Extracellular polymeric substances (EPS) are considered as the major cause of membrane fouling in membranebioreactors. Recent studies have revealed a linear relationship between fouling rate (increase of filtration resistanceover time) and polysaccharide (PS) concentration [1]. Several factors like the type of wastewater, sludge loading rate, sludge age, MLSS concentration, and mechanical stress are known to influence the concentration of dissolved EPS. Also, unsteady states like intermittent feeding or shifts in the oxygen supply have been identified as additionalfactors leading to an increase in EPS formation or to a change in its fouling propensity. However, no systematicinvestigation to quantify such influences has been undertaken so far. This study is aimed at determining the influenceof discontinuous excess sludge withdrawal often applied in small decentralised plants and resulting unsteady loadingrates on EPS formation and filtration resistance. The general trend of increased fouling rate at higher PS concentrationsis confirmed but data show a larger scatter which could be due to a change in the PS structure and hence their fouling potential. The levels of dissolved oxygen and nitrate which also fluctuate more strongly in MBRs withirregular sludge wastage appear to have a profound impact on EPS uptake rate and thereby on EPS concentration.

As part of the EU-Life ENREM demonstration project the Department of Chemical Engineering, TU Berlin, was appointed to conduct the preliminary pilot trials in a representative site for verification of basic process design and operation criteria of the full-scale MBR demonstration plant. In addition to conception and construction of the pilot plant, this investigation consisted of two successive trial phases with distinct operation conditions. The first one was dedicated to the assessment of the “irregular sludge removal strategy” (the biomass is accumulating in the reactor, which is partly emptied when the sludge concentration reaches a given value). In the second trials phase normal operation conditions with daily sludge wastage were implemented with 28,5d SRT. The major outcome of the trials was that COD removal, enhanced biological phosphorus removal and the post-denitrification performed a similar way under both operational conditions. The denitrification rate was approximately 1 mgN/(h goTS). An influence of the anaerobic sludge loading on the post-denitrification rate was observed with higher rates (up to 4 mgN/(h goTS)) corresponding to higher organic loading. An influence of storage compounds built up in the anaerobic phase is assumed. Nitrification was better in the second phase when 4 mgN/(h goTS) were constantly reached while nitrification was unstable with an average of 2 mgN/(h goTS) in the phase of irregular sludge removal. The aerobic and anoxic reactors were enlarged during the regular sludge withdrawal phase by 23% resulting in 35d SRT. This led to a better COD removal and slightly better nitrogen removal. The enhanced SRT produced possibly a deterioration of biological P removal due to overloaded poly-P storage. A second possible reason is the massive reproduction of sludge worm Tubifex tubifex, which was observed after the plant enlargement. Different strategies to reduce the worm population were attempted. Ammonium dosing had no success. Copper dosing reduced the number of worms significantly but the population grew back after the dosing was stopped. The prolongation of SRT reduced the sludge yield from 0.23 gTS/gCOD at 28.5d to 0.18 gTS/gCOD at 35d.

This study investigates a post-denitrification process without the addition of an external carbon source combined with an enhanced biological phosphorus removal (EBPR) in a membrane bioreactor (MBR). Three trial plants, with two different process configurations, were operated on two different sites, and a variety of accompanying batch tests were conducted. It was shown that even without dosing of an external carbon source, denitrification rates (DNR) much above endogenous rates could be obtained in post-denitrification systems. Furthermore, the anaerobic reactor located a head of the process had a positive impact on the DNR. Given these surprising results, the project team decided to identify the carbon source used by the microorganisms in the postdenitrification process. Batch tests could demonstrate that lysis products do not play a major role as a C-source for postdenitrification. The following hypothesis was proposed to explain the observations: the glycogen, internally stored by the substrate accumulating bacteria, if anaerobic conditions are followed by aerobic conditions could act as carbon source for denitrification in post-denitrification system. First exploratory batch tests, where the glycogen evolution was monitored, corroborate this

As part of the EU-Life ENREM demonstration project the Department of Chemical Engineering, TU Berlin, was appointed to conduct the preliminary pilot trials in a representative site for verification of basic process designs and operation criteria of the full-scale MBR demonstration plant. In addition to conception and construction of the pilot plant, this investigation consists of two successive trial phases with distinct operation conditions: the first one being dedicated to the assessment of the “irregular sludge wastage strategy” (the biomass is accumulating in the reactor, which is partly emptied when the sludge concentration reaches a given value), and the second one being planned to verify normal operation conditions with daily sludge wastage. This progress report describes implementation and results of the first phase, for which a pilot plant of 140L was operated over 6 months with waste water of a decentralized area. The influent contained high concentrations of nitrogen (100-200 mg/L), phosphorus (10- 20 mg/l) and COD (1000-2000 mg/L). Also surprising were the high VFA concentrations (100-300 mg/L) which ensured a good EBPR process. The COD and also the enhanced biological phosphorus removal (EBPR) were not impacted by the irregular sludge wastage. COD effluent concentrations were around 50 mg/L and TP effluent was 0.1 to 0.3 mg/L. The high nitrogen influent concentrations were problematic. Due to changing TS concentrations and changing nitrification rates TN effluent was 10 to 30 mg/L with a NH4-N content of 0 to 20 mg/L. Denitrification rates were measured between 1 and 3 mgN/gVS h and were depending on TS concentration, with higher rates at lower TS concentrations. Polysaccharide concentrations in the sludge water phase were higher with low TS concentrations and low oxygen concentrations. Higher PS values led to faster fouling. Results of the trials suggest that the oxygen concentration should be kept above 2mg/L to ensure both sufficient nitrification and lower fouling. Since also high TS concentrations are needed to ensure complete nutrients removal the optimum TS range is relatively small and it must be concluded that the irregular sludge wastage strategy was not beneficial in this case and the demonstration plant should be run with regular sludge removal.

Erfahrungen aus dem Betrieb einer Membranbelebung mit diskontinuierlichem Schlammabzug werden vorgestellt. Das Ziel ist, den Einfluss der Betriebsbedingungen auf Fouling und den Abbau extrazellulärer polymerer Substanzen zu bestimmen sowie mit bekannten Zusammenhängen aus Anlagen mit kontinuierlichem Abzug zu vergleichen.