Oxygen balance around MBR

Based on one-year study in a full scale municipal MBR plant in Varsseveld, the Netherlands, it was observed that 78% of compressed air was supplied for membrane scouring while only 22% for biological reaction (Giesen, 2007). But 67% of oxygen dissolution took place in aeration tank while only 33% in membrane tank. Considering the air flow rate and the actual oxygen dissolution, OTE in aeration tank was estimated 7.2 times higher than that in membrane tank. Given the fact that OTE in aeration basin with 8-12 g/L MLSS ranges around 10%, OTE in membrane tank is only 1-2%. The low OTE in membrane tank is attributed to the low driving force at high DO (4-8 ppm) and the coarse bubbles used to scour membranes.

In the same study, the amount of oxygen transferred from membrane tank to aeration tank corresponds to around 10% of the total oxygen consumption of the system. Around 1% of the dissolved oxygen was transferred from aeration tank to membrane tank. Around 76% of the total dissolved oxygen was consumed in aeration tank while 23 % was consumed in membrane tank and 1% was lost through permeate.

Though aeration demand for membrane scouring is 3.5 times higher than that for biological aeration in the study, the ratio was  significantly lower during the wet weather condition due to the increase of biological aeration. (Note: The ratio has declined continuously in recent decades as a consequence of lowering specific aeration demand (SAD) and the improvement of overall process design. For example, the ratio between biological air and scouring air can be close to 1 instead of 3.5, if flow rate stays at a design flow and newest modules are used as shown in Table 1 here).


Fig. 1. Oxygen balance of MBR equipped with vertically mounted hollow fiber membranes (GE Water) based on one-year study in Varsseveld, The Netherlands (Giesen, 2007).

As can be seen in Fig. 1, significant amount of oxygen consumption occurs in membrane tank. There are three major mechanisms of oxygen consumption shift to membrane tank, which reduces actual biological air demand in aeration tank.

  • A sizable amount of nitrogen (TKN) can pass the aeration tank due to its slow degradation. DO, HRT in aeration tank, and TKN load to the tank affect the amount of nitrogen leak to membrane tank. It can be rapidly oxidized at the high DO (4-8 mg/L) environment in membrane tank. As a result, the TKN level in the plant effluent can be buffered greatly even with a large fluctuation in aeration tank (Fig.3.6 in Brepols, 2010).
  • Endogenous respiration also promotes oxygen consumption by bacteria in the membrane tank. If SOUR of biosolids in membrane tank is 3 mg O2/g MLSS/hr and MLSS is 12 g/L, OUR becomes 36 mg/L/hr. If HRT of membrane tank is 1.5 hrs, total oxygen consumption in membrane tank becomes 54 mg/L based on feed wastewater volume. It is a significant portion of the total specific oxygen demand of the wastewater, i.e. ~300 mg/L in above case.
  • Excess dissolved oxygen in membrane tank (4-8 mg/L) is carried over to aeration tank, while the mixed liquor in the aeration tank with 1-2 ppm dissolved oxygen is transferred to membrane tank. Due to the difference in DO in the two streams, a net oxygen transfer occurs. If DO of the two streams are 8mg/L and 2 mg/L, respectively, and the flow from aeration tank to membrane tank is 4Q and the flow from membrane tank to aeration tank is 3Q, the net DO transferred back to aeration tank becomes 16 mg/L on average based on feed wastewater volume, i.e. (3Qx8 – 4Qx2)/Q.

Overall, the contribution of dissolved oxygen in membrane tank to the total oxygen demand can be 10-30% even if membrane tank is separated from aeration tank.


© Seong Hoon Yoon