The system 1 uses a cross-flow chamber, allowing fluid to flow through the hollow fiber transversely or axially. Hollow fibers of different sizes can be used, and the spacing between fibers can be adjusted. System 2 uses a feed tank to vertically place laboratory-scale hollow fiber membrane modules (about 0.3m and 0.5m long), and monitors the suction pressure at a constant flux. In these two systems, polypropylene hollow fibers with a pore size of 0.2μm and an outer diameter of 0.65-2.7mm are used ... The biomass used is dry yeast with an average diameter of 5 μ m.

1, critical flux: In 1 system, aeration can significantly increase the flow of treated water. The effect is obvious when the water output is small, but not obvious when the water output is large. The flux increases with the increase of aeration intensity and finally stabilizes at a higher value. Aeration bubbling will reduce the degree of recoverable and unrecoverable resistance. Theoretically, when the system operates under the condition of fixed effluent flux, solid particles can not accumulate on the membrane surface during operation. The outflow flux when pollutants begin to accumulate is called "critical flux" (Jcrit). Under this flux, the solid deposition caused by convection is just balanced by the shear force of fluid, and it will not cause the solid deposition on the membrane surface. Jcrit can be measured by flux stepping and observing the history of TMP. Once TMP starts to increase significantly, it means that deposition begins and the flux exceeds Jcrit. Because aeration bubbling can enhance the shear effect of liquid flow on the membrane surface, it can be expected that aeration bubbling will increase Jcrit. In the actual operation of MBR, Jcrit is not a very clear concept, because: the liquid in MBR reactor is a mixture of different species, and each substance has different response to the shear force formed by the liquid flow on the membrane surface, and flux step method can only measure JCRIT of the main species; In submerged MBR, there is pressure difference in the hollow fiber cavity along the axial direction, so there is flux distribution in the hollow fiber membrane, that is, the flux in some positions is greater than Jcrit, while the flux in some positions is less than Jcrit. Therefore, it is very realistic to consider the "sustainable flux", that is, the continuous effluent flux without membrane cleaning in an appropriate operation period. It can be considered that "sustainable flux" increases with the increase of aeration airflow speed.

2. Circulating velocity in the reactor: In practice, aeration is also an important factor affecting the performance of membrane modules. In submerged MBR, the turbulence formed by aeration can reduce the accumulation of pollutants on the membrane surface. In submerged MBR, the liquid flow formed by aeration forms an upstream zone and a downstream zone in the reactor. When the cross-sectional area of the downstream area is smaller than that of the upstream area, there is not enough liquid flowing in the bioreactor to clean the membrane surface. When the ratio is large (3.6 ~ 4.5), proper fluid flow can be formed in the bioreactor to slow down membrane surface pollution.

3. Placement mode of membrane: In the test of system 1, the performance change of submerged hollow fiber is very complicated, depending on the size of the fiber and whether there is aeration or not. No foaming: for fine fibers (id/od:0.39mm/0.65mm), the effect of horizontal placement is better than that of vertical placement; For large fibers (id/od= 1.8mm/2.7mm), vertical placement is better than horizontal placement. Bubble: the effect of vertical placement is better than that of horizontal placement. There is evidence that the horizontally placed membrane unit can retain some bubbles and cannot form enough turbulence and shear on the membrane surface.

4. Diameter of membrane fiber: In 1 system, the smaller fiber is better than the larger fiber, which may be because the finer fiber is easier to move and swing with the water flow, which is not conducive to the accumulation of dirt on the membrane surface. In the test of system 2, the best effect is achieved when the fiber with the smallest diameter is used and the fiber is loosely filled. In practical application, it should be considered that hollow fibers should not be filled too loosely, otherwise the liquid flow may lead to excessive movement of fiber filaments and damage. Fane et al. proposed a mathematical model to simulate the behavior of hollow fiber membrane module in submerged MBR, and used this model to optimize the diameter of hollow fiber membrane with a certain length to obtain the maximum output. They believe that there must be an optimal fiber diameter for a certain length of hollow fiber membrane, because the fiber with small diameter has high pressure loss, while the fiber with large diameter has low bulk density and specific surface area. Yang et al. used hydraulic calculation method to optimize the design of hollow fiber membrane module, and found that the geometric size of the membrane has a great influence on the water yield. He proposed that the membrane should be as short as possible under the condition of economic permission: at the same time, the longer the bonding length of the membrane fiber port, the lower the water yield. In the process of film making, the bonding length should be shortened as far as possible when the strength allows.