Qin Lu ,Haiyan Wu ,Haoyan Li,Dianhai Yang ,*

1 State Key Lab.of Pollution Control and Resources Reuse,College of Environmental Sciences and Engineering,Tongji University,Shanghai 200092,China

2 Xylem(China)Co.,Ltd.,Shanghai 200051,China

3 Shanghai Plant Biomass Co.,Ltd.,Shanghai 200092,China

Keywords:Modified AAO process Carbon source distribution ratio Returned activated sludge pre-concentration Biological nutrient removal

ABSTRACT A pilot-scale modified carbon source division anaerobic anoxic oxic(AAO)process with pre-concentration of returned activated sludge(RAS)was proposed in this study for the enhanced biological nutrient removal(BNR)of municipal wastewater with limited carbon source.The in fluent carbon source was fed in step while a novel RAS pre-concentration tank was adopted to improve BNR efficiency,and the effects of an in fluent carbon source distribution ratio and a RAS pre-concentration ratio were investigated.The results show that the removal efficiency of TN is mainly in fluenced by the carbon source distribution ratio while the TP removal relies on the RAS pre-concentration ratio.The optimum carbon source distribution ratio and RAS pre-concentration ratio are 60%and 50%,respectively,with an inner recycling ratio of 100%under the optimum steady operation of pilot test,reaching an average ef fluent TN concentration of 9.8 mg·L-1 with a removal efficiency of 63%and an average TP removal efficiency of 94%.The mechanism of nutrient removal is discussed and the kinetics is analyzed.The results reveal that the optimal carbon source distribution ratio provides sufficient denitrifying carbon source to each anoxic phase,reducing nitrate accumulation while the RAS pre-concentration ratio improves the condition of anaerobic zone to ensure the phosphorus release due to less nitrate in the returned sludge.Therefore,nitrifying bacteria,denitrifying bacteria and phosphorus accumulation organisms play an important role under the optimum condition,enhancing the performance of nutrient removal in this test.

1.Introduction

To prevent the eutrophication in enclosed water system and meet increasingly stringent ef fluent discharge standards,the upgrade of existing wastewater treatment plants(WWTPs)with biological nitrogen and phosphorus removal has been extensively investigated and developed [1–3].For the nutrient removal of municipal wastewater,especially with low carbon source available in most of southern China,it is significant to select the optimal BNR process and optimize the in fluent carbon source.The two points are the research focus and some modified BNR processes based on the conventional activated sludge process have been developed,including anaerobic/anoxic/oxic(AAO)and oxidation ditch.The results show that deep nitrogen removal and optimal phosphorus removal are achieved[4–7].

Due to extensive use of AAO process in municipal wastewater treatment,more researches have been carried out to obtain simultaneous nitrogen and phosphorus removal[8–10].Two strategies are usually employed in the modified AAO process:using the carbon source to a better extent and improving the situation to mitigate the competition of available organic substrates between denitrifiers and phosphorus accumulation organisms(PAOs).The optimal utilization of carbon source and advanced nitrogen removal has been achieved with step feeding in two,three and four stages[10–12].Wang et al.have indicated that the modified step-feeding anaerobic/anoxic/oxic nutrient removal system(SFA2/O)is a technically feasible and economically favorable process for simultaneous nitrogen and phosphorus removal from municipal wastewater without external carbon source[13].In this process,the optimization of feeding ratio is a key issue to make use of available organic substrate for nitrogen and phosphorus removal.Moreover,it is well known that the remainingbrought by the returned activated sludge(RAS)can change the anaerobic zone to an anoxic zone,inhibiting the biological phosphorus removal process[14,15].The remaining nitrate in the anaerobic zone leads to a competition for available carbon source between denitrifiers and PAOs,deteriorating significantly the anaerobic phosphorus release and reducing the phosphorus removal in the enhanced biological phosphorus removal system[16–18].Chang and Hao[19]observed that phosphorus removal efficiency increased with the reduction of nitrate content in ef fluent.Kazmi et al.[20]modeled the remainingshock on phosphorus removal and demonstrated that the phosphorus release was low in the anaerobic zone and TP contentwas higherin the ef fluent with the increase of the remainingconcentration.Zou et al.[21]found that anaerobic phosphorus release rate declined as soon asN was added to the anaerobic zone,resulting in low phosphorus removal efficiency.Peng et al.[22]showed that TP removal efficiency of 80%could be achieved when the content of remainingwas below 6 mg·L-1.These researches indicate that the remainingin the RAS will deteriorate biological phosphorus removal.However,some studies have verified that in the absence of exogenous carbon sources in the anaerobic zone,phosphorus removal may occur in the presence of nitrate[23–26].In addition,anoxic denitrifying phosphorus removal technology,depending on the denitrifying phosphate-accumulating organisms,can potentially save external carbon source and energy and achieve better phosphorus removal[18,27–29].Denitrifying phosphorus removal has been verified to be beneficial to enhance simultaneous nitrogen and phosphorus removal in SFA2/O system[30].

Some novel and modified step feeding processes have been developed to enhance nitrogen or phosphorus removal capability and meet satisfactory ef fluent quality[31,32].However,previous studies on the step feeding process are mainly conducted with synthetic wastewater and focus on nitrogen removal performance.Although SFA2/O system has been employed in a pilot scale for nitrogen and phosphorus removal,stable nutrient removal and real-time control need to be strengthened,especially when applied to full-scale municipal WWTPs.

The previous study of our research team shows that preconcentration of RAS will significantly improve the phosphorus removal efficiency,where the ratio of returned flow of a preconcentration tank and a RAS flow of secondary clarifier(η)plays an important role[33].In this study,a pre-anoxic tank is introduced before the anaerobic tank,followed by two stages of anoxic/oxidation(A/O)for part nitrogen removal and elimination of the effect ofin returned liquid to the following anaerobic zone for a strict anaerobic condition.Besides,a pre-concentration tank is introduced to concentrate the RAS before it returns to the pre-anoxic tank to reduce the dilution effect of in fluent carbon source and mainly mitigate the competition ofdenitrifying bacteria with PAOs for available organic carbon substrate for better phosphorus release.Moreover,the RAS preconcentration tank will enhance theremoval efficiency in the pre-anoxic zone.The objective of this study is to examine the technical feasibility of simultaneous nitrogen and phosphorus removal at different carbon source distribution ratios(λ)and RAS pre-concentration ratios(η)with a modified AAO process in a pilot scale.The effects of carbon source distribution and RAS pre-concentration on nitrogen and phosphorus removal efficiency are studied,and their optimum values are determined.Furthermore,the mechanismofnitrogen and phosphorus removal is discussed.The removalrates of nitrogen and phosphorus are determined under the optimal operation condition.Finally,some fundamentaldata are provided to the design of full-scale WWTPs to enhance BNR with this modified AAO system.

2.Materials and Methods

2.1.Experimental set-up and operational conditions

Fig.1 shows a pilot-scale modified AAO system,which consist of a pre-anoxic zone(1.1 m3)and an anaerobic zone(3.0 m3)followed by two identical pairs of anoxic[1.6 m3·(1.6 m3)-1]and aerobic zones[3.2 m3·(2.1 m3)-1]in series.A secondary clarifier was arranged at the end of the process and a pre-concentration tank was for concentrating RAS.The in fluent of 2 m3·h-1was pumped into the anaerobic zone with hydraulic retention time(HRT)of 0.5 and 1.5 h.The ef fluent of anaerobic zone was divided into two parts:one part flowing into the first anoxic chamber and the rest by-passing to the secondary anoxic chamber directly with a ratio defined as distribution ratio(λ)in this study.The following two stages of anoxic/aerobic(A/O)were with HRT of1.5,3.0,1.5 and 2.0 h in sequence followed by the secondary clarifier with HRT of 2.5 h.The RAS from the secondary clarifier was pumped back to the pre-thickener for sludge pre-thickening.Then the pre-thickened RAS was pumped into the pre-anoxic zone for eliminating NO3-N by endogenous denitrification,while the supernatant generated in the pre-concentration zone flowed back to the second stage of aerobic zone by gravity,as the returned thickened sludge from pre-thickener is less than the RAS from secondary clarifier.Mixed liquid was pumped from the secondary aerobic zone to the first anoxic zone as the internal recycle.Several mechanical stirrers were equipped in pre-anoxic,anaerobic and anoxic zones separately to ensure complete mixing ofsludge and wastewater.Two sets of fine bubble air diffusers were installed at the bottom of two aerobic zones to supply oxygen.DO probe was placed in the aerobic zones and DO concentrations were controlled automatically at the preselected set point(～1.5 mg·L-1).The sludge retention time was maintained at 15 d by controlling sludge wastage.The resulting mixed liquor suspended solid(MLSS)concentration was in the range of 3700–4000 mg·L-1.The sludge recycle ratio was constantly set at 0.75.DO,pH,ORP,MLSS and flow rate were all measured on-line.A RAS pump and a preconcentrated RAS pump were controlled by PLC.The experiment was carried out from February 2011 to August 2011(182 d).All the experiments were conducted at temperature 19–25 °C.Each run lasted for at least 30 days.

2.2.Sludge and wastewater characteristics

The pilot-scale reactor was installed in Longwangzui WWTP(150000 m3·d-1)with AAO process in Wuhan,China.The reactor was operated for about one year.The seed sludge was taken from the aerobic tank of a full-scale WWTP.After one month of acclimation periods,the process was stable and experiments were carried out.

The in fluent fed to the reactor was fetched from the grid removal tank of WWTP by in fluent pump.The MLSS in the pilot reactor were controlled to(3500 ± 500)mg·L-1.The major characteristics of the in fluent with limited carbon source are shown in Table 1.

2.3.Batch tests

To evaluate the sludge performance of nitrification,denitrification,phosphorous release and uptake,batch tests were carried out.For nitrification,the activated sludge of aerobic phase of pilot system was transferred to a 12 L Sequencing Batch Reactor(SBR)reactor.NH4Cl was fed to maintain the initial NH4-N+concentration at 30 mg·L-1.Air was supplied to keep the DO of mixed liquid at 2–3 mg·L-1and samples were taken to analyzeandat 10 min interval.For denitrification,anoxic sludge was transferred to a 12 L SBR reactor,then NaAc and KNO3was fed to obtain initial COD andconcentrations of 250 and 30 mg·L-1,respectively.DO was controlled under 0.2 mg·L-1and samples were taken with 10 min interval.For endogenous denitrification,the sludge of pre-concentration tank was transferred to a 12 L SBR reactor and KNO3was fed to obtain the initialconcentrations of 30 mg·L-1.DO was controlled under 0.2 mg·L-1and samples were taken with 20 min interval.

Fig.1.Schematic diagram ofa novelpilot-scale AAOsystem.(1)in fluent；(2)in fluentpump；(3)pre-anoxic zone；(4)anaerobic zone；(5)anoxic zone I；(6)aerobic zone I；(7)anoxic zone II；(8)aerobic zone II；(9)by-pass flow；(10)internal recycle pump；(11)secondary clarifier；(12)RAS pump；(13)pre-concentration tank；(14)pre-concentration RAS pump；(15)supernatant fluid；(16)blower.

Table 1 Major characteristics of the in fluent(mg·L-1)

To investigate phosphorous release and uptake,the anaerobic sludge of pilot system was transferred to a 12 L SBR reactor.NaAc and K2HPO4were fed to obtain the initial COD andconcentrations of 100 and 4 mg·L-1,respectively.The mixed liquid was kept in anaerobic environment for 2 h with stirring,and then transferred to two parallel 6 L SBR reactors for aerobic and anoxic phosphorus absorption.In the aerobic reactor,DO was kept at 2–3 mg·L-1.In the anoxic reactor,KNO3was fed to obtain initialconcentration of 20 mg·L-1.

2.4.Analytical methods

For analysis ofdissolved substances,wastewatersamples taken from the reactor were filtered by 0.45 mm filter papers to separate bacterial cells from liquid column and immediately cooled in order to prevent further reaction.CODCr,MLSS,MLVSS,,,,TN,and TP of wastewater samples were measured according to the standard methods(APHA,1995)[34],while TP of sludge was analyzed by the persulfate digestion followed by the vanado-molybdate colorimetric method.The DO,pH,ORP and temperature were measured continuously by on-line probes(HACH,U.S.).

3.Results and Discussion

3.1.Effects of carbon source distribution ratio on the performance of modified AAO system

Fig.2 illustrates in fluent and ef fluent concentrations and removal efficiencies of COD,,and TN at different carbon source distribution ratios(λ)with an inner recycle ratio of 100%.Although the in fluent COD fluctuated between 100.4 and 249.0 mg·L-1,the removal performance was relatively stable with an average ef fluent concentration of 19.9 mg·L-1.The carbon source distribution ratio(λ)did not have significant effect on COD removal.Most in fluent COD was consumed in anaerobic or anoxic zones.It is considered to be advantageous for the nitrification without the inhibitory effect caused by residual COD[35].Forremoval,the in fluent varied distinctly in the range of 0.6 mg·L-1to 3.4 mg·L-1,the ef fluent and its removal efficiency fluctuated accordingly with an exception of λ=80%,achieving a relatively stable and highremoval efficiency of 96%on average and the mean ef fluent concentration of 0.1 mg·L-1.Thus the carbon source distribution ratio(λ)did not have a significant effect on phosphorus removal either.For TP removal,the ef fluent TP concentration decreased as λ increased from 40%to 60%,and then increased as λ increased continuously to 100%.At λ of 40%–100%,the average concentration of ef fluent TP was 0.4 mg·L-1,satisfying the first-A wastewater discharge standard in China(TP ＜0.5 mg·L-1).The TP removal efficiency was in fluenced significantly by carbon source distribution ratio,decreasing first and then increasing withλ increasing from 40%to 60%.At λ of 60%,TP removal efficiency was relatively stable and higher,with an average of94%[Fig.2(b)].Forand TNremovalas shown in Fig.2(c),the ef fluent was relatively stable and an averageconcentration of 0.5 mg·L-1was achieved with an average removal efficiency of 97%when the in fluent fluctuated between 6.3 and 24.2 mg·L-1,which agrees well with the COD.The bestremoval occurred at λ of 100%with the average removal efficiency of 99%.The carbon source distribution ratio did not exhibit distinct effect onremoval.For TN removal,however,λ of 60%led to more stable and higher removal with an average ef fluent TN concentration of 9.8 mg·L-1and a removal efficiency of 63%.Under other conditions,especially for λ＞60%,the TN removal efficiency decreased obviously and fluctuated intensely.

Therefore,the carbon source distribution ratio(λ)has a significanteffect on TN and TP removal but little effect on COD,andremoval.At λ of 60%,the removal efficiencies of TN and TP were significantly higher,so it is regarded as the optimum value.

3.2.Effects of RAS pre-concentration ratio on the performance of modified AAO process

Fig.2.Removal of COD(a),, and TN(c)at different λ.

Fig.3.Removal of COD(a),TN(b)and TP(c)at different η.

Fig.3 shows the effects ofRASpre-concentration ratio(η)on ef fluent concentration and removal efficiencies of COD,,TN and TP with an inner recycle ratio of 100%and carbon source distribution ratio of 60%.The COD removal efficiency with sludge pre-concentration was slightly higher than that without pre-concentration.No matter the sludge was concentrated or not,to intensely fluctuated in fluent COD,the removal performance of COD was relatively stable with the average ef fluent concentration of 32.9 mg·L-1and average removal rate of 85.8%.Obviously,the RAS pre-concentration ratio(η)does not have a significant effect on COD removal.However,with similar trend in concentrations ofand TN in the in fluent,the ef fluent concentration and removal rate ofand TN exhibited diverse egulations atηof 50%(run 2),70%(run 3)and zero(run 1).The ef fluent concentration and removal efficiency ofat different η were almost the same.The averageef fluent concentration was about 0.31 mg·L-1and in the order of run 3＜run 1＜run 2.The averageremoval rate was about 98.6%and in the order of run 3＞run 1＞run 2.For TN removal,the ef fluent concentration and removal rate with RAS pre-concentration were more stable than those without RAS pre-concentration.The RAS pre-concentration ratio of 50%gave slightly higher TN removal efficiency,57.3%,than that without RAS preconcentration,56.8%,and that with η of 70%,56.1%.Thus the RAS preconcentration ratio(η)does not have a significant in fluence onand TN removal.In Fig.3(c),with severely fluctuated in fluent TP,the ef fluent TP concentration was relatively stable especially with the RAS pre-concentration ratio of 50%.Compared to that without RAS preconcentration,the TP ef fluent concentration with pre-concentration decreased significantly especially at η=50%,with an average value of 0.38 mg·L-1.This value was much lower than that without RAS preconcentration(an average of 0.81 mg·L-1).The removal efficiency of TP showed almost the same trend.At η of 50%,the removal efficiency of TP increased with time and an average of 91.4%was achieved,higher than that with η of 70%,87.9%,and without RAS pre-concentration,81.7%.It can be concluded that the degradation and removal of TP are affected significantly by RAS pre-concentration ratio and preconcentrated sludge is beneficial for phosphorus removal,probably due to improved anaerobic environmentin anaerobic zone and less inhabitation for nitrate in RAS to form PAOs.Accordingly,50%is regarded as an optimal RAS pre-concentration ratio in this pilot test with an inner recycling ratio of 100%.

3.3.Analysis for stable operation and removal mechanisms of organics and nutrients

Under the optimal operational conditions of carbon source distribution ratio of 60%,RAS pre-concentration ratio of 50%,inner recycling ratio of 100%,in fluent of 70 m3per day,sludge retention time of 15 d,DO of 1.5 mg·L-1in aerobic tank,and discharging sludge amount of 1.08 m3·d-1,the pilot system operated steadily for one month at temperature of(21±2)°C,and the results are shown in Fig.4.Excellent and steady performance was achieved.The ef fluent concentrations of COD,,TN and TP were stable and the removal efficiencies were high,the average removal efficiency reaching 82.8%,98.8%,59.6%and 91%,respectively.In particular,the removal of nitrogen was enhanced with the optimum carbon source distribution ratio.The phosphorus removal capacity of the system was also enhanced remarkably and the TP ef fluent concentrations were all below 0.5 mg·L-1with the RAS pre-concentration tank.The whole system demonstrated good performance for contaminate degradation and the ef fluent preliminarily attained level A discharge standard steadily,which agrees well with other observations[34,35].

The removal characteristics of organics and nutrients and mechanism of the pilot system under the optimum operation condition are shown in Fig.5.The removal of COD mainly occurred in the anaerobic and anoxic zones.When the wastewater entered anaerobic zone,the COD was diluted and the concentration of COD decreased rapidly due to mixing and reflux of activated sludge,then further decreased by adsorption of activated sludge and anaerobic phosphorus release of PAOs.Meanwhile,granular COD was adsorbed by bacteria and less soluble COD in wastewater in anoxic zone was degraded by denitrifying bacteria.When the wastewater entered the aerobic zone,COD concentration did not have major changes with less utilization by bacteria.The,,andconcentrations in several bio-chemical reaction tanks were measured[Fig.5(b)].was oxidized primarily in the aerobic tank by nitrifying bacteria and transformed toand,which returned with mixed liquor to the anoxic zone and removed by denitrification.Most ofin the returned sludge was removed by the endogenous denitrification in a pre-concentration tank and a preanoxic tank and less of them,about 0.03 mg·L-1of NO3-N,entered the anaerobic zone and mixed with the in fluent.The accumulation ofwas not observed in steady operation.With sufficient carbon source for nitrifying bacteria and denitrifying bacteria to develop ammoniation,nitrification and denitrification and the optimum carbon source distribution ratio under the optimum operation condition,good nitrogen removal effect was achieved in this test,which is consistent with other studies[33,35].The changes of phosphate concentration with time are shown in Fig.5(c).Under the optimum condition,a good phosphorus release effect in anaerobic zone was achieved and phosphorus uptake in the aerobic zone also met the corresponding requirement.Denitrifying phosphorus removal was observed in the anoxic zone,decreasing the soluble phosphate concentration.No phosphate release was found in returned sludge and the phosphate concentration was controlled at a low level,which could be in favor of phosphorus removal effect[28].

Fig.4.Changes of COD,,TN and TP with time under optimum operation conditions in pilot system.

Fig.5.Removal characteristics of organics and nutrients and mechanism of the pilot system under the optimum operation conditions.

3.4.Removalrate fornitrogen and phosphorus under the optimaloperation condition

In orderto investigate the effectofprocess condition on sludge property and further explain the good nutrient removal performance,sequencing batch experiments were conducted and rates of sludge anaerobic phosphorus release,aerobic phosphorus uptake,nitrifying and denitrifying were studied at 20°C after the pilot process was operated steadily for 20 d under the optimum operating condition.The results are shown in Fig.6 and Table 2.Normally,more complete aerobic nitrifying,anoxic denitrifying,anaerobic phosphorus release and anoxic and aerobic phosphorus uptake are regarded as a guarantee to achieve better nitrogen and phosphorus removal.Wang et al.[13]investigated phosphorus removal characteristic in different phases of modified step-feeding nutrient removal system and achieved a specific anaerobic P release rate of anaerobic zone as 0.0542 g PO4-P·(g VSS·d)-1.

Higher nitrifying rate and phosphorus release rate were achieved and the normal denitrifying rate was far above the endogenous denitrifying rate.Phosphorus uptake involves two ways ofaerobic phosphorus uptake and denitrifying phosphorus removal,contributed to a certain percentage of phosphorus removal.The maximum extent of nitrifying,denitrifying,phosphorus release and phosphorus uptake was obtained under the optimal operation condition,achieving good nitrogen and phosphorus removal efficiency.Additionally,the operation of entire pilot system was stable and reliable.

4.Conclusions

The pilot-scale modified carbon source division AAO process with RAS pre-concentration was confirmed to be a technically feasible and economical favorable process for nitrogen and phosphorus removal.With 100%inner recycle ratio,50%RAS pre-concentration ratio and 60%carbon source distribution,this novel process exhibited a significant effect on TN and TP removal with the removal efficiency of 63%and 94%,respectively.The distributed carbon source was utilized adequately by denitrifying bacteria in different anoxic zones.The competition for carbon source from denitrifiers in the anaerobic zone was avoided by RAS pre-concentration so that the anaerobic phosphorus release,carbon source uptake and the aerobic phosphorus uptake by PAOs were promoted.It was found that much more PAOs accumulated in the system with RAS pre-concentration.The nitrifying,denitrifying,endogenous denitrifying ratio,anaerobic phosphorus release,and aerobic phosphorus uptake ratio with a specific ratio of denitrifying phosphorus removal were higher,ensuring good performance of nutrient removal in this pilot modified AAO system.

Fig.6.Sludge biochemicalreaction rates for nitrification(a),denitrification(b),phosphorus removal(c)and endogenous denitrification(d)under the optimaloperation condition in pilot system.

Table 2 Sludge biochemical reaction rates under the optimal operation condition in pilot system