Page History

...

Figure 2.

...

Process flow diagram of experimental setup.

Procedure

Process Controller, a software based on LabView and written by Dr. Monroe Weber-Shirk, is used to automate the experiment and record the data. The automation routine, written by James Leung, controls the power to the stirrers and the speed of the pumps, calculated from the plant flow and residence time parameters. All other parameters, such as alum dose and pump tubing sizes, are fixed and preset in the routine. Process Controller also records turbidity data at 5 s intervals to a tab-delimited text file.

...

The demo plant was designed for a plant flow of 100 mL/min, with a residence time of 650 s. The above test was run for plant flows between 25 mL/min and 175 mL/min, at 25 mL/min intervals.

Results & Discussion

The graph Figure 3 below shows the effluent turbidities for various plant flows, plotted against dimensionless time, which is real time divided by the residence time of the plant at that plant flow. This is done for ease of comparison between different experimental flow rates. The effluent turbidities plotted are the average effluent turbidity values obtained through several experimental trials.

Figure 3. Average effluent turbidity

...

vs. number of residence times elapsed

...

.

Most of the plots show a large spike in effluent turbidity near the beginning of each experiment. This is due to the experimental procedure and can be ignored. Between each experiment, the water in the turbidimeter was not cleaned out, temporarily resulting in an artificially raised turbidity at the beginning of each experiment. Thus, these spikes can be neglected in our analysis of the results.

The graphs show that as plant flow rate decreases, the effluent turbidity also decreases, until we reach 50 mL/min. At 25 mL/min, we found observed that the effluent turbidity is unstable, but not lower than that of 50 mL/min.

From Figure 3, it was determined that running the plant at 50 mL/min resulted in the lowest effluent steady-state turbidity. The empirical optimal plant flow is much lower than the 100 mL/min design flow. This is most likely due to design flaws in the sedimentation tank. We thought that the parallel lamellas design would equalize the flow paths, and thus the flow rates, in each lamella, but this was not the case. We still observed faster flow rates through some of the lamellas, resulting in a settling time that was shorter than that required in the design. Thus, when plant flow rate was decreased, resulting in a longer settling time, the flocs had a more adequate amount of time to settle out, and effluent turbidity decreased. In addition, at flow rates less than 50 mL/min the degree of mixing was insufficient to produce sizable flocs; at greater flow rates, especially above 100 mL/min, the maximum shear attained in the flocculator exceeds the shear limit of the flocs, leading to floc break-up. Both these cases lead to small flocs exiting the sedimentation tank without settling, thus increasing turbidity.

Figure 4 below shows the total degree of mixing and the degree of mixing contributed by the 180° bends only, against plant flow. It also shows the maximum shear attained at the 180° bends and the average shear throughout the plant, againt plant flow. All values are calculated using a slightly modified version of the flocculator design program.

Image Added

Figure 4. Degree of mixing & peak shear vs. plant flow.

The graph shows an increase in the degree of mixing and shear with increasing flow rate. It is also important to note that the degree of mixing in the Demo Plant flocculator depends on both the loss in the 180° bend sections as well as in the vertical sections. This is due to the fact that the flocculator channels are very narrow compared to those in the actual plant. Thus, the frictional interactions with the channel walls are significant and cannot be neglected.

To test if the lamellas in the sedimentation tank experienced very non-uniform flow rates, we added red dye to the last channel of the flocculator, and observed its progress through the sedimentation tank. We saw that the red dye proceeded mostly through the last 3 lamellas of each side of the sedimentation tank, and that it moved the fastest through the last ones. Also, upon further observation, we saw that at 100 mL/min, medium and small sized flocs were being carried up these channels of the sedimentation tank to the effluent turbidity meter.

Simply by observing the Demo Plant as experiments were being run, we noted we noted that the largest flocs were created at a flow rate of 50 mL/min. These results also show us that the degree of mixing provided by the flocculator is sufficient at 50 mL/min to create very large flocs. However, at 25 mL/min, the flocs were very small, showing that the lower the lower shear had reduced the degree of mixing so much that the increased residence time could not compensate for it.

...

From the data acquired and our direct observations, we conclude that the optimal flow rate of this flocculator and sedimentation tank performance of the Demo Plant improves with lower plant flow. The performance limit of the plant is approximately 50 mL/min. This is much lower than the 100 mL/min design flow rate. This is most likely due to slight inaccuracies in construction of the sedimentation tank that caused the flow rates in each of the lamellas to be different, resulting in a settling time that was lower than expected. Thus, when plant flow rate was decreased, resulting in a longer settling time, the flocs had a more reasonable amount of time to settle out, and effluent turbidity decreased, below which there is no decrease in effluent turbidity. We hypothesize that above this flow limit, the shear is too high and therefore causes floc break up; below this limit the flocs don't fully form. With small floc size in both extremes, proper floc settling in the Sedimentation Tank does not occur, yielding a higher output turbidity.

In addition, it is possible that the uneven flow in the Sedimentation Tank provides significantly less settling time than predicted. This effect is magnified at higher flow rates and thus further increases effluent turbidity.

Finally, the experimental calculations indicate that the frictional shear in the vertical sections of the flocculator contributes significantly to flocculation, and cannot be neglected at this small scale.