Hydraulic Turbine Design Project
June 5, 2008
California Polytechnic State University- San Luis Obispo
Mechanical Engineering Department
The objective for the Hydraulic Turbine project was to create a turbine, powered solely
by a water reservoir, to pull a one pound can of Garbanzo beans ten feet in the shortest
possible time. The turbine was limited to one gallon of water emitted from the water
reservoir over the course of each run. The water reservoir height could be no higher than
4 feet 3 inches, capping the maximum total head of the system. The water reservoir also
required a mechanism to start and stop the water flow to the turbine.
Our team brainstormed possible solutions to the design challenges listed. We discussed
the tradeoffs between Crossflow turbines and Pelton turbines, taking into account ease of
manufacturing, performance, and cost. Crossflow turbines were readily available for
purchase while a Pelton wheel could be manufactured using the rapid prototyping
machine in the Mechanical Engineering Department. Our team desired a self designed
turbine that was efficient, leading to the selection of a Pelton turbine. A Solidworks
drawing of our final turbine assembly can be seen below.
Figure 1. Final CAD rendering of the turbine assembly.
We recognized that the system piping and nozzle design would require more time than
the construction of the reservoir housing. The reservoir support was to be simple and
non-intensive in labor. Overall, our team recognized the importance of adjustability in
design. While we planned to calculate the system as well as we could, translating that
into a precise actual design can be difficult and adjustments would be needed to
maximize performance. We valued testing to determine the full performance of our
The design of the Pelton buckets was
performed using Solidworks. This step
required more time than originally
thought, as drafting the bucket to the
specifications provided proved difficult.
From our system optimization using
Engineering Equation Solver (EES), we
selected a nozzle diameter of 0.8 inches.
This diameter was selected based on
minimal losses and high efficiencies,
while staying below a gallon of water.
From our Pelton bucket spreadsheet, the
buckets were to be 2.8 inches wide and
2.02 inches tall. 20 buckets were
required for a wheel ratio of 14. To
minimize turbine weight, a bucket
thickness of 0.05 inches was selected.
A square arm, 1 inch in length that
extended from the bucket, was to be
clamped to a turbine wheel after the Figure 2. Single Pelton bucket that will translate water
prototypes were complete. Having the jet stream to rotational power.
buckets separate from the turbine
housing allowed a detailed inspection of the buckets and easy access to any sanding
needed later. The arm was designed as a rectangular box instead of a cylinder to ensure
proper alignment when being attached to the wheel.
After rapid prototyping was completed, the physical buckets required surface finishing
inside the flow channel and reinforcement along the outer shell. A white lacquer was
used for this process, a gel coating that was sandable and water resistant. Multiple layers
were applied to strengthen buckets. After the lacquer dried, the insides of the buckets
were sanded smooth, eliminating friction losses that would lower the power output of the
Figure 3. Application of lacquer to buckets.
The nozzle was also designed in Solidworks and created using the rapid prototyping
machine. The nozzle was to be simple and would be attached to the piping by an epoxy
adhesive. In order to have a constant diameter jot stream exiting the nozzle, a ! inch
horizontal section was added to the end of the nozzle to eliminate any vena contracta.
The nozzle could be cut later if an increase in jet diameter was desired. The dimensions
of the nozzle can be seen in Figure 4 below.
Figure 4. Schematics of exit nozzle that will feed the Pelton turbine.
Originally our team planned to use
a wheel similar to the one modeled
in SolidWorks, seen to the right,
which consisted of two discs with
slots cut into them every 18
degrees. The discs were to
sandwich the buckets into place.
Material purchased was " inches
thick MDF board, from which two
discs of approximately 9 # in
diameter were cut. After the slots
were drilled using a mill and
rotary apparatus, it was determined
that the wheel would be too heavy
and thus have too much rotational
After this idea was scrapped, we Figure 5. Initial concept for wheel housing of buckets.
consulted with Martin Koch of the
IME department. Martin suggested
cutting two disks out of a thin Plexiglas to sandwich the buckets. He provided a printer
that laser cuts the Plexiglas based on an AutoCAD drawing provided by the user.
We decided to cut two identical discs with slots aligned to the bucket placements. The
slots were to hold support platforms for the buckets, also made out of Plexiglas cut from
the laser printer. Problems occurred with alignment of the supports on the adjacent wheel.
After carefully alignments, both discs were cured together using the Plexiglas supports
and epoxy resin. Overall, we greatly reduced the turbine weight and rotational inertia,
allowing quicker initial acceleration to our operational speed.
Figure 6. Final Plexiglas wheel with attached supports prior to encloser.
Figure 7. Careful sanding before adhering buckets to the turbine wheel.
We chose a five-gallon drinking water container commonly seen on water coolers as our
water reservoir. Since the neck of the bottle had a gradually converging profile (high r/D
value), the entrance losses and head loss would be greatly reduced.
The diameter of the bottle’s neck was approximately 1 " inches, leading our team to
select PVC piping of equal diameter. Based on talks with members of previous teams, it
was decided that the pipe should lie vertically straight, perpendicular to the ground. From
a fluid mechanics standpoint, this would reduce the length of pipe, and therefore have the
least head loss (smallest L/D value). Also, the largest diameter for flexible tubing
available from Home Depot was 1 " inches. This was desired so the nozzle angle could
be set correctly to 21 degrees from the horizontal, optimal for power output.
Figure 8. PVC connector attached to water reservoir to link reservoir to PVC piping.
At the end of the straight length of pipe a 45 degree elbow was attached. To this a ! turn
ball valve was attached. While the valve was initially difficult to turn, we practiced with
the system numerous times to have a familiarity for the time needed to turn the valve.
To the outlet of the ball valve, a small section of flexible 1 " inch tubing was attached.
The flexible tubing had a natural curve to it. By cutting the tubing at precisely the right
location, a nozzle angle of 21 degrees from the horizontal was obtained. The rapid
prototyped nozzle was attached to flexible piping using epoxy and sealed with plumber’s
epoxy putty, an extremely strong adhesive.
The support of the reservoir system was not critical, however stability and strength were
desired. We purchased four 4 foot long, 1/16 inch thick steel L shaped beams with 3/8
inch holes cut every inch to use as supports. Holding this together were 1/16 inch
straight sections of steel and bolts. For extra stability, feet made of short 2X4’s were
attached to the bottom of the support.
Figure 9. Full view of turbine housing, reservoir support, and pipe in completed assembly.
For ease of manufacturing, the housing consisted of two sheets of clear Plexiglas held in
place by four long (approximately 1 " ft) bolts. By using Plexiglas the wheel could be
seen in operation, while the bolts allowed adjustments to the width of the housing. The
pulley was placed on the outside of the housing as not to interfere with the water exiting
the buckets of the wheel. The housing was not attached to the reservoir support to allow
rooms for adjustments and modifications, as well as ease of transportation to the
Figure 10. Assembly of the turbine and pulley inside the turbine housing.
Through the center of both sheets we drilled a 0.875-inch diameter hole, where we
permanently fixed ball type bearings from a roller skate wheel. A 5/16-inch diameter
shaft was allowed to freely rotate inside the bearings. The wheel and pulley were secured
onto the shaft on the inside and outside of the housing respectively.
Figure 11. Wheel and nozzle as seen through Turbine Housing.
We decided that at full speed, a pulley diameter of nine inches would suffice. However,
the required starting torque for a nine-inch pulley was a concern. Therefore, we designed
a variable radius pulley such that the initial diameter was about 2 inches and increased
incrementally after about ! a turn to 9 inches. The diameter remained constant thereafter.
The design was a composite material consisting of a plywood disk, short lengths of
dowels inserted in to drilled holes, masking tape, Gorilla Glue and plumbers epoxy putty.
Figure 12. Variable diameter pulley allowed less torque required to start the turbine.
After assembly of the system, we completed dozens of test trials to correct any problems
with the system. Results from our own testing before the competition averaged around
1.6 seconds to 1.8 seconds, a decent time.
Figure 13. Test run picture used to analyze path of flow with turbine. From this picture we
concluded that the nozzle was positioned correctly.
In our initial run on Thursday, we placed with a time of 1.75 seconds and lost in our first
match. While we were pleased with the consistency of performance, we brainstormed
modifications to improve performance of the system we already had.
The water reservoir’s height was increased one inch to the maximum height level of 4
feet 3 inches. It was evident during test runs that we used far less than one gallon of
water. To increase the power output of the system, we cut the nozzle back to an inside
diameter of 0.9 inches. The vena contracta was analyzed, using Figure 14 below, to
reveal a 12.5 percent reduction in diameter at location D2. From this figure we also found
that the closer the nozzle is to the Pelton bucket, the less reduction happens to the flow
diameter. We moved the nozzle as close as possible to the Pelton buckets by adjusting
the width of the turbine housing. Several more trials were conducted with our updated
system. We improved our quickness in opening the valve with each practice run. Our
trial times increased significantly to between 1.2 to 1.4 seconds.
Figure 14. Test run showing the vena contracta, where the water jet diameter is reduced as it travels
further from the nozzle exit.
On the second day of competition, our turbine won the first round matchup with a time of
1.39 seconds. Our second round matchup was lost when the string, attached from pulley
to garbanzo bean can, fell off the pulley track before the finish line. Upon further
examination, the cause of the pulley malfunction may be due to how the string was tied
down on the pulley. The string is tied to a peg about two inches from the shaft driven by
the turbine. After each run, the string is cut to break the can from the turbine. Several
loops of string had accumulated on the peg from several trial runs. This resulted in our
team attaching the new loop for the last run too high on the peg, leading the string off the
pulley. In order to correct this problem, the remaining loops should be cut and the lip on
the pulley should be heightened.
Overall we have learned quite a bit this quarter working on the hydraulic turbine project.
It has been interesting to see the path of design in respect to turbines, especially having
the ability to take an idea from conception to realization in a physical product that must
meet specification. The project has taught us how to work as a team and how to utilize
each other’s strengths as mechanical engineers to effectively complete the project.