- Email: [email protected]
Development of a new non-aeration-based sewage treatment technology: Performance evaluation of a full-scale down-flow hanging sponge reactor employing third-generation sponge carriers
Water Research 102 (2016) 138e146
Contents lists available at ScienceDirect
Water Research journal homepage: www.elsevier.com/locate/watres
Development of a new non-aeration-based sewage treatment technology: Performance evaluation of a full-scale down-flow hanging sponge reactor employing third-generation sponge carriers Tsutomu Okubo a, *, Kengo Kubota b, Takashi Yamaguchi c, Shigeki Uemura a, Hideki Harada d a
Dept. of Civil Engineering, National Institute of Technology, Kisarazu College, 2-11-1 Kiyomidaihigashi, Kisarazu, Chiba, 292-0041, Japan Dept. of Civil and Environmental Engineering, Tohoku University, 6-6-06 Aza-Aoba, Aramaki, Aoba-ku, Sendai, Miyagi, 980-8579, Japan Dept. of Civil and Environmental Engineering, Nagaoka University of Technology, 1603-1 Kamitomioka, Nagaoka, Niigata, 940-2188, Japan d New Industry Creation Hatchery Center, Tohoku University, 6-6-04 Aza-Aoba, Aramaki, Aoba-ku, Sendai, Miyagi, 980-8579, Japan b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 26 February 2016 Received in revised form 2 June 2016 Accepted 14 June 2016 Available online 16 June 2016
A practical-scale down-flow hanging sponge (DHS) reactor using third-generation (G3) sponge carriers was applied for treatment of the effluent from an up-flow anaerobic sludge blanket (UASB) reactor treating municipal sewage. The process performance of the DHS reactor filled with G3 sponge carriers (DHS-G3) was evaluated by conducting an on-site experiment in India over one year. The performance of the DHS-G3 for removal of organic matter and ammonium-nitrogen at a relatively short hydraulic retention time (HRT) of only 0.66 h satisfied the Indian effluent quality standards except for fecal coliform. The removal rate constants for total biochemical oxygen demand (BOD) and fecal coliform determined based on the water quality profiles along the DHS-G3 almost reached equilibrium approximately four months after the start of operation, i.e., 2.45 h1 for BOD and 2.30 h1 for fecal coliform, respectively. The oxygen utilization activity of retained sludge was determined to assess the distribution of heterotrophic and autotrophic bacteria along the DHS-G3. Nitrification was promoted in the lower portion of the DHS-G3 reactor in the duration with low organic load, while it decreased when the organic load was increased, probably due to proliferation of heterotrophic bacteria. © 2016 Elsevier Ltd. All rights reserved.
Keywords: Sponge carrier Municipal sewage treatment Up-flow anaerobic sludge blanket reactor Down-flow hanging sponge reactor Oxygen utilization rate
1. Introduction The activated sludge process, currently the most popular method for wastewater treatment, consumes enormous quantities of energy, 70% of which is used for aeration (Fayolle et al., 2007). In contrast, the up-flow anaerobic sludge blanket (UASB) reactor is an economical treatment process that requires no power for aeration, and is becoming a core process in developing countries in tropical and subtropical regions (Sperling et al., 2008). UASB reactors built in India are designed for a hydraulic retention time (HRT) of 8.5 h and have ponds downstream for polishing the UASB-treated water (designated in India as “Final Polishing Units”, FPUs; design
* Corresponding author. E-mail addresses: [email protected] (T. Okubo), kengo.kubota.a7@ tohoku.ac.jp (K. Kubota), [email protected] (T. Yamaguchi), uemura@ wangan.c.kisarazu.ac.jp (S. Uemura), [email protected] (H. Harada). http://dx.doi.org/10.1016/j.watres.2016.06.035 0043-1354/© 2016 Elsevier Ltd. All rights reserved.
HRT ¼ 24 h). However, water treatment by the UASB-FPU combined systems has been unable to satisfy current Indian effluent quality standards (Okubo et al., 2015). The FPU also requires very large areas of land (Sato et al., 2006). Thus, the FPU process cannot really be called an optimal post-treatment technology. Our research group has been conducting basic research since the latter 1990s on the down-flow hanging sponge (DHS) process as an efficient post-treatment technology that does not detract from the economic superiority of UASB. To date, after improvements to the shape of the sponge carrier and to the filling method, the concept has evolved from its first generation (G1) to the sixth generation (G6) (Tandukar et al., 2007; Onodera et al., 2014). Parallel to those experiments, full-scale DHS reactors have been constructed in a sewage treatment plant in India, and verification tests have been conducted under local environmental conditions (Okubo et al., 2015; Onodera et al., 2016). DHS is an aerobic biological treatment method that employs polyurethane sponges, which can be manufactured in developing
T. Okubo et al. / Water Research 102 (2016) 138e146
countries, as the carrier for retaining the bacteria. The reactor is filled with these carriers in a way that exposes them to air. The water then trickles down through the sponges, facilitating aerobic conditions. In a continuous long-term experiment of a DHS reactor using the second-generation, curtain-type sponge carrier (G2) in India, this reactor (DHS-G2) was shown to be very economical and to achieve high performance in removal of organic matter and nitrogen (Okubo et al., 2015; Onodera et al., 2016). However, DHS-G2 had some disadvantages that made it unworkable; (1) to make a sponge module, the triangle sponge bars must be tilled on both sides of the plastic sheets, (2) a hook attached to the upper part of each sponge module must be hung on a bar that is equipped to the upper part of the reactor column, (3) multiple iterations of tilling sponges and hanging the hooks are time-consuming and increase costs, and (4) because the bars and hooks are always exposed to UASB-treated sewage, corrosion is inevitable, leading to failure of the sponge modules. To overcome these drawbacks of the G2 carrier, a thirdgeneration (G3) carrier was developed, which is cylindrical in shape and is composed of polyurethane sponge supported by a polyethylene plastic net to prevent compaction of the sponge. The DHS-G3 reactor only requires filling the reactor column with the G3 carriers (Tawfik et al., 2008). Thus, in this study, a full-scale DHS-G3 reactor was installed at an Indian sewage treatment plant site and continuously monitored. The characteristics of the DHS-G3 reactor at the start of operation were identified, and the reduction rate constants of biochemical oxygen demand (BOD) and fecal coliform (FC) were calculated using the profile along the main axis of the reactor with respect to time elapsed. The distribution of heterotrophic and autotrophic bacteria along the DHS-G3 was clarified by determining the oxygen utilization activity of retained sludge on different substrates. The reactor was also operated with recirculation during an experimental period at increased load, and its capabilities under these conditions were compared and evaluated. 2. Material and methods 2.1. Full-scale DHS-G3 reactor Prior to this experiment, a long-term continuous experiment was carried out using DHS-G2 constructed at a sewage treatment plant in Karnal, India (Okubo et al., 2015; Onodera et al., 2016). The reactor was then replaced with G3 carriers, and operation was restarted. Fig. 1 shows an overall view of the full-scale reactor and a
139
view of the reactor interior after it had been filled with the carriers. The reactor portion of the DHS reactor consisted of a concrete cylinder (5.5 m in diameter and 5.31 m in height) with an empty capacity of 126 m3. A clarifier was installed under the reactor. The reactor interior was vertically divided into four layers by reinforcement frames and nets. The four layers were equally filled with the G3 sponge carriers. In a previous study, we incorporated 33 and 38 ventilation windows measuring 0.2 m 0.3 m into the middle and lower sections of the concrete cylinder, respectively. In this study, we increased the number of ventilation holes to 104, which were incorporated between the sponge layers to provide more ventilation to the reactor interior. The DHS influent (i.e., UASB effluent) was pumped to the top of the reactor, and the water was uniformly sprayed onto the sponge carriers at the top of the reactor from a moving sprinkler, which rotated by the water head differential. The rotational speed of the sprinkler head was set to 8.5 rpm by adjustment of a valve in the sprayer nozzle. After sprinkling, the UASB effluent flowed downward through the sponge carriers under the force of gravity, passed into the clarifier, and flowed out as DHS effluent after it passed through the basin. Fig. 2a is a diagram of the structure of the G3 carrier showing its dimensions. The G3 sponge carrier is cylindrical, 32 mm in diameter and 32 mm long, and is composed of polyurethane sponge supported by a polyethylene plastic net to prevent compaction of the sponge (specific surface area: 1.87 cm2 cm3). Fig. 2b shows a close-up photograph of the polyurethane sponge. The long and short axes of 100 of the sponge pores were measured, and the mean value was 0.46 (standard deviation: ±0.10) mm. A total of 1,076,000 of the G3 sponges were loaded into the reactor to equally fill the four layers (269,000 layer1). The height of each layer was approximately 600 mm. Gaps measuring 450e700 mm were left between each layer to allow ingress of oxygen into the down-flowing water. The total effective sponge volume was 27.7 m3 (91.1 m2 m3 reactor working volume). Thus, relative to the empty volume (126 m3) of the DHS reactor, the sponge filling accounted for 22.3% (the sponge filling ratio relative to the effective reactor volume in each layer was 48.6%). Table 1 presents the operating conditions for the DHS reactor. Three conditions were employed to observe the effect of recirculation and increased load on treatment performance. Since the UASB effluent is mixed with a part of the treated water at a reservoir just before the sprinkler due to recirculation, the influent of the DHS is expected to contain much more dissolved oxygen (DO). Therefore, it was predicted that this would have a beneficial effect
Fig. 1. Photographs of the entire reactor (a), the distribution device (b), the reactor filled with sponge carriers (c), and a schematic diagram of the DHS reactor (d).
140
T. Okubo et al. / Water Research 102 (2016) 138e146
Fig. 2. Structure and dimensions of DHS-G3 (a) and close-up photo of polyurethane sponge (b).
on nitrification and FC removal in the reactor. Five hundred cubic meters per day of UASB effluent was treated from day 1e305 of operation without recirculation (Phase 1, HRT based on sponge volume: 1.33 h). Next, a volume of 500 m3 d1 was treated from day 306e325 of operation with effluent recirculation at a recirculation ratio of 1.0 (Phase 2, HRT based on sponge volume: 1.33 h). A volume of 1000 m3 d1 was then treated from day 326e365 of operation without recirculation (Phase 3, HRT based on sponge volume: 0.66 h). The mean amount of wastewater inflow into the Karnal sewage treatment plant during the test period was 25,450 m3 d1; the calculated HRT values from this flow are 13.5 h for the UASB reactor and 38.5 h for the FPU.
glucose, filtered sewage, filtered UASB effluent, ammoniumnitrogen (NHþ 4 -N) as ammonium chloride, and nitrite-nitrogen (NO 2 -N) as sodium nitrite. The filtered sewage and UASB effluent were prepared using a 0.45-mm membrane filter. A BOD bottle with sludge and without any substrate was also prepared as a control. The initial substrate concentrations of the test were adjusted to 80 mg chemical oxygen demand (COD) L1 for glucose, filtered sewage and filtered UASB effluent, and to 25 mg N L1 for NHþ 4 -N and NO 2 -N. The OUR measurements were conducted from a DO concentration near saturation down to approximately 0.5 mg L1 (or down to 3 mg L1 in the control bottle). These measurements took approximately 30 min (~90 min for the control). The DO decreased almost linearly throughout the measurement. Sodium bicarbonate (420 mg L1) was added at the start of the experiment to prevent the pH from being lowered due to nitrification in the experiments with NHþ 4 -N and NO2 -N. The pH in all experiments was in the range 7.01e7.19 at the beginning and end of the measurements. To suppress any nitrification, a nitrification inhibitor (allylthiourea, ATU) was added to give a concentration of 2.0 mg L1 for the test slurries in all of the systems other than NHþ 4 -N and NO2 N.
2.2. Oxygen utilization rate test
2.3. Analytical methods
An oxygen utilization rate (OUR) test was performed using DHSretained sludge in Phase 1 (sludge retrieved on days 252e255) and in Phase 3 (sludge retrieved on days 356e359). The sludge for use in the test was obtained by random selection of a total of 1604 sponge carriers from all four layers in the DHS reactor. Because the amount of biomass retained in the sponge carriers was largest in the top layer and decreased toward the lower layers of the reactor, we collected 10e20 sponge carriers from the top layer and 100e500 sponge carriers from the lower layers. The number of removed carriers corresponded to less than 0.15% of all the carriers in the DHS reactor. The sludge was squeezed from the sponges, washed in 10 mM phosphate buffered saline (PBS buffer), concentrated by centrifugation (6000 rpm, 10 min), and finally the test slurry was adjusted to a concentration of 4e6 g of volatile suspended solids (g VSS) L1. The test temperature was set to 20 ± 1 C. The substrates were placed in a 100-mL BOD bottle and agitated with a stirrer. A DO sensor (GU-Z, Iijima Electronics Corp.) was then immediately inserted and measurements were initiated. Table 2 shows the conditions for each experiment. The substrates were
The samples provided for analysis during continuous operation were sewage after passage through a bar screen (15 mm) and two mesh screens (25 mm and 10 mm) and a grid chamber, UASB effluent and DHS effluent. In addition, once a month, samples were retrieved via the ventilation holes in the DHS reactor, and a profile measurement was conducted along the main axis of the DHS reactor. Glass fiber filter paper (pore diameter: 0.45 mm, ADVANTEC-GB140) was used to obtain the soluble samples. The analysis items were BOD, CODCr, suspended solids (SS), VSS, total nitrogen (TN), NHþ 4 -N, NO3 -N, and FC. BOD was measured by adding ATU, based on the Standard Methods for the Examination of Water and Wastewater (APHA, 2005). CODCr, TN, NHþ 4 -N, and NO3 N were determined using the potassium dichromate method, the nitrogen tube test, the salicylic acid process, and the cadmium reduction method, respectively. All those tests were conducted using a DR/890 water quality analyzer (HACH). FC concentrations were determined by the most probable number (MPN) method using Difco-A1 medium (Becton, Dickinson and Company) according to APHA (2005).
Table 1 Operating conditions. Period (days)
Phase
Flow (m3 d1)
Recirculation ratio (e)
HRT (h)
0e305 306e325 326e365
1 2 3
500 500 1000
0 1 0
1.33 1.33 0.66
HRT: hydraulic retention time.
Table 2 Experimental set-up for oxygen utilization rate test.
OUR with organic substrates OUR by nitrification Endogenous OUR
Organic substrates
ATU
Sludge
PBS buffer (10 mM)
80 mg COD L1 of glucose, filtered sewage, or filtered UASB eff. 25 mg N L1 of NH4Cl or NaNO2 and 420 mg L1 of NaHCO3 e
Added e Added
4e6 g VSS L1 4e6 g VSS L1 4e6 g VSSvL1
e Added Added
ATU: allylthiourea, PBS: phosphate buffered saline.
T. Okubo et al. / Water Research 102 (2016) 138e146
3. Results and discussion 3.1. Organic matter removal characteristics Fig. 3a shows the time course of total BOD concentrations of sewage, UASB effluent and DHS-G3 effluent, during the whole experimental period. On the other hand, Fig. 3b shows the variation in BOD removal rate over time during this experiment. The average total BOD concentration over the experimental period was 130 (±38) mg L1 for sewage and 62 (±20) mg L1 for the UASB effluent. Within the first 24 h, the DHS-G3 effluent was 32 mg L1. The quality of the effluent soon remained stable, i.e., the average total BOD concentrations during Phase 1, Phase 2, and Phase 3 were 10 (±6) mg L1, 4 (±3) mg L1, and 14 (±9) mg L1, respectively. Thus, the final effluent satisfied the Indian effluent quality standards (total BOD standard: 30 mg L1) under all operating conditions. To evaluate the treatment performance with elapsed time, the total BOD profiles along the main axis of the DHS-G3 were taken 24 h, and 66, 126 and 250 days after the start of operation (Fig. 4a). The BOD concentrations of the influent to DHS-G3 (i.e., UASB effluent) vary with fluctuations because they depend on the composition of the incoming sewage and the treatment performance of the UASB reactor. Therefore, we considered that we could fit the logarithm of the BOD reduction rate to the first-order reaction equation for the region where it could be approximated as linear (Fig. 4b). The slopes of the regression lines in the figure are the reduction rate constants for total BOD (N0: BOD concentration at time 0, Nt: BOD concentration at time t). These results indicate that after 24 h of operation, most of the BOD removal was occurring in the first sponge layer but almost none in the second or lower layers. It was considered that the solid BOD components contained in the UASB effluent were trapped in
141
the sponge carriers in the first layer, and that no further biological treatment proceeded in the following layers because there was no active biomass. It was subsequently confirmed that the BOD was removed in the second and third layers as the effluent moved down through the reactor with the passage of time during operation. The overall performance of DHS-G3 was comparable to that of DHS-G2 and superior to that of TF treating UASB effluent (Okubo et al., 2015). The reduction rate constants at 24 h, and 66, 126 and 250 days from the start of operation were 1.33 h1 (regression line from DHS influent to first sponge layer), 2.04 h1 (regression line from DHS influent to second sponge layer), 2.45 h1 (regression line from DHS influent to third sponge layer), and 2.52 h1 (regression line from DHS influent to third sponge layer). The difference between the reduction rate constant after 24 h and the other constants was conjectured to be caused by biodegradation of organic matter due to sludge buildup. No great difference in BOD concentration between the third and fourth layers was found at day 126 and day 250, indicating that nearly all the BOD components had broken down by the time the sewage reached the third sponge layer. The role of the fourth sponge layer in this reactor was more for nitrification than for removing organic components, as discussed below. Since the reduction rate constants were nearly identical on day 126 and day 250, we deduce that the reduction rate constant was approximately steady from approximately four months after the start of operation. 3.2. Nitrification and nitrogen removal characteristics The changes of NHþ 4 -N in sewage, UASB effluent and DHS effluent with elapsed time are shown in Fig. 5. The mean NHþ 4 -N concentrations during the experimental period were 23
Fig. 3. Temporal changes in total biochemical oxygen demand (BOD) concentration (a) and BOD removal rate (b) during whole experimental period (n ¼ 127).
142
T. Okubo et al. / Water Research 102 (2016) 138e146
Fig. 4. Total biochemical oxygen demand (BOD) profiles in main axis direction of down-flow hanging sponge (DHS) (a) and reduction rate constants of total BOD by measurement day (b). N0: BOD concentration at time 0, Nt: BOD concentration at time t.
þ Fig. 5. Temporal changes in NHþ 4 -N concentration (a) and NH4 -N removal rate (b) during whole experimental period (n ¼ 54).
(±6) mg L1 in sewage and 26 (±6) mg L1 in UASB effluent. One of the reasons for the increased level in the UASB effluent was probably generation of NHþ 4 -N derived from organic nitrogen compounds. The levels in the DHS effluent were 12 (±6) mg L1 in Phase 1, 8 (±5) mg L1 in Phase 2, and 19 (±8) mg L1 in Phase 3; the best quality of treated water was obtained in Phase 2, when recirculation was implemented. The effluent quality standard for NHþ 4 -N level is 50 mg L1 in India, but it is important to remove NHþ 4 -N in order to mitigate its effect on the water environment in the final destination of the effluent. Just after starting Phase 1, nitrification was not observed, which was likely due to insufficient nitrifying bacteria; however, NHþ 4 -N
removal reached 40% by day 20 of operation and ranged from 40% to 80% thereafter. The level of NO 3 -N in the DHS effluent was 7 (±4) mg L1 (data not shown). Theoretically, since approximately the same amount of NHþ 4 -N is converted to NO3 -N under aerobic conditions, this lower NO3 -N concentration in the DHS effluent suggested the occurrence of denitrification inside the sponge carriers where anoxic conditions might be maintained (Uemura et al., 2010; Ikeda et al., 2013). Other than denitrification, assimilation of NHþ 4 -N to the bacteria was also considered to contribute to nitrogen removal, resulting in a 32% reduction in TN. Generally, for the nitrification process using a trickling filter, it is recommended to operate at volumetric organic loads of 0.2 kg
T. Okubo et al. / Water Research 102 (2016) 138e146
BOD m3 carrier d1 when the carriers are crushed rock (flow speed: 0.4e0.8 m h1), and 0.2e0.4 kg BOD m3 carrier d1 when the carriers are plastic (flow speed: 0.6e1.5 m h1) (Henze et al., 2000). However, the volumetric organic load in this DHS-G3 during Phase 1 was 1.1 kg BOD m3 sponge d1 and the flow speed was 1.7 m h1. Thus, the nitrification proceeded in this DHS-G3 reactor under more stringent operating conditions than those recommended for the trickling filter process. It was considered that almost all organics were removed in the upper portion of the DHSG3 reactor irrespective of its high volumetric organic load, and then an environment allowing the growth of nitrifying bacteria is established in the lower portion of the DHS reactor, which had a reduced organic load (Tandukar et al., 2005). Actually, in the DHS reactor examined here, the organic volume load in the first sponge layer was 4.28 kg BOD m3 sponge d1, but this was reduced to 0.35 kg BOD m3 sponge d1 in the fourth layer. The nitrification reactions in the third and fourth layers associated with the decreased load were observed in the profiles on day 250 (Phase 1), as shown in Fig. 6a, and were verified to progress in the downward direction of the reactor as the organic volume load decreased. The long SRT of 90e120 days for DHS-retained sludge (Tandukar et al., 2007) creates conditions favorable for proliferation of nitrifying bacteria and their retention in the carriers. Moreover, we assume that installing the sponges in four layers facilitated aeration of the wastewater, and that the high DO promoted the proliferation of nitrifying bacteria, thereby enhancing nitrification. In Phase 2, DHS-G3 was operated with recirculation at a recirculation ratio of 1.0. Therefore, the UASB effluent was mixed with the DHS effluent containing high DO just before entering the DHSG3. Thus, it is considered that the high DO-containing wastewater flowed in from the upper portion of the reactor, which meant that conditions favored the proliferation of nitrifying bacteria, contributing to the high NHþ 4 -N oxidation (70%). It was reported that when the soluble BOD concentration in treated water fell below 5 mg L1, the rate of nitrification in the trickling filter nitrification process €s, 1982). Since the concentration of was maximized (Harremoe organic matter in DHS effluent was also low in Phase 2 (soluble BOD was 3 (±2) mg L1), if removal of NHþ 4 -N is a priority, then recirculation is an effective approach, even when using a DHS reactor. Thus, the high NHþ 4 -N oxidation was shown by recirculation. On the other hand, although high NHþ 4 -N oxidation was shown by recirculation in Phase 2, TN removal worsened in Phase 2 compared to Phase 1, i.e., 22 (±9) mg N L1 and 38 (±5) mg L1 in Phases 1 and 2, respectively. With DHS carriers, the nitrificationdenitrification reactions proceed due to the unique properties of the sponge carriers, i.e., one carrier may provide an aerobic environment near the carrier surface (the nitrification reaction portion), and an anaerobic environment deeper inside the carrier (the
143
denitrification reaction portion) (Araki et al., 1999; Uemura et al., 2010). For this reason, it is possible that during Phase 2, high DOcontaining wastewater penetrates the interior of the carriers, suppressing the denitrifying bacteria and reducing the TN removal. Suppression of denitrification by effluent recirculation was also reported in a trickling filter process (Parker and Richards, 1986) and in a biological nitrogen removal process (Ekman et al., 2006). However, in the treatment of industrial wastewater containing a large amount of organics and NHþ 4 -N by a DHS reactor, effluent recirculation was advantageous for enhancing denitrification in the range of the recirculation ratio up to 2.0 (Ikeda et al., 2013). This was due to the fact that DO in the recirculated water was consumed very quickly to degrade highly concentrated organics in the upper part of the reactor. Fig. 6b shows the profiles along the main axis of the DHS reactor on day 346 (Phase 3). The increase in organic load increased the concentration of organics in the DHS effluent. The NHþ 4 -N oxidation decreased to approximately 30%, and the NO 3 -N concentration decreased to values below those seen during Phases 1 and 2. These findings suggest that the increased organic load favored heterotrophic bacteria over nitrifying bacteria in the lower portion of the reactor. However, the average concentration of DO in the DHS effluent was 5.6 (±0.8) mg L1, demonstrating that DHS-G3 was capable of taking up sufficient oxygen even at an HRT of 0.66 h. 3.3. Oxygen utilization rate (OUR) of DHS-retained sludge Measurements of the OUR in the DHS-retained sludge provide the oxygen consumption activity of microorganisms for each substrate. Fig. 7 shows the results from the OUR tests for each substrate, performed during Phase 1 and Phase 3. The OUR during Phase 1 was 1.65e2.49 mg O2 g1 VSS h1 in the glucose substrate. No marked differences were observed along the vertical axis of the DHS-G3 reactor. In Phase 3, however, values increased along the DHS-G3 reactor (varying from 1.03 to 6.53 mg O2 g1 VSS h1 in the glucose substrate). The OUR of the glucose substrate in Phases 1 and 3 was about one-tenth that of the activated sludge (Huang et al., 1985); about the same level as sludge samples from a membrane filtration reactor (Witzig et al., 2002) and from a DHS-G4 reactor operated in Japan (Tandukar et al., 2006). The OUR of the sewagesoluble fraction and the UASB effluent-soluble fraction did not differ markedly along the DHS-G3 reactor in Phase 1, but in Phase 3, just as seen in the OUR of the glucose, the OUR showed higher values in the sludge taken from increasing depths in the DHS-G3 reactor (sewage-soluble fraction substrate: 6.87e27.38 mg O2 g1 VSS h1, UASB effluent-soluble fraction substrate: 7.87e27.98 mg O2 g1 VSS h1). The probable reason for the low OUR in the upper portion of the reactor was that, even though the VSS concentration
Fig. 6. Profiles of NHþ 4 -N and dissolved oxygen (DO) along the main axis of the down-flow hanging sponge (DHS) on day 250 (Phase 1) (a) and on day 346 (Phase 3) (b).
144
T. Okubo et al. / Water Research 102 (2016) 138e146
Fig. 7. Measured OUR values in Phase 1 (sludge from days 252e255) and in Phase 3 (sludge from days 356e359) per substrate.
was high in the retained sludge in the upper portion, the percentage of microorganisms was relatively low due to the biologically inert particulate organics derived from the UASB effluent, resulting in the decreased oxygen consumption per unit of VSS. The OUR of the retained sludge on NHþ 4 -N and NO2 -N substrates in Phase 1 tended to increase from the higher to lower part of the DHS reactor, showing maximum values in the third sponge layer (16.25 mg O2 g1 VSS h1 on the NHþ 4 -N substrate and 6.98 mg O2 g1 VSS h1 on the NO 2 -N substrate). Notably, maxima also occurred in the third sponge layer in Phase 3 (12.18 mg O2 g1 1 VSS h1 on the NHþ VSS h1 on the 4 -N substrate and 7.42 mg O2 g þ NO2 -N substrate), but the OUR on the NH4 -N substrate decreased by 25% in Phase 3 compared to that in Phase 1. According to the results of the profile experiment, the nitrification was promoted in the lower portion of the DHS-G3 reactor in the duration with low organic load (Phase 1), and that the NHþ 4 -N oxidation decreased when the organic load was increased (Phase 3). Thus, those results were also clearly supported by the OUR tests. The oxygen consumption rate associated with NHþ 4 -N oxidation calculated from the reactor profiles in each sponge layer was 1%e 60% lower than the value obtained by the OUR test. A possible reason for this was that, since the OUR test was performed with agitation by a stirrer, the contact efficiency between the sludge and the substrate was much better in the OUR test than in the reactor. The OUR fed with no substrate (only sludge) showed no variation along the DHS-G3 reactor in Phase 1, while in Phase 3, however, the OUR increased in the downward direction in the reactor. This is primarily because of the high OUR due to heterotrophic bacteria in the lower portion of the DHS reactor, which proliferated with increased organic load.
3.4. Fecal coliform removal characteristics Fig. 8 shows the first-order reaction constants for FC removal only in Phase 1. The constants were predicted by regression of the values of the retention time that was calculated from the distance along the DHS-G3 reactor and FC counts at each sponge layer (N0: FC counts at time 0, Nt: FC counts at time t). There was no clear difference between the first-order reaction constants at 24 h after the start of operation and at 66 days, implying that there was no improvement in the ability to remove FC. However, the constants gradually increased from day 87 and reached greater than 2.0 h1 on day 126, indicating that high and stable FC removal was
Fig. 8. Reduction rate constants for fecal coliform by measurement day. N0: FC counts at time 0, Nt: FC counts at time t.
established in the DHS-G3 reactor after four months of operation. These results clearly indicate that FC removal is also associated with biomass build-up, as described later. It has been reported that FC removal in the DHS reactor is achieved by inactivation after adhesion onto the retained sludge, or predation by protozoa (Tawfik et al., 2008). Therefore, development of stable microfauna might also be a key factor for FC removal (Kubota et al., 2014). Fig. 9 shows the variation of the retained sludge concentration in the DHS-G3 reactor after the start of operation. Two or three carriers were randomly collected from each sponge layer and squeezed repeatedly in distilled water to obtain samples of the DHS-retained sludge. Because the sludge had not built up yet just after the start of operation, the FC removal at the early stage of operation might be established by a physical process, such as adhesion or filtration. The concentration of retained sludge on day 66 was approximately 3.0 g VSS L1 sponge, but no increase in the first-order reaction constants for FC removal was observed, suggesting that the sludge concentration had not reached the level needed for FC removal. The retained sludge concentration was 5.0 g VSS L1 sponge on day 87 when FC removal began to increase. At that time, the amount of sludge per volume was approximately the same as with the activated sludge process. This concentration
T. Okubo et al. / Water Research 102 (2016) 138e146
145
experiment (Onodera et al., 2016). However, in Phase 3, possibly because the contact time between the wastewater and sludge was insufficient due to the reduced HRT, the log reduction was lower (0.9 log). The mean log decrease during the latter part of Phase 1 (day 150 and thereafter) was 2.0 log, comparable to the range 1.4e3.5 log (mean, 2.1 log) determined for the activated sludge process in India’s capital city, New Delhi (Kazmi et al., 2007). However, the DHS-G3 effluent unsatisfied the effluent quality standard of FC for irrigation use in India (104 MPN 100 mL1). Therefore, additional treatment will be necessary to meet the standard, i.e., the water must be disinfected with chlorine. On the other hand, Uemura et al. (2010) reported that the oxygen uptake and inactivation of FC in a DHS-G2 reactor were improved by reducing the size of the carriers. More work is necessary to develop a method to improve the efficiency of FC removal in a DHS reactor. 3.5. Comparison of the performance between DHS-G2 and DHS-G3 reactors Fig. 9. Changes in retained sludge concentration in the down-flow hanging sponge.
increased to 11 g VSS L1 sponge on day 250. Figs. 8 and 9 show that FC removal in the DHS-G3 reactor was related to the increase in retained sludge concentration. Namely, the gradient of the firstorder reaction constants indicated that the removal performance of the retained sludge had reached a maximum. Even though the retained sludge concentration continued to increase after day 250, no improvement in FC removal was observed, suggesting that once the retained sludge concentration reaches a certain level, further improvements in FC removal may not be possible. The geometric means of FC in the whole experimental period were 1.1 107 MPN 100 mL1 in sewage and 6.3 106 MPN 100 mL1 in the UASB effluent. The FC concentrations in the DHS effluent were 2.4 105 MPN 100 mL1 in Phase 1, 6.4 104 MPN 100 mL1 in Phase 2, and 8.1 105 MPN 100 mL1 in Phase 3. The log-based reduction of FC associated with the UASB þ DHS were relatively high, at least 1.75 log in Phases 1 and 2; these were comparable with the results of the DHS-G2 demonstration
Table 3 provides an overview of the water quality results of the DHS-G3 reactor. The summary of the DHS-G2 reactor is also given in Table S1 (Okubo et al., 2015; Onodera et al., 2016). Comparing the results of the two tests at similar experimental conditions, i.e., an influent flow rate of 1000 m3 d1 with and without recirculation, we found slightly better performance in DHS-G2 than in DHS-G3. This can be attributed to the difference in the water distribution between the two reactors. Although the water distribution inside the DHS reactor can be evaluated by a tracer experiment (Tandukar et al., 2006), it is difficult to conduct the experiment in such a largescale plant. When comparing the water distribution in the pilotscale DHS reactors (Table S2), the DHS-G2 reactor shows superior water distribution to the DHS-G3, i.e., the ratio of actual HRT per theoretical HRT ranged from 68.5% to 82.5% in DHS-G2, while it was only 24% in DHS-G3. However, it should be noted that, in addition to DHS-G3 having better workability than DHS-G2, the quality of its effluent over the course of the whole operation indicated high treatment performance, providing effluent water qualities that satisfied the Indian effluent quality standards except for FC.
Table 3 Summary of water quality parameters after treatment in DHS-G3 verification test. Parameter
Sewage
Temperature, C Total BOD, mg L1 Soluble BOD, mg L1 Total CODCr, mg L1 Soluble CODCr, mg L1 FC, MPN 100 mL1 TN, mg L1 1 NHþ 4 -N, mg L 1 NO 3 -N, mg L SS, mg L1 VSS, mg L1 DO, mg L1 Flow rate, m3 d1 Recirculation ratio, e
UASB eff.
26.2 (3.9) 130 (38) 57 (22) 380 (113) 128 (31) 1.1 107 36 (10) 23 (6)
25.7 (4.1) 62 (20) 38 (12) 168 (34) 89 (19) 6.3 106 34 (11) 26 (6)
229 (98) 161 (68)
51 (15) 39 (12)
25,450
25,450
DHS eff. Phase 1
Phase 2
Phase 3
25.3 (4.0) 10 (6) 5 (3) 40 (13) 26 (10) 2.4 105 22 (9) 12 (6) 7 (4) 11 (4) 9 (4) 6.0 (1.0) 500 0
17.9 (1.3) 4 (3) 3 (2) 42 (8) 31 (6) 6.4 104 38 (5) 8 (5) 8 (4) 10 (5) 7 (5) 5.7 (1.4) 500 1
20.3 (2.0) 14 (9) 12 (7) 52 (20) 32 (13) 8.1 105 34 (7) 19 (8) 4 (1) 20 (9) 17 (8) 5.6 (0.8) 1000 0
91 (5) 89 (5) 1.7 32 (25) 53 (19) 94 (4)
97 (2) 89 (4) 1.8 20 (15) 70 (21) 94 (5)
90 (7) 86 (7) 0.9 23 (13) 28 (19) 90 (7)
Removal Total BOD, % Total CODCr, % FC, log10 TN, % NHþ 4 -N, % SS, % Numbers in parentheses represent standard deviations.
51 (15) 54 (13) 0.3 3 (18) 74 (13)
146
T. Okubo et al. / Water Research 102 (2016) 138e146
4. Conclusions In this study, a DHS reactor filled with G3 carriers was operated downstream of a UASB reactor under practical conditions in India. (1) Full-scale verification tests of the DHS-G3 reactor have demonstrated its treatment performance, with effluent water qualities that satisfied the Indian effluent quality standards except for FC. (2) In order for this wastewater treatment technology to be more accessible to developing countries, it is important that the workability of DHS-G3 be superior to that of DHS-G2. (3) The sponge carrier can be easily manufactured in any developing country. We therefore believe that application of the DHS-G3 should be promoted in efforts to expand the practical use of the DHS process in the future. Acknowledgments This study was supported in part by research grants from the 1) Japanese Ministry of Education, Culture, Sports, Science and Technology (MEXT), 2) Japan Society for the Promotion of Science (JSPS) and the 3) Science and Technology Research Partnership for Sustainable Development (SATREPS). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.watres.2016.06.035. References APHA, 2005. Standard Methods for the Examination of Water and Wastewater, twenty-first ed. APHA/AWWA/WEF, Washington DC. Araki, N., Ohashi, A., Machdar, I., Harada, H., 1999. Behaviors of nitrifiers in a novel biofilm reactor employing hanging sponge-cubes as attachment site. Water Sci. Technol. 39 (7), 23e31. € rlenius, B., Andersson, M., 2006. Control of the aeration volume in an Ekman, M., Bjo activated sludge process using supervisory control strategies. Water Res. 40 (8), 1668e1676. duit, A., 2007. Oxygen transfer preFayolle, Y., Cockx, A., Gillot, S., Roustan, M., He diction in aeration tanks using CFD. Chem. Eng. Sci. 62 (24), 7163e7171. €s, P., 1982. Criteria for nitrification in fixed film reactors. Water Sci. Harremoe Technol. 14, 167e187.
€s, P., Jansen, J.l.C., Arvin, E., 2000. Wastewater TreatmentHenze, M., Harremoe biological and Chemical Processes, third ed. Springer. Huang, J.Y.C., Cheng, M.D., Mueller, J.T., 1985. Oxygen uptake rates for determining microbial activity and application. Water Res. 19 (3), 373e381. Ikeda, N., Natori, T., Okubo, T., Sugo, A., Aoki, M., Kimura, M., Yamaguchi, T., Harada, H., Ohashi, A., Uemura, S., 2013. Enhancement of denitrification in a DHS reactor by effluent recirculation. Water Sci. Technol. 68 (3), 591e598. Kubota, K., Hayashi, M., Matsunaga, K., Iguchi, A., Ohashi, A., Li, Y.-Y., Yamaguchi, T., Harada, H., 2014. Microbial community composition of a down-flow hanging sponge (DHS) reactor combined with an up-flow anaerobic sludge blanket (UASB) reactor for the treatment of municipal sewage. Bioresour. Technol. 151, 144e150. Kazmi, A.A., Tyagi, V.K., Trivedi, R.C., Kumar, A., 2007. Coliforms removal in full-scale activated sludge plants in India. J. Environ. Manag. 87 (3), 415e419. Okubo, T., Onodera, T., Uemura, S., Yamaguchi, T., Ohashi, A., Harada, H., 2015. Onsite evaluation of the performance of a full-scale down-flow hanging sponge reactor as a post-treatment process of up-flow anaerobic sludge blanket reactor for treating sewage in India. Bioresour. Technol. 194, 156e164. Onodera, T., Tandukar, M., Sugiyana, D., Uemura, S., Ohashi, A., Harada, H., 2014. Development of a sixth-generation down-flow hanging sponge (DHS) reactor using rigid sponge media for post-treatment of UASB treating municipal sewage. Bioresour. Technol. 152, 93e100. Onodera, T., Okubo, T., Uemura, S., Yamaguchi, T., Ohashi, A., Harada, H., 2016. Longterm performance evaluation of down-flow hanging sponge reactor regarding nitrification in a full-scale experiment in India. Bioresour. Technol. 204, 177e184. Parker, D.S., Richards, T., 1986. Nitrification in trickling filter. J. Water Pollut. Control Fed. 58 (9), 896e902. Sato, N., Okubo, T., Onodera, T., Ohashi, A., Harada, H., 2006. Prospects for a selfsustainable sewage treatment system: a case study on full-scale UASB system in India’s Yamuna River Basin. J. Environ. Manag. 80, 198e207. Sperling, M.V., Oliveira, C.M., Andrada, J.G.B., Godinho, V.M., Assuncçao, F.M.L., Junior, W.R.M., 2008. Performance evaluation of a simple wastewater treatment system comprised by UASB reactor, shallow polishing ponds and coarse rock filter. Water Sci. Technol. 58 (6), 1313e1319. Tandukar, M., Uemura, S., Machdar, I., Ohashi, A., Harada, H., 2005. A low-cost municipal sewage treatment system with a combination of UASB and the “fourth-generation” downflow hanging sponge reactors. Water Sci. Technol. 52 (1e2), 323e329. Tandukar, M., Uemura, S., Ohashi, A., Harada, H., 2006. Combining UASB and the “fourth generation” down-flow hanging sponge reactor for municipal wastewater treatment. Water Sci. Technol. 53 (3), 209e218. Tandukar, M., Ohashi, A., Harada, H., 2007. Performance comparison of a pilot-scale UASB and DHS system and activated sludge process for the treatment of municipal wastewater. Water Res. 41 (12), 2697e2705. Tawfik, A., El-Gohary, F., Ohashi, A., Harada, H., 2008. Optimization of the performance of an integrated anaerobic-aerobic system for domestic wastewater treatment. Water Sci. Technol. 58 (1), 185e194. Uemura, S., Suzuki, S., Abe, K., Yamaguchi, T., Ohashi, A., Harada, H., 2010. Removal of organic substances and ammonium oxidation by a down-flow hanging sponge (DHS) reactor under high salinity conditions. Bioresour. Technol. 101, 5180e5185. Witzig, R., Manz, W., Rosenberger, S., Krüger, U., Kraume, M., Szewzyk, U., 2002. Microbiological aspects of a bioreactor with submerged membranes for aerobic treatment of municipal wastewater. Water Res. 36, 394e402.
Contents lists available at ScienceDirect
Water Research journal homepage: www.elsevier.com/locate/watres
Development of a new non-aeration-based sewage treatment technology: Performance evaluation of a full-scale down-flow hanging sponge reactor employing third-generation sponge carriers Tsutomu Okubo a, *, Kengo Kubota b, Takashi Yamaguchi c, Shigeki Uemura a, Hideki Harada d a
Dept. of Civil Engineering, National Institute of Technology, Kisarazu College, 2-11-1 Kiyomidaihigashi, Kisarazu, Chiba, 292-0041, Japan Dept. of Civil and Environmental Engineering, Tohoku University, 6-6-06 Aza-Aoba, Aramaki, Aoba-ku, Sendai, Miyagi, 980-8579, Japan Dept. of Civil and Environmental Engineering, Nagaoka University of Technology, 1603-1 Kamitomioka, Nagaoka, Niigata, 940-2188, Japan d New Industry Creation Hatchery Center, Tohoku University, 6-6-04 Aza-Aoba, Aramaki, Aoba-ku, Sendai, Miyagi, 980-8579, Japan b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 26 February 2016 Received in revised form 2 June 2016 Accepted 14 June 2016 Available online 16 June 2016
A practical-scale down-flow hanging sponge (DHS) reactor using third-generation (G3) sponge carriers was applied for treatment of the effluent from an up-flow anaerobic sludge blanket (UASB) reactor treating municipal sewage. The process performance of the DHS reactor filled with G3 sponge carriers (DHS-G3) was evaluated by conducting an on-site experiment in India over one year. The performance of the DHS-G3 for removal of organic matter and ammonium-nitrogen at a relatively short hydraulic retention time (HRT) of only 0.66 h satisfied the Indian effluent quality standards except for fecal coliform. The removal rate constants for total biochemical oxygen demand (BOD) and fecal coliform determined based on the water quality profiles along the DHS-G3 almost reached equilibrium approximately four months after the start of operation, i.e., 2.45 h1 for BOD and 2.30 h1 for fecal coliform, respectively. The oxygen utilization activity of retained sludge was determined to assess the distribution of heterotrophic and autotrophic bacteria along the DHS-G3. Nitrification was promoted in the lower portion of the DHS-G3 reactor in the duration with low organic load, while it decreased when the organic load was increased, probably due to proliferation of heterotrophic bacteria. © 2016 Elsevier Ltd. All rights reserved.
Keywords: Sponge carrier Municipal sewage treatment Up-flow anaerobic sludge blanket reactor Down-flow hanging sponge reactor Oxygen utilization rate
1. Introduction The activated sludge process, currently the most popular method for wastewater treatment, consumes enormous quantities of energy, 70% of which is used for aeration (Fayolle et al., 2007). In contrast, the up-flow anaerobic sludge blanket (UASB) reactor is an economical treatment process that requires no power for aeration, and is becoming a core process in developing countries in tropical and subtropical regions (Sperling et al., 2008). UASB reactors built in India are designed for a hydraulic retention time (HRT) of 8.5 h and have ponds downstream for polishing the UASB-treated water (designated in India as “Final Polishing Units”, FPUs; design
* Corresponding author. E-mail addresses: [email protected] (T. Okubo), kengo.kubota.a7@ tohoku.ac.jp (K. Kubota), [email protected] (T. Yamaguchi), uemura@ wangan.c.kisarazu.ac.jp (S. Uemura), [email protected] (H. Harada). http://dx.doi.org/10.1016/j.watres.2016.06.035 0043-1354/© 2016 Elsevier Ltd. All rights reserved.
HRT ¼ 24 h). However, water treatment by the UASB-FPU combined systems has been unable to satisfy current Indian effluent quality standards (Okubo et al., 2015). The FPU also requires very large areas of land (Sato et al., 2006). Thus, the FPU process cannot really be called an optimal post-treatment technology. Our research group has been conducting basic research since the latter 1990s on the down-flow hanging sponge (DHS) process as an efficient post-treatment technology that does not detract from the economic superiority of UASB. To date, after improvements to the shape of the sponge carrier and to the filling method, the concept has evolved from its first generation (G1) to the sixth generation (G6) (Tandukar et al., 2007; Onodera et al., 2014). Parallel to those experiments, full-scale DHS reactors have been constructed in a sewage treatment plant in India, and verification tests have been conducted under local environmental conditions (Okubo et al., 2015; Onodera et al., 2016). DHS is an aerobic biological treatment method that employs polyurethane sponges, which can be manufactured in developing
T. Okubo et al. / Water Research 102 (2016) 138e146
countries, as the carrier for retaining the bacteria. The reactor is filled with these carriers in a way that exposes them to air. The water then trickles down through the sponges, facilitating aerobic conditions. In a continuous long-term experiment of a DHS reactor using the second-generation, curtain-type sponge carrier (G2) in India, this reactor (DHS-G2) was shown to be very economical and to achieve high performance in removal of organic matter and nitrogen (Okubo et al., 2015; Onodera et al., 2016). However, DHS-G2 had some disadvantages that made it unworkable; (1) to make a sponge module, the triangle sponge bars must be tilled on both sides of the plastic sheets, (2) a hook attached to the upper part of each sponge module must be hung on a bar that is equipped to the upper part of the reactor column, (3) multiple iterations of tilling sponges and hanging the hooks are time-consuming and increase costs, and (4) because the bars and hooks are always exposed to UASB-treated sewage, corrosion is inevitable, leading to failure of the sponge modules. To overcome these drawbacks of the G2 carrier, a thirdgeneration (G3) carrier was developed, which is cylindrical in shape and is composed of polyurethane sponge supported by a polyethylene plastic net to prevent compaction of the sponge. The DHS-G3 reactor only requires filling the reactor column with the G3 carriers (Tawfik et al., 2008). Thus, in this study, a full-scale DHS-G3 reactor was installed at an Indian sewage treatment plant site and continuously monitored. The characteristics of the DHS-G3 reactor at the start of operation were identified, and the reduction rate constants of biochemical oxygen demand (BOD) and fecal coliform (FC) were calculated using the profile along the main axis of the reactor with respect to time elapsed. The distribution of heterotrophic and autotrophic bacteria along the DHS-G3 was clarified by determining the oxygen utilization activity of retained sludge on different substrates. The reactor was also operated with recirculation during an experimental period at increased load, and its capabilities under these conditions were compared and evaluated. 2. Material and methods 2.1. Full-scale DHS-G3 reactor Prior to this experiment, a long-term continuous experiment was carried out using DHS-G2 constructed at a sewage treatment plant in Karnal, India (Okubo et al., 2015; Onodera et al., 2016). The reactor was then replaced with G3 carriers, and operation was restarted. Fig. 1 shows an overall view of the full-scale reactor and a
139
view of the reactor interior after it had been filled with the carriers. The reactor portion of the DHS reactor consisted of a concrete cylinder (5.5 m in diameter and 5.31 m in height) with an empty capacity of 126 m3. A clarifier was installed under the reactor. The reactor interior was vertically divided into four layers by reinforcement frames and nets. The four layers were equally filled with the G3 sponge carriers. In a previous study, we incorporated 33 and 38 ventilation windows measuring 0.2 m 0.3 m into the middle and lower sections of the concrete cylinder, respectively. In this study, we increased the number of ventilation holes to 104, which were incorporated between the sponge layers to provide more ventilation to the reactor interior. The DHS influent (i.e., UASB effluent) was pumped to the top of the reactor, and the water was uniformly sprayed onto the sponge carriers at the top of the reactor from a moving sprinkler, which rotated by the water head differential. The rotational speed of the sprinkler head was set to 8.5 rpm by adjustment of a valve in the sprayer nozzle. After sprinkling, the UASB effluent flowed downward through the sponge carriers under the force of gravity, passed into the clarifier, and flowed out as DHS effluent after it passed through the basin. Fig. 2a is a diagram of the structure of the G3 carrier showing its dimensions. The G3 sponge carrier is cylindrical, 32 mm in diameter and 32 mm long, and is composed of polyurethane sponge supported by a polyethylene plastic net to prevent compaction of the sponge (specific surface area: 1.87 cm2 cm3). Fig. 2b shows a close-up photograph of the polyurethane sponge. The long and short axes of 100 of the sponge pores were measured, and the mean value was 0.46 (standard deviation: ±0.10) mm. A total of 1,076,000 of the G3 sponges were loaded into the reactor to equally fill the four layers (269,000 layer1). The height of each layer was approximately 600 mm. Gaps measuring 450e700 mm were left between each layer to allow ingress of oxygen into the down-flowing water. The total effective sponge volume was 27.7 m3 (91.1 m2 m3 reactor working volume). Thus, relative to the empty volume (126 m3) of the DHS reactor, the sponge filling accounted for 22.3% (the sponge filling ratio relative to the effective reactor volume in each layer was 48.6%). Table 1 presents the operating conditions for the DHS reactor. Three conditions were employed to observe the effect of recirculation and increased load on treatment performance. Since the UASB effluent is mixed with a part of the treated water at a reservoir just before the sprinkler due to recirculation, the influent of the DHS is expected to contain much more dissolved oxygen (DO). Therefore, it was predicted that this would have a beneficial effect
Fig. 1. Photographs of the entire reactor (a), the distribution device (b), the reactor filled with sponge carriers (c), and a schematic diagram of the DHS reactor (d).
140
T. Okubo et al. / Water Research 102 (2016) 138e146
Fig. 2. Structure and dimensions of DHS-G3 (a) and close-up photo of polyurethane sponge (b).
on nitrification and FC removal in the reactor. Five hundred cubic meters per day of UASB effluent was treated from day 1e305 of operation without recirculation (Phase 1, HRT based on sponge volume: 1.33 h). Next, a volume of 500 m3 d1 was treated from day 306e325 of operation with effluent recirculation at a recirculation ratio of 1.0 (Phase 2, HRT based on sponge volume: 1.33 h). A volume of 1000 m3 d1 was then treated from day 326e365 of operation without recirculation (Phase 3, HRT based on sponge volume: 0.66 h). The mean amount of wastewater inflow into the Karnal sewage treatment plant during the test period was 25,450 m3 d1; the calculated HRT values from this flow are 13.5 h for the UASB reactor and 38.5 h for the FPU.
glucose, filtered sewage, filtered UASB effluent, ammoniumnitrogen (NHþ 4 -N) as ammonium chloride, and nitrite-nitrogen (NO 2 -N) as sodium nitrite. The filtered sewage and UASB effluent were prepared using a 0.45-mm membrane filter. A BOD bottle with sludge and without any substrate was also prepared as a control. The initial substrate concentrations of the test were adjusted to 80 mg chemical oxygen demand (COD) L1 for glucose, filtered sewage and filtered UASB effluent, and to 25 mg N L1 for NHþ 4 -N and NO 2 -N. The OUR measurements were conducted from a DO concentration near saturation down to approximately 0.5 mg L1 (or down to 3 mg L1 in the control bottle). These measurements took approximately 30 min (~90 min for the control). The DO decreased almost linearly throughout the measurement. Sodium bicarbonate (420 mg L1) was added at the start of the experiment to prevent the pH from being lowered due to nitrification in the experiments with NHþ 4 -N and NO2 -N. The pH in all experiments was in the range 7.01e7.19 at the beginning and end of the measurements. To suppress any nitrification, a nitrification inhibitor (allylthiourea, ATU) was added to give a concentration of 2.0 mg L1 for the test slurries in all of the systems other than NHþ 4 -N and NO2 N.
2.2. Oxygen utilization rate test
2.3. Analytical methods
An oxygen utilization rate (OUR) test was performed using DHSretained sludge in Phase 1 (sludge retrieved on days 252e255) and in Phase 3 (sludge retrieved on days 356e359). The sludge for use in the test was obtained by random selection of a total of 1604 sponge carriers from all four layers in the DHS reactor. Because the amount of biomass retained in the sponge carriers was largest in the top layer and decreased toward the lower layers of the reactor, we collected 10e20 sponge carriers from the top layer and 100e500 sponge carriers from the lower layers. The number of removed carriers corresponded to less than 0.15% of all the carriers in the DHS reactor. The sludge was squeezed from the sponges, washed in 10 mM phosphate buffered saline (PBS buffer), concentrated by centrifugation (6000 rpm, 10 min), and finally the test slurry was adjusted to a concentration of 4e6 g of volatile suspended solids (g VSS) L1. The test temperature was set to 20 ± 1 C. The substrates were placed in a 100-mL BOD bottle and agitated with a stirrer. A DO sensor (GU-Z, Iijima Electronics Corp.) was then immediately inserted and measurements were initiated. Table 2 shows the conditions for each experiment. The substrates were
The samples provided for analysis during continuous operation were sewage after passage through a bar screen (15 mm) and two mesh screens (25 mm and 10 mm) and a grid chamber, UASB effluent and DHS effluent. In addition, once a month, samples were retrieved via the ventilation holes in the DHS reactor, and a profile measurement was conducted along the main axis of the DHS reactor. Glass fiber filter paper (pore diameter: 0.45 mm, ADVANTEC-GB140) was used to obtain the soluble samples. The analysis items were BOD, CODCr, suspended solids (SS), VSS, total nitrogen (TN), NHþ 4 -N, NO3 -N, and FC. BOD was measured by adding ATU, based on the Standard Methods for the Examination of Water and Wastewater (APHA, 2005). CODCr, TN, NHþ 4 -N, and NO3 N were determined using the potassium dichromate method, the nitrogen tube test, the salicylic acid process, and the cadmium reduction method, respectively. All those tests were conducted using a DR/890 water quality analyzer (HACH). FC concentrations were determined by the most probable number (MPN) method using Difco-A1 medium (Becton, Dickinson and Company) according to APHA (2005).
Table 1 Operating conditions. Period (days)
Phase
Flow (m3 d1)
Recirculation ratio (e)
HRT (h)
0e305 306e325 326e365
1 2 3
500 500 1000
0 1 0
1.33 1.33 0.66
HRT: hydraulic retention time.
Table 2 Experimental set-up for oxygen utilization rate test.
OUR with organic substrates OUR by nitrification Endogenous OUR
Organic substrates
ATU
Sludge
PBS buffer (10 mM)
80 mg COD L1 of glucose, filtered sewage, or filtered UASB eff. 25 mg N L1 of NH4Cl or NaNO2 and 420 mg L1 of NaHCO3 e
Added e Added
4e6 g VSS L1 4e6 g VSS L1 4e6 g VSSvL1
e Added Added
ATU: allylthiourea, PBS: phosphate buffered saline.
T. Okubo et al. / Water Research 102 (2016) 138e146
3. Results and discussion 3.1. Organic matter removal characteristics Fig. 3a shows the time course of total BOD concentrations of sewage, UASB effluent and DHS-G3 effluent, during the whole experimental period. On the other hand, Fig. 3b shows the variation in BOD removal rate over time during this experiment. The average total BOD concentration over the experimental period was 130 (±38) mg L1 for sewage and 62 (±20) mg L1 for the UASB effluent. Within the first 24 h, the DHS-G3 effluent was 32 mg L1. The quality of the effluent soon remained stable, i.e., the average total BOD concentrations during Phase 1, Phase 2, and Phase 3 were 10 (±6) mg L1, 4 (±3) mg L1, and 14 (±9) mg L1, respectively. Thus, the final effluent satisfied the Indian effluent quality standards (total BOD standard: 30 mg L1) under all operating conditions. To evaluate the treatment performance with elapsed time, the total BOD profiles along the main axis of the DHS-G3 were taken 24 h, and 66, 126 and 250 days after the start of operation (Fig. 4a). The BOD concentrations of the influent to DHS-G3 (i.e., UASB effluent) vary with fluctuations because they depend on the composition of the incoming sewage and the treatment performance of the UASB reactor. Therefore, we considered that we could fit the logarithm of the BOD reduction rate to the first-order reaction equation for the region where it could be approximated as linear (Fig. 4b). The slopes of the regression lines in the figure are the reduction rate constants for total BOD (N0: BOD concentration at time 0, Nt: BOD concentration at time t). These results indicate that after 24 h of operation, most of the BOD removal was occurring in the first sponge layer but almost none in the second or lower layers. It was considered that the solid BOD components contained in the UASB effluent were trapped in
141
the sponge carriers in the first layer, and that no further biological treatment proceeded in the following layers because there was no active biomass. It was subsequently confirmed that the BOD was removed in the second and third layers as the effluent moved down through the reactor with the passage of time during operation. The overall performance of DHS-G3 was comparable to that of DHS-G2 and superior to that of TF treating UASB effluent (Okubo et al., 2015). The reduction rate constants at 24 h, and 66, 126 and 250 days from the start of operation were 1.33 h1 (regression line from DHS influent to first sponge layer), 2.04 h1 (regression line from DHS influent to second sponge layer), 2.45 h1 (regression line from DHS influent to third sponge layer), and 2.52 h1 (regression line from DHS influent to third sponge layer). The difference between the reduction rate constant after 24 h and the other constants was conjectured to be caused by biodegradation of organic matter due to sludge buildup. No great difference in BOD concentration between the third and fourth layers was found at day 126 and day 250, indicating that nearly all the BOD components had broken down by the time the sewage reached the third sponge layer. The role of the fourth sponge layer in this reactor was more for nitrification than for removing organic components, as discussed below. Since the reduction rate constants were nearly identical on day 126 and day 250, we deduce that the reduction rate constant was approximately steady from approximately four months after the start of operation. 3.2. Nitrification and nitrogen removal characteristics The changes of NHþ 4 -N in sewage, UASB effluent and DHS effluent with elapsed time are shown in Fig. 5. The mean NHþ 4 -N concentrations during the experimental period were 23
Fig. 3. Temporal changes in total biochemical oxygen demand (BOD) concentration (a) and BOD removal rate (b) during whole experimental period (n ¼ 127).
142
T. Okubo et al. / Water Research 102 (2016) 138e146
Fig. 4. Total biochemical oxygen demand (BOD) profiles in main axis direction of down-flow hanging sponge (DHS) (a) and reduction rate constants of total BOD by measurement day (b). N0: BOD concentration at time 0, Nt: BOD concentration at time t.
þ Fig. 5. Temporal changes in NHþ 4 -N concentration (a) and NH4 -N removal rate (b) during whole experimental period (n ¼ 54).
(±6) mg L1 in sewage and 26 (±6) mg L1 in UASB effluent. One of the reasons for the increased level in the UASB effluent was probably generation of NHþ 4 -N derived from organic nitrogen compounds. The levels in the DHS effluent were 12 (±6) mg L1 in Phase 1, 8 (±5) mg L1 in Phase 2, and 19 (±8) mg L1 in Phase 3; the best quality of treated water was obtained in Phase 2, when recirculation was implemented. The effluent quality standard for NHþ 4 -N level is 50 mg L1 in India, but it is important to remove NHþ 4 -N in order to mitigate its effect on the water environment in the final destination of the effluent. Just after starting Phase 1, nitrification was not observed, which was likely due to insufficient nitrifying bacteria; however, NHþ 4 -N
removal reached 40% by day 20 of operation and ranged from 40% to 80% thereafter. The level of NO 3 -N in the DHS effluent was 7 (±4) mg L1 (data not shown). Theoretically, since approximately the same amount of NHþ 4 -N is converted to NO3 -N under aerobic conditions, this lower NO3 -N concentration in the DHS effluent suggested the occurrence of denitrification inside the sponge carriers where anoxic conditions might be maintained (Uemura et al., 2010; Ikeda et al., 2013). Other than denitrification, assimilation of NHþ 4 -N to the bacteria was also considered to contribute to nitrogen removal, resulting in a 32% reduction in TN. Generally, for the nitrification process using a trickling filter, it is recommended to operate at volumetric organic loads of 0.2 kg
T. Okubo et al. / Water Research 102 (2016) 138e146
BOD m3 carrier d1 when the carriers are crushed rock (flow speed: 0.4e0.8 m h1), and 0.2e0.4 kg BOD m3 carrier d1 when the carriers are plastic (flow speed: 0.6e1.5 m h1) (Henze et al., 2000). However, the volumetric organic load in this DHS-G3 during Phase 1 was 1.1 kg BOD m3 sponge d1 and the flow speed was 1.7 m h1. Thus, the nitrification proceeded in this DHS-G3 reactor under more stringent operating conditions than those recommended for the trickling filter process. It was considered that almost all organics were removed in the upper portion of the DHSG3 reactor irrespective of its high volumetric organic load, and then an environment allowing the growth of nitrifying bacteria is established in the lower portion of the DHS reactor, which had a reduced organic load (Tandukar et al., 2005). Actually, in the DHS reactor examined here, the organic volume load in the first sponge layer was 4.28 kg BOD m3 sponge d1, but this was reduced to 0.35 kg BOD m3 sponge d1 in the fourth layer. The nitrification reactions in the third and fourth layers associated with the decreased load were observed in the profiles on day 250 (Phase 1), as shown in Fig. 6a, and were verified to progress in the downward direction of the reactor as the organic volume load decreased. The long SRT of 90e120 days for DHS-retained sludge (Tandukar et al., 2007) creates conditions favorable for proliferation of nitrifying bacteria and their retention in the carriers. Moreover, we assume that installing the sponges in four layers facilitated aeration of the wastewater, and that the high DO promoted the proliferation of nitrifying bacteria, thereby enhancing nitrification. In Phase 2, DHS-G3 was operated with recirculation at a recirculation ratio of 1.0. Therefore, the UASB effluent was mixed with the DHS effluent containing high DO just before entering the DHSG3. Thus, it is considered that the high DO-containing wastewater flowed in from the upper portion of the reactor, which meant that conditions favored the proliferation of nitrifying bacteria, contributing to the high NHþ 4 -N oxidation (70%). It was reported that when the soluble BOD concentration in treated water fell below 5 mg L1, the rate of nitrification in the trickling filter nitrification process €s, 1982). Since the concentration of was maximized (Harremoe organic matter in DHS effluent was also low in Phase 2 (soluble BOD was 3 (±2) mg L1), if removal of NHþ 4 -N is a priority, then recirculation is an effective approach, even when using a DHS reactor. Thus, the high NHþ 4 -N oxidation was shown by recirculation. On the other hand, although high NHþ 4 -N oxidation was shown by recirculation in Phase 2, TN removal worsened in Phase 2 compared to Phase 1, i.e., 22 (±9) mg N L1 and 38 (±5) mg L1 in Phases 1 and 2, respectively. With DHS carriers, the nitrificationdenitrification reactions proceed due to the unique properties of the sponge carriers, i.e., one carrier may provide an aerobic environment near the carrier surface (the nitrification reaction portion), and an anaerobic environment deeper inside the carrier (the
143
denitrification reaction portion) (Araki et al., 1999; Uemura et al., 2010). For this reason, it is possible that during Phase 2, high DOcontaining wastewater penetrates the interior of the carriers, suppressing the denitrifying bacteria and reducing the TN removal. Suppression of denitrification by effluent recirculation was also reported in a trickling filter process (Parker and Richards, 1986) and in a biological nitrogen removal process (Ekman et al., 2006). However, in the treatment of industrial wastewater containing a large amount of organics and NHþ 4 -N by a DHS reactor, effluent recirculation was advantageous for enhancing denitrification in the range of the recirculation ratio up to 2.0 (Ikeda et al., 2013). This was due to the fact that DO in the recirculated water was consumed very quickly to degrade highly concentrated organics in the upper part of the reactor. Fig. 6b shows the profiles along the main axis of the DHS reactor on day 346 (Phase 3). The increase in organic load increased the concentration of organics in the DHS effluent. The NHþ 4 -N oxidation decreased to approximately 30%, and the NO 3 -N concentration decreased to values below those seen during Phases 1 and 2. These findings suggest that the increased organic load favored heterotrophic bacteria over nitrifying bacteria in the lower portion of the reactor. However, the average concentration of DO in the DHS effluent was 5.6 (±0.8) mg L1, demonstrating that DHS-G3 was capable of taking up sufficient oxygen even at an HRT of 0.66 h. 3.3. Oxygen utilization rate (OUR) of DHS-retained sludge Measurements of the OUR in the DHS-retained sludge provide the oxygen consumption activity of microorganisms for each substrate. Fig. 7 shows the results from the OUR tests for each substrate, performed during Phase 1 and Phase 3. The OUR during Phase 1 was 1.65e2.49 mg O2 g1 VSS h1 in the glucose substrate. No marked differences were observed along the vertical axis of the DHS-G3 reactor. In Phase 3, however, values increased along the DHS-G3 reactor (varying from 1.03 to 6.53 mg O2 g1 VSS h1 in the glucose substrate). The OUR of the glucose substrate in Phases 1 and 3 was about one-tenth that of the activated sludge (Huang et al., 1985); about the same level as sludge samples from a membrane filtration reactor (Witzig et al., 2002) and from a DHS-G4 reactor operated in Japan (Tandukar et al., 2006). The OUR of the sewagesoluble fraction and the UASB effluent-soluble fraction did not differ markedly along the DHS-G3 reactor in Phase 1, but in Phase 3, just as seen in the OUR of the glucose, the OUR showed higher values in the sludge taken from increasing depths in the DHS-G3 reactor (sewage-soluble fraction substrate: 6.87e27.38 mg O2 g1 VSS h1, UASB effluent-soluble fraction substrate: 7.87e27.98 mg O2 g1 VSS h1). The probable reason for the low OUR in the upper portion of the reactor was that, even though the VSS concentration
Fig. 6. Profiles of NHþ 4 -N and dissolved oxygen (DO) along the main axis of the down-flow hanging sponge (DHS) on day 250 (Phase 1) (a) and on day 346 (Phase 3) (b).
144
T. Okubo et al. / Water Research 102 (2016) 138e146
Fig. 7. Measured OUR values in Phase 1 (sludge from days 252e255) and in Phase 3 (sludge from days 356e359) per substrate.
was high in the retained sludge in the upper portion, the percentage of microorganisms was relatively low due to the biologically inert particulate organics derived from the UASB effluent, resulting in the decreased oxygen consumption per unit of VSS. The OUR of the retained sludge on NHþ 4 -N and NO2 -N substrates in Phase 1 tended to increase from the higher to lower part of the DHS reactor, showing maximum values in the third sponge layer (16.25 mg O2 g1 VSS h1 on the NHþ 4 -N substrate and 6.98 mg O2 g1 VSS h1 on the NO 2 -N substrate). Notably, maxima also occurred in the third sponge layer in Phase 3 (12.18 mg O2 g1 1 VSS h1 on the NHþ VSS h1 on the 4 -N substrate and 7.42 mg O2 g þ NO2 -N substrate), but the OUR on the NH4 -N substrate decreased by 25% in Phase 3 compared to that in Phase 1. According to the results of the profile experiment, the nitrification was promoted in the lower portion of the DHS-G3 reactor in the duration with low organic load (Phase 1), and that the NHþ 4 -N oxidation decreased when the organic load was increased (Phase 3). Thus, those results were also clearly supported by the OUR tests. The oxygen consumption rate associated with NHþ 4 -N oxidation calculated from the reactor profiles in each sponge layer was 1%e 60% lower than the value obtained by the OUR test. A possible reason for this was that, since the OUR test was performed with agitation by a stirrer, the contact efficiency between the sludge and the substrate was much better in the OUR test than in the reactor. The OUR fed with no substrate (only sludge) showed no variation along the DHS-G3 reactor in Phase 1, while in Phase 3, however, the OUR increased in the downward direction in the reactor. This is primarily because of the high OUR due to heterotrophic bacteria in the lower portion of the DHS reactor, which proliferated with increased organic load.
3.4. Fecal coliform removal characteristics Fig. 8 shows the first-order reaction constants for FC removal only in Phase 1. The constants were predicted by regression of the values of the retention time that was calculated from the distance along the DHS-G3 reactor and FC counts at each sponge layer (N0: FC counts at time 0, Nt: FC counts at time t). There was no clear difference between the first-order reaction constants at 24 h after the start of operation and at 66 days, implying that there was no improvement in the ability to remove FC. However, the constants gradually increased from day 87 and reached greater than 2.0 h1 on day 126, indicating that high and stable FC removal was
Fig. 8. Reduction rate constants for fecal coliform by measurement day. N0: FC counts at time 0, Nt: FC counts at time t.
established in the DHS-G3 reactor after four months of operation. These results clearly indicate that FC removal is also associated with biomass build-up, as described later. It has been reported that FC removal in the DHS reactor is achieved by inactivation after adhesion onto the retained sludge, or predation by protozoa (Tawfik et al., 2008). Therefore, development of stable microfauna might also be a key factor for FC removal (Kubota et al., 2014). Fig. 9 shows the variation of the retained sludge concentration in the DHS-G3 reactor after the start of operation. Two or three carriers were randomly collected from each sponge layer and squeezed repeatedly in distilled water to obtain samples of the DHS-retained sludge. Because the sludge had not built up yet just after the start of operation, the FC removal at the early stage of operation might be established by a physical process, such as adhesion or filtration. The concentration of retained sludge on day 66 was approximately 3.0 g VSS L1 sponge, but no increase in the first-order reaction constants for FC removal was observed, suggesting that the sludge concentration had not reached the level needed for FC removal. The retained sludge concentration was 5.0 g VSS L1 sponge on day 87 when FC removal began to increase. At that time, the amount of sludge per volume was approximately the same as with the activated sludge process. This concentration
T. Okubo et al. / Water Research 102 (2016) 138e146
145
experiment (Onodera et al., 2016). However, in Phase 3, possibly because the contact time between the wastewater and sludge was insufficient due to the reduced HRT, the log reduction was lower (0.9 log). The mean log decrease during the latter part of Phase 1 (day 150 and thereafter) was 2.0 log, comparable to the range 1.4e3.5 log (mean, 2.1 log) determined for the activated sludge process in India’s capital city, New Delhi (Kazmi et al., 2007). However, the DHS-G3 effluent unsatisfied the effluent quality standard of FC for irrigation use in India (104 MPN 100 mL1). Therefore, additional treatment will be necessary to meet the standard, i.e., the water must be disinfected with chlorine. On the other hand, Uemura et al. (2010) reported that the oxygen uptake and inactivation of FC in a DHS-G2 reactor were improved by reducing the size of the carriers. More work is necessary to develop a method to improve the efficiency of FC removal in a DHS reactor. 3.5. Comparison of the performance between DHS-G2 and DHS-G3 reactors Fig. 9. Changes in retained sludge concentration in the down-flow hanging sponge.
increased to 11 g VSS L1 sponge on day 250. Figs. 8 and 9 show that FC removal in the DHS-G3 reactor was related to the increase in retained sludge concentration. Namely, the gradient of the firstorder reaction constants indicated that the removal performance of the retained sludge had reached a maximum. Even though the retained sludge concentration continued to increase after day 250, no improvement in FC removal was observed, suggesting that once the retained sludge concentration reaches a certain level, further improvements in FC removal may not be possible. The geometric means of FC in the whole experimental period were 1.1 107 MPN 100 mL1 in sewage and 6.3 106 MPN 100 mL1 in the UASB effluent. The FC concentrations in the DHS effluent were 2.4 105 MPN 100 mL1 in Phase 1, 6.4 104 MPN 100 mL1 in Phase 2, and 8.1 105 MPN 100 mL1 in Phase 3. The log-based reduction of FC associated with the UASB þ DHS were relatively high, at least 1.75 log in Phases 1 and 2; these were comparable with the results of the DHS-G2 demonstration
Table 3 provides an overview of the water quality results of the DHS-G3 reactor. The summary of the DHS-G2 reactor is also given in Table S1 (Okubo et al., 2015; Onodera et al., 2016). Comparing the results of the two tests at similar experimental conditions, i.e., an influent flow rate of 1000 m3 d1 with and without recirculation, we found slightly better performance in DHS-G2 than in DHS-G3. This can be attributed to the difference in the water distribution between the two reactors. Although the water distribution inside the DHS reactor can be evaluated by a tracer experiment (Tandukar et al., 2006), it is difficult to conduct the experiment in such a largescale plant. When comparing the water distribution in the pilotscale DHS reactors (Table S2), the DHS-G2 reactor shows superior water distribution to the DHS-G3, i.e., the ratio of actual HRT per theoretical HRT ranged from 68.5% to 82.5% in DHS-G2, while it was only 24% in DHS-G3. However, it should be noted that, in addition to DHS-G3 having better workability than DHS-G2, the quality of its effluent over the course of the whole operation indicated high treatment performance, providing effluent water qualities that satisfied the Indian effluent quality standards except for FC.
Table 3 Summary of water quality parameters after treatment in DHS-G3 verification test. Parameter
Sewage
Temperature, C Total BOD, mg L1 Soluble BOD, mg L1 Total CODCr, mg L1 Soluble CODCr, mg L1 FC, MPN 100 mL1 TN, mg L1 1 NHþ 4 -N, mg L 1 NO 3 -N, mg L SS, mg L1 VSS, mg L1 DO, mg L1 Flow rate, m3 d1 Recirculation ratio, e
UASB eff.
26.2 (3.9) 130 (38) 57 (22) 380 (113) 128 (31) 1.1 107 36 (10) 23 (6)
25.7 (4.1) 62 (20) 38 (12) 168 (34) 89 (19) 6.3 106 34 (11) 26 (6)
229 (98) 161 (68)
51 (15) 39 (12)
25,450
25,450
DHS eff. Phase 1
Phase 2
Phase 3
25.3 (4.0) 10 (6) 5 (3) 40 (13) 26 (10) 2.4 105 22 (9) 12 (6) 7 (4) 11 (4) 9 (4) 6.0 (1.0) 500 0
17.9 (1.3) 4 (3) 3 (2) 42 (8) 31 (6) 6.4 104 38 (5) 8 (5) 8 (4) 10 (5) 7 (5) 5.7 (1.4) 500 1
20.3 (2.0) 14 (9) 12 (7) 52 (20) 32 (13) 8.1 105 34 (7) 19 (8) 4 (1) 20 (9) 17 (8) 5.6 (0.8) 1000 0
91 (5) 89 (5) 1.7 32 (25) 53 (19) 94 (4)
97 (2) 89 (4) 1.8 20 (15) 70 (21) 94 (5)
90 (7) 86 (7) 0.9 23 (13) 28 (19) 90 (7)
Removal Total BOD, % Total CODCr, % FC, log10 TN, % NHþ 4 -N, % SS, % Numbers in parentheses represent standard deviations.
51 (15) 54 (13) 0.3 3 (18) 74 (13)
146
T. Okubo et al. / Water Research 102 (2016) 138e146
4. Conclusions In this study, a DHS reactor filled with G3 carriers was operated downstream of a UASB reactor under practical conditions in India. (1) Full-scale verification tests of the DHS-G3 reactor have demonstrated its treatment performance, with effluent water qualities that satisfied the Indian effluent quality standards except for FC. (2) In order for this wastewater treatment technology to be more accessible to developing countries, it is important that the workability of DHS-G3 be superior to that of DHS-G2. (3) The sponge carrier can be easily manufactured in any developing country. We therefore believe that application of the DHS-G3 should be promoted in efforts to expand the practical use of the DHS process in the future. Acknowledgments This study was supported in part by research grants from the 1) Japanese Ministry of Education, Culture, Sports, Science and Technology (MEXT), 2) Japan Society for the Promotion of Science (JSPS) and the 3) Science and Technology Research Partnership for Sustainable Development (SATREPS). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.watres.2016.06.035. References APHA, 2005. Standard Methods for the Examination of Water and Wastewater, twenty-first ed. APHA/AWWA/WEF, Washington DC. Araki, N., Ohashi, A., Machdar, I., Harada, H., 1999. Behaviors of nitrifiers in a novel biofilm reactor employing hanging sponge-cubes as attachment site. Water Sci. Technol. 39 (7), 23e31. € rlenius, B., Andersson, M., 2006. Control of the aeration volume in an Ekman, M., Bjo activated sludge process using supervisory control strategies. Water Res. 40 (8), 1668e1676. duit, A., 2007. Oxygen transfer preFayolle, Y., Cockx, A., Gillot, S., Roustan, M., He diction in aeration tanks using CFD. Chem. Eng. Sci. 62 (24), 7163e7171. €s, P., 1982. Criteria for nitrification in fixed film reactors. Water Sci. Harremoe Technol. 14, 167e187.
€s, P., Jansen, J.l.C., Arvin, E., 2000. Wastewater TreatmentHenze, M., Harremoe biological and Chemical Processes, third ed. Springer. Huang, J.Y.C., Cheng, M.D., Mueller, J.T., 1985. Oxygen uptake rates for determining microbial activity and application. Water Res. 19 (3), 373e381. Ikeda, N., Natori, T., Okubo, T., Sugo, A., Aoki, M., Kimura, M., Yamaguchi, T., Harada, H., Ohashi, A., Uemura, S., 2013. Enhancement of denitrification in a DHS reactor by effluent recirculation. Water Sci. Technol. 68 (3), 591e598. Kubota, K., Hayashi, M., Matsunaga, K., Iguchi, A., Ohashi, A., Li, Y.-Y., Yamaguchi, T., Harada, H., 2014. Microbial community composition of a down-flow hanging sponge (DHS) reactor combined with an up-flow anaerobic sludge blanket (UASB) reactor for the treatment of municipal sewage. Bioresour. Technol. 151, 144e150. Kazmi, A.A., Tyagi, V.K., Trivedi, R.C., Kumar, A., 2007. Coliforms removal in full-scale activated sludge plants in India. J. Environ. Manag. 87 (3), 415e419. Okubo, T., Onodera, T., Uemura, S., Yamaguchi, T., Ohashi, A., Harada, H., 2015. Onsite evaluation of the performance of a full-scale down-flow hanging sponge reactor as a post-treatment process of up-flow anaerobic sludge blanket reactor for treating sewage in India. Bioresour. Technol. 194, 156e164. Onodera, T., Tandukar, M., Sugiyana, D., Uemura, S., Ohashi, A., Harada, H., 2014. Development of a sixth-generation down-flow hanging sponge (DHS) reactor using rigid sponge media for post-treatment of UASB treating municipal sewage. Bioresour. Technol. 152, 93e100. Onodera, T., Okubo, T., Uemura, S., Yamaguchi, T., Ohashi, A., Harada, H., 2016. Longterm performance evaluation of down-flow hanging sponge reactor regarding nitrification in a full-scale experiment in India. Bioresour. Technol. 204, 177e184. Parker, D.S., Richards, T., 1986. Nitrification in trickling filter. J. Water Pollut. Control Fed. 58 (9), 896e902. Sato, N., Okubo, T., Onodera, T., Ohashi, A., Harada, H., 2006. Prospects for a selfsustainable sewage treatment system: a case study on full-scale UASB system in India’s Yamuna River Basin. J. Environ. Manag. 80, 198e207. Sperling, M.V., Oliveira, C.M., Andrada, J.G.B., Godinho, V.M., Assuncçao, F.M.L., Junior, W.R.M., 2008. Performance evaluation of a simple wastewater treatment system comprised by UASB reactor, shallow polishing ponds and coarse rock filter. Water Sci. Technol. 58 (6), 1313e1319. Tandukar, M., Uemura, S., Machdar, I., Ohashi, A., Harada, H., 2005. A low-cost municipal sewage treatment system with a combination of UASB and the “fourth-generation” downflow hanging sponge reactors. Water Sci. Technol. 52 (1e2), 323e329. Tandukar, M., Uemura, S., Ohashi, A., Harada, H., 2006. Combining UASB and the “fourth generation” down-flow hanging sponge reactor for municipal wastewater treatment. Water Sci. Technol. 53 (3), 209e218. Tandukar, M., Ohashi, A., Harada, H., 2007. Performance comparison of a pilot-scale UASB and DHS system and activated sludge process for the treatment of municipal wastewater. Water Res. 41 (12), 2697e2705. Tawfik, A., El-Gohary, F., Ohashi, A., Harada, H., 2008. Optimization of the performance of an integrated anaerobic-aerobic system for domestic wastewater treatment. Water Sci. Technol. 58 (1), 185e194. Uemura, S., Suzuki, S., Abe, K., Yamaguchi, T., Ohashi, A., Harada, H., 2010. Removal of organic substances and ammonium oxidation by a down-flow hanging sponge (DHS) reactor under high salinity conditions. Bioresour. Technol. 101, 5180e5185. Witzig, R., Manz, W., Rosenberger, S., Krüger, U., Kraume, M., Szewzyk, U., 2002. Microbiological aspects of a bioreactor with submerged membranes for aerobic treatment of municipal wastewater. Water Res. 36, 394e402.