For the verification, the model was operated in a dynamic simulation mode over several 31 day periods was evaluated against observations from a full scale IFAS system. The evaluation was performed for the semi-empirical and the biofilm 1D versions. While both versions were able to predict the average effluent ammonium-N, oxidized-N, reactor MLSS, MLVSS and waste sludge production accurately (within 10% for MLSS, MLVSS, WAS and N, 20% at N concentrations less than 1 mg/L), the biofilm 1D model was able to simulate the day-to-day variations in nitrogen forms better. Also, the biofilm 1D model was able to predict the biofilm thickness and growth to within 20% in each of the aerobic cells operated with media. Because of the significantly longer time required to run Biofilm 1D models and similarity in the results for steady state and average 31 day averages in a dynamic simulation mode, the semi-empirical model was used to analyze alternatives several alternatives for tertiary removal together with with limited number of runs of Biofilm 1D model.
Several configurations were evaluated for an existing tertiary activated sludge system. The conditions evaluated were similar to those observed at the DCWASA (Blue Plains) plant. The system was operated with 40 to 60% aerobic volume and 40% post-anoxic volume (with methanol addition) at MLSS MCRTs (mean cell residence time) of 8 to 15 days. The plant can achieve near complete nitrification but suffers from partial loss of denitrification in winter (NOxN increases from 4 to 12 mg/L). Winter temperatures drop to a range between 13 and 15 C. The TSS, VSS and BOD5 loadings to the tertiary system can increase during wet weather conditions. This necessitates evaluation of both normal and wet-weather conditions for tertiary systems that are coupled to the performance of upstream secondary systems. The configurations evaluated included (a) activated sludge, (b) activated sludge with media added to the anoxic cells, and (c) activated sludge with media added to both anoxic and aerobic cells. Following the application of media with a biofilm surface area of 350 m2/m3, the model showed that the 201 effluent oxidized-N could be reduced from 11 mg/L to 5 mg/L, as simulated under winter wet weather conditions. These conditions were evaluated at denitrification rates observed for methanogenic bacteria (0.7 to 0.75 d-1 at 15 C). Additionally, methanol dosing can be increased
by 20 % during winter wet weather conditions without increasing the effluent COD to enhance the denitrification rates without impacting the effluent soluble COD (SCOD).
Several additional alternatives were considered for plants where a new tertiary tank can be added downstream of a high rate secondary treatment system similar to Harrisburg, PA. These alternatives include a new MBBR dedicated to nitrification and discharging directly to a denitrification filter (no clarifiers downstream of a MBBR), nitrification and denitrification cells within the same MBBR volume (40% aerobic, 40% anoxic, 20% reaeration), and nitrification and denitrification in an IFAS operated at a MLSS MCRT of 4 days within the same volume. In the MBBR, the effluent MLSS increased from 25 to 35 mg/L when methanol was added and the effluent oxidized-N decreased from 12 to 6 mg/L, while maintaining an effluent ammonium-N below 1.5 mg/L at 10 C. The biofilm specific surface area had to be increased from 150 to 225 m2/m3. The IFAS system, which can nitrify and denitrify in both the biofilm and the mixed liquor had 1000 mg/L MLSS and achieved better performance (<0.5 mg/L NH4N, 6 mg/L NOxN, lower effluent SCOD) at the same volume and media specific surface area as the nitrification MBBR. However, unlike the nitrifying MBBR which can discharge directly to a denitrification filter, it requires clarifiers instead of the denitrifying filter.