Reduction de la production de boues activees

The secondary treatment of wastewaters by the activated sludge process generates large quantities of excess sludge. Many biological, physical and chemical processes are being studied to reduce excess sludge production. The CannibalRTM and the sludge ozonation processes are two promising sludge reduction processes. The general objective of this thesis was to determine the mechanisms for the reduction of excess sludge production from the activated sludge in processes operated at a very long sludge retention time (SRT) and from the sludge ozonation process.;Complete sludge retention experiments were conducted in a demonstration scale membrane bioreactor (MBR) fed with a real municipal wastewater and subjected to growth and famine periods to determine the stoichiometric and kinetic parameters related to sludge production over wide-ranging loading rates (food to microorganism ratio; F/M) and SRTs. The default model parameter values (YH, bH and ƒ) from two well-accepted models were compared to determine which parameter set better fitted the COD, VSS and active biomasse accumulation. These two models were selected because of their broad application by practitioners in water resource recovery facilities (WRRFs, the new designation for wastewater treatment plant, WWTPs) design and very similar sludge production and oxygen demand predictions despite dissimilar default parameter values, but different active biomass proportions. The goodness of fit between the measurements and the model predictions indicated that the General ASDM (GM) parameter set better fitted the experimental data than the Metcalf and Eddy (ME) parameter set. The fit with the active biomass concentration and the volatile suspended solids (VSS) decrease during famine periods were significantly better, as the ME parameter set poorly estimated the active biomass component and overestimated the VSS decrease. The General ASDM model default parameter values for YH, b H and f of 0.666 g COD/g COD, 0.24 d-1 and 0.20 g COD/g COD, respectively, were appropriate and sufficient to model the prediction of excess sludge production in the activated sludge process for sludge retention up to 70 d. A slow degradation rate for X U,Inf and XE over the sludge accumulation period of 70 days did not improve model predictions and led to some fractionation discrepancies. The COD to VSS ratio (fCV) and nutrient content (nitrogen and phosphorus) of the sludge components determined were in accordance with past work on synthetic activated sludge.;A full scale CannibalRTM facility was characterized based on 3 years of historical operation data and simulations to determine the mechanisms for reducing excess sludge production in this process. The activated sludge model used allowed to individually consider the different mechanisms expected to take part in the reduction of the excess sludge production in the CannibalRTM process. Processes considered the slow degradation of the unbiodegradable organic particulates from the influent (XU,Inf) and of the endogenous residue ( XE), and the selective removal of trash (XU,Inf) and grit (inorganic suspended solids; ISS) by microscreens and hydrocyclones, respectively. A mass balance around the facility was produced based on a relatively stable 3 years period of operation, allowing evaluating the SRT, the activated sludge yield (0.08 kg TSS/kg COD) and the trash and grit yield (0.06 kg TSS/kg COD). The slow degradation of XE, and possibly of XU,Inf appeared necessary to predict the excess sludge production under very long SRTs, as found in the Cannibal RTM process and as reported in literature. First order degradation kinetics for XE et XU,Inf (bE et bU) of 0.007 d-1 appeared appropriate.;Ozonation experiments on synthetic activated sludge containing well defined proportions of XH and X E prior and after aerobic digestion were used to characterize the mechanisms of ozonation on these sludge components. Solubilization, biodegradability, biomass activity and viability were assessed after transfering various ozone doses.;The conclusions drawn from this thesis should allow to improve model predictions for excess sludge production in activated sludge models and to improve the design and implementation of sludge reduction technologies making use of a long SRT and sludge ozonation. Pilot testing of the physical separation of trash and grit under typical operating conditions and of sludge ozonation at different dosing points would allow to better quantify the respective benefits of these sludge reduction technologies. (Abstract shortened by UMI.).