Methods
To determine the effects of inner-diameter, flow rate, and floc blanket height on settling efficiency, we completed an experiment to vary these three parameters. Based on the two-inch lamella spacing currently used in AguaClara plants, we chose a range of diameters less than two-inches to push the lower-limit of lamella spacing. Similarly, the range of flow rates that we tested was based on the current capture velocity used in the plants. The final variable in this experiment, the floc blanket height, was set to both fully submerge the inlet of the tube settlers and to leave the inlets resting above the top of the floc blanket. Table 1, below, illustrates the parameters for the experiment.
| Wiki Markup |
|---|
{float:right|border=2px solid black|width=400}
{excel:file=^DataAnalysis_aguaclara.xls |sheet=Capture Velocity Table}
*Table 1. Parameters: The Test Critical Velocities for Each Tube Diameter*
{float} |
As described on our apparatus design page, this experiment used the process controller program to automate the system. The main process controller states for the experiment were floc blanket formation, two flow states for high and low floc blanket heights, a floc blanket equilibrium state, and two states devoted to incrementing the flow rates. Several other states, such as drain, were used for system maintenance.
Preliminary experiments were conducted to determine the appropriate time interval for the flow and increment states. These experiments consisted of running the system for an exorbitant length of time to identify the time frame necessary to reach constant effluent turbidity. Ultimately, we decided to form the floc blanket for three hours; run each flow rate for six hours; stop effluent flow for ten seconds between flow rates; and allow about twenty minutes for the floc blanket to grow from its low to high depths. The floc blanket was reformed each time a new diameter was tested.
Results
Let us begin by discussing the failure. As stated in the methods section above, this experiment primarily tested tube settlers of three diameters. That is, three diameters that did not exhibit utter failure under nearly every condition. In fact, tubes of 6.35mm 9.5 mm inner diameter were tested as well. These narrow tubes exhibited failure readily, as indicated in table 2, below. Certainly, the highlighted values far exceed the desired effluent turbidity range, most above the initial turbidity of 100 NTU.
| Wiki Markup |
|---|
{float:margin=50px|border=2px solid black|width=1000}
| ||6.35 mm low FB || 6.35 mm high FB ||9.5 mm low FB||9.5 mm high ||15.1 mm low||15.1 mm high||17.36 mm low|| 17.36, high||23.8 mm low||23.8 mm high||
| Average Effluent Turbidity NTU || 13.97 || 1039.04 | 0.11 | 1.20 | 0.27 |3.39 | 0.39 | 0.17 | 0.51 | 0.15
| || 26.67 || 759.96 | 0.11 | 0.13 | 0.22 | 0.15 | 0.25 | 0.33 | 0.17 | 0.15|
| || 97.13 || 506.92 | 0.13 | 0.18 | 0.30 | 0.30 | 0.40 | 0.25 | 0.23 | 0.23|
| || 11.31 || 221.05 | 0.14 || 202.74 | 0.51 | 0.35 | 0.56 | 0.62 | 0.18 | 0.32|
| || 193.14 || | 0.21 | | | | 0.67 | 1.08 | 0.30 | 0.58|
*Table 2. Results: Average Effluent Turbidities for the Varying Tube Diameters at low and High Floc Blanket Levels*
{float} |
Observation of the failing tubes revealed that the tubes would clog with flocs and then, eventually, the water flow would push the clog through to the effluent. This process of clogging led to a few interesting observations about the settling of flocs in tube settlers.
To begin, we observed a "rolling" movement of the floc as it traveled up the tube. This movement can be described by a quantitative drag analysis based on the varying velocity gradients within the settling tubes. Flocs in the smaller tubes are more likely do experience a greater drag force which can overpower the settling force due to gravity. From this drag analysis it was determined that velocity gradients vary with radius of tube and Vα. A limiting velocity gradient of 2.4 1/s was determined from results. Tubes with average velocity gradients above this number experienced failure.
The graph below displays data from one of one of these failures. The 6.35 mm tubes were run at four separate flow rates in a high floc blanket, and in every case the effluent turbidity is extremely high. Compare this graph to the larger tube size of 15 mm, again at a high floc blanket.
Graph 1. Effluent Turbidity vs. Time 6.35 mm
Graph 2. Effluent Turbidity vs. Time 15.0 mm
Results for the 6.35 mm tubes and the high floc blanket of the 9.5 mm tubes were omitted from the results graph due to their high average effluent turbidity values. The remaining discussions focuses on the tube and floc blanket setting that did not exhibit failure.
The wider tube settlers, as expected, provided slightly more sensitive data. That is, the effluent turbidity was measurably impacted by changes in flow rate, floc blanket height, and tube diameter. Table 2 and the Graph 3 below illustrate the average effluent turbidity for the three diameters described in the methods section above.
Graph 3. Effluent Turbidity vs. Critical velocity
As the results show, the second lowest capture velocity for each diameter tube resulted in the optimum settling efficiency at the low floc blanket heights. The results also show a general trend of increasing effluent turbidity once the capture velocity exceeds 11.0 m/day. it appears that the range of acceptable critical velocities is around 9 to 13 m/day. This indicates that the capture velocity designed to in the current plants is appropriate.
However, as the graph clearly illustrates, there are no well-defined trends between diameters and floc blanket height at the capture velocities where effluent turbidities are under ~0.5 NTU. Perhaps this is because the effluent turbidity is exceptionally sensitive to disturbances when it is at such a low level. Overall the average effluent turbidity values for the three largest tube diameters did not exceed 1.2 NTU. This small range of values indicates that it may be possible to build an efficient sedimentation tank with a small plate settler spacing.
Conclusions and Future Work
There is a trend of increasing average effluent turbidity with increasing critical velocity for each tube diameter. Based on these results, the optimal capture velocity is 9-13 m/day for the plate settler spacings tested in this experiment. Also notable, is that the range of average effluent turbidities is relatively small over the range of critical velocities tested.
As mentioned above the limiting average velocity gradient for a tube was determined to be 2.4 1/s. Additional analysis on the effects of drag on a floc must be conducted in order to confirm these results. The next set of experiments will determine whether maintaining geometric similitude by holding L/D constant will result in consistently low effluent turbidities.