Technically speaking

In the 1980s, attempts to enhance the activated sludge process by integrating biofilm carriers gave rise to submerged rotating contactors, plastic media packing, ringlace woven fibres, or sponge cubes (Captor and Linpor). None found widespread acceptance because the gain in reactor size was often off-set by the cost and operational challenges of the carriers.

The moving bed biofilm reactor (MBBR) was developed at the Norwegian University of Science and Technology under the direction of Professor Hallvard Odegaard, and the first plant was installed in Norway by Kaldnes in 1989. The process utilises a cylindrical plastic carrier to provide an environment in which bacterial populations and protozoa can grow very effectively. The carriers are submerged in tanks and aerated to permit biofilm growth. Aeration also supplies the energy to disperse the carriers and obtain a completely mixed liquid. The initial medium for the support of biofilm growth, K1, consists of high-density polyethylene (HDPE) cylinders approximately 10mm in diameter and 7mm long (Figure 1). The cylinders are finned and have a cross piece in the centre. This media has an effective surface area for biofilm growth of 500m2/m3 and is used in reactors at fill rates of up to 67%, giving a biofilm surface area of approximately 350m2/m3 of reactor. There are about 1M individual pieces of media per m3.

A recent development has been a larger media, Kaldnes2 (K2), which is 15mm in diameter and 15mm long with an inner cylinder, which can offer operation and economic benefits in some circumstances. This media has an effective surface area of 350m2/m3 and is used in reactors at fill rates of up to 55% giving a biofilm surface area of approximately 193m2/m3 of reactor. Both of these media have an actual surface area well in excess of that quoted, but external surface area is not accounted for as being available for biofilm growth, although microscopy photographs suggest external surfaces are well colonised in some cases.

MBBR tanks are equipped with an outlet sieve to retain the media, the media itself, and a means of aeration or mixing. Within the reactor the media, effluent and air are completely mixed resulting in very efficient contact between the biofilm and substrates within the liquid. One of the important features of the process is that biofilm thickness is controlled by the movement of the media so oxygen diffusion through the biofilm is encouraged. Detached biofilm is suspended within the reactor and leaves the reactor with the effluent.

Tanks can be any shape and existing tanks have been used in a number of cases. Ideal depths are between 4-6m which maximises oxygen transfer efficiency while making blower choice simple and keeping typical superficial air velocities in reasonable limits. Square cross sections are preferable. Each tank is considered as a continually stirred tank reactor (CSTR) so in order to differentiate between process conditions, a number of reactors in series can be installed. For example a process for BOD removal and nitrification would typically consist of four reactors with the first two for BOD removal, and the second two for nitrification. Reactor three would be oxygen substrate limited and achieve say 5mg/l ammonia as an outlet

concentration whereas reactor four would be ammonium substrate limited and achieve 1 mg/l ammonium. A reactor for nitrification and denitrification could

consist of nine reactors in series, as at in Lillehammer, Norway.

Mechanical mixing is required in anoxic reactors, usually provided by slow speed submersible mixers. Aeration is by a coarse bubble system typically using 4mm holes in stainless steel laterals of 30mm diameter. These laterals and associated headers are usually fabricated in rigid tube and fixed to the base of the reactors. The transfer efficiency of this system approaches membrane aerators, due to the extended bubble path length and dispersing action of the media. Field transfer tests have resulted in values between 7-12g O2/m/m3/h, comparable to membranes with 11g O2/m/m3/h. Sieves are used to retain the media and design has progressed from the flat stainless steel mesh sieves originally used to three current designs:

• horizontally-mounted wedge wire pipe sieves with appropriate wire spacing (5mm for K1 and 10mm for K2) are often used in the UK. There have been no problems with blockage, etc, even when used in conjunction with typical 6mm preliminary screening,

• in Scandinavia vertically-mounted perforated plate sieves are preferred,

• in anoxic reactors flat, perforated plate sieves are often used so there are no intrusions into the reactor to interfere with the mixing patterns.

  Solids separation following the reactors is achieved by

  conventional means, settlement and DAF having been used

  successfully. In Scandinavia the process is often combined with chemical dosing for phosphorus removal.

The process can be used for carbonaceous BOD removal, nitrification and denitrification. Denitrification can be by pre or post-denitrification using sewage or an external carbon source. Average results for the treatment of municipal sewage demonstrate a BOD/COD removal around 95% with resulting effluent qualities of 14mg/l and 30mg/l respectively. For dairy wastewater 85% removal of COD was obtained at a volumetric loading rate of 12kg COD/m3/d. Even higher loading rates of up to 50kg COD/m3/d were used for pulp and paper wastewater, when the removal efficiency is limited to only 65-70% of total COD.

The process can be used as a replacement for ‘conventional’ treatment and in certain circumstances this can be an attractive solution, such as when an existing tank can be converted, or space considerations are particularly tight. However the most significant advantages have been obtained when the process is used to enhance an existing treatment plant in order to add nitrification or denitrification, or increase capacity. Existing activated sludge plants can have an MBBR added either as the first reactor of a two-stage plant or as a hybrid with the MBBR as a first stage but without intermediate settlement.

In the original concept, there is no recycle of biomass and no backwashing of the reactor, but in a few upgrades of existing activated sludge, recycling of solids has been maintained to establish a hybrid biomass of suspended and fixed bacteria.

Results of a pilot study in Broomfield, Colorado, confirmed that conversion of the existing aeration basin volume by introducing plastic carriers would double the aerobic SRT without extra volume, and therefore allow the works to meet the new performance requirements of nitrification at winter design flows and loadings. The pilot system was operated at a relatively high residual dissolved oxygen concentration, which favours the nitrification in the attached growth portion of the system. Nevertheless, a significant amount of simultaneous denitrification occurred in the aerobic reactors even though a high residual dissolved oxygen concentration was maintained. About one half of the total amount of denitrification occurred in the aerobic basins. The pilot system had no problem meeting the effluent total nitrate objective using a combined RAS/recycle rate of 1.75 times the influent flow.

The full-scale system, to upgrade an existing 20,400m3/d plant, is just under start-up and can be operated at a lower recycle rate especially considering that the IFAS basins will operate at lower residual dissolved oxygen concentration. Moving bed biological reactors have been installed at more than 100 sites split between industrial and municipal applications. For example the plant at Sande paper mill in Norway treats a load of 60t COD/d. Two hybrid MBBR/AS plants treat sewage in New Zealand, one for 200,000 PE at Moa Point serving Wellington.

In the UK, an MBBR treats sludge liquor at Shoreham, and a tertiary unit nitrifies a PE of 80,000 at Bury St Edmunds. Full treatment MBBR systems are installed on several sites:

• Anwick WwTW, with a PE of around 6,500, a meat processing load of 4,000 PE and a poultry processing factory with a PE currently around 65,000, needed to be upgraded for tighter ammonia consents,

• Braintree WwTW was extended from a connected population of 17,000 to 24,000 and had to meet new consent conditions with respect to nutrient removal (2mg/l total phosphorus, and 60% nitrogen removal),

• the Corby scheme of 250,000 PE with mostly food processing effluent was required to meet a tightening effluent consent of 10mg/l BOD, 20mg/l suspended solids, 5mg/l ammonium and 1 mg/l total phosphorus,

• Grimsby WwTW, serving a population equivalent of 340,000, with approximately 70% being industrial load had to switch from a long sea outfall to secondary treatment.

In the US, the MBBR process is only now finding applications, due to the tightening of effluent limits to prevent eutrophication in water bodies such as Chesapeake Bay, Long Island Sound and the Gulf of Mexico, particularly at the mouth of the Mississippi River. In Cheyenne, Wyoming, two plants with a total capacity of 60,000m3/d were upgraded for total nitrogen requirements of less than 10mg/l, and the cost of upgrading the biological plants with MBBRs is in the range of £100-150/m3, of which the core of the MBBR: media, sieves and aeration grid accounts for less than half