The primary purpose of using gas/air in tubular membrane is to enhance turbulence in the membrane tube, but the gas also induces liquid movement in the tube and can replace the circulation pump. This process can be divided into three categories depending on the driving force of liquid circulation: 1) process relying on airlift pump effect, 2) process relying on both airlift effect and mechanical pumping, and 3) process relying on mechanical pumping.
Mixed liquor circulation by airlift effect
When gas bubble rises in liquid, the surrounding liquid tends to move together with the bubble by so called “airlift pump effect”. The airlift effect can be used to drive liquid flow tangentially along the membrane surface while gas bubbles enhance membrane surface scouring (Imasaka, 1989). The air pressure can be simultaneously used to provide the TMP for the filtration operation by throttling the gas/liquid outlet using valve (Cui, 1997).
The first known attempt was made to enhance the flux of tubular ceramic membranes while reducing the liquid circulation energy in membrane coupled anaerobic digestion (Imasaka, 1989, 1993). Compressed N2, H2, or recycled headspace gas of the digester was injected to the inlet of the membrane module to enhance turbulence in the membrane module as shown in Fig.1a. Since the liquid circulation relied on the airlift effect without circulation pumps, crossflow velocity was variable depending on the gas flow rate. The digester was placed at higher position than membrane to utilize the static pressure for the filtration.
The linear velocity of the liquid in the ceramic membrane (ID=3.8 mm, length=0.5 m) ranged between 0.27 m/s and 2.7 m/s when gas flow rate varied between 0.12 L/min and 5.12 L/min. At an MLSS of 5,000 mg/L and a static pressure of 100 kPa, flux increased from ~25 LMH to ~75 LMH while gas flow increased from 0.12 L/min to 1.51 L/min (Imasaka, 1989). Overall the specific energy consumption was estimated at 1.78 kWh/m3.
Although the flux and the energy saving can be improved simultaneously by using airlift effect, operational flexibility is limited in this process since crossflow velocity and gas (or air) scouring are coupled. The gas flow rate required to optimize membrane surface scouring is not necessarily same as the gas flow rate to obtain the optimum crossflow velocity. In addition, a large gas recirculation capacity is required.
Cui et al. (1997) performed a comparison study between a pumped and an air lift system. They found that the two phase airlift system had about 30 % higher fluxes when compared with those with single-phase flow as shown in Fig. 2.
Mixed liquor circulation by airlift effect and pumping
The liquid circulation can be decoupled from air circulation by adding circulation pumps. The added pump can decouple circulation rate (or crossflow velocity) from membrane scouring effect by bubbles (Fig. 1b). As a result the system can be adjusted to obtain the maximum overall energy efficiency by optimizing the compressed gas flow and the pumping rate (Futselaar, 2007).
Though a large head loss occurs when two-phase flow moves through the tube, t is somewhat compensated by the net positive static pressure caused by the higher bubble holdup in the module than in the aeration tank, which forces liquids flow forward. In Norit’s Airlift™ MBR, the head pressure in the discharge side of the circulation pump can be controlled below 100 kPa at a moderate crossflow velocity of < 1m/s in the relatively short channel length (~3 m). The low head pressure results in significant energy savings in Airlift™ relative to the conventional crossflow tubular membrane processes, where the head pressure ranges 300-600 kPa. Vacuum is used to obtain permeate flow in this process. The nominal flux of Airlift™ is known at 30-50 LMH, which is around 50% higher than that for immersed membranes.
Fig. 1. Schematic diagrams of membrane bioreactors running with multi-tube module using airlift effect: a) no circulation pump and no permeate pump (Imasaka, 1989, 1993), b) with circulation pump and permeation pump (Futselaar, 2007).
Fig. 2. Comparison between the airlift and the single-phase flow system (Cui et al., 1997).
© Seong Hoon Yoon