h1. Experiment 2: March 2, 2010
h2. Procedure
Our manifold setup was the same as in Experiment 1, but the way we took our data changed. Our experiment consisted of a very simple setup.
The manifold we designed is a 10' long, 6" PVC pipe with 1" diameter holes drilled every 5cm. The manifold had water pumped through it at a rate of 3.8 L/sec (roughly 1 gallon/min) and the water flows through a whole 10' section of 6" PVC pipe before it gets to the manifold to ensure that the effects of the pump have dissipated in the pipe. The manifold is suspended 14" above the bed of the flume by U-clamps and the manifold is spaced 7" from the flume wall to make sure that it runs straight in the flume. The ports of the manifold are positioned so that the jets exiting from them run parallel to the bottom of the tank.
The ADV used to take velocity readings was mounted to a beam running across the width of the flume. The ADV was positioned so that it was aimed head on into the ports (so it also lies parallel to the bed of the flume) at a fixed distance of 17 cm from the port openings.
The measurements were taken atevery 4 different points along the manifold, separated into at close to fourths as possible given the interference of bolts protruding from the walls of the flume5-6 ports, which gave us 10 different data points along the manifold. For each port, we maneuvered the ADV into the frontedge of eachthe port untilhole. We then took measurements as we thoughtmoved wethe wereADV inacross the theport peakin portionsteps of the flow0.5cm. We recorded data for approximately 1 minute and then moved the ADV 1 cm to the left and 1 cm to the right of our first recording point to ensure that we 0.5cm further and measured again until we were sure we had captured the peakentire flow. We collected data at these points for 1 minute alsojet profile.
{float}
!VelocityProfilePort52VelocityProfilePort3.png!
Example graph of a velocity profile across one of the ports
{float}
In the analysis of our data, we took the mean of the velocities at each port for all 3 (and sometimes 4) measurements. Then we plotted the velocity profile for each port, assuming a Gaussian profile, and estimated the maximum flow rate at each port. These calculations were than plotted along the length of the manifold to give a velocity profile for the uniform manifold setup.
h2. Results & Discussion
The results of our firstsecond experiment forseem a uniform manifold were not what we expected. Due to the expectation of pressure recovery dominating major losses (friction inside the manifold) we had expected the velocity coming out of the ports to actually increase along the length of the manifold. However, once the maximum velocity for each port was plotted against its distance down the manifold (see graph) it seemed that just the opposite trend was true. The velocity appeared to have peaked early on in the to reaffirm the results that we found in the first experiment. The flow starts low then peaks in the first quarter of our manifold and then gradually decreaseddecreases after that.
!Experiment1ManifoldProfileExp2.png!
WeThe good reasonednews thatwith somethese sortresults ofis headlossthat mustwe befeel dominatelike overthe pressureflow recovery,is butsufficiently afteruniform, discussiononly with Monroevarying +-0.05m/s, wefor determinedit thatto ourwork procedurein didthe notAguaClara giveplants. usIt thealso mostachieves accuratethe datagoal andof wenot woulddropping needbelow tothe makescour changesvelocity (see Experiment 2). We also realized that we wanted to get more data points alongof 0.15m/s, so we can be confident that flocs will not settle out in the manifold. inWe orderstill todo seenot ifunderstand the trend we got with our first set of data was accuratefluid mechanics of what is happening in the manifold and need to investigate that further. | 