- Email: [email protected]
Nutrient removal and energy recovery from high-rate activated sludge processes – Impact of sludge age
Accepted Manuscript Nutrient removal and energy recovery from high-rate activated sludge processes – impact of sludge age Huoqing Ge, Damien J. Batstone, Morgan Mouiche, Shihu Hu, Jurg Keller PII: DOI: Reference:
S0960-8524(17)31429-3 http://dx.doi.org/10.1016/j.biortech.2017.08.115 BITE 18724
To appear in:
Bioresource Technology
Received Date: Revised Date: Accepted Date:
3 July 2017 17 August 2017 18 August 2017
Please cite this article as: Ge, H., Batstone, D.J., Mouiche, M., Hu, S., Keller, J., Nutrient removal and energy recovery from high-rate activated sludge processes – impact of sludge age, Bioresource Technology (2017), doi: http://dx.doi.org/10.1016/j.biortech.2017.08.115
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Nutrient removal and energy recovery from high-rate activated sludge processes – impact of sludge age
Huoqing Ge,† Damien J. Batstone,†,§,*, Morgan Mouiche,‡ Shihu Hu,† Jurg Keller†,§
†
AWMC, Advanced Water Management Centre, The University of Queensland, St Lucia,
Queensland, 4072, Australia ‡
§
Engineering School EI. CESI, France CRC for Water Sensitive Cities, PO Box 8000, Clayton, Victoria 3800, Australia
*Corresponding author: Damien Batstone
Advanced Water Management Centre (AWMC), The University of Queensland, St Lucia, Queensland, 4072, Australia Phone: +61 7 3365 4727 Fax: +61 7 3365 4726 Email: [email protected]
Abstract: This study evaluated high-rate activated sludge treatment across a broad range of short solids retention times (SRT)s (0.5-3d) and found a strong SRT-outcome dependence for performance and subsequent anaerobic degradability of the sludge. Up to 50% total nitrogen, and 35% ammonia removal was also achieved at the longer SRTs, via partitioning rather than reaction. The aerobic SRT significantly affected the anaerobic degradability of the sludge produced (p<0.001), with degradability increasing from 66% to over 80% while reducing the SRT from 3d to 0.5d. This is higher than predicted by conventional models, likely due to additional mechanisms such as adsorption and storage, not included in these.
Keywords: A-stage, anaerobic, high-rate, energy recovery, SRT, model Highlights:
High-rate activated sludge enhances energy and nutrient recovery through short SRT Tested range of SRTs from 0.5d-3d for nutrient removal and energy recovery Energy recovery maximised at low SRT while nutrient removal maximised at long SRT Sludge degradability significantly higher than predicted through standard model
1. INTRODUCTION The goals of wastewater treatment are currently being expanded from the traditional removal of organic matter and nutrients (i.e. nitrogen and phosphorus) to include also energy (carbon) recovery and nutrient recovery (Batstone et al., 2015; Verstraete et al., 2016). The typical method of energy recovery in wastewater treatment plants (WWTPs) is to employ anaerobic digestion to treat waste sludge and produce biogas (methane) for onsite heat and energy generation, thereby compensating energy demands from plants. However, conventional biological nutrient removal (BNR) processes, treating domestic sewage at normal strength results in oxidation of a large fraction of organic carbon contained in wastewater due to the
need to remove nitrogen biologically, and due to long solids retention times (SRTs) (e.g. 1020d) which reduces biogas potential and sludge degradability (Gossett & Belser, 1982). A key approach to enhance energy efficiency wastewater treatment processes, is to redirect carbon and nutrients, particularly to the waste activated sludge stream, to enable energy recovery, while maintaining current treatment quality.
The most popular option to achieve this is high-rate activated sludge (HRAS), which maximises native solids capture and minimises biological oxidation. A-stage biological treatment is a popular option for this, which utilises a very short retention time activated sludge stage, with HRTs of 0.25-0.5 hours and SRTs of 0.5-3d days (Jimenez et al., 2015). This process requires approximately 70% less energy input compared to conventional BNR processes (e.g. with 10-15d SRT),(Ge et al., 2013) and focuses on the capture of carbon in a solids stream through a combination of adsorption, particulate emeshment, bioflocculation, accumulation (storage), and assimilation (growth), rather than oxidation (Jimenez et al., 2005). Energy-rich short-SRT sludge with inherently high degradability is then wasted from the A-stage process and digested anaerobically as a concentrate to produce methane. In this way, most of organic carbon in wastewater is made available for energy recovery. Recently, biological phosphorus (Bio-P) removal has been demonstrated to be feasible also at such short SRTs (i.e. 2d) (Ge et al., 2015), indicating that phosphorus in wastewater can be effectively captured and biologically concentrated in a solids (sludge) stream, concurrently with significant COD capture in HRAS. This phosphorus can be subsequently released during anaerobic sludge digestion and recovered through struvite crystallisation. All these advantages could potentially enable WWTPs to transform from major energy consumers to net energy generators as well as resource recovery/production plants.
So far, A-stage processes have been applied to full-scale WWTPs in Europe (e.g. Strass and Vienna WWTPs in Austria or Rotterdam-Dokhaven WWTP in Netherlands, etc.) and USA (e.g. Chesapeake-Elizabeth WWTP) (Jetten et al., 1997; Jimenez et al., 2015; Wett et al., 2007). Effective carbon removal is being achieved in all cases, and the removed COD is largely recovered through anaerobic sludge digestion in the form of methane that can be used for energy production. For example, in the Strass WWTP, efficient COD capture and conversion has resulted in an energy self-sufficiency of 108% (after implementation of deammonification for side-stream treatment) (Wett et al., 2007). Operation at short SRTs is well known to improve overall carbon capture and sludge degradability (Ge et al., 2013; Meerburg et al., 2015), but the extent to which capture and degradability improve across the broad range of SRTs commonly applied in high-rate activated sludge has not been assessed. In addition, the ability of the process to remove nutrients (particularly nitrogen) via assimilation (growth), has not been systematically assessed. This study addresses these limitations in detail for domestic wastewater treatment by operating a high-rate laboratory activated sludge system, and identifying how carbon and nitrogen capture, and degradability change in the 0.5-3d range, with relatively high resolution.
2. METHODS AND MATERIAL 2.1. Wastewater The feed wastewater used in this study was the wastewater effluent generated from a sewer biofilm reactor. This biofilm reactor was fed with real municipal wastewater collected on a weekly basis from a local sewage pumping station in Brisbane, Australia, and operated to mimic an anaerobic sewer pipe section for monitoring the production of methane and hydrogen sulfide. Therefore, the biofilm reactor effluent exhibited similar characteristics as the raw wastewater, but with a slightly lower COD level (approximately 10% less than the
raw wastewater). Regular analysis was performed to determine the characteristics and consistency of the feed wastewater. Averages of 20 samples taken over the 6 months of the study were TCOD – 393 mg L-1, SCOD – 290 mg L-1, pH 7-7.8, TKN - 65 mgN L-1, NH4+ 54 mgN L-1, TKP – 12 mgP L-1, and PO43- 9 mgP L-1. Standard deviations in all measures across 20 samples were on the order of 10% relative.
2.2. Reactor set-up and operation A lab-scale high-rate system used in this study consisted of an aerated bioreactor (300 mL working volume) followed by an intermediate clarifier (20cm diameter, 15cm depth) and was operated in a temperature controlled room (20-22°C) under continuous flow conditions. In this system, the sludge mixed liquor was directed from the bioreactor to the clarifier, where the mixed liquor was separated to generate an effluent stream for discharge and a thickened activated sludge stream that was returned to the bioreactor. The ratio of the return activated sludge (RAS) and the influent flow was maintained at 2:1. This high RAS ratio was required to make the small clarifier effective. The HRT in the bioreactor was maintained at 30 min, while the SRT was controlled by periodically wasting sludge from the bioreactor (three times per day), which was balanced by the solids discharged through the clarifier effluent. Air was continuously supplied to the bioreactor and the dissolved oxygen (DO, measured by an YSI DO membrane probe) level was maintained at 3-3.5 mg O2 L-1 (see details in Table 1). The pH was monitored by using a glass body pH probe (TPS, Australia), but not controlled. At start-up, the bioreactor was inoculated with sludge collected from a full-scale BNR plant treating domestic wastewater in Brisbane, Australia.
The high-rate system was operated for over 6 months. During this time, the SRT of the bioreactor was altered to create different operating periods, which are summarised in Table 1.
Real (observed) SRT was generally slightly different to targeted SRT due to variations in waste sludge concentration and losses in effluent, but values were generally close, and analysis considers real SRT. Mixed liquor suspended solids averaged 0.6 g L-1, but was heavily dependent on SRT, dropping as low as 0.2 g L-1 at 0.5d SRT (0.6g L-1 at 2d SRT). Each period was maintained for at least 7-8 SRTs to ensure characteristic operation was achieved at each operating point. Taking the solids concentration of the clarifier effluent into account, the real SRT of the bioreactor in some periods differed slightly from the target SRT. 2.3. Anaerobic sludge digestion batch tests Biochemical methane potential (BMP) tests were conducted at 37°C to assess the anaerobic degradability of the waste activated sludge produced in the high-rate bioreactor during Periods 2-6 and 10-11, corresponding to a sludge age of 0.5d, 0.75d, 1d, 1.5d, 2d, 2.5d and 3d, respectively. Methane production potential and sludge degradability (based on model based analysis of the experimental results, see below) were used as performance indicators.
BMP tests were performed in 160 mL non-stirred glass serum bottles (100 mL working volume) as previously reported (Jensen et al., 2011). Each bottle contained the pre-calculated volumes of the substrate and inoculum to maintain the substrate: inoculum ratio as approximately 0.75 (volatile solids (VS) mass basis). Inoculum used in the tests was collected from a full-scale anaerobic digester (35°C, 20d HRT) located in Brisbane, Australia. Bottles were then flushed with high purity nitrogen gas for 3 min (1 L min-1), sealed with a rubber stopper retained with an aluminium crimp-cap and stored in a temperature controlled incubator. Blanks only contained inoculum and MilliQ water to measure the background methane produced from the inoculum and this was subtracted from the test prior to parameter estimation. All tests were carried out in triplicates, and all error bars indicate 95% confidence in the average of the triplicates based on two-tailed t-tests.
2.4. Analysis 2.4.1. The high-rate wastewater treatment system To monitor the high-rate bioreactor, mixed liquor samples were collected regularly for chemical analyses, including total COD (TCOD), soluble COD (SCOD), total suspended solids (TSS), volatile suspended solids (VSS), total Kjeldahl nitrogen (TKN), total Kjeldahl phosphorus (TKP), inorganic nitrogen species (NH4+, NO3-, NO2-), phosphate (PO43-) and volatile fatty acids (VFAs). SCOD, inorganic nitrogen species, PO43-, and VFAs were measured after filtering the mixed liquor samples through Millipore filter units (0.45 µm pore size) and based on the previously reported method (Ge et al., 2013). TSS and VSS were analysed based on Standard Methods (APHA et al., 2005). TKN and TKP were measured using a Lachat Quik-Chem 8000 Flow Injection Analyser (Lachat Instrument, Milwaukee). The thickened sludge samples were periodically collected from the bottom of the clarifier for dewaterability analysis according to (Higgins et al., 2014).
2.4.2. Anaerobic sludge digestion batch tests In the BMP tests, the biogas volume was recorded at regular intervals by measuring the biogas pressure in the bottle. The pressure was measured by a manometer filled with dilute acidified water at the start of each sampling event. Accumulated volumetric biogas production was calculated from the pressure increase in the headspace and expressed under standard conditions (25°C, 1 atm). The biogas composition was determined by a PerkinElmer gas chromatograph (GC) equipped with a thermal conductivity detector (GC-TCD). The accumulative methane production was calculated by multiplying the biogas volume and methane concentration with the subtraction of methane production of the blanks (Jensen et
al., 2011). At the start and end of each test, the substrate, inoculum and combined slurry samples were analysed for TCOD, SCOD, total solids (TS), VS, NH4+, PO43- and VFAs.
The methane production curve for each batch test was fitted to a first order kinetic model (Jensen et al., 2011), which was implemented in Aquasim 2.1d (Reichert, 1994). Estimation of key sludge degradability properties, degradability extent (extent of degradation, fd) and apparent hydrolysis coefficient (rate of degradation, khyd), was based on cumulative methane production. Alternatively, the same model was used with Matlab 2015b with the optimisation routine lsqcurvfit(). The same method was used for other non-linear regression procedures (including the Gosset and Belser model fitting) with confidence intervals generated using lsqcurvefit(). A 95% confidence interval (5% significance threshold) was applied for all parameter and model error bars.
3. RESULTS AND DISCUSSION 3.1. Performance of the high-rate aerobic process Figure 1 (top) shows TCOD and SCOD measured in the influent and effluent of the high-rate process during all operational periods, and the COD removal performance is summarised in Table 2. The TCOD removal efficiency was approximately 62% in the high-rate bioreactor with 1d SRT (Period 1) and decreased to 54% when reducing the SRT to 0.75d and 0.5d (Periods 2-3). The SCOD removal efficiency was maintained at approximately 48% at these three SRTs, which was confirmed by repeating the reactor operating conditions at 0.5d SRT (Period 7) and 1d SRT (Period 4). This indicates that SRT changes affect the removal efficiency of different COD fractions (Jimenez et al., 2005), e.g. decreasing the removal efficiencies of particular and/or colloidal COD fractions at SRTs of <1d, but with limited impact on the soluble fraction (i.e. SCOD). Exocellular polymeric substance (EPS)
production levels have been previously shown to be lower at lower SRTs, which negatively affects bioflocculation thought to be responsible for removing particulate and colloidal COD from wastewater (Faust et al., 2014; Jimenez et al., 2007). When the operating SRT was above 1d, there was a progressive improvement in the efficiency of TCOD removal with increasing SRT, rising from 62% at 1d SRT to 78% at 1.5d SRT (Period 5), and further to 85% at 2d SRT (Period 6). However, there was no further improvement at 2.5d and 3d SRTs (Periods 10-11). This trend was also evident in the increasing SCOD removal between 1.5 to 3d SRTs. In addition, the DO level in the high-rate bioreactor was temporarily lowered from 3-3.5 to 1-1.5 mg O2 L-1 in Periods 8-9 (0.5d and 2d SRTs), in which periods, COD removal efficiencies dropped compared to the performance achieved at same SRTs in Periods 6-7. Again, this could be attributed to lower biomass yield, confirmed by VSS measurements (data not shown) and based on others results, decreased bioflocculation due to a decrease in EPS formation (Jimenez et al., 2007).
The COD removal in the high-rate bioreactor was achieved via two processes; (a) biomass assimilation/accumulation and (b) oxidation. The contribution of each process to the total COD removal was influenced by the SRT, as shown in Figure 2. Generally, biomass assimilation/accumulation was the main contributor to COD removal (>70%), with a small fraction of COD being oxidised, particularly at <1d SRT. This low COD oxidation extent suggests that the required aeration demand can be substantially reduced in practise, which will significantly reduce energy inputs.
Concentrations of total nitrogen (N) and total phosphorus (P) in the influent and effluent of the high-rate bioreactor during all operational periods are shown in Figure 3. The removal efficiency of the total N (mainly organic N and NH4+ in this case) achieved in the bioreactor
was substantially impacted by SRT (p=0.001), improving progressively from 22% at 0.5d SRT to 49% at 3d SRT (Table 2). The NH4+ removal efficiency followed a similar trend against the SRT varying from 8% at 0.5d to 35% at 3d SRT (Table 2) (p=0.0001). Nitrogen removal is not through nitrification, but instead by assimilation during microbial growth. At longer SRTs (2-3d) this assimilative nitrogen uptake is higher due to higher biomass yield (10-13 gVSS gCOD-1) compared to very short SRT conditions (0.5-1d) (3-6 gVSS gCOD-1), where growth is limited, and adsorption is dominant over assimilation. This was also supported by a N balance conducted in this study, which suggested that approximately 3550% of the influent N was removed via biomass assimilation/adsorption at SRTs of 1.5-3d, compared to 20-29% at SRTs of 0.5-1d. However, this partial nitrogen removal means the bioreactor effluent will likely require further N elimination to meet most of the discharge standards to sensitive environments, but would likely be adequate for (controlled) irrigation or ocean discharge.
In addition to N removal, the bioreactor consistently removed approximately 16% of the incoming total P when the SRT < 1d, as shown in Figure 3 and Table 2. A gradual increase of the SRT from 1d to 3d resulted in an improvement of the total P removal efficiency. The PO43- removal efficiency was limited to <10% at SRTs of < 2d, but improved somewhat to 15% at 2.5d SRT and 18% at 3d SRT (Table 3). Overall, both phosphate and TP removal were significantly impacted by SRT (p<0.001). The removal efficiencies of total N and total P were suppressed again during Periods 8-9, indicating low DO may have a negative impact on assimilation and adsorption of nutrients from wastewater.
The dewaterability of the waste activated sludge generated from the high-rate bioreactor was evaluated, and was comparable with the typical waste activated sludge generated from
conventional long SRTs BNR processes, ranging from 13%-14%, with no significant effect of SRT (p=0.98).
3.2. Anaerobic digestion sludge degradability vs SRT The sludge generated from the high-rate bioreactor during Periods 2-6 and 10-11 (with different SRTs) was tested for degradability speed and extent via biochemical methane potentials. Figure 4 shows the confidence regions of khyd and fd (degradability) for each sample. Generally, the regions moved slightly downwards (decreasing hydrolysis rates) and substantially to the left (decreasing degradability extent) as SRT increased. The effect of SRT on degradability was significant (p=0.0002), ranging from 83% degradable at 0.5d SRT to 65-71% at 2-3d SRT. Speed (khyd) was also significantly influenced by age (p=0.008), decreasing from 0.28 and 0.22 d-1 as age increased, but essentially comparable. In summary, sludge degraded most rapidly, and to the highest extent at 0.5d SRT. Speed outcomes have not previously been reported, but degradability outcomes are consistent with the study of (Ge et al., 2013) which reported that increasing the SRT of an aerobic activated sludge process treating abattoir wastewater from 2d to 4d resulted in a decrease of sludge degradability from 85% to 63%. There is limited other data on domestic wastewater, but (Meerburg et al., 2015) reported similar outcomes on synthetic wastewater, finding a progressive, and asymptotic decrease in methane potential at 0.4, 1.2, and 10(-12)d SRT.
The age-degradability data from the current paper covers a broader range than previously assessed, and can be compared with the classic relationship observed (for example, by (Gossett & Belser, 1982)). This comparison is shown in Figure 5, with the data from (Gossett & Belser, 1982) and model (
(1 f nd , A ) 1 f nd , AbA A
), where fnd,A is the non-degradable fraction in the
influent (0.317), bA is the active fraction decay rate, and A is the sludge age. As can be seen, the data from this study deviates substantially from, but converges towards the model of (Gossett & Belser, 1982). This is because the previous model considers only particulate and non-degradable fraction. The higher degradability achieved in this study is due to adsorption and assimilation of degradable organics (SS), as well as the lack of biomass decay to inerts, and additionally, the adsorption and incorporation of solubles into the particulate fraction. This is clearly supported not only by stochastic results (such as found here), but also the recalcitrant profiles of short vs long SRT sludges (Trzcinski et al., 2016). Existing activated sludge models (Henze et al., 2000) consider assimilation (as growth), but not adsorption, storage, and flocculation, and hence specific models such as Jimenez et al., (2015) are required to describe operation at very low sludge ages.
3.3. Analysis of the integrated high-rate aerobic process and anaerobic sludge digestion The aim of the A-stage process is to reduce aeration demands and concentrate the influent COD into the waste solids stream to achieve a low-carbon effluent while maximising the carbon redirection into the following anaerobic digestion for energy recovery. To assess the overall effect of the SRT changes on this objective, a COD balance was conducted based on the results achieved in this study to investigate the distribution of the influent COD in this integrated system (A-stage wastewater treatment combined with anaerobic digestion) and shown in Figure 6. This is comparable to the analysis of (Meerburg et al., 2015), but across a broader range of SRTs, indicating the trends are progressive. Generally, the extent of COD oxidation was relatively small at all tested SRTs (<25%), particularly when the SRT was <12d. However, the low COD removal efficiencies at 0.5-1d SRTs (50-60%) resulted in a large quantity of COD being lost in the A-stage effluent. Ultimately, less than a half of the total influent COD (<41%) was converted to methane in anaerobic digestion at these short SRTs
of 0.5-1d, although the degradabilities were very high (76-83%). As seen in Figure 6, increasing SRT further enables capture (and conversion to methane) of over 55% at an optimum SRT of 2d, and increased removal efficiencies, with decreased capture due to oxidation losses at up to 3d. None of the SRTs investigated allowed nitrogen or even carbon removal to the extent to which A-stage treatment alone could form a complete wastewater treatment solution. Therefore, downstream treatment for nitrogen removal is required, either through B-stage nitrification and denitrification (minimising energy consumption) (Wett et al., 2007), or through mainstream anammox (Laureni et al., 2016), and the latter requires a minimal organic carbon load (and hence operation at longer A-stage SRTs).
4. Conclusions Operation of high-rate activated sludge (implemented in this case as A-stage treatment) is heavily dependent on SRT, with energy recovery and conversion to methane optimised at very short SRT and performance in terms of nitrogen and COD removal optimised at longer SRTs. At short SRTs, organics are transferred to waste sludge and digested to methane, while at longer SRTs, a larger fraction of organics is oxidised in the activated sludge reactor. Standard IWA Activated Sludge Models do not predict this relationship well, particularly at very low SRTs, and modified HRAS specific models should instead be used. E-supplementary data of this work can be found in the online version of the paper. This provides methane additional operational data, methane production curves, and dewaterability testing.
Acknowledgements This study was funded by the Australian Water Recycling Centre of Excellence under the Commonwealth Water for the Future Initiative, Queensland Urban Utilities, Melbourne Water, GHD and Wide Bay Water. We thank the Analytical Service Laboratory (ASL) in the Advanced Water Management Centre, The University of Queensland for chemical analysis.
References 1. APHA, AWWA, WEF. 2005. Standard Methods for the Examination of Water and Wastewater, 21st ed. American Public Health Association, Washington, DC. 2. Batstone, D.J., Hulsen, T., Mehta, C.M., Keller, J. 2015. Platforms for energy and nutrient recovery from domestic wastewater: A review. Chemosphere, 140, 2-11. 3. Faust, L., Temmink, H., Zwijnenburg, A., Kemperman, A.J.B., Rijnaarts, H.H.M. 2014. High loaded MBRs for organic matter recovery from sewage: Effect of solids retention time on bioflocculation and on the role of extracellular polymers. Water Research, 56, 258-266. 4. Ge, H., Batstone, D.J., Keller, J. 2015. Biological phosphorus removal from abattoir wastewater at very short sludge ages mediated bynovel PAO clade Comamonadaceae. Water Research, 69, 173-182. 5. Ge, H., Batstone, D.J., Keller, J. 2013. Operating aerobic wastewater treatment at very short sludge ages enables treatment and energy recovery through anaerobic sludge digestion. Water Research, 47(17), 6546-6557. 6. Gossett, J.M., Belser, R.L. 1982. Anaerobic digestion of waste activated sludge. Journal of Environmental Engineering ASCE, 108(EE6), 1101-1120. 7. Henze, M., Gujer, W., Mino, T., van Loosdrecht, M. 2000. Activated Sludge Models ASM1, ASM2, ASM2d, and ASM3. IWA Scientific and Technical Report. 8. Higgins, M., Bott, C., Schauer, P., Beightol, S. 2014. Does Bio-P impact dewatering after anaerobic digestion? in: Water Environment Federation Residuals and Biosolids Conference. Austin, Texas. 9. Jensen, P.D., Ge, H., Batstone, D.J. 2011. Assessing the role of biochemical methane potential tests in determining anaerobic degradability rate and extent. Water Science and Technology, 64(4), 880-886. 10. Jetten, M.S.M., Horn, S.J., van Loosdrecht, M.C.M. 1997. Towards a more sustainable municipal wastewater treatment system. Water Science and Technology, 35(9), 171-180.
11. Jimenez, J., Miller, M., Bott, C., Murthy, S., De Clippeleir, H., Wett, B. 2015. High-rate activated sludge system for carbon management – Evaluation of crucial process mechanisms and design parameters. Water Research, 87, 476-482. 12. Jimenez, J.A., La Motta, E.J., Parker, D.S. 2005. Kinetics of removal of particulate chemical oxygen demand in the activated-sludge process. Water Environ Res, 77(5), 437-46. 13. Jimenez, J.A., Motta, E.J.L., Parker, D.S. 2007. Effect of Operational Parameters on the Removal of Particulate Chemical Oxygen Demand in the Activated Sludge Process. Water Environment Research, 79(9), 984-990. 14. Laureni, M., Falås, P., Robin, O., Wick, A., Weissbrodt, D.G., Nielsen, J.L., Ternes, T.A., Morgenroth, E., Joss, A. 2016. Mainstream partial nitritation and anammox: long-term process stability and effluent quality at low temperatures. Water Research, 101, 628-639. 15. Meerburg, F.A., Boon, N., Van Winckel, T., Vercamer, J.A.R., Nopens, I., Vlaeminck, S.E. 2015. Toward energy-neutral wastewater treatment: A high-rate contact stabilization process to maximally recover sewage organics. Bioresource Technology, 179, 373-381. 16. Reichert, P. 1994. Aquasim - a tool for simulation and data-analysis of aquativ systems. Water Science and Technology, 32(2), 21-30. 17. Trzcinski, A.P., Ganda, L., Yan Ni, A.S., Kunacheva, C., Qing, Z.D., Lin, L.L., Tao, G., Lee, Y., Ng, W.J. 2016. Identification of recalcitrant compounds in a pilot-scale AB system: An adsorption (A) stage followed by a biological (B) stage to treat municipal wastewater. Bioresource Technology, 206, 121-127. 18. Verstraete, W., Clauwaert, P., Vlaeminck, S.E. 2016. Used water and nutrients: Recovery perspectives in a 'panta rhei' context. Bioresource Technology, 215, 199-208. 19. Wett, B., Bucauer, K., Fimml, C. 2007. Energy self-sufficiency as a feasible concept for wastewater treatment systems. in: IWA Leading Edge Technology Conference, IWA. Singapore.
List of figure captions:
Figure 1. The COD removal performance during each period in the high-rate bioreactor. Red and blue lines represent the TCOD removal efficiency and the SCOD removal efficiency, respectively. (S represents SRT and the DO level was reduced from approximately 3-3.5 mg O2 L-1 to 1-1.5 mg O2 L-1 during Periods 8-9). Figure 2. Total COD removal and total COD oxidation impacted by the SRT in the high-rate bioreactor. Figure 3. The nitrogen (N) and phosphorus (P) removal efficiencies during each period in the high-rate bioreactor. Red and blue dashed lines represent the total N removal efficiency and the total P removal efficiency, respectively. (S represents SRT and the DO level was reduced from approximately 3-3.5 mg O2 L-1 to 1-1.5 mg O2 L-1 during Periods 8-9). Figure 4. Confidence regions of khyd and fd for mesophilic digestion treating the waste activated sludge with 0.5-3d SRT generated from the high-rate bioreactor. Figure 5. Data from this study vs the model and data of Gossett and Belser (1982). Solid line indicates best-fit model. Dashed line indicates model 95% confidence interval Figure 6. The distribution of the influent COD in the integrated high-rate system.
Start -up
1000
Period 5 Period 8 Period 1 Period 2 Period 3 Period 4 Period 6 Period 7 Period 9 Period 10 (0.7dS)
(1dS)
(0.5dS) (0.9dS)(1.5dS) (1.9dS) (0.5dS)(0.6dS) (2dS)
(2.4dS)
Period 11
100
(2.9dS)
TCOD influent
mg COD L
-1
TCOD effluent SCOD influent
600
SCOD effluent
60
400
40
200
20
COD removal efficiency (%)
80
800
0
0 0
20
40
60
80
100
120
140
160
180
200
Time (days)
Figure 1. The COD removal performance during each period in the high-rate bioreactor. Red and blue lines represent the TCOD removal efficiency and the SCOD removal efficiency, respectively. (S represents SRT and the DO level was reduced from approximately 3-3.5 mg O2 L-1 to 1-1.5 mg O2 L-1 during Periods 8-9).
100
COD removal/oxidation (%)
Total COD removal from the influent COD Total COD oxidation of the influent COD 80
60
40
20
0 0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
SRT (day)
Figure 2. Total COD removal and total COD oxidation impacted by the SRT in the high-rate bioreactor.
Period 5 Start -up
100
Period 1 Period 2 Period 3 Period 4 (1dS)
(0.7dS)
Period 8
Period 6Period 7 Period 9 Period 10
(0.5dS) (0.9dS) (1.5dS)(1.9dS) (0.5dS)(0.6dS) (2dS)
(2.4dS)
Period 11 (2.9dS)
50
Total N influent Total N effluent Total P influent
-1
mg N/P L
40
Total P effluent
60
30
40
20
20
10
N/P removal efficiency (%)
80
0
0 0
20
40
60
80
100
120
140
160
180
200
Time (days)
Figure 3. The nitrogen (N) and phosphorus (P) removal efficiencies during each period in the high-rate bioreactor. Red and blue dashed lines represent the total N removal efficiency and the total P removal efficiency, respectively. (S represents SRT and the DO level was reduced from approximately 3-3.5 mg O2 L-1 to 1-1.5 mg O2 L-1 during Periods 8-9).
Degrades faster
0.5
Mesophilic anaerobic digestion
1 day SRT 0.75 day SRT
0.4
-1
khyd (d )
2 days SRT 0.3
0.5 day SRT
0.2
1.5 days SRT 3 days SRT 2.5 days SRT 0.1 0.5
0.6
0.7
0.8
fd
0.9
1.0
Degrades more
Figure 4. Confidence regions of khyd and fd for mesophilic digestion treating the waste activated sludge with 0.5-3d SRT generated from the high-rate bioreactor.
1 0.9 This study
Anaerobic degradability (fd)
0.8 0.7
Model
0.6 0.5 0.4 Gossett and Belser data
0.3
0.2 0.1 0 0
10
20
30
Activated Sludge Age (d)
Figure 5. Data from this study vs the model and data of Gossett and Belser (1982). Solid line indicates best-fit model. Dashed line indicates model 95% confidence interval
Distribution of the influent COD (%)
100
80
60
40
20
0
0.5
0.75
1
1.5
2
2.5
3
SRT (day) Residual COD in the A-stage effluent
Residual COD in the digestion effluent
COD loss via oxidation
COD converted to methane
Figure 6. The distribution of the influent COD in the integrated high-rate system.
Table 1. Summary of the high-rate bioreactor operating conditions in this study Operating period
Target SRT (d)
Real SRT (d)
DO level (mg O2 L-1)
Start-up (22d)
1
1.1
3-3.5
Period 1 (17d)
1
1.0
3-3.5
Period 2 (19d)
0.75
0.7
3-3.5
Period 3 (13d)
0.5
0.5
3-3.5
Period 4 (11d)
1
0.9
3-3.5
Period 5 (10d)
1.5
1.5
3-3.5
Period 6 (18d)
2
1.9
3-3.5
Period 7 (8d)
0.5
0.5
3-3.5
Period 8 (8d)
0.5
0.6
1-1.5
Period 9 (15d)
2
2.0
1-1.5
Period 10 (18d)
2.5
2.4
3-3.5
Period 11 (24d)
3
2.9
3-3.5
Table 2. Summary of the high-rate bioreactor performance during each operating period. NH4+ removal (%) 13.6 ± 2.3
Total P removal (%) 15.3 ± 3.0
PO43removal (%) 7.8 ± 2.0
Period 1
1.0d
TCOD SCOD Total N removal removal removal (%) (%) (%) a 65.3 ± 3.3 49.2 ± 4.0 31.8 ± 4.3
Period 2
0.7d
54.5 ± 2.6
48.5 ± 2.7 26.2 ± 5.1
10.5 ± 1.6
17.7± 2.5
5.4 ± 2.1
Period 3b
0.5d
52.2 ± 3.1
48.2 ± 2.4 22.3 ± 4.6
8.2 ± 2.1
15.2 ± 2.1
5.1 ± 2.3
Period 4b
0.9d
60.4 ± 4.2
54.8 ± 3.1 28.6 ± 3.7
12.8 ± 1.4
14.3 ± 2.6
7.1 ± 1.9
Period 5
1.5d
77.8 ± 3.6
69.3 ± 4.2 35.2 ± 5.0
16.3 ± 2.0
20.6 ± 4.3
8.3 ± 3.0
Period 6
1.9d
84.6 ± 3.4
80.9 ± 2.4 38.6 ± 4.8
25.4 ± 3.1
24.2 ± 6.7
10.2 ± 2.1
Period 7
0.5d
50.4 ± 4.8
47.3 ± 3.8 21.5 ± 3.4
8.0 ± 1.6
14.6 ± 3.7
4.8 ± 1.8
Period 8c
0.6d
35.6 ± 5.0
30.8 ± 4.1 14.6 ± 4.7
5.2 ± 2.1
8.1 ± 3.5
3.3 ± 1.4
c
Period 9
2.0d
62.7 ± 3.1
60.7 ± 3.6 22.4 ± 3.9
15.1 ± 2.5
15.1 ± 3.2
6.3 ± 1.5
Period 10
2.4d
82.8 ± 1.9
80.6 ± 2.7 42.6 ± 4.1
29.6 ± 3.0
28.3 ± 4.7
14.7 ± 2.0
SRT
Period 11 2.9d 84.3 ± 3.4 81.3 ± 3.7 49.6 ± 4.8 36.5 ± 3.2 34.2 ± 4.6 18.2 ± 1.8 a : Error bars indicate 95% confidence intervals across different measurements over each period. b
: Data collected during Day 65-69 (Period 3) and Day 75-79 (Period 4) were not included in
the performance analysis due to large variations in the wastewater feed. c
: The DO level in the high-rate bioreactor was reduced from approximately 3-3.5 mg O2 L-1
to approximately 1-1.5 mg O2 L-1 during Periods 8-9.
Clarifier Effluent
Discharge
High-rate Bioreactor (30 min HRT, 0.5-3 days SRT) Wastewater
Sludge mixed liquor
Air
Wasting
Return activated sludge
100
1
Anaerobic degradability (fd)
COD removal/oxidation (%)
Total COD removal from the influent COD Total COD oxidation of the influent COD 80
%COD removed
60
40
%COD oxidized
20
0.9 This study
0.8 0.7
Model
0.6 0.5 0.4
Gossett and Belser data
0.3
0.2 0.1 0
0 0.0
0.5
1.0
1.5
2.0
SRT (day)
2.5
3.0
3.5
0
10
20
Activated Sludge Age (d)
30
S0960-8524(17)31429-3 http://dx.doi.org/10.1016/j.biortech.2017.08.115 BITE 18724
To appear in:
Bioresource Technology
Received Date: Revised Date: Accepted Date:
3 July 2017 17 August 2017 18 August 2017
Please cite this article as: Ge, H., Batstone, D.J., Mouiche, M., Hu, S., Keller, J., Nutrient removal and energy recovery from high-rate activated sludge processes – impact of sludge age, Bioresource Technology (2017), doi: http://dx.doi.org/10.1016/j.biortech.2017.08.115
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Nutrient removal and energy recovery from high-rate activated sludge processes – impact of sludge age
Huoqing Ge,† Damien J. Batstone,†,§,*, Morgan Mouiche,‡ Shihu Hu,† Jurg Keller†,§
†
AWMC, Advanced Water Management Centre, The University of Queensland, St Lucia,
Queensland, 4072, Australia ‡
§
Engineering School EI. CESI, France CRC for Water Sensitive Cities, PO Box 8000, Clayton, Victoria 3800, Australia
*Corresponding author: Damien Batstone
Advanced Water Management Centre (AWMC), The University of Queensland, St Lucia, Queensland, 4072, Australia Phone: +61 7 3365 4727 Fax: +61 7 3365 4726 Email: [email protected]
Abstract: This study evaluated high-rate activated sludge treatment across a broad range of short solids retention times (SRT)s (0.5-3d) and found a strong SRT-outcome dependence for performance and subsequent anaerobic degradability of the sludge. Up to 50% total nitrogen, and 35% ammonia removal was also achieved at the longer SRTs, via partitioning rather than reaction. The aerobic SRT significantly affected the anaerobic degradability of the sludge produced (p<0.001), with degradability increasing from 66% to over 80% while reducing the SRT from 3d to 0.5d. This is higher than predicted by conventional models, likely due to additional mechanisms such as adsorption and storage, not included in these.
Keywords: A-stage, anaerobic, high-rate, energy recovery, SRT, model Highlights:
High-rate activated sludge enhances energy and nutrient recovery through short SRT Tested range of SRTs from 0.5d-3d for nutrient removal and energy recovery Energy recovery maximised at low SRT while nutrient removal maximised at long SRT Sludge degradability significantly higher than predicted through standard model
1. INTRODUCTION The goals of wastewater treatment are currently being expanded from the traditional removal of organic matter and nutrients (i.e. nitrogen and phosphorus) to include also energy (carbon) recovery and nutrient recovery (Batstone et al., 2015; Verstraete et al., 2016). The typical method of energy recovery in wastewater treatment plants (WWTPs) is to employ anaerobic digestion to treat waste sludge and produce biogas (methane) for onsite heat and energy generation, thereby compensating energy demands from plants. However, conventional biological nutrient removal (BNR) processes, treating domestic sewage at normal strength results in oxidation of a large fraction of organic carbon contained in wastewater due to the
need to remove nitrogen biologically, and due to long solids retention times (SRTs) (e.g. 1020d) which reduces biogas potential and sludge degradability (Gossett & Belser, 1982). A key approach to enhance energy efficiency wastewater treatment processes, is to redirect carbon and nutrients, particularly to the waste activated sludge stream, to enable energy recovery, while maintaining current treatment quality.
The most popular option to achieve this is high-rate activated sludge (HRAS), which maximises native solids capture and minimises biological oxidation. A-stage biological treatment is a popular option for this, which utilises a very short retention time activated sludge stage, with HRTs of 0.25-0.5 hours and SRTs of 0.5-3d days (Jimenez et al., 2015). This process requires approximately 70% less energy input compared to conventional BNR processes (e.g. with 10-15d SRT),(Ge et al., 2013) and focuses on the capture of carbon in a solids stream through a combination of adsorption, particulate emeshment, bioflocculation, accumulation (storage), and assimilation (growth), rather than oxidation (Jimenez et al., 2005). Energy-rich short-SRT sludge with inherently high degradability is then wasted from the A-stage process and digested anaerobically as a concentrate to produce methane. In this way, most of organic carbon in wastewater is made available for energy recovery. Recently, biological phosphorus (Bio-P) removal has been demonstrated to be feasible also at such short SRTs (i.e. 2d) (Ge et al., 2015), indicating that phosphorus in wastewater can be effectively captured and biologically concentrated in a solids (sludge) stream, concurrently with significant COD capture in HRAS. This phosphorus can be subsequently released during anaerobic sludge digestion and recovered through struvite crystallisation. All these advantages could potentially enable WWTPs to transform from major energy consumers to net energy generators as well as resource recovery/production plants.
So far, A-stage processes have been applied to full-scale WWTPs in Europe (e.g. Strass and Vienna WWTPs in Austria or Rotterdam-Dokhaven WWTP in Netherlands, etc.) and USA (e.g. Chesapeake-Elizabeth WWTP) (Jetten et al., 1997; Jimenez et al., 2015; Wett et al., 2007). Effective carbon removal is being achieved in all cases, and the removed COD is largely recovered through anaerobic sludge digestion in the form of methane that can be used for energy production. For example, in the Strass WWTP, efficient COD capture and conversion has resulted in an energy self-sufficiency of 108% (after implementation of deammonification for side-stream treatment) (Wett et al., 2007). Operation at short SRTs is well known to improve overall carbon capture and sludge degradability (Ge et al., 2013; Meerburg et al., 2015), but the extent to which capture and degradability improve across the broad range of SRTs commonly applied in high-rate activated sludge has not been assessed. In addition, the ability of the process to remove nutrients (particularly nitrogen) via assimilation (growth), has not been systematically assessed. This study addresses these limitations in detail for domestic wastewater treatment by operating a high-rate laboratory activated sludge system, and identifying how carbon and nitrogen capture, and degradability change in the 0.5-3d range, with relatively high resolution.
2. METHODS AND MATERIAL 2.1. Wastewater The feed wastewater used in this study was the wastewater effluent generated from a sewer biofilm reactor. This biofilm reactor was fed with real municipal wastewater collected on a weekly basis from a local sewage pumping station in Brisbane, Australia, and operated to mimic an anaerobic sewer pipe section for monitoring the production of methane and hydrogen sulfide. Therefore, the biofilm reactor effluent exhibited similar characteristics as the raw wastewater, but with a slightly lower COD level (approximately 10% less than the
raw wastewater). Regular analysis was performed to determine the characteristics and consistency of the feed wastewater. Averages of 20 samples taken over the 6 months of the study were TCOD – 393 mg L-1, SCOD – 290 mg L-1, pH 7-7.8, TKN - 65 mgN L-1, NH4+ 54 mgN L-1, TKP – 12 mgP L-1, and PO43- 9 mgP L-1. Standard deviations in all measures across 20 samples were on the order of 10% relative.
2.2. Reactor set-up and operation A lab-scale high-rate system used in this study consisted of an aerated bioreactor (300 mL working volume) followed by an intermediate clarifier (20cm diameter, 15cm depth) and was operated in a temperature controlled room (20-22°C) under continuous flow conditions. In this system, the sludge mixed liquor was directed from the bioreactor to the clarifier, where the mixed liquor was separated to generate an effluent stream for discharge and a thickened activated sludge stream that was returned to the bioreactor. The ratio of the return activated sludge (RAS) and the influent flow was maintained at 2:1. This high RAS ratio was required to make the small clarifier effective. The HRT in the bioreactor was maintained at 30 min, while the SRT was controlled by periodically wasting sludge from the bioreactor (three times per day), which was balanced by the solids discharged through the clarifier effluent. Air was continuously supplied to the bioreactor and the dissolved oxygen (DO, measured by an YSI DO membrane probe) level was maintained at 3-3.5 mg O2 L-1 (see details in Table 1). The pH was monitored by using a glass body pH probe (TPS, Australia), but not controlled. At start-up, the bioreactor was inoculated with sludge collected from a full-scale BNR plant treating domestic wastewater in Brisbane, Australia.
The high-rate system was operated for over 6 months. During this time, the SRT of the bioreactor was altered to create different operating periods, which are summarised in Table 1.
Real (observed) SRT was generally slightly different to targeted SRT due to variations in waste sludge concentration and losses in effluent, but values were generally close, and analysis considers real SRT. Mixed liquor suspended solids averaged 0.6 g L-1, but was heavily dependent on SRT, dropping as low as 0.2 g L-1 at 0.5d SRT (0.6g L-1 at 2d SRT). Each period was maintained for at least 7-8 SRTs to ensure characteristic operation was achieved at each operating point. Taking the solids concentration of the clarifier effluent into account, the real SRT of the bioreactor in some periods differed slightly from the target SRT. 2.3. Anaerobic sludge digestion batch tests Biochemical methane potential (BMP) tests were conducted at 37°C to assess the anaerobic degradability of the waste activated sludge produced in the high-rate bioreactor during Periods 2-6 and 10-11, corresponding to a sludge age of 0.5d, 0.75d, 1d, 1.5d, 2d, 2.5d and 3d, respectively. Methane production potential and sludge degradability (based on model based analysis of the experimental results, see below) were used as performance indicators.
BMP tests were performed in 160 mL non-stirred glass serum bottles (100 mL working volume) as previously reported (Jensen et al., 2011). Each bottle contained the pre-calculated volumes of the substrate and inoculum to maintain the substrate: inoculum ratio as approximately 0.75 (volatile solids (VS) mass basis). Inoculum used in the tests was collected from a full-scale anaerobic digester (35°C, 20d HRT) located in Brisbane, Australia. Bottles were then flushed with high purity nitrogen gas for 3 min (1 L min-1), sealed with a rubber stopper retained with an aluminium crimp-cap and stored in a temperature controlled incubator. Blanks only contained inoculum and MilliQ water to measure the background methane produced from the inoculum and this was subtracted from the test prior to parameter estimation. All tests were carried out in triplicates, and all error bars indicate 95% confidence in the average of the triplicates based on two-tailed t-tests.
2.4. Analysis 2.4.1. The high-rate wastewater treatment system To monitor the high-rate bioreactor, mixed liquor samples were collected regularly for chemical analyses, including total COD (TCOD), soluble COD (SCOD), total suspended solids (TSS), volatile suspended solids (VSS), total Kjeldahl nitrogen (TKN), total Kjeldahl phosphorus (TKP), inorganic nitrogen species (NH4+, NO3-, NO2-), phosphate (PO43-) and volatile fatty acids (VFAs). SCOD, inorganic nitrogen species, PO43-, and VFAs were measured after filtering the mixed liquor samples through Millipore filter units (0.45 µm pore size) and based on the previously reported method (Ge et al., 2013). TSS and VSS were analysed based on Standard Methods (APHA et al., 2005). TKN and TKP were measured using a Lachat Quik-Chem 8000 Flow Injection Analyser (Lachat Instrument, Milwaukee). The thickened sludge samples were periodically collected from the bottom of the clarifier for dewaterability analysis according to (Higgins et al., 2014).
2.4.2. Anaerobic sludge digestion batch tests In the BMP tests, the biogas volume was recorded at regular intervals by measuring the biogas pressure in the bottle. The pressure was measured by a manometer filled with dilute acidified water at the start of each sampling event. Accumulated volumetric biogas production was calculated from the pressure increase in the headspace and expressed under standard conditions (25°C, 1 atm). The biogas composition was determined by a PerkinElmer gas chromatograph (GC) equipped with a thermal conductivity detector (GC-TCD). The accumulative methane production was calculated by multiplying the biogas volume and methane concentration with the subtraction of methane production of the blanks (Jensen et
al., 2011). At the start and end of each test, the substrate, inoculum and combined slurry samples were analysed for TCOD, SCOD, total solids (TS), VS, NH4+, PO43- and VFAs.
The methane production curve for each batch test was fitted to a first order kinetic model (Jensen et al., 2011), which was implemented in Aquasim 2.1d (Reichert, 1994). Estimation of key sludge degradability properties, degradability extent (extent of degradation, fd) and apparent hydrolysis coefficient (rate of degradation, khyd), was based on cumulative methane production. Alternatively, the same model was used with Matlab 2015b with the optimisation routine lsqcurvfit(). The same method was used for other non-linear regression procedures (including the Gosset and Belser model fitting) with confidence intervals generated using lsqcurvefit(). A 95% confidence interval (5% significance threshold) was applied for all parameter and model error bars.
3. RESULTS AND DISCUSSION 3.1. Performance of the high-rate aerobic process Figure 1 (top) shows TCOD and SCOD measured in the influent and effluent of the high-rate process during all operational periods, and the COD removal performance is summarised in Table 2. The TCOD removal efficiency was approximately 62% in the high-rate bioreactor with 1d SRT (Period 1) and decreased to 54% when reducing the SRT to 0.75d and 0.5d (Periods 2-3). The SCOD removal efficiency was maintained at approximately 48% at these three SRTs, which was confirmed by repeating the reactor operating conditions at 0.5d SRT (Period 7) and 1d SRT (Period 4). This indicates that SRT changes affect the removal efficiency of different COD fractions (Jimenez et al., 2005), e.g. decreasing the removal efficiencies of particular and/or colloidal COD fractions at SRTs of <1d, but with limited impact on the soluble fraction (i.e. SCOD). Exocellular polymeric substance (EPS)
production levels have been previously shown to be lower at lower SRTs, which negatively affects bioflocculation thought to be responsible for removing particulate and colloidal COD from wastewater (Faust et al., 2014; Jimenez et al., 2007). When the operating SRT was above 1d, there was a progressive improvement in the efficiency of TCOD removal with increasing SRT, rising from 62% at 1d SRT to 78% at 1.5d SRT (Period 5), and further to 85% at 2d SRT (Period 6). However, there was no further improvement at 2.5d and 3d SRTs (Periods 10-11). This trend was also evident in the increasing SCOD removal between 1.5 to 3d SRTs. In addition, the DO level in the high-rate bioreactor was temporarily lowered from 3-3.5 to 1-1.5 mg O2 L-1 in Periods 8-9 (0.5d and 2d SRTs), in which periods, COD removal efficiencies dropped compared to the performance achieved at same SRTs in Periods 6-7. Again, this could be attributed to lower biomass yield, confirmed by VSS measurements (data not shown) and based on others results, decreased bioflocculation due to a decrease in EPS formation (Jimenez et al., 2007).
The COD removal in the high-rate bioreactor was achieved via two processes; (a) biomass assimilation/accumulation and (b) oxidation. The contribution of each process to the total COD removal was influenced by the SRT, as shown in Figure 2. Generally, biomass assimilation/accumulation was the main contributor to COD removal (>70%), with a small fraction of COD being oxidised, particularly at <1d SRT. This low COD oxidation extent suggests that the required aeration demand can be substantially reduced in practise, which will significantly reduce energy inputs.
Concentrations of total nitrogen (N) and total phosphorus (P) in the influent and effluent of the high-rate bioreactor during all operational periods are shown in Figure 3. The removal efficiency of the total N (mainly organic N and NH4+ in this case) achieved in the bioreactor
was substantially impacted by SRT (p=0.001), improving progressively from 22% at 0.5d SRT to 49% at 3d SRT (Table 2). The NH4+ removal efficiency followed a similar trend against the SRT varying from 8% at 0.5d to 35% at 3d SRT (Table 2) (p=0.0001). Nitrogen removal is not through nitrification, but instead by assimilation during microbial growth. At longer SRTs (2-3d) this assimilative nitrogen uptake is higher due to higher biomass yield (10-13 gVSS gCOD-1) compared to very short SRT conditions (0.5-1d) (3-6 gVSS gCOD-1), where growth is limited, and adsorption is dominant over assimilation. This was also supported by a N balance conducted in this study, which suggested that approximately 3550% of the influent N was removed via biomass assimilation/adsorption at SRTs of 1.5-3d, compared to 20-29% at SRTs of 0.5-1d. However, this partial nitrogen removal means the bioreactor effluent will likely require further N elimination to meet most of the discharge standards to sensitive environments, but would likely be adequate for (controlled) irrigation or ocean discharge.
In addition to N removal, the bioreactor consistently removed approximately 16% of the incoming total P when the SRT < 1d, as shown in Figure 3 and Table 2. A gradual increase of the SRT from 1d to 3d resulted in an improvement of the total P removal efficiency. The PO43- removal efficiency was limited to <10% at SRTs of < 2d, but improved somewhat to 15% at 2.5d SRT and 18% at 3d SRT (Table 3). Overall, both phosphate and TP removal were significantly impacted by SRT (p<0.001). The removal efficiencies of total N and total P were suppressed again during Periods 8-9, indicating low DO may have a negative impact on assimilation and adsorption of nutrients from wastewater.
The dewaterability of the waste activated sludge generated from the high-rate bioreactor was evaluated, and was comparable with the typical waste activated sludge generated from
conventional long SRTs BNR processes, ranging from 13%-14%, with no significant effect of SRT (p=0.98).
3.2. Anaerobic digestion sludge degradability vs SRT The sludge generated from the high-rate bioreactor during Periods 2-6 and 10-11 (with different SRTs) was tested for degradability speed and extent via biochemical methane potentials. Figure 4 shows the confidence regions of khyd and fd (degradability) for each sample. Generally, the regions moved slightly downwards (decreasing hydrolysis rates) and substantially to the left (decreasing degradability extent) as SRT increased. The effect of SRT on degradability was significant (p=0.0002), ranging from 83% degradable at 0.5d SRT to 65-71% at 2-3d SRT. Speed (khyd) was also significantly influenced by age (p=0.008), decreasing from 0.28 and 0.22 d-1 as age increased, but essentially comparable. In summary, sludge degraded most rapidly, and to the highest extent at 0.5d SRT. Speed outcomes have not previously been reported, but degradability outcomes are consistent with the study of (Ge et al., 2013) which reported that increasing the SRT of an aerobic activated sludge process treating abattoir wastewater from 2d to 4d resulted in a decrease of sludge degradability from 85% to 63%. There is limited other data on domestic wastewater, but (Meerburg et al., 2015) reported similar outcomes on synthetic wastewater, finding a progressive, and asymptotic decrease in methane potential at 0.4, 1.2, and 10(-12)d SRT.
The age-degradability data from the current paper covers a broader range than previously assessed, and can be compared with the classic relationship observed (for example, by (Gossett & Belser, 1982)). This comparison is shown in Figure 5, with the data from (Gossett & Belser, 1982) and model (
(1 f nd , A ) 1 f nd , AbA A
), where fnd,A is the non-degradable fraction in the
influent (0.317), bA is the active fraction decay rate, and A is the sludge age. As can be seen, the data from this study deviates substantially from, but converges towards the model of (Gossett & Belser, 1982). This is because the previous model considers only particulate and non-degradable fraction. The higher degradability achieved in this study is due to adsorption and assimilation of degradable organics (SS), as well as the lack of biomass decay to inerts, and additionally, the adsorption and incorporation of solubles into the particulate fraction. This is clearly supported not only by stochastic results (such as found here), but also the recalcitrant profiles of short vs long SRT sludges (Trzcinski et al., 2016). Existing activated sludge models (Henze et al., 2000) consider assimilation (as growth), but not adsorption, storage, and flocculation, and hence specific models such as Jimenez et al., (2015) are required to describe operation at very low sludge ages.
3.3. Analysis of the integrated high-rate aerobic process and anaerobic sludge digestion The aim of the A-stage process is to reduce aeration demands and concentrate the influent COD into the waste solids stream to achieve a low-carbon effluent while maximising the carbon redirection into the following anaerobic digestion for energy recovery. To assess the overall effect of the SRT changes on this objective, a COD balance was conducted based on the results achieved in this study to investigate the distribution of the influent COD in this integrated system (A-stage wastewater treatment combined with anaerobic digestion) and shown in Figure 6. This is comparable to the analysis of (Meerburg et al., 2015), but across a broader range of SRTs, indicating the trends are progressive. Generally, the extent of COD oxidation was relatively small at all tested SRTs (<25%), particularly when the SRT was <12d. However, the low COD removal efficiencies at 0.5-1d SRTs (50-60%) resulted in a large quantity of COD being lost in the A-stage effluent. Ultimately, less than a half of the total influent COD (<41%) was converted to methane in anaerobic digestion at these short SRTs
of 0.5-1d, although the degradabilities were very high (76-83%). As seen in Figure 6, increasing SRT further enables capture (and conversion to methane) of over 55% at an optimum SRT of 2d, and increased removal efficiencies, with decreased capture due to oxidation losses at up to 3d. None of the SRTs investigated allowed nitrogen or even carbon removal to the extent to which A-stage treatment alone could form a complete wastewater treatment solution. Therefore, downstream treatment for nitrogen removal is required, either through B-stage nitrification and denitrification (minimising energy consumption) (Wett et al., 2007), or through mainstream anammox (Laureni et al., 2016), and the latter requires a minimal organic carbon load (and hence operation at longer A-stage SRTs).
4. Conclusions Operation of high-rate activated sludge (implemented in this case as A-stage treatment) is heavily dependent on SRT, with energy recovery and conversion to methane optimised at very short SRT and performance in terms of nitrogen and COD removal optimised at longer SRTs. At short SRTs, organics are transferred to waste sludge and digested to methane, while at longer SRTs, a larger fraction of organics is oxidised in the activated sludge reactor. Standard IWA Activated Sludge Models do not predict this relationship well, particularly at very low SRTs, and modified HRAS specific models should instead be used. E-supplementary data of this work can be found in the online version of the paper. This provides methane additional operational data, methane production curves, and dewaterability testing.
Acknowledgements This study was funded by the Australian Water Recycling Centre of Excellence under the Commonwealth Water for the Future Initiative, Queensland Urban Utilities, Melbourne Water, GHD and Wide Bay Water. We thank the Analytical Service Laboratory (ASL) in the Advanced Water Management Centre, The University of Queensland for chemical analysis.
References 1. APHA, AWWA, WEF. 2005. Standard Methods for the Examination of Water and Wastewater, 21st ed. American Public Health Association, Washington, DC. 2. Batstone, D.J., Hulsen, T., Mehta, C.M., Keller, J. 2015. Platforms for energy and nutrient recovery from domestic wastewater: A review. Chemosphere, 140, 2-11. 3. Faust, L., Temmink, H., Zwijnenburg, A., Kemperman, A.J.B., Rijnaarts, H.H.M. 2014. High loaded MBRs for organic matter recovery from sewage: Effect of solids retention time on bioflocculation and on the role of extracellular polymers. Water Research, 56, 258-266. 4. Ge, H., Batstone, D.J., Keller, J. 2015. Biological phosphorus removal from abattoir wastewater at very short sludge ages mediated bynovel PAO clade Comamonadaceae. Water Research, 69, 173-182. 5. Ge, H., Batstone, D.J., Keller, J. 2013. Operating aerobic wastewater treatment at very short sludge ages enables treatment and energy recovery through anaerobic sludge digestion. Water Research, 47(17), 6546-6557. 6. Gossett, J.M., Belser, R.L. 1982. Anaerobic digestion of waste activated sludge. Journal of Environmental Engineering ASCE, 108(EE6), 1101-1120. 7. Henze, M., Gujer, W., Mino, T., van Loosdrecht, M. 2000. Activated Sludge Models ASM1, ASM2, ASM2d, and ASM3. IWA Scientific and Technical Report. 8. Higgins, M., Bott, C., Schauer, P., Beightol, S. 2014. Does Bio-P impact dewatering after anaerobic digestion? in: Water Environment Federation Residuals and Biosolids Conference. Austin, Texas. 9. Jensen, P.D., Ge, H., Batstone, D.J. 2011. Assessing the role of biochemical methane potential tests in determining anaerobic degradability rate and extent. Water Science and Technology, 64(4), 880-886. 10. Jetten, M.S.M., Horn, S.J., van Loosdrecht, M.C.M. 1997. Towards a more sustainable municipal wastewater treatment system. Water Science and Technology, 35(9), 171-180.
11. Jimenez, J., Miller, M., Bott, C., Murthy, S., De Clippeleir, H., Wett, B. 2015. High-rate activated sludge system for carbon management – Evaluation of crucial process mechanisms and design parameters. Water Research, 87, 476-482. 12. Jimenez, J.A., La Motta, E.J., Parker, D.S. 2005. Kinetics of removal of particulate chemical oxygen demand in the activated-sludge process. Water Environ Res, 77(5), 437-46. 13. Jimenez, J.A., Motta, E.J.L., Parker, D.S. 2007. Effect of Operational Parameters on the Removal of Particulate Chemical Oxygen Demand in the Activated Sludge Process. Water Environment Research, 79(9), 984-990. 14. Laureni, M., Falås, P., Robin, O., Wick, A., Weissbrodt, D.G., Nielsen, J.L., Ternes, T.A., Morgenroth, E., Joss, A. 2016. Mainstream partial nitritation and anammox: long-term process stability and effluent quality at low temperatures. Water Research, 101, 628-639. 15. Meerburg, F.A., Boon, N., Van Winckel, T., Vercamer, J.A.R., Nopens, I., Vlaeminck, S.E. 2015. Toward energy-neutral wastewater treatment: A high-rate contact stabilization process to maximally recover sewage organics. Bioresource Technology, 179, 373-381. 16. Reichert, P. 1994. Aquasim - a tool for simulation and data-analysis of aquativ systems. Water Science and Technology, 32(2), 21-30. 17. Trzcinski, A.P., Ganda, L., Yan Ni, A.S., Kunacheva, C., Qing, Z.D., Lin, L.L., Tao, G., Lee, Y., Ng, W.J. 2016. Identification of recalcitrant compounds in a pilot-scale AB system: An adsorption (A) stage followed by a biological (B) stage to treat municipal wastewater. Bioresource Technology, 206, 121-127. 18. Verstraete, W., Clauwaert, P., Vlaeminck, S.E. 2016. Used water and nutrients: Recovery perspectives in a 'panta rhei' context. Bioresource Technology, 215, 199-208. 19. Wett, B., Bucauer, K., Fimml, C. 2007. Energy self-sufficiency as a feasible concept for wastewater treatment systems. in: IWA Leading Edge Technology Conference, IWA. Singapore.
List of figure captions:
Figure 1. The COD removal performance during each period in the high-rate bioreactor. Red and blue lines represent the TCOD removal efficiency and the SCOD removal efficiency, respectively. (S represents SRT and the DO level was reduced from approximately 3-3.5 mg O2 L-1 to 1-1.5 mg O2 L-1 during Periods 8-9). Figure 2. Total COD removal and total COD oxidation impacted by the SRT in the high-rate bioreactor. Figure 3. The nitrogen (N) and phosphorus (P) removal efficiencies during each period in the high-rate bioreactor. Red and blue dashed lines represent the total N removal efficiency and the total P removal efficiency, respectively. (S represents SRT and the DO level was reduced from approximately 3-3.5 mg O2 L-1 to 1-1.5 mg O2 L-1 during Periods 8-9). Figure 4. Confidence regions of khyd and fd for mesophilic digestion treating the waste activated sludge with 0.5-3d SRT generated from the high-rate bioreactor. Figure 5. Data from this study vs the model and data of Gossett and Belser (1982). Solid line indicates best-fit model. Dashed line indicates model 95% confidence interval Figure 6. The distribution of the influent COD in the integrated high-rate system.
Start -up
1000
Period 5 Period 8 Period 1 Period 2 Period 3 Period 4 Period 6 Period 7 Period 9 Period 10 (0.7dS)
(1dS)
(0.5dS) (0.9dS)(1.5dS) (1.9dS) (0.5dS)(0.6dS) (2dS)
(2.4dS)
Period 11
100
(2.9dS)
TCOD influent
mg COD L
-1
TCOD effluent SCOD influent
600
SCOD effluent
60
400
40
200
20
COD removal efficiency (%)
80
800
0
0 0
20
40
60
80
100
120
140
160
180
200
Time (days)
Figure 1. The COD removal performance during each period in the high-rate bioreactor. Red and blue lines represent the TCOD removal efficiency and the SCOD removal efficiency, respectively. (S represents SRT and the DO level was reduced from approximately 3-3.5 mg O2 L-1 to 1-1.5 mg O2 L-1 during Periods 8-9).
100
COD removal/oxidation (%)
Total COD removal from the influent COD Total COD oxidation of the influent COD 80
60
40
20
0 0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
SRT (day)
Figure 2. Total COD removal and total COD oxidation impacted by the SRT in the high-rate bioreactor.
Period 5 Start -up
100
Period 1 Period 2 Period 3 Period 4 (1dS)
(0.7dS)
Period 8
Period 6Period 7 Period 9 Period 10
(0.5dS) (0.9dS) (1.5dS)(1.9dS) (0.5dS)(0.6dS) (2dS)
(2.4dS)
Period 11 (2.9dS)
50
Total N influent Total N effluent Total P influent
-1
mg N/P L
40
Total P effluent
60
30
40
20
20
10
N/P removal efficiency (%)
80
0
0 0
20
40
60
80
100
120
140
160
180
200
Time (days)
Figure 3. The nitrogen (N) and phosphorus (P) removal efficiencies during each period in the high-rate bioreactor. Red and blue dashed lines represent the total N removal efficiency and the total P removal efficiency, respectively. (S represents SRT and the DO level was reduced from approximately 3-3.5 mg O2 L-1 to 1-1.5 mg O2 L-1 during Periods 8-9).
Degrades faster
0.5
Mesophilic anaerobic digestion
1 day SRT 0.75 day SRT
0.4
-1
khyd (d )
2 days SRT 0.3
0.5 day SRT
0.2
1.5 days SRT 3 days SRT 2.5 days SRT 0.1 0.5
0.6
0.7
0.8
fd
0.9
1.0
Degrades more
Figure 4. Confidence regions of khyd and fd for mesophilic digestion treating the waste activated sludge with 0.5-3d SRT generated from the high-rate bioreactor.
1 0.9 This study
Anaerobic degradability (fd)
0.8 0.7
Model
0.6 0.5 0.4 Gossett and Belser data
0.3
0.2 0.1 0 0
10
20
30
Activated Sludge Age (d)
Figure 5. Data from this study vs the model and data of Gossett and Belser (1982). Solid line indicates best-fit model. Dashed line indicates model 95% confidence interval
Distribution of the influent COD (%)
100
80
60
40
20
0
0.5
0.75
1
1.5
2
2.5
3
SRT (day) Residual COD in the A-stage effluent
Residual COD in the digestion effluent
COD loss via oxidation
COD converted to methane
Figure 6. The distribution of the influent COD in the integrated high-rate system.
Table 1. Summary of the high-rate bioreactor operating conditions in this study Operating period
Target SRT (d)
Real SRT (d)
DO level (mg O2 L-1)
Start-up (22d)
1
1.1
3-3.5
Period 1 (17d)
1
1.0
3-3.5
Period 2 (19d)
0.75
0.7
3-3.5
Period 3 (13d)
0.5
0.5
3-3.5
Period 4 (11d)
1
0.9
3-3.5
Period 5 (10d)
1.5
1.5
3-3.5
Period 6 (18d)
2
1.9
3-3.5
Period 7 (8d)
0.5
0.5
3-3.5
Period 8 (8d)
0.5
0.6
1-1.5
Period 9 (15d)
2
2.0
1-1.5
Period 10 (18d)
2.5
2.4
3-3.5
Period 11 (24d)
3
2.9
3-3.5
Table 2. Summary of the high-rate bioreactor performance during each operating period. NH4+ removal (%) 13.6 ± 2.3
Total P removal (%) 15.3 ± 3.0
PO43removal (%) 7.8 ± 2.0
Period 1
1.0d
TCOD SCOD Total N removal removal removal (%) (%) (%) a 65.3 ± 3.3 49.2 ± 4.0 31.8 ± 4.3
Period 2
0.7d
54.5 ± 2.6
48.5 ± 2.7 26.2 ± 5.1
10.5 ± 1.6
17.7± 2.5
5.4 ± 2.1
Period 3b
0.5d
52.2 ± 3.1
48.2 ± 2.4 22.3 ± 4.6
8.2 ± 2.1
15.2 ± 2.1
5.1 ± 2.3
Period 4b
0.9d
60.4 ± 4.2
54.8 ± 3.1 28.6 ± 3.7
12.8 ± 1.4
14.3 ± 2.6
7.1 ± 1.9
Period 5
1.5d
77.8 ± 3.6
69.3 ± 4.2 35.2 ± 5.0
16.3 ± 2.0
20.6 ± 4.3
8.3 ± 3.0
Period 6
1.9d
84.6 ± 3.4
80.9 ± 2.4 38.6 ± 4.8
25.4 ± 3.1
24.2 ± 6.7
10.2 ± 2.1
Period 7
0.5d
50.4 ± 4.8
47.3 ± 3.8 21.5 ± 3.4
8.0 ± 1.6
14.6 ± 3.7
4.8 ± 1.8
Period 8c
0.6d
35.6 ± 5.0
30.8 ± 4.1 14.6 ± 4.7
5.2 ± 2.1
8.1 ± 3.5
3.3 ± 1.4
c
Period 9
2.0d
62.7 ± 3.1
60.7 ± 3.6 22.4 ± 3.9
15.1 ± 2.5
15.1 ± 3.2
6.3 ± 1.5
Period 10
2.4d
82.8 ± 1.9
80.6 ± 2.7 42.6 ± 4.1
29.6 ± 3.0
28.3 ± 4.7
14.7 ± 2.0
SRT
Period 11 2.9d 84.3 ± 3.4 81.3 ± 3.7 49.6 ± 4.8 36.5 ± 3.2 34.2 ± 4.6 18.2 ± 1.8 a : Error bars indicate 95% confidence intervals across different measurements over each period. b
: Data collected during Day 65-69 (Period 3) and Day 75-79 (Period 4) were not included in
the performance analysis due to large variations in the wastewater feed. c
: The DO level in the high-rate bioreactor was reduced from approximately 3-3.5 mg O2 L-1
to approximately 1-1.5 mg O2 L-1 during Periods 8-9.
Clarifier Effluent
Discharge
High-rate Bioreactor (30 min HRT, 0.5-3 days SRT) Wastewater
Sludge mixed liquor
Air
Wasting
Return activated sludge
100
1
Anaerobic degradability (fd)
COD removal/oxidation (%)
Total COD removal from the influent COD Total COD oxidation of the influent COD 80
%COD removed
60
40
%COD oxidized
20
0.9 This study
0.8 0.7
Model
0.6 0.5 0.4
Gossett and Belser data
0.3
0.2 0.1 0
0 0.0
0.5
1.0
1.5
2.0
SRT (day)
2.5
3.0
3.5
0
10
20
Activated Sludge Age (d)
30