You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 22 Next »

Unknown macro: {float}

Figure 1: Sedimentation Tanks at the Cuatro Communidades plant

Introduction and Objectives

The Cuatro Communidades plant has two sedimentation tanks designed for an upflow velocity of 70 m/day. The tanks are shallower than those of previously constructed Aguaclara plants. Like the other plants, the tanks use lamella to increase their efficiency. The sedimentation tanks were designed to accommodate a floc blanket. A floc blanket is suspension of flocs that are too large to rise to the top of the tank and two small to settle out. A floc blanket should act as a filter trapping flocs as they rise to the top of the sedimentation tanks. In the laboratory, floc blankets have been show to greatly increase effluent water quality in a flocculation-sedimentation system. The sedimentation tanks in the Cuatro Communidades plants are a much larger less controlled system and it was unclear whether it is possible to form and maintain a floc blanket in the plant.
No previous observations of the tanks showed conclusive evidence of a floc blanket. The Cuatro Communidades plant was designed with a plant flow rate of 380 liters per minute but the available head in the conduction line set a plant flow rate of only 190 -270 liters per minute. At these flow rates no effects of a floc blanket were observed and the upflow velocity may not have been great enough to keep flocs in suspension. One tank was shut off to increase the upflow velocity through the sedimentation tanks and to attempt to form a floc blanket.

Methods

It should be kept in mind that during these tests, the main purpose of the plant was to provide potable water for the communities. These tests were a lower priority and this created several initial challenges to shutting off one sedimentation tank. When all the flow passed through one tank, the head loss through the original effluent launders was high enough to flood the plant. New effluent launders were created to run one tank. They were designed with a head loss of four centimeters using the equation:

Unable to find DVI conversion log file.

where:

  • Q is the plant flow rate
  • A is the total area of the effluent launder orifices
  • K is a minor loss coefficient
  • h is the head loss

To increase the area and decrease the head loss the effluent launder holes were increased from 5/16 in to 7/16 in.

Another problem was that in the first three attempts to run one tank the effluent turbidity increased from below 10 NTU to above 20 NTU. At this point it was unacceptable to send the effluent to the distribution line and the water was sent to waste. This procedure could only be maintained for 2-3 hours during the day without draining the water storage tank for the communities. No improvement in effluent quality was observed in these trials. The plant was recently cleaned before the second and third of the three trials. During these initial trials the incoming turbidity was above 100 NTU.

In order to shut off the plant for a longer period of time, the fourth trial was run at night when the communities were not consuming water. The sedimentation tank was cleaned at 9 pm before the plant was started. The plant was successfully running by 12:00 pm with an influent turbidity of 40 NTU and effluent turbidity of 4 NTU. It should be noted that the three hours in between were spent cleaning the sedimentation tank and filling the plant. After thirty six hours, the lamella were removed from the tank to see if there was floc blanket beneath them. After the lamella were replaced, the plant continued to run with only one sedimentation tank for the next four days. The incoming turbidity was consistently between 20 and 40 NTU.

Results and Discussion

As was mentioned, in the first three attempts, the effluent turbidity spiked initially. The trials could not be run for longer than three hours without disrupting the communities water supply. In the third of these initial trials, it was hard to visually distinguish between the water at the end of the flocculator and the end of the sedimentation tank closest to the flocculator after an hour. The turbidity measured at the end of the flocculator was 119.1 NTU and in the back of the sedimentation tank the turbidity was 102.4 NTU indicating that flocs were directly rising as they entered the back of the sedimentation tanks. However, close to the effluent channel the turbidity measured in the sedimentation tank was 27.12 NTU indicating that most of the flocs were rising in the back of the tank. This third test was only one and a half hours and the effluent quality from the sedimentation tank was consistently decreasing. It is unclear whether this upflow velocity was too high for the tanks or if, given more time, a floc blanket would have formed.

In the fourth test, there was no spike in effluent as had been expected from the first three trials. The plant was run for thirty six hours with no significant changes in effluent quality (figure 2). No evidence of a floc blanket was seen in the plant sludge judge.

Unknown macro: {float}

Figure 2

Although the plant was not monitored for the second night, it did not rain, no changes in the plant conditions were observed in the morning and the influent conditions were constant before and after the test. It is reasonable to assume that the influent and effluent did not significantly change in the night.
Finally to determine if there was a floc blanket, water was siphoned from different heights in the sedimentation tank. The turbidity was consistently below 12 NTU below the lamella. However the exact height of the hose was never determined because the hose was slightly buoyant and could not be seen below the lamella. If the test if repeated, the hose should be weighted down or attached to a rigid pole to measure the turbidity at an exact depth.

Unknown macro: {float}

Figure 3: half the lamella were removed from the sedimentation tank

After thirty six hours, the lamella were removed from the tanks (figure 3). Other than the initial disturbance of removing the lamella from the tank the water was fairly clear. It was possible to see to the top of the plates resting in the bottom of the sedimentation tanks. Flocs that had settled in the channel entering the sedimentation tanks were swept into the tank in an attempt to see where they were going. It was not possible to see what happened to these flocs. However, in general it seemed as though more flocs were rising in the middle of the sedimentation tanks than the sides. When the lamella were replaced and one tank was shut off for the following four days, there was no measured improvement in effluent quality. During this period of time the plant influent was consistently below 40 NTU and the flocculator was working well.

These results indicate that flocs were not reaching the sedimentation tanks but settling out in the flocculator, and the channel in between the flocculator and the sedimentation tanks during the thirty six hour test. Figure 4 shows the sludge build up underneath the plates in the sedimentation tanks after a month without cleaning. Flocs were clearly reaching the sedimentation tanks when the incoming turbidity was higher in the first few trials. It was unclear whether or not a floc blanket could have formed at high incoming turbidities. One tank should be shut off at night when higher turbidities are coming to the plant to see if a floc blanket can formed in a system as complex as the Cuatro Communidades sedimentation tanks given more time.

Unknown macro: {float}

figure 4: sludge settled below the plates in the bottom of the sed tank

However, if a floc blanket can be formed at this plant given ideal conditions, it may unreasonable to expect that one can be maintained at this plant. The amount of flocs reaching the sedimentation tanks would need to be increased to ensure that the floc blanket could be sustained at low incoming turbidities. Also the plant often needs to be shut off for cleaning or minor malfunctions, at this point the floc blanket would collapse. If the effluent turbidity must initially spike before improving, the first three attempts to run one tank show that it takes longer for the floc blanket to form than the residence time of the communities' water storage tank and their water demand allows.

  • No labels