Water Reuse in Brazil Advanced Wastewater Treatment with integrated MBR/RO

To help solve dwindling water sources in São Paulo, water reuse is being applied to large industrial areas outside the city. Dr Dirk Herold and Deepak Raina discuss the application of Tertiary Membrane Bioreactor technology as part of the largest project in the Southern Hemisphere and trial data showing how a MBR/RO system would meet the project's goals.

Pennwell web 400 223

To help solve dwindling water sources in São Paulo, water reuse is being applied to large industrial areas outside the city. Dr Dirk Herold and Deepak Raina discuss the application of Tertiary Membrane Bioreactor technology as part of the largest project in the Southern Hemisphere and trial data showing how a MBR/RO system would meet the project's goals.

The demand for advanced wastewater treatment is drastically increasing due to the need for water reuse in all parts of the world. Membrane Bioreactors (MBR) have become state-of-the-art when it comes to recycling municipal wastewater for unrestricted irrigation purposes. Modern water quality demands from industrial customers, however, is requiring not only removal of particular substances, but also a reduction of dissolved solids and conductivity. Reverse osmosis (RO) has been used by many plants as part of a water reuse scheme. Unfortunately it is often not as simple as just adding RO after a conventional wastewater treatment plant to meet reuse water quality parameters.

Pennwell web 400 223

As the system was still a membrane bioreactor it was designated TMBR – Tertiary Membrane Bioreactor.

Many RO systems use membrane filtration as pre-treatment to remove suspended solids. The membrane system does an excellent job of providing water with low suspended solids to feed an RO. However, the membrane system requires additional space and does not effectively reduce the amount of dissolved contaminants such as organics that are fed to an RO system. To remove dissolved organics biological wastewater treatment is required.

Instead of separating biological treatment and ultrafiltration (UF), it is more efficient to combine both processes in an MBR. MBRs can reduce COD and nutrient levels far below those achieved by conventional biological treatment only. To achieve high removal of COD and nutrients, a proper biological design is essential. The design considerations should include plug flow reactors, multiple recirculation flows, pre-denitrification, anaerobic zones, de-oxygenation tanks, etc.

Pennwell web 400 223

Brazillian behemoth: the Aquapolo Ambiental water reuse project is not only the first of its kind in Brazil, but with a potential capacity of 86,400 m3/day, also the largest in the southern hemisphere and the fifth largest in the world

When an existing conventional wastewater treatment plant needs to be upgraded for partial water reuse, a Tertiary Membrane Bioreactor (TMBR) may also be an option. The TMBR is a polishing MBR. It treats water coming from the existing clarifier and is further treating this in a newly built biological step using membranes as the separation process.

Design Considerations for MBR/RO Integration

Requirements for an MBR module and the integrated system are very specific for industrial applications because wastewater treatment or water reuse is never the primary task of an industry but a necessary evil. Therefore it is prerequisite for the membrane module and the system to be reliable and easy to operate. In this context it is important to notice that costs of such a system are not only capital costs, influenced by membrane price, but mainly operational costs in form of power demand, chemical consumption, membrane replacement and last but not least maintenance and daily operation.

Prevention of Fouling

In the combination of MBR and RO for water reuse, the protection of the RO against components that can cause fouling is most critical. There are four categories of fouling that need to be considered:

1. Particles that might block the brine spacers of the RO

2. Fouling of the RO by adsorption of residual soluble organics in the RO feed

3. Biofouling caused by bacteria growing downstream of the permeate side of the UF

4. Inorganic fouling (scaling) by exceeding the solubility level of salts.

It should be noted that the rejection of particles is the biggest advantage of using a combination of biological treatment and Ultrafiltration, either in form of conventional wastewater treatment plus tertiary treatment by UF, or in form of MBR, (in comparison to biological treatment with sedimentation and sand filtration). Using Ultrafiltration on biologically treated water generates a permeate where suspended solids are less than 1 mg/l and a particle size < 0.1µm.

Reducing the amount of soluble organics in the wastewater before entering the RO is the second major advantage of an MBR. In comparison to applying UF as tertiary filtration, the combination of biological treatment and UF in the integrated process of an MBR, increases the efficiency of biodegradation significantly. It has been proven in the past that MBRs can reduce COD and nutrient levels far below that achieved by conventional biological treatment only. To achieve high removal of COD and nutrients, a proper design of the biology is essential. The design considerations should include plug flow reactors, multiple recirculation flows, pre-Denitrification, selectors, anaerobic zones, de-oxygenation tanks etc.

Biofouling and how to reduce it is the most critical issue for RO. Several parameters on system design and plant operation have been proven to be effective. Biofouling requires both the presence of microorganisms and a nutrient source. Even though both have been significantly reduced by the biological treatment in the MBR process, small amounts of bacteria and soluble nutrients are still present in the permeate of the UF. In combination with factors like residence time, elevated temperatures and sometimes light, ideal conditions have been created to promote bio-growth. To reduce bio-growth the following measures have proven to be effective:

• Piping between MBR and buffer tank should be as short as possible without any dead zones. Unnecessary long piping increases residence times and gives time and area for bio-films to grow. If these bio-films detach from the pipe, either naturally or by chemical cleaning, they might block the brine spacers of the RO. Safety filters on the RO inlet can protect the RO but must be frequently changed/cleaned.

• The MBR buffer tank should be large enough to feed the RO system, but small enough to keep retention time below 30 minutes. The requirements of an RO might be in opposition to this, but long residence times and high temperatures will lead to bio-growth. It is also important to choose a tank material that does not allow light penetration. Otherwise algae will grow.

• Chemical cleaning of the MBR membranes should be carried out frequently. Experience shows that a more frequent dosing of low concentration chlorine prevents biogrowth in the permeate system of the MBR. Koch membrane Systems optimized this approach to the degree that without increasing the overall amounts of chemicals for maintenance cleaning the system is kept completely clean and at the same time the chemicals are applied in a way that does not jeopardize the downstream RO membranes.

• Another measure to reduce biogrowth is the use of chloramines to disinfect the water downstream the MBR. Experience has shown that the best place of dosing is upstream of the feed tank to the RO, see figure 1.

Pennwell web 600 225

Upgrade scenarios

Considering much of São Paulo's wastewater treatment plant's capacity was available for reuse, there was no need to convert the entire existing installation into an MBR. Three scenarios that were evaluated made use of a single train that was not in operation. This consisted of a single primary clarifier, aerated tank and secondary clarifier. The difference between these options was the new designation of each of the old structures. All three approaches have a common denominator - they use raw sewage as a source. This methodology was chosen as the best nutrient balance was available from the raw sewage and the ratios of C:N:P were favorable for a balanced Biological Nutrient Removal (BNR) MBR design. The latter also maintained an alkalinity balance and required less additional chemicals for alkalinity control and P removal.

Pennwell web 400 210

In the combination of MBR and RO for water reuse, the protection of the RO against components that can cause fouling is most critical.

Retrofitting existing concrete structures as required for system 1, 2, and 3, can throw the odd surprise during the demolition and re-building phases. Interaction with existing structures must be carefully calculated and must fit into the plants operational philosophy for the future population expansion. As a consequence of this, a completely different approach was taken for the fourth option.

The idea behind this was to make use of the existing plant operation and infrastructure and use the effluent of the secondary clarifier as source for the new installation. Clearly this nutrient deficient design would be a challenge biologically, but could be developed as a Greenfield enhancement to the existing site operation without a direct interaction with the site's operation. As the system was still a membrane bioreactor it was designated TMBR – Tertiary Membrane Bioreactor.

Pennwell web 600 365

Figure 2 shows the concept of adding a polishing MBR or tertiary MBR (TMBR) to the existing plant. Around 56,000 m3/d coming from a number of selected secondary clarifiers is pumped to the new TMBR consisting of a denitrification zone to recover alkalinity, an aerated zone for nitrification and carbon removal and membrane modules for solid liquid separation.

TMBR pilot trials and results

A standard pilot plant was used to exactly simulate the TMBR design concept. Membrane modules used were hollow fiber submerged UF membranes. All flows were according to the feasibility design and the control was maintained for Dissolved Oxygen, pH and MLSS. The pilot was run between 2 and 6 g/L MLSS and start up and daily plant operation was simulated. The feed flow was controlled to a fixed value and all Maintenance Cleaning (MC) and down time was reflected in the operational flux of the membranes. The plant output was 775 L/h net flow to reflect the net full-scale production of 650 L/s. The membrane flux was <25 L/m2h net as yearly average with an operational flux of ~30 l/m2h. N-1 operational points were also tested. The biological system was exactly simulated with a hydraulic retention time (HRT) of ~3 hours and a Sludge Retention Time (SRT) of between 12 and 20 days.

Pennwell web 600 330

The variation in the conventional plant COD discharge was extreme to say the least, as shown in Figure 3, and certainly not suitable for direct RO polishing. This is largely associated to TSS that is washed out of the secondary clarifiers. The frequency of wash out was unpredictable and often combined with heavy rain, despite the sewer system being separated from rain collection. As the organic load to the TMBR varied so drastically, it was difficult to control the biological performance. The system had to process huge COD peak loads followed by periods of almost zero load.

The latter transformed the TMBR sludge from a well flocculated sludge into a fine dispersed sludge often associated with industrial MBR systems. This had a direct impact on membrane performance as dispersed sludge normally yields lower average fluxes. In the period up until July primary feed was occasionally by-passed to the TMBR and occasional ethanol was dosed to the TMBR. These "operational tools" proved useful to stabilize the biological system, but more importantly to stabilize the nitrifier population and alkalinity balance. After July a further "operational tool" was added in the form of ferric dosing. This had a two advantages; first, that the total residual phosphorus was precipitated out to allow the correct discharge criteria and secondly, the biomass became marginally more flocculated and easier to control for the nutrient removal processes. The target COD of 20 mg/L was achieved with the "operational tools".

RO pilot trials

A standard RO pilot was used to exactly simulate the RO design concept. The scale up factor was 538x and used to simulate the operation of the proposed full scale system for 200 L/s feed water. The pilot trials were set up in three well defined phases, a simple batch mode, a modified batch mode and a continuous mode.

The simple batch mode was used to establish the range, quality and basic operation window of the RO fed with TMBR permeate. The simple batch mode results were analysed and the results fed into the RO design program to estimate the correct configuration and operation. The modified batch modes further fine tuned the design of the RO system and a continuous mode was calculated and finally tested on the pilot.

Case Study – Aquapolo (Brazil)

Today water sources for drinking water in São Paulo are becoming increasingly scarce. Therefore it becomes increasing important to apply reuse water to the big industrial areas surrounding the city and to safeguard drinking water for the city's inhabitants. The Aquapolo Ambiental water reuse project is not only the first of its kind in Brazil, but with a potential capacity of 86,400 m3/day, also the largest in the southern hemisphere and the fifth largest in the world. This followed the state government recently issuing new regulations to restrict the industrial use of potable water, forcing factories to look for ways to reuse their wastewater, or obtain recycled water from another source.

The goal is to use water from a municipal wastewater treatment plant, the ABC Sewage Treatment Plant (ABC STP) and by additional treatment steps increase the water quality to a degree that allows reuse by industrial consumers, e.g. in local petrochemical industries. The initial 2011/12 phase will produce 56,160 m3/day of reuse water. The volume of first-use water that will no longer be consumed by the industries is enough to continuously supply drinking water to a population of 350,000 inhabitants, with the potential capacity to reach 600,000 if extended for other reuse clients.

The ABC STP at the boundary between São Paulo City and São Caetano do Sul was built in the 1990's and has an installed capacity of 260,000 m3/d. It consists of headworks, eight primary clarifiers, eight aerated tanks of 18,000 m³ each and twelve secondary clarifiers of 6,500 m³ each. Today the plant is only loaded to 50% of its design capacity with four primaries functioning, three to four of the aerated tanks in operation, and five secondary clarifiers. More than 50% of the plant is currently not required to operate. The design of the ABC STP biological treatment system only incorporates carbonaceous removal although periodic partial nitrification does occur. Denitrification is not foreseen in current design.

The continuous mode was further optimized with anti-scalant dosing, but relatively little anti-scalant was required to stabilize the RO performance. Overall the RO ran surprising well with little external interference and fine tuning, which was attributed to the performance of the TMBR system as pre-treatment for the RO. Cleaning tests were also carried out and the RO elements were inspected after the trial. Fouling was minimal.

Overall the RO yielded a recovery of 75% at a flux of 21.5 L/m2h. The produced RO permeate was of a super quality and ideal for blending back into the TMBR permeate to produce a final reuse water quality of <720 μS/cm2.

Summary

The feasibility studies led to a number of viable options for polishing the wastewater at the ABC plant. The most viable solution was to polish the conventional system secondary effluent with a tertiary MBR system or TMBR. Despite being more expensive the operational aspects were more favorable. The TMBR solution was simulated in a pilot and all data and operational aspects confirmed in the feasibility study. The TMBR was not easy to operate due to the highly viable feed water quality, but a number of "operational tools" were developed to stabilise the operation and improve the biomass quality.

The submerged membranes profited from the optimized biological system. Average fluxes of >25 L/m2h were achieved and Maintenance Cleaning on a daily basis was found to enhance the UF performance. Recovery Cleaning returned the UF membrane to original process permeabilities. The RO system was able to produce an excellent water quality with a 75% recovery and a flux of 21.5 L/m2h. The final water quality for reuse was achieved via blending the TMNR permeate and the RO permeate. All aspects of the feasibility study were proven and yielded a viable full scale solution for water reuse under difficult operational conditions.

The State of São Paulo is the world's seventh most populous urban area and is considered the economic, financial, and technical hub of Brazil. The region contains nearly one-fourth of the country's population but less than 2% of Brazil's water.

Upon completion, the Aquapolo Ambiental water reuse facility will free up enough drinking water to continuously supply a population of 350,000 inhabitants, with the potential capacity to reach 600,000.

Construction of the full-scale installation started in April 2010; start-up is scheduled for April 2012. WWi

Author's note:Dr. Dirk Herold is the European process engineering manager and Deepak Raina is the regional manager for the Middle East, Koch Membrane Systems. Co-authors on the article include Christoph Kullmann, European business manager PURON, Koch Membrane Systems and Emyr Costa, senior project director, Odebrecht Group. Email: infoeurope@kochmembrane.com.

More Water & WasteWater International Current Issue Articles
More Water & WasteWater International Archives Issue Articles

More in Wastewater