Biological treatment of water contaminated by hydrocarbons in three-phase gas-liquid-solid fluidized bed

Biological treatment has been carried out in two different systems: aerated closed and threephase fluidized bed reactors for hydrocarbons removal from refinery wastewaters. For the two systems, hydrodynamic study allowed the determination of operating conditions before treatment experiments. Then, in a second time, biological treatments have been conducted in the same operating conditions. The obtained results showed that in the three-phase fluidized bed we can degrade hydrocarbons more rapidly than in a closed aerated bioreactor. Among the different appropriate techniques available to create efficient contacts between phases, the three-phase fluidization G/L/S where carrier particles are moving inside the reactor seems very interesting. It allows an intimate contact between phases and present many advantages concerning hydrodynamic and mass transfer phenomena. In fact, depending on operating conditions and the bubble flow behaviour, the three-phase fluidized bed could display different flow regimes In these systems called bioreactors the solid particles covered with a biofilm are fluidized by two ascending flows of air and contaminated water. With favourable operating conditions, from a hydrodynamic and mass transfer point of view, the pollutant can be biologically degraded up to 90%. Until this date, the three-phase bioreactors modelling remains very complex because it required taking into account several factors: the pollutant biodegradation rate in the biofilm, the bioreactor hydrodynamic characteristics, and the reactant interfacial gas-liquid and liquidsolid mass transfer. Thus the essential purpose of modelling is to integrate the microbial kinetics with the reactor hydrodynamics. We can notice that a few models have incorporated both bioreactor hydrodynamics and microbial kinetics. For the steady state bioreactor model, we generally assume that the particles are uniform in size, the biofilm is uniform in thickness, and the biofilm can be considered as homogeneous matrix through which oxygen and substrate diffuse and are consumed by the microbes. The liquid phase in the bioreactor substrate is considered to be axially dispersed while the gas phase is assumed to be in plug flow [2]. Rittmann (1997) proposed a model based on wake theory for predicting bed expansion and phase hold-ups for three-phase fluidized bed bioreactors. In this model he modified the correlation for the computation of the bioparticles drag coefficient CD [3]. He also attempted to explain the biofilm detachment which can occur with three broad patterns: erosion, sloughing and scouring and assumed that the factors affecting detachment rates can be grouped into two categories (physical forces and microorganisms physiology in the biofilm).