The sedimentability of the activated sludge can be affected by the presence of a large variety of coagulants and polymers from a previous physical-chemical process. In this paper, the activated sludge settling process in industrial wastewater treatment plants where the sludge does not settle in a conventional way is studied. The two observed constant hindered settling velocity stages and the instant the intermediate sludge acceleration period occurs are described. A variation of the Richardson and Zaki model is used to characterize the two stages of constant settling velocity. The concentration of suspended solids, where a sudden decrease of hindered settling velocity was observed, is calculated. Finally, a new hypothesis to explain the processes triggering the collapse of the initial homogeneous sludge structure and the existence of an acceleration period is formulated.

Introduction

Many secondary settling tanks design methods are based on the experience acquired in the design of urban wastewater treatment plants. It is also common to use models to simulate settling tanks by calculating the hindered settling velocity from standard values or using empirical formulas obtained from urban wastewater treatment plants. However, in some urban wastewater treatment plants, and especially in industrial ones, the sedimentability of activated sludge can be affected by the presence of coagulants and polymers used in a previous physicalchemical process. In a batch settling test with activated sludge it is usual to find an initial induction stage of a few minutes. The increase of the settling velocity during the induction stage is due to the dissipation of the turbulence generated in the filling of the settling column [1]. The presence of cationic polymers in the sludge reduces the length of this induction period and increases the hindered settling velocity of the sludge [2,3]. The existence of an induction period in settling tests with flocculate suspensions in the compression range has also been described. This induction period is due to the initial increase in the permeability of the suspension resulting from the formation of channels through which the water rises [4]. Chen et al. [5] studied the activated sludge settling velocity from an industrial wastewater treatment plant in Taiwan. In the settling tests they observed, for certain solid concentrations, two stages with constant hindered settling velocity and an intermediate stage where the acceleration of sludge occurs. They also observed the appearance of large flocs during the speed-up period starting from a homogeneous sludge. Zhao [6,7] also described the same acceleration process with primary sludge coagulated with aluminium sulphate and with an anionic organic polymer. Zhao [6] observed that the sludge acceleration was higher in the underdose range of the polymer (< 10 mg/l) and that the increase of settling velocity could be due to the sludge flocculation during sedimentation. According to Zhao [6] in the underdose range of the polymer the sludge sedimentability was controlled by the formation of large flocs and by the progressive decrease of the viscosity. While in the overdose range of the polymer (> 10 mg/l) the sludge sedimentability was controlled by the formation of a networked sludge structure surrounded by an excess of polymer. Zhao [7] considered that this networked structure links flocs and seems to cause a packing regime. An increase in the hindered settling velocity in colloidal suspensions with weak gel structure was also reported. In a weak gel, the excess of unabsorbed polymer generates weak bonds between the colloidal particles due to the attractive forces induced by the depletion mechanism.