cloak.toggle.exclusive=false

A new AguaClara drinking water treatment plant has been designed for the town of Támara, Honduras. The plant has a maximum flow rate of 740 liters per minute and features a vertical flocculation tank with one turn, three sedimentation tanks, and a new plant leveling tank.

Keywords: AguaClara, Design, Támara

AguaClara is a team of students from Cornell University who work to design sustainable water treatment plants in Honduras. The goal of the team is to design and disseminate water treatment plants globally that are easy to build and maintain, that are economical to operate, and that can be prepared using locally
available materials.

To date, two AguaClara water treatment plants have been built in Honduras. The first was built under the supervision of Fred Stottlemeyer in La 34, Honduras (Figure 1). This plant featured a horizontal hydraulic flocculator. The second AguaClara plant was built in Ojojona and its initial construction was completed in Fall 2006. The Ojojona plant (Figure 2) was designed by the AguaClara team, and Ted Segal completed the structural design. This was an experimental plant with both a vertical and horizontal flocculator.
null
Figure 1. La 34 AguaClara Plant.

null
Figure 2. Ojojona Plant.

The next AguaClara plant was planned to be built in the town of Moroceli, Honduras. However, due to transmission pipeline problems, design of this plant was delayed so that it would not sit idle as the transmission line is repaired. The next AguaClara plant will be built in Támara, Honduras.

The Támara design team was only responsible for the environmental engineering design consideration for this plant because Agua Para el Pueblo has hired a civil engineer to deal with all structural aspects.

The Támara plant was partially designed using programs created by the AguaClara team during previous semesters. In Fall 2006, Monroe Weber-Shirk's CEE 454 class created algorithms for the unit processes involved in the plant. In Spring 2007, these algorithms were combined into a Main Program that would begin to design a plant. The program is written in MathCAD, and accepts user inputs such as flow rate and tank width to calculate output design parameters such as tank length and baffle spacing. In Fall 2007, the CEE 454 class created new programs to design pipes and flow measurement structures. These programs were used by the Támara design team as well.

The team also made use of the automated drawing capabilities developed by the Fall 2007 automated design team. This allowed the team to produce AutoCAD commands for drawing of the sedimentation tank by inputting design constraints to a MathCAD program.

The Támara plant flow rate was determined using an initial town population of 5,500 people. The flow rate was based on the projected population and associated water demand in 20 years. Population was calculated using a linear growth model with a 3.5% growth rate. Water demand was estimated as 114 Liters per person per day. The minimum plant flow rate was initially estimated as half of the maximum flow rate. However, after discussion with John Erickson and Carol Serna, the AguaClara engineers in Honduras, it became clear that the town did not currently have a population of 5,500. At the time of design, 570 homes were hooked up to the transmission line. With an average of 6 people per house, this population was 3420 people. This would result in a water demand of 463 L/min in 20 years. The potential for 120 new connections in the near future exists.
However, the engineers in Honduras believe that source at Támara can provide at least 740 L/min. It was decided that 740 L/min should remain the maximum plant flow rate because the actual demand for water is uncertain, and the extra size of the plant would not be too costly. Initial baffle spacing in the flocculation tank will be designed in a way to ensure that the current flow rate, possibly as low as 270 L/ min could be handled.

The Támara plant layout has a few major changes from the layout of the previous Ojojona plant. 
 

1. The horizontal flocculator was removed.  Testing in Ojojona confirmed that the vertical flocculator worked well  The vertical flocculator is less costly to build because it does not need to be elevated, so the horizontal flocculator will no longer be used in AguaClara plants. 
2. At the request of plant operators in Ojojona, a "bodega" was added to the plant.  This walled-in, roofed area is to be used for chemical storage, mixing of chemicals, and possibly a cot for night-time operators.   
3. A plant leveling tank was added to replace the pipe elbow level control feature in Ojojona. 
4. Chemical barrels were separated.  The alum and chlorine barrels will be located on different tables to allow for open walkways without tubes stretching across them, and to allow the tables to be at different heights as necessary. 
5. Tanks will be built of brick.  APP is more familiar with brick, they believes it is less porous than concrete, and they can build more cheaply with brick. 
6. The plant will be more protected.  There are two possible scenarios. The first is to have a fence around the plant with a roof over the platform. The second is to have open walls that start at the platform and stretch partway up to a roof. These walls would eliminate the need for a railing around the plant or extra sunlight protection for thetanks. Overall, this decision and design will be left to the plant operators and engineers inHonduras.  
7. A catwalk may be added to span the flocculation and sedimentation tanks.  Because brick walls will be thin, a catwalk would allow for safer access to these tanks.  The AguaClara team envisions a sturdy, metal catwalk that could slide from one end of the tank to the other.  .The tank walls need to have a ledge inside the external walls of the building covering the plant so there is a place for the catwalk to rest.
   null
   Figure 3. Támara Plant Layout

The new plant layout will be ideal for scaling of future AguaClara plants.  As long as baffle materials remain available to allow for the same width of tanks, the width of the plant should not change.  When a new plant is designed, the only necessary change to the layout will be the length of the tanks.  This will allow for more rapid design, easier automated design, and more uniform future plants. Another important feature of the new layout is that many of the pipes will be under the platform.  While this may be somewhat more difficult to build, it allows for easier access with no obstructions to walkways. The area under the platform should be secured so that all of the valves and pipes are protected
 Although not shown in the above drawing, all four major tanks will have drains that leave from the entrance side of the tanks.  These pipes will go to a waste collection tank under the platform where the operator will be able to see the water that is leaving the plant.  In this way he may know when all the dirty water has been flushed out.  This will be especially useful when cleaning out the sludge from the bottom of the sedimentation tanks.
null
Figure 4. Támara Plant Layout, modified from a drawing by APP engineer Ing Serrano. Magenta pipes in sedimentation tank represent sedimentation tank inlet pipes. Green Pipes are for the sludge drainage.

A major part of designing an AguaClara plant is sizing the pipes in the plant. A pipe design program was created so that all pipes could be designed quickly and correctly. The program returns a pipe diameter given a length, allowable headloss, and number of elbows, tees, valves, etc. A table of the K values used and a table of input pipe sizes used can be found in the Appendix as Table 2 and Table 3.

The program uses functions defined in the Fluids Functions program developed in the Fall 2007 CEE 454 class. In order to find the proper diameter the program iteratively solves for the flow that can fit through a pipe of the given diameter with the given headloss until a large enough diameter is found.

Headloss values are calculated using the following set of equations.
Equation 1
Re = Reynolds Number, to determine if flow is turbulent or laminar
Q = Flow Rate
D = Pipe Diameter
¿ = Kinematic viscosity of water
Equation 2
Equation 3
f = Friction factor, equation for either turbulent or laminar flow
¿ = Roughness factor of PVC
Equation 4
Equation 5
hmajor = Major head loss
hminor = Minor head loss

Values for all constants used may be found in the Appendix

A simple new design is in place for the rapid mix.  It was decided that the mixing elements used in past designs were not necessary, and that if the rapid mix pipe from the grit chamber to the flocculation tank has at least 2 90° elbows enough mixing would occur.  This was determined using the rapid mix pipe parameters and determining the G¿ value for the rapid mix.  A G¿ value of at least 600 was desired.  The G¿ and G values were determined using the following set of equations.

Equation 6

Equation 7 

h_l = headloss in pipe
Assuming that major losses may be neglected:

Equation 8

V = Velocity in pipe = Q / (¿ D2/4)
L = Length of pipe

Equation 9

Equation10

This two 90° elbow minimum after alum addition is easily met.  The current design has three 90° elbows: the turn from vertical to horizontal along the underside of the platform (1), the turn down along the side of the flocculation tank (2), and the turn back to horizontal to release the water into the flocculation tank(3).  See figure number 5 for a depiction of these turns.  The first baffle in the flocculation tank will be resting on the bottom of the tank and will direct water upwards.  The pipe from the grit chamber will enter the flocculation tank near the bottom  The first baffle will be a bottom baffle so the flow must turn over it.  This means the pipe should enter near the bottom of the floc tank so the water will have further to travel.
 
null
Figure 5. Turns in Rapid Mix after alum addition. Alum addition occurs through the white pipe outside of Grit Chamber.

null
Figure 6. Grit Chamber with one riser pipe leading to the rapid mix. Picture from CEE 454 Homework 4, Plant Measurement written by Monroe Weber-Shirk, Fall 2007.

The pipes for the grit chamber and rapid mix were designed using a program developed by the CEE 454 class in Fall 2007. This flow measurement program designed a grit chamber with two riser pipes that have orifices to allow for flow measurement.

The two riser pipes in the grit chamber contain orifices that will cause the water level to build up above them. This means that the plant flow rate may be determined based on the height of water above these orifices. Then water will free fall for a space and go down through the rapid mix. The open air pipe above the rapid mix, with alum flow holes, will allow for alum addition.

For the rapid mix pipe a maximum headloss, ¿horifice, of 20 cm was set for a pipe with 4 total turns, including 2 turns after alum addition and two turns before alum addition. The diameter of the pipe was determined using methods discussed in the "Pipe Design" section of this paper.

The riser pipe design algorithm set the height of water above the orifices at the design flow rate to be 20 cm. The diameter of the orifices in the risers was set to 1.5 inches, the same value as in Ojojona, and the area of the orifices was limited to 0.3 times the area of the pipe to ensure that orifice flows did not interfere.

The total area of orifices needed was determined using the above constraints and the orifice equation below. The Korifice is 0.63.

Equation 11

The number of riser pipes needed was determined from the total orifice area needed, the area of the pipe, and the ratio of pipe to orifice area. Finally, the total number of orifices was determined by dividing the total area of orifices by the area of one orifice and rounding up.

The footprint of the grit chamber was set to the same size of the plant leveling tank. The height of the walls was set to the depth of water in the grit chamber, as detailed in the "Water Levels" section of this paper, plus 10 cm of freeboard. This grit chamber tank will sit on the platform of the plant. The final parameters for the grit chamber are below.

¿ Height of Water Above Orifices: 20 cm
¿ Head Loss through Rapid Mix Pipe = 20 cm
¿ Total Number of Rapid Mix Elbows: 4
¿ Number of Rapid Mix Elbows After Alum Addition: 2
¿ Length of Rapid Mix Pipe: 3 m
¿ Rapid Mix G¿ Value: 1662
¿ Rapid Mix Average G Value: 375 /sec
¿ Diameter of Orifices: 1.5"
¿ Area of Orifices/ Area of Pipe ratio: 0.3
¿ Diameter of Riser Pipes and Rapid Mix: 6"
¿ Number of riser pipes: 2
¿ Total Number of Orifices: 9 => 4 on one pipe and 5 on the other
¿ Footprint of Grit Chamber: 1m x 0.70 m
¿ Height of Grit Chamber Walls: 1.25 m

The initial flocculation tank design was completed using the "Main Program" developed by the Spring 2007 design team.  However, due to changes in the lamella design and length of the sedimentation tanks, it had to be revised.  The current total flocculation tank length is twice the length of the sedimentation tanks.
 The flocculation tank is 21 inches wide, half the width of the sedimentation tank and doubles back to form a u-shape.  This means that the polycarbonate plastic roofing material sheets will need to be cut in half in order to make baffles.  The reason for the shape of the flocculation tank is that increasing the length of the tank and decreasing the width causes the tank to need less baffle material.  This is ideal because in the past baffles have bent and broken.  Additionally, smaller baffles will be more stable, and the design will require the baffles to be spaced farther apart. 
 Baffle spacing will be determined in Spring 2008 when a better understanding of baffles and flocculation has been achieved An assessment of the actual water demand for Tamara should be made early next year and if possible  the flocculator should be designed to handle the range of flow from the current demand to the plant design.. 
 Two drainage pipes have been added to the flocculation tank for emptying the tank.  These pipes are located at the very beginning and very end of the tank and will flow into the waste collection tank under the platform.  Due to the nature of the baffles in the flocculation tank, water will not be able to freely flow out of the tank.  A system will need to be set up to raise these baffles while the tank is draining.  One suggestion is tying string to the PVC pipes that attach the baffles to each other.  The string would be tied on each side of each baffle.  This string could loop over a long pole spanning the length of each tank.  When the tank was being emptied, a brick could be placed under each end of this pole to keep the baffles raised off the bottom of the tank and allow water to flow out.
 
Sizing of the flocculation drainage pipes was based on a time to drain tank of 30 min.  The initial flow rate was found as two times the average flow rate.  This means that:

Equation 12
Where L_tank is the length of one-half of the tank, and the average depth is found as 2 m plus half the expected headloss, or 2.1 m.  The drain size was determined using the pipe sizing program with a headloss of 200 cm and flow rate of Q_initial.
 
The wall heights of the flocculation tank are based on the assumption that there will be 20 cm of headloss in the flocculation tank.  The wall heights were set to approximately 10 cm above the water level in the tanks.  The first half of the tank will have walls 20 cm higher than the second half of the tank because the majority of the headloss comes from the first half of the flocculation tank where the baffles are closer together 
The turn in the flocculation tank was modeled as one baffle. The opening in this turn was calculated to have the same area as the opening left by a baffle at the end of the flocculation tank with a G value of 15/sec.  This baffle opening area was determined the Fall 2007 CEE 454 flocculation solution.  The program iterates to solve the following equations.

Equation 13
R_h = Hydraulic Radius
w = Width
b = Baffle spacing
Equation 14
 In this equation it is assumed that dissipation occurs over approximately the length of baffle spacing, b.
 
V_max = Maximum velocity needed to produce proper G
 f = Friction factor, new equation for f_turbulent because not using pipes
 Equation 15

Equation 16

Equation 17

Equation 18
= Distance between baffle and floor for top baffle
Equation 19
 Area = Area left by a bottom baffle
 
The opening will be left at the bottom of the wall separating the two halves of the tank.  This will allow water to drain out of the tank more easily.  The width of this opening was set to be the space between two baffles with G = 15/sec, the same as b calculated above.  Because the opening is at the bottom, the top baffle next to the opening must be removed
null
Figure 7. Drawing of second half of Flocculation Tank. Notice the opening in the dividing wall on the right. Drainage pipes will leave from the bottom left side of the tank in this view.

Design parameters for the flocculation tank are below.
¿ Initial Inside Dimensions: 4.2 m long, 21" wide on each side
¿ Updated Inside Length (based on sed tank length): 4.6 m
¿ Maximum Head Loss: set at 20 cm
¿ Depth at End of Tank: 2 m
¿ Gap for Baffle at end of tank: 0.445 m
¿ Area Left by Baffle at end of: 0.445 m * 21 in = 0.238429m2
¿ Tank Turn Opening Dimensions: 0.30 m x 0.80 m
¿ Wall Height: 2.1 m in second half and 2.3 m in first half
¿ Initial Drainage Flow Rate: 343.3 L/min
¿ Drainage pipes: 2 inches

The design of the channel connecting the flocculation tank to the sedimentation tank, while a seemingly simple connection, has many constraints.

1. The most important constraint: the channel cross-sectional area must be large enough that it will not break up flocs at max flow.
2. The channel must be shallow enough that the operator can reach the caps to the sedimentation inlet pipes.
3. Ideally the channel would fit in the otherwise unoccupied triangular space above the first plate settlers.
4. The bottom of the channel will be the level of the platform.  The height of the platform will dictate whether a bucket may be filled from the plant leveling tank in order to mix chemicals. 

More information on the bucket problem is included in the "Plant Leveling Tank" portion of this report.   However, easy solutions to constraints #2 and #3 were found.  In order to provide easier access to the sedimentation inlet pipe caps in the bottom of the channel, a new design for the caps will be used.  A conceptual drawing of the caps is shown in Figure 8 below. The top of the sedimentation inlet pipes (shown as the tall pipes in the figure) will be flush with the bottom of the channel so there will be a smaller entrance loss.   The caps will use slip fittings rather than threaded fittings.  The threaded fitting at Ojojona are beginning to break, so slip fittings will provide a more sustainable design.  Thin handles, made of smaller PVC pipe, will be attached to the tops of the caps so that the operator can easily pull and push the caps off and on.  This will eliminate the need for the operator to reach his hands into the water, and this will eliminate the depth constraint.

null
Figure 8. Drawing of Caps for Pipes leaving the Channel

The lamella constraint was solved by simply lengthening the sedimentation tank to allow for replacement of lamella that the channel would interfere with. However, it was desired to keep the tank as short as possible, so an attempt was made to interfere with as few lamella as possible.

A range of channel dimensions that would not break up flocs was determined using the "Channel_Dimensions" program created for the Marcala plant design. This program looked at the shear caused by water making the 90 degree turn into the channel. First the velocity required to achieve the given value of G, 15/sec, was determined.

Equation 20

Where is simply the entrance loss coefficient of 0.5. The length of dissipation, L, was set to 2 times the width of the channel. This estimated value was used because the actual dissipation length was unknown. The determined velocity was used to find a depth for the input channel width given the continuity equation below.

Equation 21

A range of channel dimensions that would not disturb the lamella was determined based on the geometry of the tank. The geometry of the sedimentation tank may be seen in Figure 9 and Figure 10.

An equation was developed to describe the total width of the channel that would not hit the lamella given a height of the channel. The total channel height included 20 cm of freeboard plus 12 cm for the thickness of the platform beneath the channel. The total channel width included 15 cm for the width of the brick wall.

Equation 22

Depthlamella is the distance from the top of the tank walls to the lamella, 120 cm. ¿ is the angle of the lamella, 60 degrees.

null
Figure 9 . Side view of a Sedimentation Tank. The blue outline represents the water level at 2 m of depth and the green lines represent lamella. Dimensions are in meters

null
Figure 10. Side view of one Lamella in the Sedimentation Tank. Lamella shown as green line.

A graph was made by combining the dimensions determined from each method.

null
Figure 11. Graph comparing Channel Dimension Constraints from the floc break-up and lamella constraints (Constraints #1 and #3)

Preliminary design of the plant leveling tank suggested that bucket would be able to be filled if the channel water depth was approximately 50 cm; with a platform thickness of 12 cm and a freeboard of 20 cm, this would correspond to a channel height of 75 cm. Figure 11 shows that a height of 82 centimeters is right near the ideal range of the graph where the floc breakup line and lamella line are nearest to each other.

For the designed channel, the difference between the width of the channel that would fit in the triangle above the lamella (23.35 cm) and the actual outer width of the channel (45 cm) is 23.35 cm. This value must be added to the length of the sedimentation tank so that the same amount of lamella may be included.

Design parameters for the channel are below.
¿ G value: 15 / sec
¿ Dissipation Length: 2 x width
¿ Channel Inner Dimensions: 30 cm wide x 50.5 cm deep (depth of water)
¿ Channel Outer Dimensions: 45 cm wide x 82.5 cm deep (Depth from top of wall to bottom of platform)