Troubleshooting

Over the course of the semester, our team experienced many technical challenges that impaired our ramp state experimental results. We have outlined three major challenges and possible solutions below.

Initial Turbidity Peaks in Data

Problem

Analyzing the effluent turbidity vs. flow rate plots collected following ramp state experiments, a single sharp initial peak in effluent turbidity was observed for most trials regardless of the tube diameter used.

Proposed Solutions

Our team has proposed two reasons for this observed peak in turbidity. One possible reason is that the tube settlers in our apparatus do not begin "pulling up" water until the ramp state initiates. A second possible reason is that the floc blanket formation time set in the process controller method is not long enough to ensure that a thick enough floc blanket is formed before the ramp state initiates. Our team has proposed experimenting with increased floc blanket formation times, approximately 8 hours as opposed to the original 6 hours.

Influent Water and Clay Mixer

Problem

During the semester, our influent water-clay mixer malfunctioned and needed to be replaced. However, our team continued to proceed with experiments until a new mixer was ordered and received. Without the mixer, our system ran through our concentration clay stock solution at a much quicker rate in order to maintain an influent water-clay solution of 100 NTU without a mixer.

Proposed Solution

In the event that future teams experience a similar problem, we suggest assigning a team member to be responsible for monitoring the rate at which the system runs through the clay stock using process control software and/or adjusting the concentration of clay in the solution as necessary. Although the problem we experienced went away once we received a new mixer, assigning a team member to such a role will be helpful in the event that future teams decide to experiment with influent turbidities greater than 100 NTU.

Ramp State Time

Problem

In order to get a complete full run for an experiment, we decided to temporarily decrease the ramp time to two hours for our ramp state functions. This created a problem with the smaller diameter tubes because tubes with smaller diameters use lower flow rates in the ramp state function. Because these flow rates are so low, the residence time in both the tube settler and the turbidimeter becomes significant. (Can you show these calculations?)

Proposed Solution

When designing future experiments in process controller, it is important to consider these residence times when determining the ramp time. In order to determine the ramp state time that should be used (it should be minimized to reduce potential error with the apparatus), a successful experiment needs to be completed. We are currently working to collect data for a successful run. The ratio of the experiment time to the residence time in the tube settler and the effluent turbidimeter will be applied to the residence times for all plate settler spacings to determine the best ramp time to be used.