Full-Scale application of the UASB technology for sewage treatment

By :  A. Schellingkout* and C.J. Collazos**

Tomado de: Water Science and Technology


In recent years, many pilot-scale investigations were reported on the application of the UASB for sewage treatment. Information about large-scale experience is scarce, however. This paper describes the experience obtained with the design, construction and initial operation of experience obtained with the design, construction and initial operation of a 160,000 PE (31,000 m3/day, 8 MGD) plant in Colombia, consisting of UASB reactors and a facultative pond in series. It describes the possibilities and limitations of the use of prefab concrete as a building material for UASB reactors. The real cost of erection of the plant was USD 17 per capita; the cost for operation and maintenance amounted to USD 1.50 per capita.


Sewage; wastewater; anaerobic treatment; UASB; lagoons; sludge; design; construction; costs


Since 1982, in several parts of the world experiments have been undertaken to assess the applicability of the UASB system - the Upflow Anaerobic Sludge Blanket system, described in detail by several authors (Lettinga et al., 1980, Lettinga and Hulshoff Pol, 1986) for the direct treatment of sewage in warm climates. Experiences in Brazil (Souza, 1986). Indonesia (National Institute for Public Health et al. 1988), India (Siddigi, 1990) and Colombia (Schellinkhout et al., 1985) showed that a BOD reduction of 75% is feasible under tropical conditions and somewhat lower in colder areas (Vieira and Souza, 1986). It was also indicated that considerable cost reductions could be achieved in comparison to other treatment systems. In 1990, the thus far largest sewage treatment plant based on UASB technology was built in the Colombian city of Bucaramanga (600 000 inhab, 900 m + sea level). Smaller plants are being operated in Kanpur, India (5 000 m3/day) and Cali, Colombia (20,000 population equivalent (PE) ).

Bucaramanga, as many cities in South America, deals with a rapid growth of the population that has resulted in a dramatic deterioration of the many littler streams passing its urban area. To cope with the increased pollution, in 1983 an integrated plan for restoration of the streams was presented. A major component of the plan was the connection of the sewers to interceptors and the erection of sewage treatment plants (STP's). At this moment, 90 % of the interceptor system is in use and the biggest of the STP's is in operation. Having completed the first stage of the sanitation plan, Bucaramanga is the most advanced city in sanitation of Colombia.

Pilot plant results

Over the years 1983 - 1987 a pilot plant was operated on raw sewage. The plant consisted of two parallel primary treatment units, an anaerobic pond and a UASB reactor (volume 35 m3 each), and two facultative ponds of 107 m2 as a secondary treatment. The results (Table 1, Fig. 1) indicated that applying a UASB reactor as a primary treatment step would reduce the total Hydraulic Retention Time (HRT) by a factor of 4 to 5 in comparison to a system with only ponds, while a better effluent quality could be obtained as well (Jakma et al., 1986; Schellinkhout et al., 1988). These results served as an argument to select the UASB technology for the full-scale STP.

The removal of pathogens was 85%, which is insignificant, confirming results obtained in Cali (Haskoning et.al., 1985). Nutrient removal and gas production were not monitored.

                                    Table 1  Performance Primary Treatment Pilot Plant 1983 - 1987

                                       system                                  anaerobic                                  UASB
                                                                                    lagoon                                     reactor


                                     HRT                          [h]               21                                    19            5

                            COD effluent                  [g/m3]            210                                    166         145
                            BOD effluent                  [g/m3]              70                                      57           39

                            COD removal                   [%]                60                                     72            66
                            BOD removal                   [%]                67                                      79           80
                            TSS removal                     [%]                69                                     70           69

                            BOD: biochemical oxygen demand; COD: chemical oxygen demand; TSS: total suspended solids

gra1.jpg (23675 bytes)


Design of full-scale plant

Based on the results of the pilot plant, design criteria for full-scale STP's were developed, as presented in table 2.

                                                     Table   2  Design Criteria Full-Scale Plants


                                        average HRT                                                           [h]          5.2
                                        minimum HRT                                                          [h]          3.0
                                        depth                                                                       [m]         4.0
                                        expected BOD removal                                           [%]          70
                                        max. COD loading rate                                 [kg/m3.d}]
                                        BOD loading rate                                          [kg/(ha.d)]         550  
                                        depth                                                                       [m]          1.5
                                        expected BOD removal                                            [%]          50
                                        average HRT                                                            [h]           30
                              total expected BOD removal                                               [%]          85

The first plant, designed for a capacity of 160,000 PE (31,000 m3/day, 8 MGD), was completed in September 1990.
The composition and dimensions of the plant are given below (see also Fig. 2):

-  flow restriction device
-  coarse and fine screens, 2 parallel units
-  grit chamber, 3 parallel units
-  UASB reactors, 2 units of 3,300 m3 each                  
-  facultative lagoon of 2.7 ha

The excess sludge from the UASB reactors is de-watered on 48 drying beds of 120 m2 each. The gas is stored in a gasholder and flared.

gra2.jpg (30499 bytes)

Materials and general design

The tank of the UASB reactor of the pilot plant was made of gabions, sealed with cement. The inside structures (gas-liquid-solids (GLS) separators) were made of galvanised steel. This construction method was not quite applicable in the full-scale plant, since corrosion of the metal parts would occur and the tank would be insufficiently water tight. Possible  materials that were considered for the UASB tanks included terracrete (a mixture of cement and soils, poured on site). For the internal parts polyester (GRP), in situ poured concrete, prefab concrete and stainless steel were viable possibilities. After performing a cost comparison, concrete tanks with prefab concrete inside structures were selected.

Special attention was given to guarantee a smooth surface of the concrete. Additives were applied to reduce the susceptibility of the concrete for corrosion and its permeability for gas.

The UASB tanks are equipped with one inlet per 2.9m2 , resulting in 288 inlets per tank, which is in the range recommended by several authors (Souza, 1986; Lettinga and Hulshoff Pol, 1986). To obtain such a minute distribution, the incoming wastewater is distributed over three splitter boxes in series, the first consisting of one box with four outlets, the second step is formed by four boxes of 9 outlets each and the 36 final boxes have 16 outlets each. The design of the final influent splitter boxes is such that dragging of air by the influent into the reactor, which would be disatrous for the anaerobic biomass (Vieira, 1988), is prevented.

The angles of the gas collectors is 52°. This resulted in a settling compartment only 1.40 m deep, thus leaving 2.60 m for the digestion compartment. Slopes more gradual than 50° are suspected to accumulate sludge on their surfaces, especially when applying a relatively rough material like concrete. Steeper angles would lead to unnecessarily deep settlers.

More details on the design were presented by others (Collazos, 1990).

gra3.jpg (100566 bytes)       

Scum control

In the Bucaramanga pilot plant it was experienced that heavy scum formation occurred in the (open) gas compartment of the UASB reactors, while in the closed gas chambers of the Cali pilot plant only a slight accumulation was noticed (Schellinkhout, personal observation, unpublished). Based on these observations, inthe full-scale plant, a two-step approach was chosen.
First, it was assumed that if the sum was frequently mixed with the reactor contents and protected from desiccation, degradation of its grease, proteins and solids would eventually occur. Second, in case this mixing would happen to be insufficiently effective, the gas collectors should be accessible to remove the scum. Therefore, a gas injection system was installed on each gas collector to allow periodic injection of gas just below the scum, at both extremes of the gas collectors removable covers were placed. The suggerstion fo sprinklers (Lettinga and Hulshoff Pol, 1986) was found not feasible, since sprinklers are susually only effective for the abatement of foam rather than scum, and they could hinder a free passage from one extreme to the other of the collector dome.

Scum formed in the settling compartment is allowed to leave the reactors and is eventually collected in the inlet distribution channel of the lagoons, the latter thus functioning as an Integrated Scum Accumulation Device (ISAD), from where it is removed periodically.

The gas collectors prove to be quite gas tight; the porosity of the concrete is almost nil. The concrete GLS separators show some problems though, since escape of gas takes place through the joints, causing turbulence in thje settler.

Behaviour and experiences during start up

Loading and performace

From October 30, 1990, one reactor was operated at full load, viz. HRT of 5.2 hours during approximately 16 hours in the day and at lower loadings at night. The second reactor was taken into operation on January 12, 1991, at the same loading rate. The start up was performed without inoculum, since experience had demonstrated that was well possible (Schellinkhout et al., 1985) and since no seed sludge was available.

The first four months of operation produced results that are still below the ones obtained in the pilot plant. The BOD removal efficiency of the UASB reactors varied form 17 to 48%, the figure for COD between 18 and 44%.
The laggon, on the other hand, demostrated a great flexibility to the periodical overloadings, so a final effluent with a BOD between 48 - 66 g/m3 and a COD between 160 and 210 g/m3 could still be achieved.

These somewhat unexpected results should be attributed to sludge wash-out, caused by turbulence in the settling compartments. The origins of this phenomenon will be described below.

Coliorms and nutrients were not analyzed; neither methanogenic activity nor gas production were determined, unfortunately, since appropriate equipment was not available yet. It is expected that these activities can start in March 1991; the results will be published later.

Performance of the GLS separator

The gas-liquid-solids (GLS) separator is the key element of the UASB reactor, together with the influent distribution system. When using prefab concrete, two aspects resulted to cause difficulties: the thickness of the material and the joints between elements.

The joints between the elements of the upper dome of the gas collector were sealed with 15 mm neoprene sheets. This proved to be adequate. The lower elements, however, were not sealed in this way, and gas bubbles seeping through the joints passed into the settling compartment, causing heavy turbulence and moving sludge to the outlet weirs. In March, these joints were sealed with a kit and the consult leading the gas to the upper dome was improved.

In the design of the GLS separator, an overlap in the aperture between the digesting and the settling comparment of 15 cm was provided. Nevertheless, gas bubbles were noted to pass thorough the aperture into the settler, withdrawing sludge to the effluent weirs. Ths phenomenon proved to be caused by the thickness of 10 cm of the GLS separator plates used. Small gas bubbles reaching the rim of the upper plate (see Fig. 4) are guided perfectly into the gas collector. Large bubbles, however, cannot pass quickly due to the obtuse angle of the rim. The bubble will break up and part of it will pass into the settler. This problem will be solved by installing an extra sheet, directing the gas away from the aperture.

Odur nuisance

A major problem of all treatment plants, including anaerobic plants, is the emission of odour. In the Bucaramanga plant, especially during start up when the UASB effluent was not well stabilized, odour emission occurred.
Also, the biogas when disposed of into the air represented much odour. To reduce smell nuisance it was experienced that flaring the (excess) biogas was effective. In future plants that have to be built near dwelling areas of Bucaramanga, preliminary systems, the influent distribution systems and the effluent channels of UASB's should be covered and ventilated, the extracted air should be treated in a biofilter.

It is experienced that in the sunny, warm climate of Bucaramanga, sludge drying beds hardly emit any small, but during the filling procedure some emission occurs. Therefore, the beds will not be covered (an action that would hamper the de-watering process seriously). Still it is recommended to locate sludge dewatering on the part of the site away from the houses.

Sludge handling

After four months of operation, no excess sludge was produced. Nevertheless, some experiments were undertaken with drying some sludge samples on the beds. It was experienced that when applying a layer of 25 cm of sludge of 8% dry solids, after 7 days the sludge volume had reduced by more than 80% and the sludge was sufficiently solid to be handled by spade, even in a period in which it rained heavily during 4 nights.

Cost and personnel

The cost of the whole project, including extension of the sewage intercepter system by 30 km, is USD 5.8 million. The plant itself costs USD 2.5 million or approximately USD 17 per capita. One UASB reactor costs USD 600,000, or USD 181 per m3 of reactor volume. These are real costs under Colombian conditions, where equipment is more expensive than in Europe or North America, but labour considerably cheaper. These actual costs for the UASB reactor are even lower than presented by others; Souza (1986) gave an estimate of USD 280 - 350 per m3 for Brazil, Vieira and Souza (1986) reported real cost for a 120 m3 steel reactor of USD 300 per m3.

The plant is operated by a staff of 4 operators, 4 assistants and 1 head of the plant. Running costs are USD 1.50 per annum per capita for personal, including lab staff for water analysis. There are some extra costs for energy for illumination and for renning the small maintenance equipment.


The expectation that anaerobic technology, and more specifically the UASB concept, would contribute to a cost-effective and efficiente solution of environmental problems in tropical climates has only been reinforced by the experiences of the full-scale plant in Bucaramanga, in spite of some initial problems. At relatively low costs a removal of 85% on a BOD basis will be achieved. The CDMB therefore is determined to continue its sanitation task, applying UASB technology as a primary treatment. Other cities in Colombia are preparing similar solutions.

Prefab concrets is an attractive material for constructing UASB reactors in many South American Countries, since it is durable, widely available and relatively cheap. The Bucaramanga experience has demostrated, however, that this material:

=  requires more skill than may be available with local contractors
=  causes difficulties in levelling of channels and gutters
=  requires special additives to be sufficiently gas tight
=  needs special attention to join elements gas tight
=  requires care and experience in upscaled designs.


The project, including the sewage treatment plant, has been designed, built and operated by the Corporación de Defensa de la Meseta de Bucaramanga. The project was funded by own resuorces and loans of the Colombian government and of the Inter American Development Bank. Technical assistance in the disign, construction and start up was provided by DHV Consultants and by Paques BV of the Netherlands, sponsored by the Netherlands Ministry of Foreign Affairs (DGIS).


Collazos, C.J. (1990). Tratamiento de Aguas Residuales Domésticas en Bucaramanga (Colombia) Mediante Reactores UASB y Lagunas Facultativas. Taller y Conferencia Sobre Tratamiento Anaeróbio de Aguas Residuales en América Latina., UNAM, México DF.

Haskoning, Agricultural University, Universidad del Valle and INCOL (1985). Anaerobic Treatment and Re-Use of Domestic Wastewater. Final Report Pilot plant Study, Cali - Colombia. Ministry of Foreign Affairs (DGIS/DPO/OT), The Hague, the Netherlahds.

Jakma, F.F.G.M., Collazos, C.J. and Schellinkhout, A. (1987). Anaerobic System for Sewage Treatment in (Sub)tropical Countries Successful. Hp 20, (1) 2-5, Rijswijk, the Netherlands (in Dutch, English summary). Lettinga, G., Velsen, A.F.M. van, Hobma, S., Zeeuw, W.J. de and Klapwijk, A. (1980). Use of the Upflow Sludge Blanket (USB) Reactor Concept for Biological Wastewater Treatment. Biotech. and Bioeng. 22, 699-734.

Lettinga, G. and Hulshof f Pol, L (1986) . Advanced Reactor Design, Operation and Economy. Wat. Sci. Tech. 18, (12) 99-108.

National Institute for Public Health and Environmental Hygiene (RIVM), Agricultural University (LUW) and General Hospital St.Borromeus (1988). Upflow Anaerobic Sludge Blanket low cost sanitation project in BandunglIndonesia. Ministry of Foreign Affairs, DGIS/DPO/OT, the Hague, Netherlands.

Schellinkhout, A., Lettinga, G., Velsen, A.F.M. van, Louwe Kooijmans, J. and Rodríguez, G. (1985). The Application of the UASB-Reactor for the Direct Treatment of Domestic Waste Water Under Tropical Conditions. In: Proceedings of the SeminarlWorkshop on Anaerobic Treatment of Sewage, M.S. Switzenbaum (Ed)., University of Massachusetts, Amherst, MA, USA. Schellinkhout, A. , Jakma, F. F. G. M. and Forero, G. (1988) . Sewage Treatment the Anaerobic Way Advancing in Colombia. Sth International Symposium on AD, Posters Papers, Tilche and Rozzi (Ed)., Monduzzi Editori, Bologna, Italy.

Siddiqi, R.H. (1990). Treatment of Municipal Wastewater by UASB Reactor, Performance of 5 MLD Plant at Kanpur. Singular publication. Dept. of Civil Engineering, Aligarh Muslim University, Aligarh, India.

Souza, M.E. (1986). Criteria for the Utilization, Design and Operation of UASB Reactors. Wat. Sci. Tech. 18, (12) 55-69.

Vieira, S.M.M. (1988). Anaerobic Treatment of Domestic Sewage in Brazil Research Results on Full-Scale Experience. 5th International Symposium on AD, Posters Papers, Tilche and Rdzzi (Ed)., Monduzzi Editori, Bologna, Italy.

Vieira, S.M.M. and Souza, M.E. (1986). Development of Technology for the Use of the UASB Reactor in Domestic Sewage Treatment. Wat. Sci. Tech. 18, (12) 109~121.

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