Reruns of the Previous Fluidized Bed Experiments

Parameters:

For both Experiment 1 and Experiment 2, the general procedure had been followed using values for parameters as listed and compared below:

{float:left|border=12px solid white|width="200"}
h5. Table 1: Comparison of Parameters Used.
||Parameters:||Experiment 1||Experiment 2||Control Experiment||
|Sand Grain Type|Sand 40|Sand 30|No Sand|
|Sand Grain Diameter|0.42 mm - 0.59 mm|0.59 mm - 0.84 mm|-|
|Sand Bed Depth|60 cm|60 cm|-|
|Sand Bed Expansion|50%|50%|-|
|Aerator Air Pressure|100 KPa|100 KPa|100 KPa|
|Flow Rate|225 mL/min|485 mL/min|530 mL/min|
{float}

Results and Discussion:

All three experiments were run for more than ten hours, during which time several data collection periods ("runs") were completed.

The water level in the bubble collector behaves like a periodic function. Each period represents a specific time interval during which the water level in the bubble collector gradually falls from its maximum to the set minimum point. In Figure 1. and Figure 2., this period is represented by each of the slanting lines on the graph. As soon as the minimum water level is reached, the bubble collector refills with water to begin the next run. For this reason, at the beginning of each run, the water outflow valve is closed until the water level reaches the set maximum point. These refilling periods are represented on the graph by the vertical lines. More detailed information on the bubble collector setup can be found here.

{float:left|border=10px solid white}[!figure 1.02.png|width="440", height="312"!|Gas Removal Preliminary Graphs]
{float}

{float:left|border=15px solid white}[!figure 2.02.png|width="432", height="307"!|Gas Removal Preliminary Graphs]
{float}

Once the change in water level in the bubble collector was recorded, we added a linear fit line to each of the runs to see the rate of change of the water level inside the bubble collector with respect to time. Figures 3. and 4. show the linear fit line for the second data collection period in both experiments, and more detailed graphs can also can be found here.

{float:left|border=15px solid white}[!figure 3.04.png|width="418", height="252"!|Gas Removal Preliminary Graphs]
{float}

{float:left|border=12px solid white}[!figure 4.02.png|width="404", height="254"!|Gas Removal Preliminary Graphs]
{float}

The value of the linear fit is very close to 1, indicating that the data can be modeled accurately using a linear relationship. Once the slope of the fitted line was known, we calculated the content of gas removed per liter of water sent through the sand filter using the formulas:

$$
\frac{\Delta Volume}{\Delta Time} = slope * \pi * r _{collector} ^2
$$

$$
\frac{mL\:gas\:removed}{L\:water\:treated} = \frac{\frac{\Delta Volume}{\Delta Time}}{Q_{water}}
$$

where the radius of the bubble column was 1.9 cm.

The calculations for the amount of gas removed during each data collection periods gave us the results summarized in Tables 2 and 3 below:

{float:left|border=12px solid white|width="200"}
h5. Table 2: Gas Removal vs. Collection Periods; Experiment 1.
||Run||Slope (cm/min)||R ^2^ value||Dissolved Gas Removed (mL/L)||
|2|0.1013|.9948|5.0909|
|3|0.0986|.9920|4.9397|
|4|0.0861|.9933|4.3348|
|5|0.0739|.9945|3.6795|
|6|0.0659|.9921|3.2763|
|7|0.0616|.9872|3.0747|
{float}

{float:left|border=12px solid white|width="200"}
h5. Table 3: Gas Removal vs. Collection Periods; Experiment 2.
||Run||Slope (cm/min)||R ^2^ value||Dissolved Gas Removed (mL/L)||
|2|0.0854|.9944|1.9970|
|3|0.0856|.9482|2.0017|
|4|0.0847|.9952|1.9806|
{float}

{float:left|border=12px solid white}[!figure 5.04.png|width="382", height="446"!|Gas Removal Preliminary Graphs]
{float}

For further reference, please download the experimental data logs from Experiment 1., Experiment 2., and Control Experiment.

The data from Experiment 1 is very similar to the results obtained last semester, when the measured content of gas removed was 5.07 mL/L. Yet the data log revealed a gradual decrease in the gas removal rate for each of the subsequent runs. It is likely that this resulted from a clogging problem in the sand filter. The believed cause of clogging in the sand filter is the relatively small diameter of the sand column. Large bubbles can form in the sand bed and push segments of sand up to the top of the filter. While we did not directly observe this problem during the experiment, sensor data collected through Process Controller suggests that clogging might have occurred.

Unlike in Experiment 1, the data from Experiment 2 shows consistent gas removal rate for each run. The majority of the experiment monitored visually and no clogging was observed. While the results were very consistent, the gas removal rate was still a bit lower than the results obtained last semester of 3.23 mL/L. The cause of the difference in results might be the experimental setup, which has been modified since last semester. For further observation, we measured the dissolved oxygen content at three sampling points in the system. More detailed information can be found here.

Additionally, these experimental results may be modeled as gas removal efficiency when subjected to two different sand grain sizes. Data suggests that the sand with larger media diameter might be less effective at gas removal. Particles with larger diameter have relatively lower surface area to volume ratio and thus provide less extra surface area to which the bubbles can adhere to in the sand column. Although the media size is just one of the factors that affect the gas removal rate, the results indicate that perhaps using particles with higher surface area to volume ratio might further optimize the conditions for dissolved air removal and thus facilitate the process under certain conditions.

The control experiment data showed that more amount of dissolved gas was removed in the absence of sand (7.47 mL/L vs. 5.09 mL/L for Sand 40 and 1.99 mL/L for Sand 30). As a result, sand might be inhibiting the gas removal as opposed to helping in the process.