Intermittent aeration

With defined cycle time

Periodic increase and decrease of scouring air flow rate is effective in reducing overall scouring air demand. The first patent granted to Zenon Environmental Inc. (currently a part of GE Power) covered a broad range of cyclic aerations with repeated cycles of 10-120 seconds, where air flows can be zero at low air cycles (Côté, 2000). Subsequently a patent with more detailed control mechanisms was granted to the same company (Rabie, 2001).

The effectiveness of the intermittent aeration is demonstrated experimentally as shown in Fig. 1, where membrane fouling rates at different air cycle times are plotted against flux. When average air flow rates were maintained at 0.36 m3/m2/hr,  membrane fouling rates decreased with intermittent aerations except the case with very short aeration cycles at 5s – 5s (5s on and 5 s off) under the experimental condition (Guibert, 2002). This result suggests overall aeration rate can be reduced with the intermittent aeration without increasing membrane fouling rate. Currently GE membranes are relying on intermittent aerations with 10 s – 10s cycle mostly but the cycle times can vary by location. By performing aeration for 50% of the total operating time, around 30% of net aeration savings (not 50%) can be achieved while not affecting membrane performance as will be discussed below. The cycle time can be modified to 10 s – 30s (10 s aeration and 30 s pause) to save more air under favourable conditions (Buer, 2010).

Presentation110Fig. 1. Improvements over continuous aeration obtained with intermittent aeration. Fouling characteristic obtained at an average aeration rate of 0.36 m3/m2/h. ZW500a module with 46.5 m2 surface area was used to filter 1.5 g/L bentonite solution (Guibert, 2002).

The intermittent aeration can be performed in two different ways: 1) on/off pulse between cassettes and 2) alternating between two groups of diffusers in a cassette. All the diffusers in one cassette are either on or off in the former method while only a half of diffusers are on in each moment in the latter method.

The experiment performed by Fulton (2010) reveals the comparison of shear stresses on membrane surface generated by the two intermittent aeration methods and the continuous aeration. Three commercial scale GE ZW500c modules were used in clean water at a gross air flow rate of 0.17 m3/m2/hr. As shown in Fig. 2, surface shear stress was directly measured by electrochemical sensors embedded in five different vertical columns parallel to the fiber bundles (c), four different levels (d), and three different vertical columns vertical to the fiber bundles (e) in the cassette. For intermittent aeration modes, two different aeration cycle times were tested at 3s-3s (fast) and 6s-6s (slow). In this experiment, surface shear stress should have a positive correlation with membrane performance since cake layer formation tends to be discouraged by higher surface shear stresses.

As shown in Table 1, average shear stress remained almost identical for all intermittent aeration modes regardless of the cycle time and the method of intermittent aeration. It appears that only the average air flow rates are important while the cycle time is not. Meanwhile average surface shear stress with continuous aeration was measured 30-40% higher than that in intermittent aeration modes. It is important to notice that average shear stress did not increase 100% although average air flow rate is 100% higher for continuous aeration than for intermittent aeration. If the same average shear stresses are desired in intermittent aeration, gross air flow rate must be raised somewhat. These observations suggest that the intermittent aeration at a 50% on time can save air demand by less than 50%, perhaps ~30%.

Other remarkable finding from the same experiment was that aeration pattern greatly affected the distribution of surface shear stresses along the fibers. The median shear stresses in the planes defined in Fig. 2c – Fig.2e are summarized in Table 2, where shear stresses are well distributed along the fiber in intermittent aeration modes while they are not in continuous aeration modes. In continuous aeration modes, shear stresses are much higher in the horizontal plane in the top (horizontal plane 1) and in the vertical plane in the center (Plane CFIL) than other matching planes. The median shear stress in the bottom plane (horizontal plane 4) was much lower than those in other three horizontal planes.

  • Median shear stress in the bottom plane (#4) was only 20% higher with a continuous aeration
  • Median shear stress in the top plane (#1) was 270% higher with continuous aeration
  • Median shear stress in the side plane (plane ADGJ) was only 20% higher with a continuous aeration
  • Median shear stress in the center plane (plane CFIL) was 100% higher with continuous aeration

The mal-distribution of shear stress causes inefficient air use since the fouling started from the low shear area can spread to adjacent area. Once the low shear area is fouled, the fluxes in the rest of the area must increase to compensate the loss, which eventually expedites membrane fouling. Therefore, the more evenly distributed shear stress obtained in intermittent aeration modes allows more efficient scouring air use.

image002asdfghjklFigure 2. Membrane system used to characterize surface shear stresses induced by sparging. (a) cassette frame with three membrane modules (only outer front module is visible), (b) locations where shear forces were measured on each vertical plane (rows ABC, DEF, GHI, and JKL at ‘y’ of 1.57, 1.08, 0.59, and 0.10m , respectively, and columns ADGJ, BEHK and CFIJ at ‘x’ of 0, 0.18 and 0.36m, respectively), (c) – (e) shaded vertical, horizontal and probe planes, respectively, in quarter of interest within the three membrane module cassette. Three ZW500c modules (GE) were used at an air flow of 0.17 m3/m2/hr (Fulton, 2010).

Table 1. Average surface shear stress at 90% confidence level at different aeration modes.

Sparging Conditions Average shear stress (Pa)
Slow alternating (3s-3s)

Slow pulse (6s-6s)



Fast alternating (3s-3s)

Fast pulse (6s-6s)



Continuous 1.41±0.18

Table 2. Comparison of the average median shear stresses at different planes in a membrane cassette

Location Continuous aeration

Median shear ( kg/m/s2)

Slow pulse aeration

Median shear ( kg/m/s2)

Horizontal plane 1 2.24 0.61
Horizontal plane 2 1.20 0.74
Horizontal plane 3 0.94 0.68
Horizontal plane 4 0.80 0.66
Plane ADGJ 0.63 0.52
Plane BEHK 1.20 0.73
Plane CFIL 1.66 0.83
Vertical plan 1 1.39 0.73
Vertical plan 2 1.67 0.47
Vertical plan 3 1.13 0.67
Vertical plan 4 1.43 0.78
Vertical plan 5 0.83 0.71


© Seong Hoon Yoon