...
The sedimentation rate velocity and the residual turbidity of flocculated suspension are important properties in a flocculator. The Spring 2009 team evaluated quantitatively the effect of shear velocity on these parameters. To do so, they used the flocculation residual turbidity meter (FRETA) developed by the AguaClara team and a data processor (MathCAD file) to analyze these parameters automatically.
Our goal for these first experiments was to familiarize ourselves with the apparatus and the data processor (MathCAD file) made by the previous team (Spring 2009) and to try to replicate one of their last experiments to make sure that the apparatus and the MathCAD file were working properly.
...
The experiment was conducted with the parameters showed shown in table 1. The parameters were based on one of Ian's experiment (data from 5/13/2009) conducted with the same parametersinputs, except for the flow rate varying rates which vary from 4 to 19 mL/s and the turbidity which was set at 50 NTU(see MathCAD file attached). Figure 3 shows the plot of the effluent turbidity , during the settling , state as a function of time and flow rate. The flow is, in fact, changing at each state cycle and we can see that the turbidity is decreasing at each settling state.
...
The figures 4,5,6,7,8,9,10, and 11 show the plots, given by the data processor, with the data from our experiment (9/24/09) and the data from the Spring 2009 experiment (5/13/09). Comparing graphs at each state, we can observe the same behavior on the result. The values are not the same on the graph because we did not run the experiment exactly in the same conditions but to the experiment conducted closely replicated one of the Spring 2009 experiments. Date smoothing and normalization are the first transformation observed similar results when compared with the previous years experiments. Our experimental values; seen on the graphs below, are not exactly the same as Ian's Spring 2009 data because the experiments were not performed under the same conditions, but they closely resembled one another. Data smoothing and normalization were the first transformations performed on the raw turbidity data. The smoothing allows excluding outliers, us to exclude outlying data points caused by really large flocs that cause important turbidity fluctuation passing in front of the light sensor of FRETA, which create turbidity fluctuations. The normalization allows comparing us to compare data sets with different varying influent turbidityturbidities. The settling velocity (Vs) was calculated by dividing the distance between the ball valve and the zone illuminated by the infrared LED of FRETA by the time elapsed. The plot of normalized turbidity vs. Vs can be interpreted as a cumulative distribution function (CFD) of turbidity with respect to Vs. A CFD describes the probability that a variable is less than or equal to some value. To make the analysis more robust, the experimental data was fit to a gamma distribution. Then a derivative of the CFD of the gamma distribution gives a probability distribution of the particle population with respect to Vs.
Figure 4 and 5 show the normalized effluent turbidity vs. time during the settling state for respectively Fall Spring 2009 team's experiment and Spring Fall 2009 team's experiment, respectively. In both casecases, the turbidity of the water decreases rapidly. BesideMoreover, we can observe that for low flow raterates, the turbidity decreases more rapidly and the residual turbidity is lower in both experiments. Figure 6 and 7 represent the normalized turbidity as a function of Vs. As Vs is the inverse of time, we observe the opposite trend in the data. Figure 8 and 9 show the cumulative distribution function and figure 10 and 11 show the resulting probability density function (PDF) of settling velocities obtained from the fitted gamma distribution for Fall 2009 experiment and Spring 2009 experiment. We can observe on the PDF curves that in both cases the mean velocities are decreasing with the flow rate (trace 1 being the lowest flow rate and trace 8 the highest). For more details on the data processor see Ian Tse's M.S. thesis (put a link).