Hydrodynamique d'un filtre biologique en systeme insature avec nitrification d'un effluent septique

Abstract

Hydrodynamics of a biological unsaturated filtration system with nitrification of a septic effluent: Most of the on site waste-water treatment plant designed by means of empirical relations, do not answer to the European quality guide-line for waste-water effluent. Further more, the classical process based on a pretreatment combining settling and liquefaction of suspended solids followed by an aerobic filtration in natural or artificial soil beds is a source of nitrate. Recently, process implementing a saturated bed filter (denitrificant reactor) and a streaming filter (nitrificant reactor) removed nitrates efficiently (Rousselle, 1990, Doctorate Thesis, University of Chambery, France; Cochet et al., 1990; Wat. Sci. Tech. 22, 109–116). The denitrification filter is located upstream of the nitrification filter and the effluent is recirculated in order to achieve the denitrification process without any extra carbon source. The purpose of this work is in a first stage to study the hydrodynamics of the unsaturated reactor and in a second stage to determine the kinetic constants of carbon removal and ammonia oxidation. These two parts are necessary for establishing the functional equations of this kind of process. The hydrodynamic description of this type of reactor is discussed in the first part of this work. Material and methods: Hydraulic loads applied at the inlet reactors were from 0.41 to 54cmh−1. For continuous flow reactor, the column had a 0.20 or 0.30m diameter for hydrodynamic study and 0.09m for nitrification constants determination. Kinetic investigations were performed in a batch reactor including a filtering column which is linked to an effluent tank (Fig. 1 ). The pulsed periodic flows were obtained with an inlet flow during 5minh−1. Pulse technique was used for the determination of reactor behaviour. Fresh water applied at a given rate was enriched with a potassium iodide tracer impulse. The iodide concentration measurements are done at the outlet of the packed bed. Various flow rates were tested to understand the hydrodynamic behaviour. The filtering media used in these investigations was a mixture of sand and flocks textile. Table 1Table 1Caractérisitiques physiques du garnissage utilisé: mélange sable-géotextile. Physical caracteristics of filtering media: sand-flocks textile mixtureMatériauxDiamètre effectif (mm)coef. uniformitéDensité relativePorosité externeSable0,232,32,620,3–0,4Mélange0,252,82,410,61–0,65 reports the characteristics of this filtering media. Hydrodynamic studies: Hydrodynamic studies were carried out for both permanent flow and pulsed periodic flow. A first approach of residence time was obtained with a theoretical equation including the thickness of the liquid film (Eqs. (27)-(29)). Results were close to those determined from the experimental residence time distribution (RTD) except for the high values. The unsaturated flow in the nitrification reactor was described with a two phase flow (stagnant and dynamic) with initial and limit closed reactor conditions. Transformation of the coupled equations system in Laplace domain leads to a transfer function which can be defined as the ratio of the Laplace transform of the output divided by the Laplace transform of the input. Three parameters are defined in this matter exchange with dispersed flow model (EMD model). P, N, f, represent respectively dispersion in dynamic flow (Peclet number), number of transfer units and ratio of dynamic phase volume divided by total volume of two phases. Unsaturated flow could be well represented by EMD model (Table 3). Model parameters (P and f) were not significantly affected by hydraulic load variation. For a pulsed periodical flow the filtering media behaves as a flow shock adsorber. For a 0,95m packed bed, the outlet flow was in steady state conditions. Residence time distributions are well described by EMD model. (Table 5). Kinetics studies: Elimination reaction of ammonia occurs with order 0 and it is about equal to 4–4.5mgL−1h−1. For COD, results are in agreement with a first order reaction with a constant equal to 0.12–0.14h−1. Experiments with different preliminary functioning durations were carried out: 2, 6, 12months (Table 6). After six months of operating, kinetic constants was stable.