Experiment 1: Average Effluent Turbidity vs Velocity - Low Floc Blanket Formation
During Experiment 1, the following capture velocities were used: 5 m/day, 10 m/day, 15 m/day and 20 m/day (convert to mm/s). This experiment was set on low floc blanket formation.
Figure 1: Average Effluent Turbidity vs. Capture Velocity
(Do you by chance have the datalog uploaded for this experiment?)
Conclusions
As expected, the effluent turbidity is much higher than that of the control experiment. The presence of saturated water in the influent has a negative effect on the effluent turbidity. In order to quantitatively examine the performance of the experiment with saturated water, data from the control ramp experiment was referenced. The average difference in error between the results was found to be 3.857 NTU. (What error and which results are you referring to on the graph? You can report standard deviation or relative standard deviation, but if you want to do this, I would suggest adding error bars for each point on the graph), with the effluent turbidity of the saturated water experiment generally higher in all cases than that of the control experiment. The bubbles released as a result of a pressure drop in the system disturbed floc formation, allowing less floc particles to settle and more, smaller, lighter particles to leave with the effluent. (Note floc break-up as a cause since you noted cloudy water from observation before)
Experiment 2: Average Effluent Turbidity vs Velocity - High Floc Blanket Formation
During Experiment 2, the following velocities were used: 5 m/day, 10 m/day, 15 m/day and 20 m/day. This experiment was set on high floc blanket formation.
Figure 2: Average Effluent Turbidity vs. Capture Velocity
Conclusion
Similar to the data collected for the saturated water experiment on low floc blanket formation, the effluent turbidity for the experiment on high floc blanket formation is generally higher than that of the control experiment. In order to quantitatively examine the performance of the experiment with saturated water, data from the control ramp experiment was referenced. The average difference in error between the results was found to be 0.173 NTU. (Again, we need to also quantify the standard deviation in the data points to see if this is significant. Perhaps adding error bars would be helpful on the graph.), with the effluent turbidity of the saturated water experiment generally higher than that of the control experiment. The difference isn't as great as that found in the experiment with the low floc blanket formation due to the intitially low effluent turbidity for this particular experiment. High floc blanket formation already results in an increase in effluent turbidity.
(I disagree with your conclusion here. I do not see a large statistical difference for influent water here. Can you speculate as to why there was less change here? Is it because the tube settler is already in the floc blanket and this experiment did not reflect acutal conditions because some air bubbles would go through the system but not perturb performance locally in the tube settler? It is possible that saturated air had little effect and there was a problem with another problem in the experiment causing the worsened performance at the beginning? Also, why was performance significantly worse for 5 m/day saturated compared to 10 m/day? Let's look at the data and add error bars to see if this is consistent with what we would expect or if there was a large pertubation over a small period of time.)