Performance evaluation of a UASB activated sludge system treating municipal wastewater

M. von Sperling*, V.H. Freire and C.A. de Lemos Chernicharo

*Department of Sanitary and Environmental Engineering, Federal University of Minas Gerais Av. Contorno 842 7o andar, 30110-060 Belo Horizonte, Brazil (E-mail: marcos@desa.ufmg.br) Abstract Recent research has indicated the advantages of combining anaerobic and aerobic processes for the treatment of municipal wastewater, especially for warm-climate countries. Although this conguration is seen as an economical alternative, is has not been investigated in sufcient detail on a worldwide basis. This work presents the results of the monitoring of a pilot-scale plant comprising of an UASB reactor followed by an activated sludge system, treating actual municipal wastewater from a large city in Brazil. The plant was intensively monitored and operated for 261 days, divided into ve different phases, working with constant and variable inows. The plant showed good COD removal, with efciencies ranging from 69% to 84% for the UASB reactor, from 43% to 56% for the activated sludge system only and from 85% to 93% for the overall system. The nal efuent suspended solids concentration was very low, with averages ranging from 13 to 18 mg/l in the typical phases of the research. Based on the very good overall performance of the system, it is believed that it is a better alternative for warm-climate countries than the conventional activated sludge system, especially considering the total low hydraulic detention time (4.0 h UASB; 2.8 h aerobic reactor; 1.1 h nal clarier), the savings in energy consumption, the absence of primary sludge and the possibility of thickening and digesting the aerobic excess sludge in the UASB reactor itself. Keywords Municipal wastewater; anaerobic processes; aerobic processes; UASB; activated sludge; post-treatment

Water Science and Technology Vol 43 No 11 pp 323328 2001 IWA Publishing and the authors

Introduction

The activated sludge process has been widely applied for the treatment of domestic and industrial wastewaters, due to its high efficiency, operational flexibility, possibility for nutrient removal, among other advantages. However, there are some disadvantages, namely: high mechanisation level, high construction and operational costs, sophisticated operation and the need for treating a substantial amount of sludge (von Sperling, 1997). In the past few years, the high-rate anaerobic processes, especially the UASB (Upflow Anaerobic Sludge Blanket) reactor has shown to be a technology capable of overcoming some of the disadvantages of the mechanised aerobic systems, especially because of the absence of energy consumption and lower excess sludge generation. Nevertheless, the treated effluent is usually unable to comply with most existing discharge standards (Chernicharo, 1997). Aiming at balancing the advantages and disadvantages of both systems, recent research has indicated the benefits in combining the processes, with the anaerobic stage being followed by the aerobic stage. The advantages of the combination are: (a) lower energy consumption; (b) lower chemical consumption for dewatering; (c) less sludge to be disposed of; (d) less equipment required; (e) higher operational simplicity (von Sperling and Chernicharo, 1998). Figure 1 presents the flowsheets of a conventional activated sludge plant and the UASBactivated sludge system. In either case, the sludge lines are also taken into account. In the alternative of UASB-activated sludge, an important feature of the system is the return of the aerobic excess sludge to the anaerobic reactor, where the solids undergo stabilisation, considerably simplifying the plant flowsheet and the sludge processing. This approach of

323

M. von Sperling et al. Figure 1 Comparison between the owsheets of a conventional activated sludge system and an UASBactivated sludge system
324

returning the activated sludge wastage to the UASB reactor has been tested experimentally by Souza and Foresti (1996) and Charleston et al. (1996) and showed to give good results. In spite of the various advantages of the combined interaction of the UASB and the activated sludge systems, it has been the subject of little investigation for municipal wastewater treatment, even at a worldwide level. Some of the research undertaken is described below. Colleti et al. (1997) investigated under laboratory scale an activated sludge plant, treating the effluent from a compartmentalised UASB reactor, which received municipal wastewater. The main objective was to determine the stoichiometric and kinetic coefficients of the activated sludge process working as a post-treatment of the UASB reactor. The system achieved removal efficiencies of 95% (BOD) and 88% (COD). Souza and Foresti (1996) worked with a system composed of a UASB reactor followed by two aerobic Sequencing Batch Reactors in parallel. The system received synthetic wastewater. The removal efficiencies averaged 95% for COD and 85% for TKN. Silva et al. (1995) investigated a pilot-scale UASB-activated sludge system, receiving an influent comprised of approximately 90% of industrial wastewater flow. The UASB

reactor achieved removal efficiencies around 70% for COD and 80% for BOD. The activated sludge process was predominantly very unstable and subjected to filamentous bulking. In the more stable periods the removal efficiencies for the aerobic stage alone averaged 42% for COD and 63% for BOD. The instability was attributed to the high percentage of industrial wastewater flow. Passing et al. (1999) presented one-year operational data from a full-scale UASBactivated sludge plant treating domestic wastewater for a design population of 100,000 inhabitants in Brazil. The average effluent concentrations were 174 mg/l (COD) and 47 mg/l (BOD). The plant had not yet needed wastage of the aerobic excess sludge. The present paper investigates the performance of a pilot-scale UASB-activated sludge plant operated for a period of 261 days. The plant received actual municipal wastewater pumped from the main intercepting line of a large city in Brazil (Belo Horizonte, 2.0 million inhabitants). The plant is situated in the Pilot Installations Laboratory of the Department of Sanitary and Environmental Engineering, Federal University of Minas Gerais, Belo Horizonte.
Methodology

M. von Sperling et al.

The UASB reactor has a volume of 416 L, feeding, besides the activated sludge system, other post-treatment lines. The activated sludge reactor has a volume of 23 L. The average frequency of sampling was 2 to 3 times per week. The aeration was by diffused air, and the DO level was kept around 2.0 mg/l. The operating period of 261 days was divided into the following five phases (see also Table 1): Phase I: Unseeded start-up of the activated sludge system (UASB reactor had been already previously started-up). Constant inflow. Phase II: Reduction of the Hydraulic Retention Time (HRT) in all units. Constant inflow. Phase III: Same average HRT as in previous phase. Variable inflow, simulating a typical diurnal pattern (peristaltic pump adjusted to follow a polynomial curve). Phase IV: Reduction of the HRT (volume) of the settler in the UASB reactor. This reduction was to force a lowering in the COD removal efficiency of the UASB reactor, aiming at exciting the activated sludge system more. Variable inflow (same flow and pattern as in previous phase). Phase V: 20% of the inflow by-passing the UASB reactor and entering the activated sludge system, in order to further excite the activated sludge system. Variable inflow (same flow and pattern as in previous phase).
Table 1 Main characteristics of the ve operational phases
Average hydraulic retention time (h)

Duration Phase (days)

UASB reactor

Aerobic reactor

Final settler

Total HRT Remarks

I II III IV V

62 48 69 52 30

6.0 4.0 4.0 4.0 4.0

3.7 2.8 2.8 2.8 2.8

1.5 1.1 1.1 1.1 1.1

11.2 7.9 7.9 7.9 7.9

Constant inow Constant inow Variable inow Reduction in volume of UASB settler 20% raw sewage by-pass to activated sludge

Results and discussion

The summary statistics of the results from each operational phase are presented in Table 2. The average COD removal efficiencies are summarised in Table 3. Figure 2 presents the

325

Table 2 Summary statistics from the ve operational phases

Operational phases

Phase I

Phase II

Phase III

Phase IV

Phase V

Parameter

Statistics

In

InAS

Ef

In

InAS

Ef

In

InAS

Ef

In

InAS

Ef

In

InAS

Ef

M. von Sperling et al.

326

COD (mg/L) COD ltered (mg/L) SS (mg/L) VSS (mg/L) MLSS (mg/L) MLVSS (mg/L)

average stand.dev number average stand.dev number average stand.dev number average stand.dev number average stand.dev number average stand.dev number

958 881 14 269 140 16 443 385 15 329 278 14

137 63 15 64 46 16 25 10 16 20 7 16

105 66 16 60 37 16 23 17 16 17 12 16 562 401 7 430 363 7

734 542 7 255 179 7 251 379 8 205 291 8

114 47 8 58 30 8 32 12 8 28 10 8

50 555 16 239 8 17 40 262 22 112 8 6 8 12 6 8 1414 317 8 1235 280 8 16 67 18 99 48 18 13 115

85 18 18 53 15 18 21 5 20 19 4 20

56 386 119 25 174 19 16 18 12 19 10 19 1415 528 18 1218 441 18 14 83 15 44 15 38 15 30 206 22 15 62 24 15 34 11 15 31 13 15

58 557 26 135 15 17 15 10 15 9 15 1335 469 13 1158 418 13 13 85 13 58 13 51 13 30 243

180 69 13 66 14 12 53 26 12 46 22 12

128 40 13 63 17 13 99 59 13 48 24 13 1055 656 13 928 573 13

15 120

18 141

14 107

16 129

In: inuent to UASB; InAS: inuent to activated sludge (efuent from UASB); Ef: nal efuent, from activated sludge

Table 3 Average COD removal efciencies in four operational phases

Phase UASB Activated sludge UASB activated sludge

II III IV V

84 % 85 % 69 % 68 %

56 % 43 % 51 % 50 %

93 % 91 % 85%

box-and-whisker plot of the final effluent COD distribution in the five operational phases. The box-and-whisker plot shows the minimum and maximum values (whiskers), together with the 25, 50 and 75 percentiles (box). The stepwise reduction in the average COD and SS concentrations, from the influent, effluent from UASB and final effluent are shown in Figures 3 and 4, respectively. Figure 5 plots the time series of the MLVSS concentration in the reactor of the activated sludge system. Other parameters have been monitored and analysed, being covered in detail in Freire (1999). BOD data has been collected, but the

Figure 2 Box-and-whisker plot of the distribution of the nal efuent COD in four operational phases

M. von Sperling et al.

Figure 3 Average COD values in the various stages of the treatment line, according to the operational phase

Figure 4 Average SS values in the various stages of the treatment line, according to the operational phase

Figure 5 Time series of the MLVSS concentration in the reactor of the activated sludge system

number of samples is insufficient to support the same level of interpretation. Due to practical difficulties, sludge settleability tests and microbiological analyses have not been undertaken. In Figures 2, 3 and 4, only Phases II to V have been considered. Phase I was not included, because it did not represent a typical operation of the plant, being mainly associated with the specific condition of the start-up of the activated sludge plant. Based on the tables, figures and additional analysis, the following comments are made: The average final effluent COD concentrations in Phases II, III and IV were around 55 mg/l, well below the discharge standard of 90 mg/l, set out by the State Environmental Agency. The Analysis of Variance undertaken with the final effluent COD indicated that there were no significant differences between Phases II, III and IV (Phase V is untypical,

327

because of the raw sewage by-pass to the activated sludge). An important consequence of this analysis is the conclusion that the variable inflow did not affect the performance of the activated sludge process. The reduction of the HRT in the UASB settler (Phase IV) resulted in a decrease in the COD removal efciency of the UASB reactor, from 85% to 69%. However, the activated sludge was able to cope with the load increase, and the nal efuent quality was not affected (as shown by the Analysis of Variance mentioned above). Phase I showed lower COD removal efciencies in the activated sludge system, resulting from the acclimation of the biomass during the start-up period. The nal efuent concentrations of SS during Phases II, III and IV were able to satisfy commonly applied international discharge standards. The higher values of SS in the nal efuent during Phase V, which was characterised by the by-pass of 20% of raw sewage to the activated sludge, was due to sludge bulking. The bulking was only occasional during the other phases, but was systematic in Phase V. The MLVSS concentration in the reactor of the activated sludge system had average values around 1000 mg/l, but was highly variable during the operational phases. It took approximately 60 days for the biomass to reach the average value. This average value of 1000 mg/l was half of the design value adopted for the plant (2000 mg/l). However, the F/M ratio in Phases II, III and IV was within typical design values of conventional activated sludge plants.

M. von Sperling et al.

328

Conclusions

In the more typical operational phases (II, III and IV), the overall performance was very good, as seen by the following results: Average COD removal efficiencies between 85% and 93%. Average final effluent COD concentrations between 50 and 58 mg/l. Average final effluent SS concentrations between 13 and 18 mg/l. The combined system (UASB - activated sludge) showed to be a very good alternative for the treatment of municipal wastewaters, based on the performance of the system and the compactness of the treatment units (total hydraulic retention time of 7.9 h, computing the UASB reactor, activated sludge reactor and final clarifier).
References
Charleston, L.O., Robles, A.N. and Bohrquez, S.S. (1996). Efectos de lodos activados de purga sobre el funcionamiento de un reactor UASB piloto y las caractersticas del lecho de lodo granular. XXV Interamerican Congress on Sanitary and Environmental Engineering (Proceedings). Mexico, 37 November 1996 (in Spanish). Chernicharo, C.A.L. (1997). Princpios do Tratamento Biolgico de guas Residurias Reatores Anaerbios. Vol 5. Departamento de Engenharia Sanitria e Ambiental UFMG (in Portuguese). Coletti, F.J., Povinelli, J. and Daniel, L.A. (1997). Ps-tratamento por lodos ativados de efluentes provenientes de processos anaerbios de tratamento de esgoto sanitrio; Determinao de constantes cinticas. Anais: 19o Congresso Brasileiro de Engenharia Sanitria e Ambiental, Foz do Iguau, set/97 (in Portuguese). Freire, V.H. (1999). Avaliao do desempenho de um sistema combinado UASB-lodos ativados no tratamento de efluentes sanitrios e modelagem do sistema de lodos ativados. MSc Dissertation, Federal University of Minas Gerais, Brazil (in Portuguese). Passing, F.H., Vilela, L.C.H. and Ferreira, O.P. (1999). ETE Piracicamirim - nova concepo de sistema de tratamento de esgotos sanitrios partida, operao e monitoramento de desempenho. Anais: 20 Congresso Brasileiro de Engenharia Sanitria e Ambiental, Rio de Janeiro, Maio/99 (in Portuguese). Souza, J.T. and Foresti, E. (1996). Domestic sewage treatment in an upflow anaerobic sludge blanket sequencing batch reactor system. Wat. Sci. Tech., 33(3), 7384. Silva, S.M.C.P. da, Alm, S.P.; Jr., A.S.G. (1995). Avaliao do Sistema Reator UASB e Processo de Lodos Ativados para Tratamento de Esgotos Sanitrios com Elevada parcela de Contribuio Industrial. Anais: 18 Congresso Brasileiro de Engenharia Sanitria e Ambiental, Salvador, Bahia, set/95 (in Portuguese). Von Sperling, M. (1997). Princpios do tratamento biolgico de guas residurias. 4. Lodos ativados. Departamento de Engenharia Sanitria e Ambiental UFMG. 415 p. (in Portuguese). Von Sperling, M. and Chernicharo, C.A.L. (1998). Selection of wastewater treatment systems in urban areas. Comparison between conventional aerobic systems (activated sludge) and anaerobic-aerobic systems (UASBactivated sludge). Anais, V Taller Seminario Latinoamericano de Digestion Anaerobia, Via del Mar, 2730 Out 1998.