The use of immersed membranes for solid-liquid separation in biological nutrient removal activated sludge (BNRAS) systems was investigated at lab scale. Two laboratory-scale BNR activated sludge systems were run in parallel, one a MBR system and the other a conventional system with secondary settling tanks. Both systems were in 3 reactor anaerobic, anoxic, aerobic UCT configurations. The systems were set up to have, as far as possible, identical design parameters such as reactor mass fractions, recycles and sludge age. Differences were the influent flow and total reactor volumes, and the higher reactor concentrations in the MBR system. The performances of the two systems were extensively monitored and compared to identify and quantify the influence of the membranes on system response.
The MBR UCT system exhibited COD, FSA, TKN, TP and TSS removals that were consistently equivalent or superior to the conventional system. Better P removal in the MBR was attributed to lower observed P uptake in the anoxic zone. High nitrate loads to the anoxic reactor appeared to be the determining factor in stimulating P uptake.
The MBR UCT system had a greater sludge production than the conventional system. This was partly attributable to the retention of all solids in the MBR reactor. For steady state design this increase is accommodated by increasing the influent unbiodegradable particulate COD fraction. Additionally an attempt was made to determine the Alpha values in the oxygen transfer rate.
This paper briefly summarises and compares the results from both systems, and the conclusions that can be drawn from these results.