The filtering yield of the pilot N lay between 79.1 and 95.3 %, and should be related to its height (60 cm of sand). During the first week, a remarkable curve ascending has been observed, which was linked to the maturation of filters [20]. The filtering yield values of pilot N4 was quite good, varying in the range 82.94 to 94.6 %.The first measurements of pilot N4 were greater than those of pilot N (85.3%), which should be most likely related to the adsorption of the organic matter on the activated carbon [21]. Yields values of pilots N8 and N12 lay between 82.9 % (at the beginning of filter functioning) and 95.3% . An intermediate optimal value of 94.6 % was registered after the second week. For the pilot N16, promising yields were observed in the beginning of the operating time (84.4%), which remained nearly constant (82.94%) till the experience was stopped. The same results have been noted for pilot N20 (85.27 and 96.1 %); while the higher yields were realized with pilot N24 (89.10 % – 97.7 %). This result should be assigned to the complete maturation of the sand layer (36cm). Overall, it should be noticed that the filtering yields of the different pilots increased with the quantity of activated carbon [22-24] The observed decrease of the filtering efficiency with time may be explained by the reduction of sand porosity [20] and the saturation of the activated carbon [25]. It is to be notified that the filtering yields were expressed in terms of the eliminated quantity of COD in Kg/m2/day .These variation are shown in figure 4.
70