CONSTRUCTED WETLANDS FOR TREATMENT OF BIOGAS PLANT EFFLUENTS M. Blumberg Ingenieurbüro Blumberg, Bovenden, Germany Contact: Michael Blumberg, MSc. (Agriculture), Ingenieurbüro Blumberg, Gänsemarkt 10, D-37120 Bovenden, Germany, Tel.: + 49 5593 93 77 50, Email: contact@blumberg-engineers.de, www.blumberg-engineers.com

EXECUTIVE SUMMARY Different combinations of microflotation, anaerobic treatment (EGSB) and constructed wetlands were tested in a pilot plant to treat wastewater of the starch industry within the scope of an ongoing Vietnamese-German joint research project (Pick et al., 2011) funded by the German Federal Ministry of Education and Research (BMBF). In this paper the results of a treatment cascade consisting of the main elements microflotation, biogas plant (EGSB reactor) and a constructed wetland are presented. A second line using a simplified process consisting of a single-step pretreatment and a two-stage biological treatment via constructed wetlands (vertical subsurface flow), adapted to the possibilities of small and medium-sized enterprises, is also described. A removal efficiency of 99 % was obtained with respect to chemical oxygen demand (COD). As another example of using constructed wetlands to treat high-strength industrial wastewater from the food industry, the pilot plant has demonstrated its general suitability as final treatment stage downstream of a biogas plant (or without conventional stages) for primary and secondary treatment. Key words: Constructed wetlands, treatment of biogas plant effluents, tapioca starch industrial wastewater

1 Introduction In Asian countries like Vietnam, China, Thailand, India and others the manioc plant is used for the production of tapioca starch which is used e.g. in the food industry for monosodium glutamate, as fuel additive (alcohol), for biopolymeres, for soap, medicines, cosmetics and so on. In China the area for cultivation of tapioca increased in 2007 from 500,000 ha to 800,000 ha.

Figure 1 Tapioca farmer in Vietnam

Figure 2 Tapioca tubers

Figure 3 Starch production in China

The processing of 4 t of manioc tubers (see Figure 4) into 1 t of tapioca starch produces 12 to 20 m³ of wastewater (see Figure 5).

Figure 4 Tapioca starch processing scheme

Figure 5 Typical concentration of wastewater from tapioca starch production

In Vietnam the cultivable area for tapioca increased by 60% in 5 years (2001 - 2006) due to the rising global demand for starch products (Kato, 2007). The lack of adequate treatment of industrial and domestic wastewater resulted in a dramatic water pollution in the catchment area of the river Dong Nai and caused massive environmental damage. A medium-sized company with a tapioca processing capacity of 100 tons/d produces a wastewater volume of about 2.000 m3 per day, which is attributable to starch extraction, equipment cleaning and root washing. Wastewater from the production of tapioca starch contains high organic concentrations and has a chemical oxygen demand (COD) of 11.000 to 24.000 (mg/l) (Annachhatre & Amatya, 2000). In addition to their organic loads these effluents are a significant risk to the environment because of the content of toxic cyanogenic glycoside of plant origin (cassava, manioc). It appears that the effective treatment methods that are available today - upflow anaerobic sludge blanket reactor (UASB), or expanded granular sludge bed digestion (EGSB) - are too difficult to handle for the majority of the small family-owned enterprises in Vietnam. To solve the sewage problem, these enterprises usually rely on simple sedimentation and storage lagoons with insufficient cleaning performance. In the rural areas of the Vietnamese Tay Ninh province a great number of businesses is currently in urgent need of an economical low-maintenance wastewater treatment method, in order to obtain the licence for continued operation. A licensable wastewater treatment plant cannot be achieved by simply extending the existing treatment pond systems (Rajbhandari & Annachhatre, 2004). Therefore, two combinations have been tested in addition to the main treatment via cascade microflotation, biogas plant and constructed wetlands, which is presented in more detail by Pick et al. (2011):

• Microflotation downstream of a two-stage constructed wetland.

• Sedimentation pond with floating islands downstream of a two-stage constructed wetland.

The research project was launched in March 2009 and is due to be completed in October 2012 by:

• University of Braunschweig (project coordination)

• University of Ostwestfalen-Lippe

• Institute for Environment and Resources Vietnam National University of HCMC

• Ingenieurbüro Blumberg

• Hager + Elsässer Company

• Enviplan Company

2 Structure of the experimental plant A pilot plant was built in 2010 on the factory premises of Thanh Vinh Company in the Vietnamese province of Tay Ninh (see Figure 6).

Figure 6 Schematic layout plan of the pilot plant

This medium-sized company has a processing capacity of about 500 tons of cassava roots per day, which is associated with a volume of 1500 m3/d of wastewater. The raw sewage has an average chemical oxygen demand (COD) of approximately 10.500 mg/l and a total suspended solids (SS) content of about 1.800 mg/l. The pilot wastewater treatment process consists of two separate lines. Each line is designed for a hydraulic load of 12.5 m3/d of wastewater. One treatment line consists of a microflotation unit (Enviplan® Ingenieurgesellschaft), an EGSB biogas plant (Hager+Elsässer) and a downstream artificial wetland. In the second line raw wastewater enters a sedimentation pond with floating islands followed by a two-stage constructed wetland in series (Blumberg Engineers). The first vertical flow constructed wetland accommodates the high solids load of the primary treated wastewater and operates on the principle of sewage sludge humification. The second wetland basin is charged with the filtered wastewater from the first vertical filter in a tertiary treatment step. Hence, the tested designs differ in the type of pretreatment and hydraulic load. Different marsh plants have been tested under tropical conditions. In the experimental design 1 the microflotation is used as a pretreatment to reduce the suspended solids content and some organic carbon compounds. In the experimental design 2 a sedimentation pond with planted floating islands has been tested for its capacity to reduce settleable solids and the toxic cyanide, without any technical treatment units. A number of first results of these two tests lines from 2011 and 2012 are presented below.

Figure 7 Experimental design 1 with microflotation and biogas plant as pretreatment stages

Sedimentation

pond with floating

islands 95 m³

Constructed

wetland 1

(OKA 2)

416 m²

2.

line

up

to

20

m³/

d

Constructed

wetland 2

(OKA 3)

312 m²

Raw wastewater

after extraction

Microflotation

EGSB

1.

line

up

to

12

m³/

d

Constructed

wetland 1

(OKA 1)

416 m²

Raw wastewater

after extraction

Figure 8 Experimental design 2 with sedimentation pond as pretreatment stage

Sludge Biogas

Sludge

The main elements of the pilot plants downstream of the buffering tank and neutralisation stage are a microflotation unit (see Figure 9) and an EGSB reactor (see Figure 11), both imported from Germany, container based (see Figure 10).

Figure 9 Principal structure of microflotation

Figure 11 Schematic diagram of EGSB reactor

Figure 10 Container based microflotation and EGSB

With its high load of organic substances, the sewage from tapioca processing offers an opportunity to use the cleaning process for the generation of power. The required technology is provided in the form of an EGSB reactor of the company Hager+Elsässer. EGSB reactors have already proven their suitability for the treatment of this kind of wastewater in other Asian countries. The biogas generated in the reactor can be used directly on site for the generation of heat for the starch production. The innovative concept within the framework of the project is the combination of the anaerobic stage with an upstream microflotation unit and a downstream eco-technical treatment system. To ensure that the run-off water quality is good enough to permit discharge into a river, the fermentation residues of the EGSB reactor must pass through an aerobic stage for further biological decomposition of organic substances. To this end, the biogas system has a downstream connection to a vertical subsurface flow constructed wetland system for final cleaning (see Figures 12 and 13, OKA 1).

Figure 12 Constructed wetlands . schematic cross section (realized as OKA 1, 2 and 3)

Figure 13 Layout plan of the constructed wetlands in Tay Ninh

The specific problem regarding this particular sewage is the undissolved content of fibrous cellulose components of the manioc roots and starch. In order to limit the associated colmation risks for the reed-covered soil filters, a very low hydraulic load of 30 l/m² x d has been specified, which accounts for a required surface area of the filter basin of 416 m² (given a projected daytime wastewater volume for the cleaning cascade of Qd = 12.5 m³). By means of concrete fiber slabs, the marsh plant pond (OKA 1) was divided into four partial ponds, each with a surface area of 104 m². These partial ponds are intermittently charged with sewage via a time-controlled pump station and a superterrenean post-mounted distribution system (see Figure 14). Alternating operation (in a rotational cycle, one of the four ponds is not charged for a few days) is also possible.

Figure 14 Distribution pipes for artificial wetlands

The constructed wetland cells OKA 2 and 3 are very similar in their design and used to treat the raw wastewater without technical pretreatment. The construction work (see Figure 16) was carried out in the years 2010/2011. The basins for the sedimentation pond and the constructed wetlands were sealed with foil. The filters are designed as shown below:

The sand was washed in order to obtain a grain size distribution which complies with German standards (FLL/IÖV 2008). A post-mounted distribution system with perforated PVC pipes was installed on the subsurface flow constructed wetlands. After the ponds had been filled with the filter substrates, a total of 3,750 plants was taken from natural habitats in Vietnam near Ho-Chi-Minh-City and transported to the Tay Ninh province for re-planting. Wastewater feed and recirculation including the control is provided by six pump stations (see Figure 6 - Layout diagram). The trial runs for the main test cascade Line 1 (see Figure 7) started in November 2010 and are due to continue until August 2012.

Wetland 1 and wetland 3 Volume Thickness

Freeboard - 30 cm

Sand 1/4 mm 235 m3 50 cm

Fine gravel 2/5 mm 66 m3 15 cm

Coarse gravel 16/32 mm 142 m3 20 cm

Wetland 2 Volume Thickness

Freeboard - 35 cm

Sand 1/4 mm 73 m3 15 cm

Fine gravel 2/5 mm 126 m3 35 cm

Coarse gravel 16/32 mm 186 m3 30 cm

Figure 15 Filter substrates of the wetland units

Figure 16 Wetland cells OKA 1, 2 and 3 under construction

In treatment variant II (without anaerobic stage) the pre-treatment units for sedimentation, pH-increase and cyanide detoxification are connected to a downstream sedimentation basin in order to reduce the load and colmation risk for the downstream ecotechnical plant which is associated with settleable and filtratable substances. This covered pre-clarification pond is designed to force the run-off water through the densely rooted floating mats made of textile knitted fabric, as shown in the diagram below (see Figure 17, Blumberg, 2009).

Figure 17 Outflow scheme of sedimentation pond with floating islands

Optionally, pre-cleaned sewage from the constructed wetland OKA 2 can be recirculated via a return pipe to another dilution and denitrification cycle during the several-month scheduled downtime of the tapioca company (recirculation, see Figure 6).

Figure 18 Sedimentation pond with floating islands

The flow measurement is provided by magnetic-inductive flow metering in the feed pipe to the EGSB reactor and in the drain pipe of the three soil filters. Tipping counters of the company UFT are used.

Figure 19 Fluid tipper for flow measurement

The feed quantity of inflows from the microflotation unit (bypassing of EGSB reactor and in raw sewage mode) was determined by multiplying the defined pump receiver volume of 50 l with the time-controlled pump intervals (e.g. 10 intervals per day).

3 Conclusions The expanded granular sludge bed reactor operates up to now with a specific biogas production of 0.44 m3/kg COD

eliminated. The specific methane yield is 0.31 m3 CH4/ kg COD eliminated. The composition of the biogas (median values) is 70% methane (CH4) and 30% carbon dioxide (C02).

At this stage of the project, the decomposition results so far achieved have been well above the expected values; moreover, in respect of the COD parameter they are below the legal discharge limit of 80 mg/l (see Figure 20).

Figure 20 COD removal line 1 In the current operational phase, ammonia-nitrogen concentrations averaging 2 mg/l are achieved in the processing line 1 (with technical pre-treatment stages). The nitro-nitrogen efflux is 120 mg/l. In the 2nd treatment sequence (also without filling of the filter section) 11 mg/l of ammonia-nitrogen and 24 mg/l of nitro nitrogen were measured at the drain pipe. In other words, the nitrification performance is not as high as in trial line 1. The COD run-off values for the trial line 2 (raw sewage treatment) - only through wetland system without any kind of technical pretreatment - are shown in Figure 20 below.

Figure 21 COD removal line 2 The hydraulic load of the eco-technical processing stages (without recirculation volumes) has been increased continuously - at consistently high concentrations of substances contained in the incoming wastewater - for processing line 2 (sedimentation pond, OKA 2, OKA 3) and is now 20 m³ per day (equivalent to a hydraulic loading rate of 48 l/m²d). So far, we have neither observed a major drop in the cleaning performance, nor have any clogging effects occurred on the filter surfaces. The removal rate bases on mass units for line 1 and the ongoing sample period starting on February 12th are 96 % for COD, 98 % NH4-N and 27% PO4-P respectively. For testing line 2 the treatment efficiency is 98 % COD, 76 % NH4-N and 17 % PO4-P, respectively. In the constructed wetlands cells 1 and 2 the following plant species have been used: Dracaena sanderiana Typha angustifolia Phragmites communis Arundo donax Pandanus humilis Saccharum officinarum Pennisetum purpureum Miscanthus sacchariflorus Under high strength wastewater conditions Phragmites, Arundo, Typha and Miscanthus perform best. For the plant survival rate the influence of the wastewater is less important than the condition of the plants after they have been transported from their habitats near HCMC. A single stem and few roots have been found to be typical indications of a poor state. Potted precultivated helophyte plants from nurseries are not available in South Vietnam, except Cyperus papyrus, Arundo donax and Colocasia esculenta, and a number of other species used as foodstuffs. Another key factor for a good development is the water level inside the basins, which is adjusted by the variable depth weir in the final control shafts.

When poorly developed plants are used, the basins must be dammed in for months in order to achieve a dense vegetation carpet. The associated anaerobic conditions in the filter substrate of the normally aerobic vertical subsurface flow constructed wetlands have a toxic effect on some plants which do not have the aerenchyma (air conducting tissue) typical of swamp plants, such as Arundo donax or Dracaena sanderiana, and therefore do not tolerate a permanent bank-up in the root area for prolonged periods. This applies e.g. to the genera Phragmites, Typha and some Cyperus species. The impressive Pandanus humilis, which has been planted sporadically in the basins, also appears to be not permanently flood-resistant. Through intermittent loading of the reed beds a radical change in the oxygen regime is achieved. After the water has been saturated through the distribution system a drainage network at the base collects the purified water. Then the pore space of the substrate is refilled with air by low pressure which triggers aerobic decomposition processes. As the "aeration automatism" induced by the intermittent feed was unable to set in before the bank-up operation was terminated in late January 2012, the efflux values achieved before that date do not reflect the normal operating condition but atypical anaerobic conditions that were required to provide and maintain the moisture levels needed for the new plants. Specifically, this means that operation without bank-up is associated with an increased oxidation of the organic carbon compounds as well as a much higher nitrification rate. Wetland cell number 3 is designed as final treatment step with a low pollutant load, which is why a number of food plants like banana (Musa spp.) and sugar cane (Saccharum officinarum) have been tested. The third plant used in OKA 3 is Vetiver grass (Vetiver zizanioides). All three species performed well, especially the bananas, which grow up to a height of more than 6m.

The floating islands on the sedimentation pond have been planted with more than 20 emergent and floating helophytes and hydrophytes. Following their exposure to raw wastewater of up to 10.000mg/l COD only the following species survived: Colocasia esculenta

Figure 22 Constructed wetland cell 3 (OKA 3)

Limnocharis flava Neptunia oleracea Cyperus malaccensis Ipomoea aquatica Persicaria attenuata Hydrocotyle verticillata Eichhornia crassipes 4 Acknowledgements The research project was funded in the period from 2009 until 2012 by the German Federal Ministry of Education and Research (BMBF) and the Vietnamese Ministry of Science and Technology (MOST).

Figure 23 Plant development on floating islands

REFERENCES Annachhatre, A. P., & Amatya, P. L. (2000): UASB Treatment of Tapioca Starch Wastewater. Journal of Environmental Engineering , pp. 1149-1152. Blumberg, M. (2011): Treatment of tapioca processing wastewater and sustainable water pollution control management of key economic zones in South Vietnam (2009 - 2012), Poster Presentation on 3rd international Symposium "Re-Water Braunschweig", 21. - 22. November 2011, Braunschweig. Kato, K. (2007): Tapioca wastewater treatment and clean development mechanism project. In V. L. University (Hrsg.), Conference on Tapioca-Eco-Industry Cluster in Vietnam. Pick, V., Fettig, J., Austermann-Haun, U., Fabritius, B., Stein, A., Blumberg, M., Phuoc, N. V. (2011): Eine neue Verfahrenskombination zur Reinigung von Stärkeabwasser in Vietnam. In DECHEMA (Hrsg.), Management und Behandlung industrieller Roh-, Prozess- und Abwässer, 07. - 08. November 2011, (pp. 140 - 147). Frankfurt am Main. Rajbhandari, B., & Annachhatre, A. (2004): Anaerobic ponds treatment of starch wastewater: Case Study in Thailand. Bioresource Technology , 95, pp. 135-143 Fettig, J. (2010): Treatment of tapioca wastewater and sustainable water pollution control management in key economic zones of South Vietnam In: InnovationsAllianz – Der NRW-Hochschulen E.V. (Editor): The Universities of NRW: Your partners for Europe research projects. Pick, V., Fettig, J., Austermann-Haun, U., Fabritius, B., Stein, A., Blumberg, M., Phuoc, N.-V. (2011): Eine neue Verfahrenskombination zur Reinigung von Stärkeabwasser in Vietnam. Vortrag zur DECHEMA 2011, Berichtsband „Industrietage Wassertechnik“, Frankfurt. Blumberg, M. (2011): Treatment of tapioca processing wastewater and sustainable water pollution control management of key economic zones in South Vietnam (2009 - 2012), Poster Presentation on 3rd International Symposium "Re-Water Braunschweig", 21. - 22. November 2011, Stadthalle Braunschweig, Braunschweig. Forschungsgesellschaft Landschaftsentwicklung Landschaftbau e. V. (FLL), Ingenieurökologische Vereinigung e. V. (IÖV), (2008): Empfehlungen für Planung, Bau, Pflege und Betrieb von Pflanzenkläranlagen. FLL- Regelwerk, Bonn. Blumberg, M. (2009): Treatment of tapioca wastewater and sustainable water pollution control management in key economic zones of South Vietnam. BMBF-Project 02WA0995, Execution planning Ingenieurbüro Blumberg, unpublished. Wang, W. (2007): Cassava Production for Industrial Utilization in China - Present and Future Perspectives, In: Howeler, RH, Cassava Reserach and Development in Asia: Exploring New Opportunities for an Ancient Crop, Seventh Regional Workshop, Bangkok, Thailand 28.10. - 1.11.2002, Centro International de Agricultura Tropical CIAT, Colombia, S. 33-38