Automated Flow Control Scientific Paper Rough Draft
The actual formated word version of the paper is available here, it is formatted after the papers in Water Research
Title: Linearize the Correlation between Head and Flow Rate for Use in Water Purification Plants
Abstract
Initial concepts for a non-electrical technique for the regulation of alum addition in a water treatment plant were tested. A riser pipe was fabricated with holes to mimic a sutro weir was added to the entrance tank of a water treatment plant. The entrance tank and riser pipe were evaluated on their ability to provide a linear correlation between water height in the entrance tank and the flow rate through the plant. Flow rates between 5 GPM and 36 GPM were tested in a pilot water treatment plant. The flow rate was shown to be linearly related to water height with the riser pipe implementation with a percent deviation of ten percent. This will allow creation of an alum flow control module regulated by a float system in the entrance tank. The experiment shows a significant achievement in the movement to bring clean water to communities without access to electricity.
1. Introduction
Adaptation of water purification technology for use in developing countries is a concern of growing urgency. The lack of clean drinking water provides avenues for infection by waterborne pathogens which are currently one of the major causes of mortality for children under five years old. The Agua Clara team is working to bring clean water to communities throughout Honduras with a flocculation-sedimentation water treatment design. The design requires the addition of aluminum hydroxide as a coagulation agent. Plant operators manually adjust the alum dose based on the daily flow rate of the treatment facility. A useful innovation is necessary in the regulation of aluminum hydroxide into the system as flow changes without the use of electrical devices. Currently options for the control of chemical additions to water treatment plants are either an electric pump or manual adjustment of chemical flows. The design proposed uses an entrance tank as a way to transfer information about the flow rate to the aluminum hydroxide flow control module through a float system. The system accurately delivers consistent aluminum hydroxide dosing into the incoming water based on the height of the water in the entrance tank. Without intervention the flow rate through the entrance tank is proportional to the square root of the pressure head, height of the water in the entrance tank, as shown in equation 1.
Equation 1: The Orifice Equation.
The relationship between water height and flow rate create a problem when transferring of data to other mechanisms. A predictable linear relationship between the height of the water in the entrance tank and the flow rate through the plant allows for a system of floats to meter the aluminum hydroxide flow rate. The sutro weir developed by Victor Sutro in 1915 can be used to create a linear correlation between head and flow rate. The configuration is shown below in figure 1.
Figure 1: The Sutro Weir Image.
The bottom width of the sutro weir is relatively large and the width gradually decreases as the height increases. As the height of the water in the tank increases the pressure created by the weight of the fluid causes the flow rate through the bottom of the weir to increase. By minimizing the weir width as height increases the change in the total flow through the weir as height increases is kept in a linear relationship. Implementation of sutro weirs in currently operating water treatment facilities is infeasible due to restrictions of shape and space. Grit chambers receive the inflow first and transmit the water to the rest of the plant through pipes. If a sutro weir shaped hole were cut into a pipe the pipe would become unstable and require skilled labor for construction. Therefore we are approximating the sutro weir with a riser pipe added onto the pipes that connect the initial grit chamber (entrance tank) and the flocculation tank. The pipe would be easily introduced into previously constructed plants and new plants at low cost. Holes would be drilled into the riser pipe that would mimic the sutro wier basics of design. The drilling of holes as specified heights doesn't require skilled labor. The water is forced to flow through the riser pipe in order to leave the entrance tank and enter the flocculation tank. Experimental tests were conducted to evaluate the accuracy of the simulated sutro weir in a pilot plant application.
2. Materials and Methods
Experimental set-up
The Sutro weir shape is dictated by a series of equations. Approximating the weir was achieved through a computer program in matlab that simultaneously solve the equations and designs a template for hole location and quantity on the riser pipe. The program requires input of maximum flow rate through the water treatment plant and then through iteration selects the largest drill bit size that is able to theoretically mimic the sutro weir with 10% accuracy. The large holes size is beneficial because obstructions are common in untreated water and large holes are less likely to clog. A clogged hole would produce an exaggerated water height and over estimate the flow rate. A riser pipe and entrance tank were designed for implementation in a pilot plant with a maximum flow rate of 120 L/min. The riser pipe was designed for a maximum flow rate of 125 L/min to allow for spikes in flow rate. The riser pipe is made out of three inch white PVC pipe with a consistent hole diameter of three eighths inches. Holes in the riser pipe were spaced over 20 cm and every centimeter was evaluated for necessary holes. The water from the plant flows into the entrance tank, a five gallon bucket, and through the riser pipe into the flocculation tank. In-line with the water inflow pipe a flow meter, siemens model SITRANS F M MAG 3100, is installed. The device provides flow measurements in GPM with an accuracy of +/-0.25% of flow rate.
Measurement of Water height in Entrance Tank as Flow Rate Fluctuates
The riser pipe design's accuracy was tested in Ithaca, NY. Before each trial the riser pipe is cleaned and all holes obstructions are removed. Performing the experiment requires reading the flow rate from the flow meter and measuring water height in the entrance tank with a ruler. Flow rates were varied between 5 GPM and 36 GPM by 1 GPM increments. The flow meter requires acclimation time to report accurate readings. After every manual change in flow rate there was a rest period of five minutes before water height and flow rate were measured. The recorded water height is adjusted by an offset of 3.5 inches. The offset accounts for the distance between the bottom of the bucket and the center of the first row of holes. Three trails were run on three different days.
3. Results and Analysis
The distribution of entrance tank water height was correlated directly with the flow rate through the pilot plant. The data from the three trails was plotted
on figure 2.
Figure 2: Experimental data on the three trials showing the correlation between the Flow rate and the water theight in the entrance tank. The trend line for trial one is y = 18.116x - 4.3486 with an R2 = 0.994, trial two is y = 17.527x - 2.7457 with an R2 = 0.993, trial three is y = 17.55x + 2.5598 with an R2 = 0.9974, Predicted is y = 15.814x + 0.3126 with an R2 = 0.9999.
The predicted data set from the Matlab program was also plotted on the graph to provide a comparison for the collected data. All of the collected data have trend lines with slopes of approximately 0.05 which shows high precision among the three trials. There is a deviance between the predicted values and the observed values as the flow rate increase above 90 L/min. The difference between the predicted and actual values is displayed in figure 3.
Figure 3: Plotted difference between experimental data and predicted values.
Further analysis on the percent deviation of the data shows a more significant difference at low flow rates. The percent difference levels out under ten percent as displayed in figure 4.
Figure 4: Plotted percent difference between experimental data and predicted values.
The difference between the observed and predicted values is within 10 percent for the majority of the flow rate range. The high percentage ratings in the low flow range are due to the impact of small errors on small values. The data supports the accuracy of the sutro weir approximation.
4. Conclusions
The experiments prove that the system of holes accurately simulates the sutro weir. The height readings with the entrance tank and riser pipe show error under 10 percent when in high flow rates. The trend in the data supports even lower errors as flow rate increases to flow rates of actual water treatment plants. The new riser pipe will allow for increased autonomy of the water treatment plant. Supplies for construction of the riser pipe are low cost and addition of the pipe in the grit chambers of new and existing water treatment facilities is a low labor endeavor. The communities only need to find a length of PVC pipe to improve their water purification plant. Daily monitoring of the riser pipe is necessary to remove obstructions from the holes, at least once daily. Expectations of constant monitoring of plant flow rate by the operator are unreasonable and without correct alum dosing the water either won't be purified or the plants will waste money on overdosing. The innovation utilizes advanced theory to improve the lives of the least fortunate.
5. Acknowledgements
This research was funded by the Agua Clara project through generous donations by the Rotary Foundation and the Brown Family. The help of Professor Weber-Shirk, Timothy Brock, Tom Cook, and Tom was integral.