Black and Veach's Frank Rogalla looks at issues surrounding membrane bioreactorsThe first membrane bioreactors (MBRs) for wastewater treatment were developed for industrial and commercial applications, in an attempt to generate high-quality water for non-potable reuse. In the industrialised nations of Europe, North America and Asia, increasingly stringent environmental regulations are imposed to protect water resources and human health. They require high levels of treatment with respect to nutrients as well as bacteria and viruses, but in small footprint, low-impact facilities.
Currently there are a limited number of MBR installations in municipal treatment applications. Because of high capital and operation and maintenance (O&M) costs, mostly proportional to flow, MBR installations have mainly been used in small plants of less than 5000m³/d: within the US, there are about 24 municipal MBRs, all of which use Zenon hollow fiber membranes, in addition to about ten installations in Canada. The largest plant under construction has been announced in Italy, with 38,000m³/d. In the UK, a plant in Swanage on the southern coast, built and operated by Wessex Water with Kubota plate and frame membranes, has a peak flow of 12,500m³/d, and has been operating for nearly a year.
Vivendi Water/US Filter recently introduced its MBR design, and has a facility under construction in California, which combines the technologies of the Memcor CMF-S membrane, used in drinking water, with Jet Tech aeration and pumping equipment. Similarily, Mitsubishi Rayon, which has the largest number of small units in operation in Japan, has submitted its equipment for evaluation at the Aqua 2000 testing facility in San Diego, California.
It is accepted that fine screening and grit removal is required to protect the membranes. Fine screening, of 3mm opening or lower, is preferred to prevent particles from embedding in the membranes. Mixed liquor suspended solids (MLSS), solids retention time (SRT), and membrane flux rates are key criteria for the design of the bioreactor. Initial industrial applications indicated design conditions using SRTs of 25 days or greater as well as MLSS concentrations up to 20,000mg/l. However, it has been determined through various pilot-scale facilities that operating at such high MLSS adversely impacts the flux rates and oxygen transfer. As a result, the suggested MLSS concentration has been reduced to 10,000 - 12,000mg/l for most applications, which is still three to four-times higher than conventional aeration basins.
Operating at 25 days or greater SRT provides a stable biological system, better suited for unattended operation at remote facilities. Long sludge age also minimises solids production, albeit not below that of a conventional extended aeration process. Yet, the desire to minimise facility footprint and to implement biological nutrient removal (BNR) has driven research to evaluate operations at shorter SRTs. On the other hand, reducing the SRT can increase fouling of the membranes, potentially raising the number of membranes to accommodate a loss in flux rate, leading to higher capital cost.
The selection of the appropriate membrane flux rate is critical. An aggressive flux rate can limit plant capacity, needing a high cleaning frequency, which continuously takes membranes out of service. The flux rate is influenced by temperature and biomass concentration, therefore a universal flux rate cannot be given. One manufacturer suggests an average flux rate of approximately 25 l/m²/h at 20ºC, allowing for occasional peak flux rates around twice the average. Other suppliers contend that flux rates should not exceed approximately 17 l/m²/h at any time in order to minimize fouling and maintain a reasonable chemical cleaning frequency. In order to accommodate the lower flux rates economically, upstream equalization is recommended. Satellite 'scalping' plants for water reclamation, operating at constant flow rates, are therefore a favourable application.
The O&M issue of greatest importance is the membrane cleaning operation. With the high concentration of solids within the basin, there is a need for a high intensity scouring of the membranes to maintain operating flux rates. Membrane scouring is provided using aeration, resulting in intensive energy consumption and power requirements, as well as high residual dissolved oxygen levels, complicating the design for denitrification. In addition to continuous air scouring, the membranes must be removed from service at an interval between one to four months for chemical cleaning, depending on operating conditions and wastewater characteristics. Chemical cleaning is either achieved by lifting the membranes out of the basin and dipping them into a chemical bath of a chlorine or citric acid solution, or by isolating some parts of the membrane system for chemical injection. This process can be labour intensive and poses specific safety issues for a large facility, membrane cleaning would be required on a continuous basis.
There is a bright future for MBRs for municipal wastewater where very high quality effluent is needed from a small footprint. Key issues in the design of the MBR systems which need further research are:
- SRT optimisation, i.e. how low can it go without sacrificing flux rate?
- determination of BNR design criteria and configuration to consistently achieve low effluent phosphorus and nitrogen,
- improvements in the membrane cleaning operation to reduce air scouring and minimise labour and residuals during chemical cleaning.
As MBR design criteria become better established, and as more players enter
the field, the economics will also improve. Currently, even with the trade-off
of significantly less facilities, i.e. no final clarifiers, no filtration and
less disinfection, MBRs are not cost-effective, unless a microfiltration quality
water is required for reuse. With cost reductions and improvements in membrane
cleaning, the technology becomes realistic even for larger facilities.
Black & Veatch is the parent company of Paterson Candy