Introduction
It is key to know the true flow rate from our pump for our understanding of how our experimental model can be related to a true application of an inlet manifold in an AguaClara plant. In order to evaluate the flow rate inside our manifold, the experimental set up was modified so that the Acoustic Doppler Velocimeter (ADV) could be utilized. Two pumping models were evaluated, one where the pump was directly attached to the manifold and the other where it was not connected to the manifold, this is the vacuum pump setup used to increase the flow rate.
Set up
The pump was positioned at the end of the 10-ft long PVC pipe. The ADV was positioned 17 inches away from the end of the pump. Figure 1 shows the ADV positioning. Figure 2 shows the first pump set up (connected). Figure 3 shows the vacuum pump set up. The 10-ft PVC was used so that the flow would be fully developed at the point at which measurements were taken.
Figure 1: ADV positioning to measure manifold velocity profile
Figure 2: Setup with connected pump
Figure 3: Setup with vacuum pump, sometimes a string was needed so that the force of the jet didn't push the pump out of the end of the manifold.
Procedure
The procedure is identical for the two experimental models. The ADV was positioned at the top of the pipe. A 30 second velocity measurement was taken here. The ADV was then repositioned to be 1 cm lower than the previous location, and a 30 second velocity measurement was taken. The procedure was repeated until we got to the lower end of the manifold.
Results and discussion
The results for the first experiment (with the pump directly connected to the manifold) can be seen below in figure 4.
Unknown macro: {float}
Figure 4: Measured flow rate for the experimental model where the pump is connected to the manifold.
You can see that the velocity has an unexpected profile in this set up. There are two positive peaks and one large negative peak while the rest of the profile remains essentially zero. The reason for this is hypothesized to be due to the fact that the connected pump actually physically moves the PVC pipe back and forth. This would contribute to the negative velocity measured at 5.5 inches.
The results for the second experimental model (vacuum) can be seen below in figure 5.
Unknown macro: {float}
Figure 5: Measured flow rate for the experimental model where the pump is not connected to the manifold.
Here you can see that the velocity follows a more standard profile. There is a distinct peak and a relatively symmetrical pattern. By using a method of concentric rings and summing up the flow rates for each ring, we estimated the velocity to be 3.9 L/s. This is lower than the expected 6.7 L/s for the vacuum pump set up, but it is what we were able to derive using the data we collected.
The total flow rate was calculated by estimating the center of the flow and solving for the flow rates of the concentric rings where measurements were taken. The flow rates were found by averaging the 2 velocities recorded in each ring, then finding the area of that ring and calculating the flow rate. A visual of this method is located below in Figure 6.
Figure 6: The flow rate was tabulated for each ring and then summed up to calculated the total flow rate.
Conclusions
From analyzing the set up where the pump is actually connected to the pipe, we can see that the readings are not reliable and do not present any useful information. This may change if the manifold has more mass (achieved by adding on several more 10ft sections) or is better stabilized the next time this experiment is run. However, as it stands, the connected pump data cannot be reliably used to measure flow rates.
The vacuum pump setup, however, shows a decent velocity profile and a corresponding flow rate of 3.9 L/s. This flow rate is less than what we estimated in Experiment 3, in which we estimated the flow rate to be closer to 6.7 L/s. The advertised flow rate for the pump is 1 gal/s (3.8 L/s) so one would assume that the vacuum pump should provide a velocity greater than 3.8 L/s, but we hoped it would be greater than 3.9 L/s.
The results from this experiment and the results from Experiment 5 call into question the the mechanics of the ADV. It does not seem to be recording accurate velocities anymore so before more experiments continue, the ADV should be tested using a known velocity.
The Excel Sheet with the analysis is attached.