrated and assigned to the numerical model. The calibration of the hydraulic conductivities was made based on trial and error while comparing simulated seepage flow with the observed ones (Table 1). Table 1.Fontaine des Gazelles Dam Material Properties used to develop FEM Model. Fig. 2. Mesh Formation for Fontaine des Gazelles Dam. Seepage flow was computed at ten different pond level scenarios. The SEEP/W software gives output in terms of flow-net which comprises of equipotential lines, streamlines, the velocity vectors showing dominant flow (seepage) field, and the saturation line. The results revealed that the water level and piezometric level influence considerably on the seepage flow downstream of the dam. Due to the presence of the toe drain located at the downstream of the dam, the phreatic line is controlled and had a minimum chance to cut the downstream slope face of the dam. Therefore, the drain plays a key role in the safety of the dam. The overall maximum and minimum seepage flow was recorded at the pond level 383m (5.04920 LPS) and 374m (2.100 LPS) respectively. These results are according to the indings of (Arshad et al.,2014a), who analyzed he behavior of an earthen dam (Hub Dam) while computing the seepage flow at the foundation and within the body of the dam for various pond evels. The simulated flow-net profile results at various hydraulic heads are elaborated in Figure 3 — Figure 12) respectively. Fig.4. Simulated Flow-net Profile at EL 382 m. Fig.9. Simulated Flow-net Profile at EL 377 m. Fig.6. Simulated Flow-net Profile at EL 380 m. Fig.7. Simulated Flow-net Profile at EL 379 m. MODEL VALIDATION the comparison indicates a good coincidence the model can be recommended for practice. Table 02 contains the data pertaining to predicted seepage flow and measured ones and the relative error. It is well known that the validation of any model needs the comparison between simulated and observed results at various hvdraulic heads. If Fig.12. Simulated Flow-net Profile at EL 374 m. Fig.11. Simulated Flow-net Profile at EL 375 m. Table 2.Observed and Simulated Seepage at Different Hydraulic Heads. Q, = denotes i" value of observed seepage flow and Q,adenotes average or mean of seepage flow. Fig.13. Relationship between observed and simulated Seepage at the different reservoir levels. seepage values. Thus, it is concluded that observed values of seepage flow are not much different than the simulated ones.
