Low temperature treatment of municipal sewage in anaerobic fluidized bed reactors

By: I. Sanz and F.Fdz-Polanco

Tomado de:  Water Research

Abstract

The anaerobic fluidized bed reactor (AFBR) appears to be most promising for the treatment of low strengh wastes, such as municipal sewage, at low temperature, since the process is able to maintain a large mass of active microorganisms and provides effective removal of TSS. The study is divided in three parts. The objective of the first part is to characterize the effect of decreasing temperature on the performance of two mature AFBR reactors. The second part presents the data from 220 days of operation at 10°C, and in the third part two start-ups, with and without inoculum at 15°C, are evaluated. A gradual temperature decrease from 20 to 5° C, allowing the microorganisms to acclimate to the new lower temperature, did not have a great effect on effluent quality. However a great accumulation of TSS was observed in the top of the fluidized bed. At 10° C, and a hydraulic retention time of 1.5 h. 70% of T_COD removal was achieved. It is possible to start-up the AFBR at 15°C without inoculation; however, at least 4 months is required to get good quality effluents.

Key words - anaerobic fluidized bed reactors, raw municipal sewage, low temperature treatment.

The succesful application of anaerobic processes to domestic sewage, considering its particular characteristics, is a great challenge. First, the organic strength of sewage is relatively weak. Therefore, cell growth rate is limited, since the rate of growth of the organisms and the rate of substrate uptake are proportional to substrate concentration (Jewell, 1985). In addition, domestic sewage should be treated without heating, otherwise it is not economically feasible. This means the process should be able to operate over a wide range of temperatures (10-22°C) and occasionally lower than 10°C. In addition, domestic wastewater generally contains more particulate organics than soluble organic material. This slows the reaction rate, because the degradation rate of particles is slower than for soluble species, and the initial hydrolysis reaction can be relatively slow (Rittmann and Baskin, 1985). Some domestic sewages have a high sulphate concentration (>100 mg/l). This can decrease methane production because some sulphate reducing bacteria use acelate as an electron source. The sulphide produced may cause serious odour problems.
The anaerobic fluidized bed reactor (AFBR) system appears to be most feasible for the treatment of domestic sewage. The process is able to maintain a large population of active microorganisms required to overcome the limit of slow growth of anaerobic microorganisms.
AFBR are extremely efficient in removing organic suspended solids. Most of the solids remain in the reactor until they are hydrolysed (Yoda et al. 1985).
It has been reported that the hydrolysing and acidifying communities are associated with entrapped suspended solids, whereareas the methanogens are located in the film (Jewell, 1985). This can improve both particulate hydrolysis and rate of methanogenesis.

The main objective of this work is to provide information about the stability of the AFBR process treating domestic sewage at low temperature, and most of them do not contain data over long operation periods (over 100 days). The results from an experiment at decreasing temperatures, a restarting and long operating period at low temperature (220 days at 10°C) and a comparative start-up (with-without inoculum) are presented.

Material and methods

Four laboratory anaerobic fluidized bed reactors were used to carry out the experiments. Figure 1 shows a schematic diagram of the AFBR-1 and AFBR-2 The reactors AFBR-3 and AFBR-4 are schematically represented in Fig. 2. Some of the reactor characteristics are summarized in Table 1. Arlita ^(R) is an aluminium and iron silicate with traces of calcium carbonate and titanium, calcium sodium and potassium oxides.

Fig.1 Scheme of the reactors AFBR-1 and AFBR-2

AFBR-1 and AFBR-2 were made of glass. Gas collectors were not installet AFBR-3 and AFBR-4 consisted of a plexiglass tube with dimensions of 5.4 cm internal diameter by 63 cm high. The upper part of the reactors was equipped with a combined gas collector-settler along the column at 15 cm intervals allowing samples of the media and accumulated solids to be taken.
The reactor effluent was drawn from the top and pumped into the bottom assembly at a constant flow rate to achieve a 20% bed expansion. The influent was pumped into the recycle line of the reactor using a peristaltic pump. A submerged membrane pump was used in the recycle line.
The recirculation container (0.15 ml) was opened to the atmosphere in AFBR-1 and AFBR-2, but it was closed in AFBR-3 and AFBR-4.
The inlet system installed in AFBR-3 and AFBR-4 (Fig.2) was an improvement of the system used in AFBR-1 and AFBR-2 (Fig. 1). It allowed better flow distribution, fewer possibilities for channeling and solved clogging problems in the inlet point during shut-downs which occurred very often with the system used in AFBR-1 and AFBR-2.
The temperature was controlled by using a cryostatic bath.
The influent used was raw domestic sewage obtained from two different sewers having different wastewater characteristics. The experiment at decreasing temperature was carried out with raw sewage from Sewer-1. For the other experiments, wastewater from Sewer-2 was used. Some of the influent characteristics are presented in Table 2.
Organic loading rated (B_{v tcod}) and hydraulic retention time (f) are calculated based onthe empty volume of the fluidized bed.

Fig. 2 Scheme of the reactors AFBR-3 and AFBR-4

Influet and effluent analyses were determined according to Standard Methods (APHA et al. 1985) Attached biomass was measured as indicated in Swizenbaum (1978) Attached biomass was mesured as indicated in Switenbaum (1978). Nitrate, chloride, phosphate and sulphate were performed by ion chromatography with a Waters HPLC using the column Waters IC-PACK Anion Column P/N 07355. Elemental sulphur, carbon, hydrogen and nitrogen were determined with an elemental analyst Perkin-Elmer Model 240C. The biogas composition was measured with a Hewlett-Packard model 5790 Gas Chromatograph using Poropak Q as column packing.

                                                                Table 1  Reactor characteristics

                                                                   Table 2. Wastewater characteristics

Results and discussion

To evaluate the general behaviour of the anaerobic fluidized bed reactor for treating a low-strength wastewater, such as raw municipal sewage at low temperature, three different experiments were carried out.

Performance at decreasing temperatures

This experiment was carried out with AFBR-1 and AFBR-2 which had previously been working for a period of 20 months at room temperature (20°C) achieving T_cod and T_bod5 removals of 75 and 80%, respectively, when t = 2.7h and B_v.tcod = 2.5-3.5 g/l.d with effluent T_cod = 110 mg/l, T_bod5, = 50 mg/l and TSS = 20 mg/l.
Raw domestic sewage from Sewer-1 (Table 2) was used as influent. The goal of this experiment wast to evaluate the effect of a gradual temperature decrease on process efficiency. The operating conditions of the last room temperature period were kept constant while the temperature was decreased step-by-step from 20 to 5°C.
Figures 3 and 4 show the evolution of T_cod, organic loading rates (Bv._tcod), T_cod removal, gradual temperature decrease and mean hydraulic retention time (f) for AFBR-1 and AFBR-2, respectively.
On day 68, the reactors were fed with an unusually high stregth influent (T_cod = 900 mg/l) which provided a B_{v, tcod} = 9.0 g/l.d. The effluent T cod increased to 180 mg/l for AFBR-1 and 215 mg/l for AFBR-2. However, the reactors recovered quickly after the overloading.
Sudden and short temperature decrease had no significant impact on the performance of the reactors.
They returned quickly to their habitual performance level.
When the reactors were operated at room temperature (20°C), a small amount of accumulated solids in

the top of the fluidized bed was obserbed. These solids could be either non-attached biomass or suspended solids from the influent trapped in the reactor. During the period of operation at lower temperature (10°C), the accumulated solids increased considerably. Samples taken for analyses showed levels of TS = 30 g/l and VS = 15 g/l of compacted solids.

Table 3 summarizes the operating conditions, and some influent and effluent characteristics during the period 43-135 days, when the temperature was 10°C and lower.
Effluent Tcod as well as T aod, had a tendency to increase when temperature decreased, but recovered gradually to previous concentrations after several days. With a hydraulic retention time of 2.7 h (Bvrtcod = 4.5 g/l.d), a Tcod removal greater than 75% was achieved. The effluent TSS are quite stable and usually lower than 20 mg/l.
TKN and NH₃-N in the reactors. The effluent TKN concentration was slightly lower than that of the influent (3%) due to the nitrogen requirements of the microorganisms and physical removal of particulate Org-N.
Sulphide was no found in the effluent. A consistent relationship between influent and effluent sulphate concentrations was not found. On the other hand, small white granules containing a high percentage of sulphur (78%) were observed in the recirculation tank and in the effluent. A similar phenomenon was documented previously by Coulter et al. (1957) and Brown (1985), who concluded that the sulphate reduction by sulphate reducing bacteria has sulphur as the endproduct, since insufficient organic materials are available to reduce sulphate to sulphide. Howevere, Thiothrix bacteria were found in the effluent.
These microorganisms are able to use sulphide as an energy source and transform it into sulphur, which is deposited as sulphur granules in their cells. This indicates that part of the sulphates were reduced to sulphides, which may be used for the Thiothricx bacteria and converted into sulphur. Unfortunately there are not enough data to establish whether or not the sulphates were reduced to sulphides and subsequently reoxidized to elemental sulphur. Further research is needed.
After 93 days of operation at 10°C and lower, the attached biomass was 22 g VS/1 for AFBR-1 and 25 g VS/1 for AFBR-2.

Performance at 10°C

The goals of this experiment were to study the restart of anaerobic fluidized bed reactors at low temperatures (10°C) after a shut-down of 2.5 months, and to evaluate their performance during a long operating period (235 days) at low temperatures (10°C). This was carried out with AFBR-1 and AFBR-2 when the experiment at decreasing temperatures was concluded.
Raw domestic sewage from Sewer-1 was used during the restart; but since day 98, wastewater from Sewer-2 was used. This change allowed operation with higher loading rates and higher fluctuations in T_cod,T_bod, and TSS.
Figures 5 and 6 show the evolution of effluent Tcod, Bv tcod, Tcod removal and mean hydraulic retention time, for AFBR-1 and AFBR-2, respectively.
The initial f was 13.7 h for AFBR-1 and 10.1. h for AFBR-2 After 4-days, effluent Tcod waw lower than 100 mg/l. Thereafter, hydraulic retention time was gradually decreased. After 20 days, the reactors were operating at conditions similar to those before the shut-down.
Table 4 summarizes the operation conditions and some influent and effluent characteristics when the reactors were operating at f=1.5 h with high Bvtcod fluctuations, B_{v tcod} = 6.0-35.5 g/l.d (days 60-235).
In spite of the organic loading rates applied (mean Bv tcod = 8.9 g/l.d for AFBR-1 and 10.4 for AFBR-2), and the great fluctuations in influent characteristics, the reactors performed in a vey stable manner

Fig. 5. Performance at 10° C, AFBR-1

Fig.6. Performance at 10° C, AFBR-2

Tcod removal was usually higher than 70% and Tbod, removal about 80%, which shows the high efficiency of the anaerobic fluidized bed reactors treating low strength wastes at low temperatures. However, because the effluent quality is not sufficient for direct discharge, some kind of post-treatment must follow the anaerobic process.
TKN, Org-N and NH3-N were analysed periodically. About 80% of the effluent TKN is NH3-N, which shows that 70% of the influent Org-N is converted into NH3-N in anaerobic conditions.
As in the experiment at decreasing temperatures, sulphide was never detected in the effluent, and white granules with a high percentage of sulphur (>70%) were present in the effluent.
After 230 days of operation the attached biomass was 31 g VS/1 in AFBR-1 and 37 g VS/1 in AFBR-2.

Comparative start-up

The start-up of anaerobic fluidized bed reactors is a vey pivotal period in their operation. The challenge is to develop a well attached biolayer on the carrier.
The usual procedure is to use anaerobic sludge from an operting anaerobic reactor as inoculum This is quite expensive and generally the microorganisms need time to adapt to a new substrate and possibly to a new temperature. Inoculation is an obligatory step when the wastewater, such as some industrial wastes, lacks microorganisms. However, this is not a problem in domestic sewage. There is documentation of almost unseeded reactors having been started (Schellinkhout et al., 1985).
AFBR-3 and AFBR-4 (Fig.2, Table 1) were used to compare the start-up with and without seed.
Anaerobic sludge from a UASB reactor which treats sugar beet wastewater (T_cod = 4000 mg/l) at 35°C was used as seed.
The reactors were filled with 0.81. of Arlita ^(R)  and settled domestic sewage. AFBR-3 was seeded with 50 ml of anaerobic sludge (TS = 142 g/l. VS = 36 g/l).

Fig. 7. Comparative start-up. Seeded reactor (AFBR-3)

Fig.8. Comparative start-up. Unseeded reactor (AFBR-4)

No seed was added to AFBR-4. Both reactors were maintained for 1 week in recirculating conditions before the pumping of raw domestic sewage from Sewer-2 was initiated. The temperature was controlled at 15°C.
Figures 7 and 8 show the evolution of effluent T_cod,Bv.tcod,T_cod removals and the mean hydraulic retention time for the seeded (AFBR-3) and unseeded (AFBR-4) reactors. In Fig. 7 the effluent soluble COD and soluble COD removal are represented for the first 56-day period. From day 56, T_cod and Scod values differed less than 5% and only T_cod is represented.
The initial f for both reactors was 15 h. At the beginning, the effluent of the seeded reactor had a large amount of black suspended solids (TSS = 410 mg/l) mainly sludge that had not been retained in the reactor and was responsable for the high values of Tcod. However the soluble COD was very low (50 mg/l). In the unseeded reactor, the effluent T cod was very high, 375 mg/l, slightly lower than the soluble influent COD. However, a TSS removal of 90%, and effluent TSS lower than 25 mg/l, were usually achieved.

As shown in Figs 7 and 8 (B_v,tcod = 2.0 g/ld), about 70 days were required until both effluents reached similar characteristics. T cod about 125 mg/l and TSS about 40 mg/l.
Table 5 summarizes the operating conditions and some of the influent and effluent characteristics during the period 135-153 days. Five months after the different start-up, both reactors were operating under similar conditions (f = 2,4 h for AFBR-3 and f= 2.2.h for AFBR-4) Bvtcod ranged from 5.5 to 10.0 g/l.d, achieving Tcod removal higher than 75%.
The effluent quality was slightly better in the unseeded reactor, effluent Tcod = 178 mg/l, TSS = 46 mg/l for AFBR-3 and effluent Tcod = 170 mg/l and TSS = 43mg/l for AFBR-4.
Both reactors were equipped with a combined gas collector-settler device (Fig.2) and a gas meter.
For the last period of operation, f = 2.2-2.4 h and Bv.tcod = 7.7-8.3 g/l.d, the gas production rate observed was 1.0-1.51 of total gas/d. Biogas analysis showed a methane content always higher than 60%, with a methane content always higher than 60%, with an average value of 80%, while the rest was made up of nitrogen (13%) and carbon dioxide (7%).

It would be reasonable to assume that the nitrogen gas initially dissolved in the influent was stripped from the liquid phase by methane gas in the anaerobic reactor (Lettinga et al., 1985). Since the anaerobic treatment process for municipal sewage requires operation at high hydraulic loadings, a substantial part of the methane gas leaves the reactor in the dissolved phase. It has been reported that the methane collected represents only 10-30% of the entire methane produced (Yoda et al.,1985). In this study 40-60% of the maximum possible methane produced was collected.

Conclusions

A gradual temperature decrease from 20 to 5°C. allowing the microorganisms to acclimate to the new lower temperature, had little effect on effluent quality. Though effluent T_cod and T _bod5 had a tendency to increase when temperature was decreased, they recovered their previous concentrations after several days.
At 10°C, f = 2.8 h and B_v.tcod = 2.4-3.3 g/l.d, a T _cod removal higher than 75%, T _bod5 removal = 85%, effluent T_cod=125 mg/l and TSS < 25 mg/l was achieved. However, at temperatures of 10°C or lower there was a large accumulation of suspended solids from the influent on the top of the fuidized bed.
The restarting at 10°C after a 2.5 month shut-down was very quick, achieving an effluent T _cod lower than 100 mg/l in 4 days. In spite of the high organic loading rates applied, B_v.tcod = 6.9-35.5 g/l.d (f=12.5 h), and great fluctuations in the influent characteristics, the reactors operated very stably with T_cod removal> 70% and T_bod5 removal = 80%.
With municipal sewage as the influent, it is possible to start-up anaerobic fluidized bed reactors without inoculum at 15°C. However, it requires at least 4 months to get good quality effluent.

References

APHA et al. (1985) Standard Methods for the Examination of Water and Wastewater, 16th edition. American Public Health Association, American Water Works Association and Water Pollution Control Federation, Washingyton, D.C.

Brown D.S. Mehotra S., Petrasek A.C., Eralp A.E., Bishop D.F. and Springer A.M. (1985) Evaluation of anaerobic expanded bed contactors for municipal wastewater treatment. Proc. Sem./Workshop Anaerobic Treatment of Sewage, Univ. Massachusetts, pp. 209-258

Coulter J.B. Soneda S. and Ettinger M.B.(1957) Anaerobic contact process for sewage disposal. Swge ind. Wastes 29,468.

Jewell W.J. (1985) The development of anaerobic wastewater treatment. Proc. Sem.Workshop, Anaerobic Treatment of Sewage, Univ. of Massachusetts, pp. 17-54.

Jewell W.J. ( 1987) Anaerobic sewage treatment. Envir. Sci. Technol. 21, 14-21.

Lettinga G., Roersma R. and Grin P. (1985) Anaerobic treatment of raw domestic sewage at ambient temperature, using granular bed UASB reactor. Biotechnol. Bioengng 25, 1701-1723.

Rittman B.E. and Baskin D.E. (1985) Theoretical and modeling aspects of anaerobic treatment of sewage. Proc.Sem./Workshop, Anaerobic Treatment of Sewage. Univ. of Massachusetts, pp. 55-94.

Sanz I. (1987) Depuración anaerobia de agua residual urbana. Tesis doctoral, Univ. de Valladolid.

Schllinkhout A., Lettinga G., van Velsen L., Kooijmants J.L. and Rodriguez G. (1985) The application of UASB reactor for the direct treatment of domestic of UASB reactor for the direct treatment of domestic wastewater under tropical conditions. Proc. Sem. Workshop Anaerobic Treatment of Sewage, Univ. of  Massachusetts. pp. 259-277

Switzenbaum M.S. (1978) The anaerobic attached film expanded bed reactor for the treatment of dilute organic wastes. Ph.D dissertation Agricultural Engineering Department, Cornell University.

Yoda M., Hatton M. and Miyaji Y. (1985) Treatment of municipal wastewater by anaerobic fluidized bed behaviour of organic suspeded solids in anaerobic reactor Proc.Sem/Workshop.Anaerobic Treatment of Sewage. Univ. of Massachusetts, pp.161-196.

Date upgrated Mar/23/99
Coments to Webmaster
[  Homepage CEPIS   ]