Recirculation of flocs in settling column

Unknown macro: {toggle-cloak}

Concerns

The Tube Floc team observed a slow but seemingly interminable recirculation of flocs inside the settling column during the settling state of each experimental run. This recirculation was troubling because it violated our assumption that the fluid inside the settling column was quiescent enough to allow flocs to be accelerated only by gravity. Since the experimental apparatus was designed to observe floc sedimentation inside the settling column, if flocs were actually being accelerated by a recirculating flow, the data we had collected cannot be interpreted as information about floc sedimentation. This recirculation problem may be a reason for why we have not been seeing ideal settling curves as we would have expected in our results. Changes need to be made to assure that the settling column is quiescent during the settle state, or a new approach must be taken to interpret the data.

Unknown macro: {toggle-cloak}

Possible Causes

Upon identifying the recirculation problem and bringing it up with Dr. Weber-Shirk, the Tube Floc team attempted to identify the cause of the recirculation problem. After weeks of diagnosing the experimental apparatus, the team was able to identify several potential causes for the recirculation:

1. The 90 degree elbow connectors at each end of the settling column was originally thought to be the cause of the circulation because such a sharp turn was bound to introduce some vorticity into the flow. Although the recirculation occurred after the mean flow has been stopped, the redirection of momentum caused by the 90 degree bend may still have lingered and thus could have contributed to the recirculation.
2. The connection between the 3/8 inch diameter flocculator tube and the 1/2 inch settling column has always been a source of concern. The abrupt change from small to big diameter should theoretically cause flow instabilities that could lead to the formation of turbulent eddies. Moreover, the jaggedness of the inner walls of the pipe connectors will also interrupt the stability of the flow and compound the problem. A solution to this problem would be to fabricate a cone shaped connector that could provide a smooth and very gradual transition from small to large diameter. This connector would need to have a small enough angle to the axis of the column to ensure that the adverse pressure gradient created by the increase in diameter is weak enough that it will not cause flow separation at the boundary of the pipe.
3. We noticed that the recirculation began the moment the valves were shut off. Since the solenoid valves close very abruptly, a small volume of fluid is displaced in an extremely short period of time each time a valve closes. This pulse of momentum may be strong enough that it traveled through the tubes and into the settling column where it may create instabilities leading to vorticity.
4. The tungsten lamp that generates the white light inside the turbidimeter radiates a significant amount of heat to the settling column. Some summer students had performed a heat test which concluded that the heat generated by the lamp did not affect the turbidimeter's ability to measure turbidity accurately; however, because the heat tests had been performed using mixtures containing very fine clay particles, the effects of density driven convection was not apparent in the turbidity readings. If the heat transfer between the lamp and water was large enough, the change in the local density of fluid would cause convection currents that could potentially create the type of circulation we have been observing.

Unknown macro: {toggle-cloak}

Visual Experimentation

Several visual experiments were conducted to determine what the actual cause of the reciruclation was from. First, we conducted experiments changing the pump control shutoff, then we experimented with the changing diameter of the elbow to the column, and lastly, we experimented with the heat factor.

Unknown macro: {toggle-cloak}

Pump Control Shutoff

We ran some tests to gauge what contribution the sudden stoppage of flow had on the recirculation problem. First, we introduced a new Process Controller state between the Loading state and the Settle state that allowed the pumps to ramp down before shutting off. We hypothesized that a more gradual reduction in flow rate would provide an improvement over abrupt shut-off. We also wanted to see the effects of shutting the valves verses not shutting the valves, so we ran some test without shutting the valves.

After careful observation, we concluded that the recirculation was occurring even without the valve shut-off (about 20-30 seconds into settling state.) We did observe a slight reduction in the intensity of the circulation with the increase in ramp down time, but circulation was still observed after a ramp down time of 60 seconds. The instabilities maybe a result of either the pump shutoff causing a "push/pull" in the water (even though the pump is being ramped down, there is still a slight pressure change when the pump is completely shutoff) or from the diameter change from the flocculator tube to the settling column causing eddies of turbulence in the top of the column.

Unknown macro: {toggle-cloak}

Elbow Connections

Another possible source of recirculation in the settling column is the elbow connections. The settling column is bounded by both an elbow connection coming into the column and going out of the column. Not only do these connections cause a sharp turn in flow, but there is a diameter change between the elbow and the column. The change in diameter could be causing preferential flow in the column, therefore, a recirculation current.

The recirculation occurs over the entire diameter length scale of the column; therefore, if these large eddies could be made smaller, either viscosity will dampen the smaller eddies out or the recirculation would be confined to small scales. We tried to implement a technique similar to the use of honeycomb screens in wind-tunnels to homogenize flow by placing a bundle of plastic drinking straws that had been cut to 2.5 inches in the upper portion of the settling column where the column attaches to the elbow connection. This, the team thought, would help to reduce the preferential flow pattern that could be caused by the change in diameter. After placing the straws into the settling column, we observed that the recirculation currents were reduced. It is important to note for the next experiment that the majority of these elbow experiments were conducted with the settling column outside the turbidimeter.

Unknown macro: {toggle-cloak}

Heat Induced Currents

Next, it was hypothesized that the heat produced from the tungsten lightbulb in the turbidimeter may be causing the recirculation. Comparison experiments were conducted with the settling column outside the turbidimeter, where heat from the lamp would not be an issue, and inside the turbidimeter, where the heat could have an affect on the column. When the column was placed outside the turbidimeter, little to no recirculation was observed. However, recirculation currents were clearly visible when the column was inside the turbidimeter,

Figure. Flocs in the tube showing the ideal condition when there is no heat is applied (left) and when heat is applied to the settling tube (right).

In order to ensure that heat may be the determining factor, the team wanted to test whether another type of heat source would cause recirculation. The settling column was placed outside the turbidimeter but when the experiment came to the settling state, a team member wrapped their hand around one part of the column to provide a heat source. Once again, recirculation was visible and the team determined that heat was definitely the major contributing source to recirculation in the settling column.

Unknown macro: {toggle-cloak}

Conclusions

While the recirculation problem in the settling column was finally treated by removing the heat source from the column, the team believes that the problem could have been caused by several different sources and the culmination of all of these experiments probably helped in solving the recirculation currents.

To remove the heat source, the team is working with HF Scientific to replace the IR lamp with a non-heat producing LED lamp. Additionally, the team has replaced the elbow connections with straight connections. In order to be able to use these straight connections, the team needs to figure out a way to stop the flow traveling into the settling column after the settle state has begun, in order to prevent watching the settling of a long vertical tube besides just the sample in the settling column. By introducing a slanted tube with a simple trap before the tube inlet, the team has managed to collect settling flocs in the tube from the top of the board during no-flow state.

Also, the team will determine whether or not the straws are helping to reduce the recirculation problem even after the straight connections have been added and the heat source is removed, or if they are unnecessary.