Changing the flocculator
Wiki Markup |
---|
{composition-setup}
cloak.toggle.exclusive=false
{composition-setup} |
Wiki Markup |
---|
{toggle-cloak:id=Concerns} |
Concerns
Wiki Markup |
---|
{cloak:id=Concerns} |
The Tube Floc research team has known for a while now that sedimentation was occurring within the flocculator and that such premature sedimentation was bad. The switch from a horizontally coiled flocculator to a vertically coiled one was made partly to try to overcome this issue. Throughout the Fall 2007 semester, the Lab Floc team performed numerous experiments with the flocculator apparatus despite not having completely eliminated sedimentation. It seemed, at the time, that the sedimentation accrued per run was minimal and that it would not contribute a significant amount of error into our measurements. Further investigation by the Lab Floc team this semester, however, has given new light into the significance of this premature sedimentation as well as a new problem of fluid circulation inside the settling column immediate after the termination of flow through the apparatus. These issues are significant not only because they are unexpected and are not optimal experimental conditions, but they also put into question the validity of the data obtained last semester and the method by which we have been analyzing and describing these past results. Below are more detailed discussions of the different aspects of these problems.
Wiki Markup |
---|
{toggle-cloak:id=Issues with premature sedimentation} |
Issues with premature sedimentation
Wiki Markup |
---|
{cloak:id=Issues with premature sedimentation} |
We have hypothesized that because our experimental apparatus is currently limited to generating Poiseuille flows, there are no transverse velocities with respect to the streamlines of the flow strong enough to keep large flocs in suspension once they are large enough to settle out. Therefore, large flocs are settling out on the bottom walls of the tubes where the streamline velocity is at a minimum and where there are negligible velocities normal to the wall to scour the flocs back towards the center of the pipe. There are several reasons why this is unwanted:
- If flocs are settling out, they are not making it out of the flocculator and we loose information about the performance of our flocculator. Since we are interested in how shear/strain affects floc size and the final clarity of the water, we need the flocs to stay in suspension so we can measure how well they were formed. With significant premature sedimentation, the measurements of final settling effluent turbidity clearly underestimates flocculation performance.
- Not only does premature settling underestimate flocculation, but it also makes determining how parameters like flow rate (and other measurements of shear and strain) have affected flocculation. For instance, it is nearly impossible to determine for certain whether the steeper settling curves at higher flow rates was a result of larger flocs being formed or whether the higher flow rate just kept more flocs in suspension and thus was able to settle out faster.
- The efficiency of the flocculator is compromised when big flocs settle out because the primary mechanism by which fine particles are removed from solution is adsorption onto large flocs. The poor final turbidity observed in our experiments may be due to too many large flocs settling out of the flow before they reach the end of the flocculator.
Wiki Markup |
---|
{toggle-cloak:id=Options for fixing premature sedimentation} |
Options for fixing premature sedimentation
Wiki Markup |
---|
{cloak:id=Options for fixing premature sedimentation} |
A new observation: While sedimentation may happen during the loading state, much of the settling in the flocculator is permitted to happen during the sedimentation state, when the flow is shut off to allow sedimentation in the settling column. Since there are flocs present in the flocculator during this settle state (which usually as a 10 minute duration,) those flocs will settle out into the bottom walls of the flocculator. Additionally, since the flow during the clean flocculator phase is not turbulent, the majority of settled flocs do not get resuspended into the flow and washed out of the flocculator. As these sediments continue to build up over each experimental run, it is difficult to determine if the bulk of the sedimentation occurs during loading or is simply residual flocs that have settled there during the settle state and which the flow was unable to clean out. The Lab Floc team will make a few apparatus adjustments to see if we can eliminate the amount of sediment present in the flocculator by redirecting the flow around the setting chamber during the settle state and commence the flushing-out of the flocculator at the same time as settling inside the column occurs. This may elucidate when the majority of the flocs are settling.
- Increasing the flow rate of our experimental apparatus so that we are generating turbulent flow inside the tube flocculator. While this option may reduce the amount of flocs settling out of suspension, the increase in shear stress introduced by turbulence could disrupt the formation of flocs. First, we must determine if our current apparatus can generate turbulent pipe flow. Additional pump heads and/or increasing tubing size may be required to get the Reynolds number of the flow high enough to transition into turbulence. Once the ability to generate turbulent flow is achieved, we must evaluate whether we can achieve good flocculation. An additional challenge of operating in the turbulent regime is that shear, strain and G are now not so easily calculated and one can no longer assume these parameters are constant throughout the flocculator. On the other hand, the effectiveness of laminar flow to generate flocs has also been called into question. In a perfectly straight Poiseuille flow, the parabolic velocity profile requires that the gradient of velocity to be zero at the centerline, which means that if a particle were to travel along the centerline, it will have limited opportunity to collide with other particles because there is no gradient at the centerline. By having a coiled tube flocculator, this issue may have been ameliorated by the secondary flows generated by the centrifugal forces that disrupt the symmetry of the flow. This option, however, is the easiest option to undertake and should be explored first.
- Minimizing the amount of horizontally oriented tube sections should minimize settling because there will be no wall against which flocs could settle onto. This, however, would entail orienting the tube straight up and down, which would limit the length of our flocculator. If the tube were wrapped around a long racetrack-with the bends occurring at the top and bottom of the loops being much shorter than the vertical sides, we may be able to see some improvements. While this setup may reduce floc settling out of the flow, it brings into question whether laminar straight pipe flow is the right regime to study flocculation. As mentioned earlier, there are some theoretical arguments against flocculating in Poiseuille flow. This problem could be aided by the addition of some kinks in the pipe so that minor losses can introduce horizontal velocity components which will transport particles to different radial locations. More calculations are needed to evaluate this option.
- A third and most difficult option would be to rotate the entire flocculator in such a way that the direction of gravitational acceleration will always be transient with respect to the centerline axis of the tube. Since it is gravity that pulls floc out of the flow and down onto the wall oriented below the centerline of the pipe, by varying the angle of the vector of gravitational acceleration, sedimentation might be avoided. This option is the most complicated and hardest to implement simply because it requires new mechanical equipment and non-trivial design calculations.
Wiki Markup |
---|
{toggle-cloak:id=Implementing a new setup} |
Implementing a new setup
Wiki Markup |
---|
{cloak:id=Implementing a new setup} |
In order to prevent premature sedimentation in the flocculator, we have changed the flocculator setup. The flocculator hollow PVC tube (black) has been replaced with a peg board by which the team can plug in pegs to wrap the flocculator tube around. The board is 75 inches tall by 35 inches wide and is supported by a wooden brace to stand vertical.
The flocculator tube wraps around 10 different pegs that form semi-circles on the vertical ends of the board. The 50 foot flocculator tube wraps around these pegs about three times between rapid mix and the settling column.
If the team wanted to shorten the vertical distance that the tube wraps around, it is an easy transition by just replacing the pegs to a different hole on the board. This new setup is convenient and provides a lot of different ways to achieve a vertical flocculator.
This new setup reduces the amount of horizontal pieces in the flocculator, and will hopefully reduce or remove the amount of settling altogether.
Wiki Markup |
---|
{toggle-cloak:id=Pictures of new setup} |
Pictures of new setup
Wiki Markup |
---|
{cloak:id=Pictures of new setup} |
![board-top.jpg](/plugins/servlet/confluence/placeholder/unknown-attachment?locale=en_US&version=2)
From Left to Right:
1) new tube flocculator setup on the board
2) Effluent Turbidimeter
3) The settling column with straight inlet and outlet (straws at the top)
4) Top of the new tube flocculator setup
Wiki Markup |
---|
{toggle-cloak:id=Data Analysis} |
Data Analysis
Wiki Markup |
---|
{cloak:id=Data Analysis} |
Data Analysis for the new flocculator setup can be viewed here.
Wiki Markup |
---|
{toggle-cloak:id=Changing the Apparatus Setup for Turbulent Flow} |
Changing the Apparatus Setup for Turbulent Flow
Wiki Markup |
---|
{cloak:id=Changing the Apparatus Setup for Turbulent Flow} |
In order to achieve turbulent flow in our flocculator setup, we needed to figure out if our pumps can handle very high flow rates to get high Reynolds number. For turbulent flow, our target Reynolds number is 2500, with flocculator inner diameter of 0.6cm. The calculation shows that the flow rate needs to be at least 11.8 mL/s for turbulent flow. (With our current setup, the max flow rate is 3.1 mL/s) Thus, one of our solutions was to add more pump heads, since the max flow rate of pump is limited by how fast a pump head motor can rotate.
In order to test out how many additional pump heads are needed, we added in new set points to Process Controller method - "3-Alum Number of Heads" and "3-Clay Number of Heads". We found that 2 alum pump heads, 2 clay pump heads, and 3 raw water pump heads are sufficient to provide our target flow rate of 11.8 mL/s.