for information on sedimentation, please see Plate Settler Spacing Research Fall 2008.
Click below for the following image.
Original System
Photo of Whole System
The temperature controlled water tank that provides water to the system was originally sitting on the table top. The system was not getting enough water from controlled water source. The mixture tank would run out of water after the system was running for a long time. As a solution we decided to raise a tank to account for the head loss in the 3/8 in tube leading to the the mixture tank (labeled as bucket 2 in the schematic). After raising the water tank by fourteen inches, the mixture stop running low on water. Initially we performed a head loss calculation to determine the exact height the tank would need to be elevated, however the final height of 14 inches was more than enough and based on the materials available for elevating the tank(two cement blocks). From our observations and data, it was confirmed that raising the water tank was the right decision.
The second change was the addition of a rapid mix chamber. A small cylinder with 3 in diameter was used a base to coil a 1/4 in inner diameter tube. The addition of the rapid mix section was used to improve the mixing of the alum and dirty water prior to entering pump 3, which leads directly to the flocculator. Increasing the mixing allowed for more contact time between the alum and the suspended clay particles.
Aside from the additional cylinder, the flocculator was improved. To remove bubbles and prevent their formation, the flocculator was rotated into a vertical position. An in depth description of the design of the new flocculator can be found in the design parameters section (below).
Another change to the setup is the method used to control the floc blanket. Originally valve was used to control waste from 1/2 in ID tube connected to a bulk head approximately 20 cm from the top of the settle column. The would be partially or fully opened to regulate the height of the floc blanket. This method, however, was far from precise and the floc blanket level fluctuated. To end this fluctuation in the floc blanket level, a device called the flow control weir was attached. This flow control weir is joined by black tubing. Within this black tubing, the head loss occurred. By altering height of the flow control weir, the height of the floc blanket is shifted. Essentially, it had the same functions as the original method of opening the valve. However, this flow control weir controlled the height of the floc blanket better. This was proven by observations and data collected from each experiment. As a result, a stable floc blanket was achieved. It was also determined the floc blanket must be below the settling tubes for optimum effluent turbidity.
Apart from the floc blanket, the floc formation was improved by removing a mesh in the column. The initial setup had a mesh that was locked onto the top of the cone inside the column. For the improved setup, this mesh was removed to form larger flocs. Originally,as the floc formation improved, and larger flocs were formed, the mesh began to break up the larger flocs, which inhibited floc blanket formation. Without the presence of the mesh, it was observed that bigger flocs were formed.
A peristaltic pump was added to the system to allow independent alum flow control. The original concentration of the alum varied with the flow rate of the settling tubes. Even with the use of a smaller tubing size, the alum flow rate was relatively high and required lower concentrations. As a result, the alum stock was limited did not last more than 10 hours. This made it difficult to run the system for long periods of time. Instead for the sharing the same pump as the settling tubes, the alum stock utilized the additional pump. With this additional pump, the alum stock concentration was adjusted so that system could run for extended periods of time. The concentration was to 15 g/L for 4 liters tank. With this concentration and tank size, the alum stock could last for 1.68 days running at a flow rate of 1.67 mL/min.
In conclusion, the following major changes were made:
The use of a tube flocculator allows for a more controlled flocculation environment as compared to flocculation with baffles. The flocculation design parameters are based on creating optimal conditions for floc blanket formation. The flow rate through the system of the original experimental apparatus was 9.1 mL/s. The following parameters described the original flocculator.
These values were found using the following equations:
Velocity Gradient
The velocity gradient in a curved tube is given by (Mishra & Gupta 1979)
Where De is the nondimensional Dean Number and characterizes the effect of curvature on fluid flow:
Residence Time
Several flocculator designs were used in the trial runs of the experiment. The initial flocculator was oriented horizontally across the work bench. This orientation led to a numerous small bubbles trapped throughout the system. It also made the flocculator difficult to fill. The first change therefore, was to use a vertical flocculator orientation. This allowed bubbles in the flocculator to flow upwards towards the top where a tee and vertical tube open to the atmosphere released any bubbles in the flocculator.
Upon observation, it was clear that the flocs being produced in the flocculator were too small. It was originally thought that this was due to the high flow rate through the system. Again the flow rate through the flocculator is based on an upflow velocity of 100m/day for proper floc blanket formation in the settling column. The flow rate was then halved by adding another flocculator. Since the flow through pump 3 was already split this was adjustment was quite simple. Each of the split flows were led directly to a flocculator. The flow out of the flocculator was then joined into a single tube flowing into the settling column using a tee. After these adjustments floc formation was still less than satisfactory, with the average floc size in the flocculators observed as less than 1 mm. It was later noted that the combining of the two flows (each with a flow rate of approximately 272 mL/min) created shear valves far exceeding the average shear values in the flocculators (71.4 /s), which likely resulted in the breaking up of flocs formed in the two flocculators. It is important to note that the system was producing a floc blanket, however, it was not very dense and the effluent turbidities were still quite high.
The current flocculator set-up contains one continuous flocculator placed on two adjacent columns. The flow rate through the system remains at 545 mL/min, however, the length of the flocculator has increased to 27 m. As seen in the schematic above there are now two flocculator columns with the water spiraling upward to the top of the first flocculator where there is the first bubble release T-connector. The flow then crosses over to the top of the second column where it spirals downward reaching another bubble release t-joint right before heading into the bottom of the settling column. The increased residence time in the system allows for larger floc formation, and an increase in the total number flocs forming in the flocculator. Improved floc formation allows for the formation denser floc blankets.
Final flocculator parameters (see curved tube flocculator design.xmcd)
The high G value and long residence time helps create the large amount of flocs needed for the initial formation of the floc blanket. The flow rate through the flocculator (9.1 mL/s) is generated by two equal flows of 4.54 mL/s, based on the maximum flow through the peristaltic pumps.
There are still some improvements to be made to the flocculator. Suggestions for improvement are discussed in the future goals section
{float:right|border=2px solid black}
!Plate Settler Spacing Research Fall 2008^Uniformflow.png|width=250px!
{float} |
Uniform flow in the settling column is key for having laminar flow in the tube settlers as well as the creation of the floc blanket. As seen in the image to the right, a uniform flow is needed at the inlet of the settling tube so the flow becomes laminar before reaching the end of the tube. All design assumptions and parameters for the column are based on a laminar flow through the settling tubes.
Flow in to the system from the bottom of the settling column creates a jet in the center of the column, which is far from uniform flow. A typical dissipation ratio is 1:10; therefore, it would take on the order of 10 m for jet to dissipate in a column with a diameter 10 cm, as in the current setup. A cone with a ratio of 4:6 is used a diffuse to maximize the lateral spread of the jet. In addition to the cone, there is a mesh directly on top of the cone made of a 1 cm thick plastic sheet with uniformly distributed holes of 0.5 cm diameter. The purpose of the mesh is to break up the single jet coming from the column inlet into several smaller, weaker jets that dissipate quicker. The preliminary stages of the experiment tested the effectiveness of the cone and mesh system in dissipating the jet and allowing for optimal floc blanket formation and maintenance. The mesh was eventually removed due the breaking up of flocs. Further discussion about testing the effectiveness of the mesh can be found in the future work section.
Based on the allowable space in the settling column the floc blanket height will be no more than 60 cm. The alum dosage of 45 mg/L is also important for the creation of the floc blanket. In previous experiments it was thought that 45 mg/L was too high increasing the "stickiness" of the tube settlers and leading to clogging. The same chemical properties of the aluminum sulfate actually aid in the creation and maintenance of the floc blanket, creating a stick mesh of suspended flocs.
The goal of the plate settler spacing team is to develop a more complete understanding of the effects of plate settler spacing on the sedimentation process. We seek to purify turbid water as efficiently as possible by optimizing materials and minimizing costs. Currently, our experiments vary the inner-diameter of the tube settlers. Ultimately, we will explore the interdependency of floc blanket depth, flow rates, and tube settler inner diameter.
Mishra, P. and Gupta, S. N. (1979) Momentum Transfer in Curved Pipes. 1. Newtonian Fluids. Industrial & Engineering Chemistry Process Design and Development 18(1), 130-137.