壓縮包內(nèi)含有CAD圖紙和說(shuō)明書(shū),均可直接下載獲得文件，所見(jiàn)所得，電腦查看更方便。Q 197216396 或 11970985
The effect of substrate composition on the nutrient
removalpotential of sequencing batch reactors
Rüya Tasli*, Derin Orhon and Nazik Artan
Environmental Engineering Department, Istanbul Technical University, I.T.U.
Insaat Fakültesi, 80626 Maslak, Istanbul, Turkey
Abstract
Experimental results suggest that sequencing batch reactors (SBR) are not efficient for enhanced biological phosphorus removal from domestic sewage with low/medium organic carbon content when denitrification preferentially competes for available carbon.Total COD is not a meaningful parameter to reflect available substrate for N and P removal; COD fractionation and identification of the readily biodegradable COD fraction are required for an accurate assessment of system performance. The degree of soluble COD removal in the non-aerated phase is observed to be much higher than what may be calculated from stoichiometric relationships for N removal and P release, indicating the existence of competing mechanisms such as organic carbon storage by non-polyP bacteria under anaerobic conditions.
Keyword：sequencing batch reactors (SBR)，organic carbon，phosphorus removal
1 Introduction
It is well-known that enhanced biological phosphorus removal (EBPR) requires an anaerobic/aerobic sequence. In the anaerobic stage, phosphorus release is associated with storage of organic substrate within biomass. During the aerobic stage, excess phosphorus uptake takes place at the expense of stored organics which also serve as carbon source for the growth of micro-organisms which have the ability to accumulate polyphosphates (polyP bacteria).
The nature of the organic substrate available plays a key role for effective phosphorus removal which selectively requires the existence of short-chain fatty acids. It is now commonly agreed that the major function of the anaerobic stage is to generate the necessary fermented substrate which can be utilised by polyP bacteria. In all the kinetic models so far proposed, acetate is referred to as the sole external substrate which is taken up and stored as poly? hydroxybutyrate (PHB) during anaerobic conditions. The question of how a reducing power is created for PHB synthesis constitutes the main distinction between various biochemical models (Comeau et al., 1986; Wentzel et al., 1986; Mino et al., 1987). It has been shown that other short-chain fatty acids may also be used to a limited extent for the same purpose and in such cases, other polyhydroxyalkanoates are synthesised besides PHB (Satoh et al.,1992).
Currently, the sequencing batch reactor (SBR) technology is a well promoted and tested alternative with distinct advantages over the conventional activated sludge process. Basically, the SBR is a very simple system involving a single tank (Orhon and Artan,1994). As it is temporally controlled, as contrasted to spatially controlled conventional continuous-flow processes, it offers a major advantage for the observation and the interpretation of different phenomena associated with the anaerobic/aerobic sequence of the EBPR process.
The magnitude of the available organic substrate and its readily biodegradable fraction is likely to have a decisive impact on the extent of EBPR from domestic sewage and such an impact can best be visualised within an SBR. EBPR may also be severely affected by simultaneous nitrogen removal which is often desired in systems treating domestic sewage. The presence of nitrate at the beginning of the anaerobic phase is detrimental to EBPR, since oxidised nitrogen will preferentially consume organic substrate for denitrification and this will create an anoxic phase before true anaerobic conditions can be sustained. In this case, the amount and the nature of substrate will have to be sufficient both for denitrification and PHB storage for EBPR.
The main objective of this experimental study was to investigate the effect of different substrate conditions and simultaneous nitrogen removal on the EBPR efficiency of an SBR system treating domestic sewage with a fluctuating organic content.
2 Materials and methods
The experimental work was carried out in a timer-controlled,laboratory-scale sequencing batch reactor with a total volume adjustable up to 8.8l . The reactor was equipped with both mechanical mixing and diffused aeration devices to create a sequence of anaerobic/anoxic and aerobic conditions. Wastewater feeding dur-ing
the fill period was secured by an adjustable-flow Watson-Marlow type peristaltic pump and three different exit ports, each controlled by a solenoid valve, which provided discharge of the treated effluent.
The laboratory SBR unit was operated for around 200 d in four different consecutive runs. The operation was adjusted to four cycles a day (cycle time, t c = 6 h) with the exception of run III where the cycle time was extended to 8 h (three cycles a day). Each cycle involved the regular consecutive sequence of anoxic/anaerobic fill and mixing phase, the aerated reaction phase and the settle/idle phase where the treated effluent was discharged. The operation conditions and the characteristics of the laboratory SBR unit are outlined in Table 1. The sludge age was controlled and the mixed liquor volatile suspended solids (MLVSS) concentration was kept constant for each individual run, by wasting the required amount of activated sludge, once a day, at the end of the aerobic period of the same cycle.
During the start-up period (Run I) and Runs II and III, the system was fed with domestic sewage only, collected from the Greater Metropolitan Area of Istanbul. The relatively strong sewage from the Tuzla collection area was used for Run I and sewage
from the Atak?y region was provided for Runs II and III. Run IV was conducted with four different feeding patterns, each using the domestic sewage from Atak?y, supplemented with different acetate concentrations.
The performance of the SBR system during each run was monitored continuously and evaluated at cyclic steady state conditions,in terms of daily measurements of the influent and effluent characteristics associated with a selected cycle. The evaluation also involved the fate of soluble COD, phosphate, ammonia and nitrate
concentration profiles within a selected cycle, under steady state conditions characterising different runs.
All analyses were performed in accordance with Standard Methods (1989). The soluble fraction of COD and the other parameters were defined as filtrates through Whatman CF/C glass fiber filters with a effective pore size of around 1 m. The same filters were also used to assess particulate components on a sus-pended solids (SS) or volatile suspended solids (VSS) basis.Phosphate and total phosphorus were measured using the ascorbic acid and persulphate digestion methods, respectively. Nitrate was reduced to nitrite in a cadmium reduction column and quantified with a colorimetric method. Ammonia was measured by means of the phenate method. The closed reflux titrimetric method was used for COD assessments.
3 Experimental results and evaluation
3.1 Start-up period
The study was started using a previously running SBR system.In this period, the reactor was fed with a synthetic wastewater basically prepared with Tryptose Soy Broth, (TSB) adjusted to an influent COD of 275 mg/1 and a total P concentration of 8.3 mg/1 . At steady state, the SBR was operated to sustain an MLVSS concentration of 1 070 mg/l , with a VSS/SS ratio of 0.76. Complete P removal was secured, resulting in a relatively high total P/MLVSS ratio of 8.3% for EBPR systems. The PO 4 -P and COD profiles observed during the cyclic operation of the SBR system is illustrated in Fig. 1 (Tasli et al., 1997a).
During the 76 d start-up period, the SBR system was acclimated to domestic sewage provided from the Tuzla region in Istanbul, a relatively strong wastewater with average COD, TKN and total P concentrations of 420, 72, and 11 mg/1 respectively. The system reached steady state after the
first 50 d of operation: The MLVSS concentration increased from 1 070 mg/l to 2 880 mg/1 , then stabilised at around 2 300 mg/l ; the VSS/SS ratio dropped to 0.56 as a result of sewage with a relatively high concentration of inert suspended solids. Nitrification developed leading to a final NO 3 -N level of 17 mg/1 , with a corresponding TKN concentration of 4 mg/1 . As far as EBPR was concerned, the effluent PO 4 –P concentration first increased to 7 to 8 mg/1 to finally level at 4.2mg/1 . The overall performance of the SBR during the start-upperiod is outlined in Table 2, along with the results associated with the other runs. The effluent quality obtained exhibits typical characteristics of a well-nitrifying system, with partial P removal.
3.2 Experiments with domestic sewage
The first part of the experimental study after start-up, involved investigating the EBPR potential of SBR using domestic sewage with a lower organic carbon content. The system was fed with domestic wastewater from the Atak?y area in Istanbul with an average COD content of 300 to 325 mg/1 and a moderate COD/P
ratio in the range of 33 to 39.
The study was conducted with the main objectivet oexplore the impact of simultaneous nitrogen removal on EBPR, which is often desired in biological treatment processes. Moreover, simultaneous nitrogen removal can hardly be avoided, unless the system is operated at very low sludge ages. Unlike coventional continuous processes,an anoxic phase is inherently establishedinSBRsbefore the true anaerobic phase, where the readily biodegradable COD fraction is consumed at the expense of available NO -3 -N, reducing thePHB storageandthephosphorus release associated with EBPR. Therefore, the degree of P removal that can be achieved in nutrient removing SBRsystemsdependsupon thedelicatebalancebetween organic carbon, nitrogen and phophorus levels and important operation parameters such as V 0 /V F ratio and the duration of the anoxic/aerobic conditions.
Experiments with domestic sewage were conducted in two stages. The SBR operation in the first stage (Run II), continued for 28 d, involving four cycles a day (6 h cycles). Each cycle included first a 2.5 h anoxic/anaerobic phase, where domestic sewage was fed during the first 2 h , a 2.5 h aerated phase, 0.5 h more than that applied in the start-up period, and a final 1 h settle, discharge/idle
phase, a period kept relatively short because of good settling conditions. In this run, the initial SBR volume (V 0 ) was reduced to 3.3l; a fill volume (V F ) of 2.0 _ was maintained, resulting in a V F /V T ratio of 0.37, and a V 0 /V F ratio of 1.65; the MLVSS concentration was slightly increased to 2 930 mg/l with a VSS/SS ratio of 0.6l . In this operation, the average effluent characteristics of NH 3 -N = 2.0 mg/l , NO x -N = 25 mg/l and PO 4 -P = 5.8 mg/1 , as listed in Table 2, indicate that both nitrogen and phosphorus removal efficiencies deteriorated.
Limited N and P removal in the SBR system may be attributed to either a lower readily biodegradable COD available at the beginning of the anoxic/anaerobic phase, or insufficient reaction time under anoxic/anaerobic conditions. To test the latter, the second stage operation (Run III-1), was adjusted to three cycles a day, thus creating a longer anoxic/anaerobic reaction time; the aeration phase was also set as 3.5 h. Accordingly, V F was increased to 2.6l ; V 0 was maintained as 3.3 _ , leading to V F /V T = 0.44 and V 0 /V F = 1.27. The MLVSS concentration was reduced to 2 050 mg/l with VSS/SS = 0.7. During the 20 d of operation, the influent quality was practically the same as that of the previous period though with a slightly lower TKN value and extending the anoxic/ anaerobic phase provided, as shown in Table 2, only a slight improvement in nitrogen and phosphorus efficiencies; basically,the effluent NO X -N dropped from 25 mg/l to 18.5 mg/l , inducing a moderate reduction in the corresponding PO 4 -P concentration to 4.8 mg/l .
In the following 22 d (Run III-2), the anoxic/anaerobic phase was further increased to 4 h, keeping the three 8 h cycles, also increasing the MLVSS concentration to 2 840 mg/l and VSS/SS to 0.76. This change generated an effluent with only slightly lower NO x -N and PO 4 -P concentrations of 17 mg/l and 3.5 mg/l ,respectively, providing a clear indication that the limiting factor was not the anoxic/anaerobic reaction time, but the amount of readily biodegradable COD to satisfy both the requirements of simultaneous denitrification and EBPR.
3.3 Experiments with domestic sewage supplemented by acetate
This part of the experiments (Run IV), was designed to investigate the impact of the available readily biodegradable COD on N and P removals, by supplementing the domestic sewage feed with different concentrations of acetate. During the entire run, a 6 h cycle time was adopted and the SBR volume fractions were maintained as V 0= 3.3 _ , V F = 2 _ and V T = 5.3 _ as in Run II with domestic sewage corresponding to V F /V T = 0.37 and V 0 /V F = 1.65.
During the first two parts of the experiments (Runs IV-1 and IV-2), the acetate addition was adjusted to 50 and 100 mg/l COD equivalent of acetate, increasing the influent total COD concentra-tion to 350 mg/l and 490 mg/l respectively. The corresponding COD/P ratios were calculated as 43 in the first part and 65 in the second part. As shown in Table 2, the desired EBPR performance could not be secured in both parts where the average effluent PO 4 -P concentrations were measured as 4.6 mg/l and 4.0 mg/l .
In Run IV-2, the P/VSS level was only 3.5%, a value that was also obtained in the previous experiments with only domestic sewage. In the following part (Run IV-3), the acetate addition was further increased to 300 mg/l COD equivalent, with a correspond-ing total COD concentration of 530 mg/l and a COD/P ratio of 66. As outlined in Table 2, this part of the experiments was character-ised by an effluent PO 4 -P of 0.15 mg/l and NO x -N of 12 mg/l . In the final part (Run IV-4), where the domestic sewage feed was particularly weak, the acetate concentration was lowered to 150 mg/l COD equivalent, resulting in a total COD of 365 mg/l ; the same system performance could still be maintained with an effluent NO x -N = 12 mg/l and a slightly higher PO 4 -P of 1.0 mg/l . The results in Table 2 showed that the amount of readily biodegradable COD available in Run IV-3 with an acetate adition of 300 mg/l COD equivalent was more than what was required for a complete phosphorus removal and a nitrogen removal level that is potentially achievable with a pre-denitrification scheme simu-lated by the SBR operation. In this part, the SBR was observed to sustain a MLVSS concentration of 2 040 mg/l , with a P/VSS level of 5.3%, more compatible with EBPR.
3.3 The impact of denitrification on EBPR
PHB storage and phosphorus release during the first phase of the EBPR closely relates to the extent of readily biodegradable substrate available for this process. If simultaneous nitrogen removal oc-curs,this phase will initially sustain an oxidised nitrogen pool,(NO x -N) acting as electron acceptor in the absence of dissolved oxygen and preferentially consuming the existing readily biodegradable substrate. Therefore, an anoxic phase is established before the true anaerobic conditions, significantly reducing EBPR, because the substrate fraction that is likely to be stored as PHB is diminished and the reaction time for acidification and formation of fermentation products (acetate), if present, is shortened. As a result, EBPR cannot take place or it can proceed only at a very reduced rate as reported in the literature (Satoh et al., 1994; Tasli et al., 1997b).
In this study the impact of nitrogen removal on EBPR was investigated by measuring and observing the concentration profiles of relevant parameters within a selected complete cycle during the steady state operation of the SBR under different conditions. The first cyclic experiment relates to Run II where the SBR was fed with
domestic sewage alone in four 6 h cycles, with a V 0 /V F ratio of 1.65. On the day the experiment was carried out, the sewage character-istics were measured as COD = 220 mg/l , TKN = 40 mg/l and total P = 6 mg/l , a wastewater with a low organic content and consequently a low COD/N ratio of 5.75. The concentration profiles observed within one full cycle are illustrated in Fig. 2.
As shown in this figure, the SBR performed very poorly,producing an effluent with a NO X -N concentration of 22.5 mg/l ,and a PO 4 -P concentration of 4.8 mg/l . During the aerated phasealmost no phosphorus uptake occurs as evidenced by a PO 4 -P drop from 5.5 mg/l to only 4.8 mg/l . The rational interpretation of this observation closely relates to nitrogen transformations taking place in the sequence of anoxic/aerobic phases of the cycle: The aerobic sludge age of the system was sufficient to secure complete nitrifi-cation and the NH 3 -N content of the aerated phase was fully converted to NO x -N. The V 0 /V F ratio in the SBR corresponds to the total recirculation ratio in continuous systems; consequently, a V 0 /V F ratio of 1.65 would potentially allow for an effluent NO X –N concentration of around 13 mg/l , provided that the non-aerated phase secures full denitrification. For this experiment however, this was not the case, as the non-aerated phase remained fully anoxic and the NO x -N profile persisted during the entire period, dropping from 22 mg/l to only 12 mg/l . Apparently, the COD content and its readily biodegradable fraction was not sufficient to consume all available oxidised nitrogen as electron acceptor and the polyP bacteria could not compete with denitrifiers for suitable
biodegradable COD.
4 Conclusions
The results and evaluations of this study may be summarised as follows:
（1）The main difference between continuous-flow and SBR acti-vated sludge systems for N and P removal is that an anoxic phase is inherently established before truly anaerobic condi-tions leading to PHB storage and P release. Denitrification preferentially consumes available organic carbon, thus block-ng biological processes related to EBPR. Therefore, as the experimental findings of this study confirm, SBR is not likely to perform efficiently with respect to EBPR, for domestic sewage with low/medium RBCOD content.
（2） As the process efficiency should be directly related to RBCOD availability in the influent wastewater, total COD or ratios based on total COD, such as COD/N, COD/P are incapable of reflecting the available readily biodegradable organic carbon potential within the non-aerated phase. Experimental results showed that while domestic sewage with a total COD content of 390 mg/l could not even satisfy the organic carbon demand of denitrification, full N and P removal was restored when a mixture of domestic sewage and acetate with a COD equivalent of only 300 mg/l was used.
11