DALHOUSIE UNIVERSITY DEPARTMENT OF MECHANICAL ENGINEERING MECH 4010/ 4020 WINTER TERM FINAL REPORT Prepared by: GROUP 10: ROTATIONAL IMPACT INTERNAL COMBUSTION ENGINE Adam Krajewski Aziz Martakoush Braden Murphy Brett Dickey Jean‐François Pelletier Supervisor Dr. Darrel Doman April 9, 2010
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
DALHOUSIE UNIVERSITY
DEPARTMENT OF MECHANICAL ENGINEERING
MECH 4010/ 4020
WINTER TERM FINAL REPORT
Prepared by:
GROUP 10:
ROTATIONAL IMPACT INTERNAL
COMBUSTION ENGINE
Adam Krajewski
Aziz Martakoush
Braden Murphy
Brett Dickey
Jean‐François Pelletier
Supervisor
Dr. Darrel Doman
April 9, 2010
Executive Summary
The Rotational Impact Internal Combustion Engine (RIICE) is a conceptual engine with one main concept: the pistons rotate about a center shaft. The primary difference of the RIICE concept is all four strokes of a conventional internal combustion engine occur simultaneously in one cycle.
In this project, no fuel is used, and compressed air is used to simulate fuel ignition. The high‐pressure air forces the pistons to rotate about the shaft in one direction, as the pistons are restricted from rotating in the reverse direction by the use of one‐way bearings. An additional of one way‐bearing arrangement is used to transfer power to the drive shaft when the pistons rotate forwards.
The power transmission using one‐way bearings and the piston shape had the most influence in the design process and on the final design. It was chosen to use a circular cross section piston to simplify sealing. Therefore, a toroidal shaped casing is required. The casing of the mechanism is comprised of two halves that cover and seal the upper half of the pistons, and two cranks form the lower half of toroid. To simplify the fabrication process the pistons are made of three cylindrical pieces, which fit into the toroidal shape of the cranks and casing.
The sealing between the cranks and casing is critical to maintain high compression ratios by reducing air leaks to a minimum. Several sealing methods are investigated and elastomer energized plastic seal are implemented into the design. The pistons are sealed with acetal lip seals.
The control system is intended to manage the timing of compressed air injected into the engine to ensure full rotation of the pistons and continuous engine operation. This system is intended to control the rotation of the engine and ultimately its power output.
Two prototypes have been built to prove various aspects of the concept. The first prototype was a plastic model to demonstrate the operation of the one‐way bearing arrangements. It was shown that the bearings work as intended. An open casing test was performed with compressed air on the plastic prototype. The test demonstrated that the pistons rotate as intended with no casing. The second prototype is an aluminum prototype built to test the operation of the engine with a closed casing. It is necessary to operate the engine in a closed casing, because the intent of the project is to simulate operation as an internal combustion engine.
From tests performed on the closed casing, the team could not get the engine to operate as intended. The primary reason that it did not work, is thought to be because of a balance between the sealing and friction in the closed casing engine. From tests on various seal geometries, it was chosen to use a seal with lower friction and sealability for the prototype. From the testing it was found that this balance was not the right choice to get the engine to operate. Therefore, it is recommended that further research be concentrated on determining the correct seal geometries to establish an effective low friction seal. In addition, it is recommended that a method of repositioning the pistons between tests be incorporated into the design. In addition, a method of spinning the pistons before the compressed air is implemented is recommended to start the engine.
ii
Table of Contents
Executive Summary ............................................................................................................................... i List of Figures.........................................................................................................................................iv List of Tables ........................................................................................................................................... v 1 Introduction .....................................................................................................................................1 1.1 Mechanical Efficiency..............................................................................................................................1 1.2 RIICE Concept.............................................................................................................................................3 1.3 One‐Way Bearing ......................................................................................................................................4 1.3.1 Fundamentals....................................................................................................................................4
4.8 Control System........................................................................................................................................ 28 4.9 Theory......................................................................................................................................................... 28 4.10 Development ......................................................................................................................................... 28 4.11 Final Control.......................................................................................................................................... 31 4.12 Future Recommendations............................................................................................................... 33
iii
4.13 Compressed Air System ................................................................................................................... 34 4.14 Position Sensing................................................................................................................................... 35
6 Testing............................................................................................................................................. 42 6.1 Open Casing ‐ Model ............................................................................................................................. 42 6.1.1 Manual Test ..................................................................................................................................... 42 6.1.2 Compressed Air Test ................................................................................................................... 44 6.1.3 Open Casing ‐ Prototype ............................................................................................................ 46
6.2 Closed Casing ‐ Prototype .................................................................................................................. 50 6.2.1 All Piston Seals – Old Seals ....................................................................................................... 50 6.2.2 Half Piston Seals – Initial Design............................................................................................ 51 6.2.3 Piston Seals – New Design ........................................................................................................ 52 6.2.4 Sub – Test A: Quality of Seal ..................................................................................................... 52
6.3 Sub – Test B: Eliminating Back Stopping Seal Leaks .............................................................. 52 6.3.1 Sub – Test C: Transmitting momentum from one crank to another using higher pressure............................................................................................................................................................ 53 6.3.2 Sub – Test D: Compression cycle – No impact ring ........................................................ 53
3.1.2 Option 2: Fixed Casing with indexing freewheels inside
In response to the limitations of the option one, option two has the ring gear inside the
housing of the engine that allows for a fixed casing. By having the ring gear on the inside it is
possible to select a smaller gear that decreases the cost. In addition, the design is
significantly simplified. An inside ring gear is attached to each crank. The backstopping
freewheels are attached between the inside ring gears and the housing which is fixed to the
ground.
3.1.3 Option 3: Fixed Casing with indexing and backstopping one‐way bearings
With option three it is possible to maintain the backstopping motion by placing the indexing
and backstopping freewheels coaxially in series around the shaft, seen in Figure 9. The
13
backstopping freewheels have a larger ID allowing them to be connected between the right
and left cranks and fixed to the housing. This third and final option is a simplification of
options one and two by removing the ring gears altogether. Placing the bearing in series
allows for a more compact assembly of the RIICE engine block. In the end it is possible to
use less material and build an engine with a smaller displacement.
Figure 9 Option 3: Coaxial oneway bearings in series with fixed casing (exploded view)
3.2 Pistons
Each piston is designed to transfer the energy supplied by the compressed air system to the
cranks. There are two pistons attached to each crank, for a total of four pistons in the
assembly. The design of the pistons considered the following:
• Face shape: round vs. square
• Manufacturability: toroid vs. square
The Wankel engine is a widely known, production rotary engine with square shaped
pistons. Due to the popularity of the Wankel, the first iteration of the piston design
considers a square shape. The advantage of the square shape is the simplicity of
manufacturing the casing. The other advantage of considering the Wankel’s square shaped
14
pistons is the issue of sealing has already been addressed. This is done through the use of
Apex seal, which are essentially flat bar made of hard metal pressed flat against the engine
casing by a leaf spring. If Apex seals were used in the RIICE, custom seals would be needed,
which would be an expensive option. There were also concerns for sealing the sides and the
sharp corners of square pistons. For these reasons, square pistons were not pursued any
further.
3.2.1 Piston shape and Assembly
Round pistons were chosen because a circle provides the largest area with the smallest
perimeter. A round piston requires a round cylinder and when combined with the rotary
concept of the RIICE, this leads to a toroidal expansion chamber (a hollow donut shape). The
second design of the piston matched the toroidal shape, as seen in figure 10. Manufacturing
toroidal pistons is a challenge. To overcome this challenge, a third design was selected for
the pistons, also seen in Figure 10.
Figure 10 – Toroidal, straight three piece piston and exploded three piece piston
The pistons still have the toroid geometry but are made from three pieces: the front face,
the rear face and the center section all made of aluminum 6061‐T6. The front and rear faces
are identical. These sections can be seen in Figure 10, piston face on the left, center section
on the right and an assembled piston can be seen in Figure 11. The two holes on the right
hand side of the components of Figure 11 are used to assemble the piston using screws.
The piston is screwed to the crank via ¼” shoulder screw.
Toroid Straight
15
Figure 11 – Assembled Piston
3.2.2 Piston Seals
The initial design was to use piston rings to seal around the piston in the toroidal chamber.
Pistons rings, typically made out of cast iron, come in a variety of standard sizes, are readily
available and inexpensive. These seemed to be an excellent choice; however the technicians
raised some issues with regards to wear. Having cast iron scraping against the aluminum
casing will be an issue and an alternate design is to be considered for sealing or casing and
cranks shall be manufactured out of steel. A second suggestion to is to manufacture the
pistons out of durable wear resistant polymer such as Acetal or Orlon. These polymers are
used in bearing internals and have both high wear characteristics and considerably high
tensile strengths. This would be implemented into the design of the pistons by removing
the grooves and adding a recess on the edges to form a seal. A third suggestion is to use
piston cup seals as seen in Figure 12 to be mounted on each face. Once pressure is applied,
the cup seals will expand forming a seal with the casing and acting as a wiper blade. Further
research is required in these fields and the design is changing accordingly.
Figure 12 – Piston cup seal [McMaster‐Carr]
Evaluating the three possible piston face designs, the team decided to go forwards with the
second option: Fabricating piston faces using Acetal. Conventional piston rings used in
internal combustion engine are typically made from cast iron. Since the RIICE design uses a
aluminum casing, conventional piston rings would have caused excessive wear of the casing
eventually leading to compressed air leakage or seizure.
16
The third option, piston cups, was not chosen because they are an off‐the‐shelf, non‐
modifiable component. Typically off‐the‐shelf components are preferred since they are
simpler and do not require design, only selection. In this case, it was unknown how the
engine would perform therefore having the option to modify the piston seal was critical. In
addition, there was some concern that due to the large cup area, when energized the piston
cup seal would provide a significant amount of friction.
Initially, energized piston faces were designed and manufactured. The concept of this
sealing method was that when compressed air was applied to the piston face the seal would
energize, the seal lip would expand thus preventing the compressed from leaking. It was
found that the seals had to be manufactured with very tight tolerances (± 0.0005 in) to
achieve effective sealing. It was originally thought that the Acetal could be oversized, and
would squeeze into the chamber. It was discovered the material is not as elastic as required
under the desired operation of the piston faces, and therefore a more elastic material may
be recommended for this method of manufacturing.
The piston faces were modified in order to achieve maximum sealing while minimizing the
friction of the piston when contacting the walls of the casing and crank. The pistons were
precisely manufactured to the cross sectional area of the toroidal chamber. Furthermore
the lip width was decreased significantly to minimize friction, but maintain sealing.
3.2.3 Piston Mounting
The stress on the screw was calculated with the worst‐case scenario where the friction is
neglected. This means there is no load sharing of the screw force and clamp force;
therefore, the screw takes the entire load. The direct shear stress on the screw was
calculated using:
€
τ =VA Equation 3
€
τ = 2.7 ksi
The yield stress of an SAE grade 1 screw is 36 ksi therefore shear stress is not a limiting
factor in our design.
17
4 Final Design
With the power transmission and pistons design and finalized the, the remaining
components have been designed to match. The remaining components include; the cranks,
the impact rings the bearing housings and the casings. These are all components that must
be manufactured and are identified in figure 13. There are also several components that will
not be manufactured that are crucial to the operation of the engine. They are: the control
system, air system and sealing aspects.
Figure 13 – Exploded view of final design.
4.1 Cranks
The purpose of the cranks is to transmit torque from the piston to the shaft. The cranks are
connected to the shaft via two one‐way bearings. These one‐way bearings, referred to as
indexing freewheels, engage when torque is applied to a crank. Furthermore, the indexing
freewheels allow the shaft to continue rotate if the cranks are stationary. For the location of
these components on the crank refer to drawing M200 in Appendix B.
18
Figure 14 – Crank: rear view, front view and crank assembly with pistons
During expansion, one crank is prevented from rotating backwards by an additional one‐
way bearing, known as the backstopping freewheel. The backstopping freewheels are
mounted on the crank and press fit into the bearing housing. The mounting and housing of
the indexing and backstopping freewheels require a hardened surface. Since the cranks are
made from aluminum, a steel collar is press fit onto the cranks where the backstopping
freewheels attach. The collar and crank assembly can be seen in Figure 14. This assembly
may require a keyway between the crank and collar, which has not yet been included in the
design.
4.1.1 Crank Friction
In designing the RIICE concept, the friction between the pistons and cranks was taken into
consideration. In order to approximate the friction of the pistons on the cranks, it was
assumed that the friction is similar to a conventional internal combustion engine. This is a
reasonable assumption because the RIICE concept uses the compression rings from a
conventional engine to seal the expandable chambers. It was found that an approximate
rule for estimating ring friction is that each compression ring contributes about 1 psi MEP1.
This was interpreted as requiring 1 psi of pressure on the piston face to overcome friction,
which was used to calculate the net frictional torque on each crank, as described in the
following paragraphs.
The pistons are attached to the crank with socked head cap screws as illustrated in Figure
15. This means that each piston seal moves against the opposite crank and casing surfaces,
but does not move against the surface of the crank it is attached to. Each crank takes up
19
25% of the seal area. Therefore, there is friction acting on 75% of the seal surface, which
corresponds to a 0.75 psi of friction per piston ring.
Figure 15 Friction area on piston seals
With a piston diameter of 1.3 in. centered at a radius of 1.75 in. from the drive shaft, 0.75 psi
corresponds to a resistance torque of 1.74 in‐lb per piston ring. With 4 seals per piston and
2 pistons per crank, there is a net frictional torque equal to 14 in‐lb per crank. In order to
account for friction of the seals between the cranks, and the rolling friction of the bearings,
this frictional torque was doubled to equal 28 in‐lb per crank.
Note that the compression ring material may be changed to a polymer seal material as
discussed in the sealing section. This method of sealing is not dramatically different from
the original method chosen, and the estimated frictional torque should be within the same
range.
The overall size of the mechanism is governed by two main components: the cranks and the
pistons. The limiting torque rating of the one‐way bearings dictates the maximum torque of
the RIICE mechanism. Curves for the maximum operating pressure of the mechanism have
been derived to identify the optimal piston size and crank radius. The maximum torque is
equal to the applied force through the center of the piston times the crank radius. The
maximum operating pressure term is introduced by replacing the applied force with the
product of the maximum operating pressure and piston cross sectional area.
This derivation yields the following equation:
20
€
Pmax =4Trπdp
2
Equation 4
Where:
Pmax is maximum operating pressure
Tr is the limiting torqure rating
r is the crank radius from center of shaft to center of piston
dp is the piston face diameter
A flat piston face was conservatively assumed in the derivation of the above formula.
Figure16 shows two maximum operating pressure curves from the above equation.
Figure 16 – Maximum operating pressure curve of RIICE
0
50
100
150
200
250
300
350
0 1 2 3 4 5 6
Maximum Pressure (Pmax) psi
Piston Diameter (dp) and Crank Radius (R) in
Maximum Operating Pressure as a Function of Piston Diameter and Crank
Radius
Pmax Vs. R
Pmax Vs. dp
Maximum Operating Point
Maximum Operating Point
21
The blue curve in Figure 16 represents the maximum operating pressure as a function of
crank radius with a fixed piston diameter of 1.3in. The red curve in figure 16 represents the
maximum operating pressure as a function of piston diameter with a fixed crank radius of
1.75in. The two maximum operating points are the points at which the piston diameter is
1.3in and the crank radius is at 1.75in. At these operating points the mechanism can
operate at a maximum pressure of 177.5psi. At a pressure higher than 177.5 psi, the
indexing one‐way bearing will begin to slip on the output shaft. The chosen crank radius
and piston diameter are 1.75in and 1.3in respectively.
4.2 Impact Rings
The impacts rings are designed to prevent damage to the pistons and to promote energy
transfer between them during start up as well as transferring kinetic energy between the
cranks. The impact rings will be mounted between cranks, one on each front face of each
crank. The rings shall prevent the pistons from impacting one another by and a wider arc
length than the pistons inner radius, insuring the impacts rings contact and the pistons do
not.
4.3 Bearing Housings
The bearing housings, also referred to as pillow blocks, are used to house the external radial
ball bearings of the mechanism. These radial ball bearings take the radial load applied to
the indexing and backstopping freewheels and to keep the shaft true with respect to the
cranks and freewheels. As mentioned in the previous section the bearing housings also
serve as a ground for the backstopping one‐way bearings. The bearing housings screw onto
the casing as shown in drawing M200 of Appendix A. The front and rear view of one of the
two bearing housing and radial ball (shown in green) bearing assembly is shown in Figure
17.
22
Figure 17 – Bearing Housings
4.4 Casing
The casing of the mechanism is comprised of two halves. Each half covers a crank and
piston assembly as shown in drawing M200 of Appendix A. The functions of the casing are
to ground the cranks, and both cover and seal the upper half of the pistons not covered by
the cranks. The bearing housings are bolted onto the casing, thus anchoring the
backstopping loads. Furthermore, the casing holds all the internal components of the
mechanism together. The casing will be machined from aluminum 6061‐T6 and will have a
fillet featured to match the circular face of the pistons. A series of 12 bolts will fasten the
two halves of the casing together with an o‐ring to prevent and air leaks. The casing will be
mounted to a base on 2 of these bolts. The base has yet to be finalized, but will securely hold
the RIICE mechanism in place. The casing will also have the tapped holes for the air input,
exhaust and intake, as shown in Figure 18.
Figure 18 – Casing Left Half and Right Half
Air Input
Exhaust
Intake
23
4.5 Sealing
4.6 Initial design
Figure 19 is half the cross section that shows the location of the most important areas for
sealing rotating parts. These areas are circled at the edges and in the center. Note, only one
of the two halves of the casing is shown for clarity. The sealing of these areas are critical in
order to maintain high compression ratios by reducing air leaks to a minimum.
Figure 19 – Cross section identifying areas requiring sealing between rotating parts
The first type of seal that was looked at was conventional O‐rings. O‐rings make a good
seal; however, they are bad if there is any movement in the axial direction. Movement in the
axial direction may damage the O‐ring due to twisting and excessive wear. The way O‐rings
24
work is by placing them in a groove and then deform their shape by squeezing them. This
deformed shape will fill in the groove and consequently seal. Squeezing the O‐ring is
difficult to do in this design due to the friction produced between rotating parts.
The second type of seal that was looked at was the Quad‐ring®; this is shown in Figure 20.
The Quad‐ring® does not spiral or twist due to the presence of corners. They are generally
harder than O‐rings and can fit in the same groove of the same size. Quad‐rings® seal the
same way as the O‐rings do and therefore will produce high frictions.
Figure 20 Quadring® from Engine Mechanics
[www.tpub.com]
The third type of seal that was looked at was elastomer energized plastic seals. Figure 21 is
a figure from the book Seals and Sealing Handbook by Robert Flitney. This seal is
categorized as a high‐pressure double acting seal that prevents fluid from leaking from
either way; it consists of two parts: an O‐ring and a PTFE dynamic component. The seal is
energized due to an O‐ring that acts as a spring and keeps the dynamic component in
contact with the surface. The grooves provide smaller surface contact to reduce friction and
are designed to eliminate any leakage through, by trapping the fluid.
25
Figure 21 High pressure rotary seal
[Trelleborg Sealing Solutions and Freudenberg Simrit]
Figure 22 shows how the elastomer energized plastic seal was implemented into the design.
The grooves will be filled with grease to reduce friction and trap the air. Sealing in between
the cranks is quite difficult due to the cranks rotating intermittently. While one crank is
rotating, the other is stationary and a proper seal is to be maintained. The seal in between
the cranks is one part and is required to seal the other two; therefore two O‐rings are
required to maintain seal contact with the surface. Figure 23 shows what the center seal
would look like.
Figure 22 Cross section identifying seals between rotating parts
26
Figure 23 Section of the seal in between the two cranks
4.7 Design Revisions
4.7.1 Outer Seals
The first seal tested was press fit into the casing, with an o‐ring located on the exterior face,
between the seal and a shoulder on the casing. This formed an effective static seal, but
required 3.1 ft‐lbs of torque to turn the crank in the casing. It was concluded that the reason
for this is that the seal was constrained radially and axially by the crank/ casing assembly.
This caused the seal to rub excessively on the outside diameter and the face of the shoulder
of the crank. In order to eliminate one of these constraints, it was recommended to
undersize the seal axially, and press fit the seal on the crank, with an o‐ring located on the
interior face between the shoulder on the casing and the seal. This eliminated the axial
friction from the seal, but may reduce the sealing effectiveness. However, it was estimated
that, as the pressures being used are relatively low (100 psi range), the sealing would not be
compromised.
The second seal tested was designed to press fit onto the outer diameter of the crank, with
an o‐ring between the shoulder of the crank and the seal. The outer surface of the seal forms
a dynamic seal with the casing bore, and in order to reduce the contact area between the
seal and the casing bore, two groves were added on the outer surface of the seal. The seal
design can be seen in Figure 24. The outer diameter of the seal was originally cut to 3.500
in., and the inner diameter to 3.200 in. It was observed that the seal could rotate easily in
the casing or crank when inserted independently. However, when the seal was fit between
the casing and crank, the torque required to turn the crank was excessive. Therefore, it was
recommended to undersize the outer diameter of the seal by 0.030 in. to 3.470 in.
27
The third seal tested was designed to the specifications of the previous recommendations.
The torque required to turn the crank in the casing was measured as 2.1 ft‐lbs. This method
does not fully seal the casing at higher pressures, but when 60 psi is applied to the engine,
the flow resistance is 41 psi, 1 psi above the measured benchmark value of 40 psi.
Figure 24 Revised seal in engine cross section
4.7.2 Inner Seals
The design modifications made to the outer seals were applied to the inner seals. The seals
were originally intended to have o‐rings energize the seal to the inner bore of the crank. It
was found from the outer seal testing and refinement, that this design had an excess amount
of friction. Therefore, the final design used for the outer seal, was applied to this application.
In particular, the seal is designed to press‐fit into one crank, and undersized on the other
crank. This creates a static seal on one crank, and a dynamic seal on the other. In addition,
grooves are cut into the contacting surface of the dynamic seal to further reduce the friction.
28
4.8 Control System
4.9 Theory
The purpose of the control system is to accurately deliver compressed air to push pistons
and to measure the location of the pistons as they rotate. The control system uses two
rotary encoders to measure the location of the pistons. These encoders are mounted on the
casing and connected to the cranks by a one‐to‐one gear ratio. A solenoid valve is connected
to the top of the engine to control flow from a supply of compressed air. LabView was the
programming language used to control the interaction between these components.
The main design requirement of the control system was to fire two pulses of compressed air
per rotation of each crank. Through the development of the controller, various
improvements were taken into consideration to optimize the functionality. These
developments will be explained below.
4.10 Development
The solenoid that was used was a 12V, 150psi max solenoid valve that is normally open,
meaning when not energized, the valve is open. Thus to close the vale, the control system
must continuously supply power to the solenoid.
The first task of the control system was a simple program to turn on and off of the solenoid
valve. This was achieved by the use of a relay switch to control the 12V supply to the
solenoid. Since the DAQ card only outputs 5V, a digital pin signal opened the relay,
supplying 12V to the solenoid, which in turns closes the valve. Like wise, the valve is opened
by a digital signal that switches the relay, disconnecting the 12V supply. Figure 25 shows
the control of the solenoid as described.
29
Figure 25 Schematic of solenoid control
The rotary encoders are used to measure the angular position of the pistons as they rotate.
The encoders used, output 1000 digital pulses per revolution. The output of the encoder is
read by a counter port in the DAQ card and read into LabView. In the program, angular
position in radians is calculated as follows:
Equation 5
Boolean logic, true or false statements, is a large part of LabView programming. Since on/off
programming is easiest way to operate, the firing control was designed in the same manner.
Since the piston motion is cyclic, the optimal logic for the firing control is to use a periodic
triggering method. To do this, the angular position in radians was converted to a sine wave.
Since the sine wave always oscillates between 1 and ‐1, the Boolean logic can be
programmed to fire the when the sine of the angular position is either greater than or less
than a desired value. This control method gives the desired two firings per encoder
revolution and is always 180° apart, corresponding to the two pistons that are also 180°
apart. To optimize the control of the firing, an offset was introduced into the angular
position firing control.
!
"rad
= counts#2$
1000
30
Figure 26 Sinusoidal firing control of solenoid
This offset or phase shift (seen in Figure 26 ‐ Sinusoidal firing control of solenoid allows the
Boolean control to use the positive aspects of the local maxima and minima, but the shift
eliminates the need for pistons to be at the same location. The two flat dashed lines at 0.5
and ‐0.5 in Figure 26 represent the triggering values for the solenoid. When is
greater than or less than those line, the valve opens and delivers a pulse of air. These firing
positions are sine of the offset, thus when the user manipulates the offset to fire sooner or
later, the trigger values decrease (corresponding to a longer pulse of air) or increase
(shorter pulse of air) respectively.
The last optimization of control system was the ability to reset the counts with each
increment of the index pulse from the encoder. Since LabView simply counts the rising
edges of the digital waveform graph of the encoder, and this counts starts at zero and
increase as long as the encoder is rotating, there is no way of knowing the exact location of
this piston. The location of the index is fixed on the encoder and is aligned with the firing
!
sin(")
31
position of the piston. This means that no mater where the piston starts, once the index
count increments the count will reset and the operators knows the location of the piston
and can thus trigger the air reliably, knowing that it is always triggering in the correct
position.
4.11 Final Control
The final design incorporates all of the above features to create an open loop control
system. This is represented in Figure 27 with a more detailed break down in Figure 28.
Figure 27 Open loop control diagram of control system
Figure 28 Detailed schematic of open loop control
The final system uses a manually adjusted regulator to regulate the pressure to a maximum
of 130psi. The maximum operating pressure for the solenoid valve is 150psi.
Table 4 DAQ Pinout for Solenoid Valve and Encoders
Solenoid Valve Encoder 1 Encoder 2
Relay Ground
(1)
Port 33
(Digital ground)
5V Port 35
Digital 5V
5V Port 34
(Digital 5V)
Relay Switch
(2)
Port 27
(Digital line 0)
Ground Port 33
(Digital ground)
Ground Port 24
(Digital
Ground)
32
Relay Ground
(3)
Port 33
(Digital ground)
Quadrature
A
Port 47
(Counter 0)
Quadrature
A
Port 41
(Counter 1)
Relay Switch
(4)
Port 25
(Digital line 1)
Index Port 29
(Digital line 2)
Index Port 26
(Digital line 4)
Table 4 details the pin‐out of the DAQ card to the various components required to in for the
control system. The LabView program entitled “Main Control Rev 3.vi”, processes all the
necessary data to achieve the desired control operations. A copy of this program can be
seen in Appendix A.
Trigger Values
Offset
Slider
Figure 29 Control system user interface
Figure 29 shows the graphical user interface for the main control program. The main
features of this interface are the real time interactive slider on the right side, the encoder
counts in middle in bold number and the graph. The slider is used to set the offset, which as
mentioned earlier sets the trigger values for the solenoid. There is a green LED light to show
whether the valve is open or closed. The Graph outputs the sinusoidal function of the
counts; from this the user can clear determine when the piston is in the firing position.
This control system was an effect method of pulsating air based upon the encoder counts.
The full system assembly, consisting of the air supply, solenoid valve, encoders, control
33
program and engine was only tested once. The lack of testing of the full control system was
due to the team’s dedication to achieve continuous rotation before implementation of the
control. Different controllers, comprising of the same logic, but not as extensive, were used
to complete the testing. When the full control system was implemented the same effects
occurred where the engine would stall before the back piston would move into the correct
firing location. Thus, although all the individual components of the system were successfully
tested, the full control system was never completely validated.
4.12 Future Recommendations
The largest challenge during the creation of the control system was the ability for the
controller to recognize the locations of the piston during start up. This was overcome by the
using the index pulse to reset the counter on the first rotation. The difficulty when doing
this was the DAQ card only had two counter ports, not the required 4. Wiring the index
pulse into a digital line of the DAQ card solved this problem. This method worked well at
slow rotation speed (less than 60rpm) but was not as reliable at higher speed (greater than
60rpm) as the program would miss the index pulse occasionally. A DAQ card with at least
four counter ports would greatly increase the reliability, as the index pulse can be measure
at high angular speed if measured in a counter port.
The maximum pressure of the solenoid valve also limited the control system. With a max
rating of 150psi, the valve couldn’t not be tested using pressure sources greater that 150psi.
This fact limited the upper range of pressures during the testing phase of the engine
development, forcing the team to open and close a manual valve to test at high pressures.
The final recommendation observed during the controller development was the occasional
impact of noise in the system. It was observed that when the unshielded wires of the
encoders were in close proximity of the solenoid, there would often be jumps in encoder
counts during valve actuation.
34
Figure 30 Affect of noise on encoder counts
The sharp drop in the sin(theta) value in Figure 30 was caused by the noise of the solenoid
valve. This sharp drop can have a detrimental affect on the control since the timing of the
engine is based off the encoder counts, this error could cause misfiring and unwanted
behavior of the engine. This can be solved by two methods. One, by resetting the counts
with the index, insuring no more than one rotation is affected and, two, using shielded
wiring for the encoders that will prevent noise affecting the counting signal.
4.13 Compressed Air System
The compressed air subsystem is comprised of the air supply, the air delivery equipment
and the control system, designed with the goal of bringing the operation of the engine from
start to steady‐state. This system will control the rotation of the engine and ultimately its
power output. The logic behind the controller is based upon sensing the position of the
pistons and when they are in the correct locations to inject air. The air inlet and exhaust are
fixed locations and by monitoring the pistons locations relative to these positions, the
controller can vary the amount of air injected to ensure full rotation of the pistons and
continuous engine operation. The air supply will be a scuba tank, regulated by a computer
controlled solenoid valve. The components will interface with a controller designed in
LabView.
35
4.14 Position Sensing Rotary encoders on mounted concentrically to each crank, which gives continuous data of
piston location. This allows for more precise controlling of the air input, resulting in
smoother, more efficient power output from the engine. Due to size restraints, cost and the
manufacturing complexity, a regular absolute rotary encoder with a single output shaft will
be used. This shaft of the rotary encoder will be attached to the crank through a one‐to‐one
gear train. This gives more flexibility in the mounting location and wiring connections of the
encoder. The gear train will be an added cost, but cost savings from switching from a hollow
to a shaft encoder is superior.
5 Prototypes
5.1 Plastic prototype
In order to build a model, the cranks and pistons were rapid prototyped. They were then
assembled as shown in Figure 31. The model was mounted using Lego and was attached to
gears that were controlled by motors to mimic the operation of the RIICE mechanism.
Figure 31 Prototype of RIICE mechanism
36
5.2 Final Prototype
In this section, the final built prototype of the RIICE engine will be discussed in detail. The
major components are explained in the sections below including photographs. The RIICE
engine is composed of the following major components:
• Piston Assembly – Quantity of four • Crank Assembly – Quantity of two • Casing Assembly – Quantity of two • Pillow Block Assembly – Quantity of two • Engine Mount – Quantity of two
5.2.1 Piston Assembly
A total of four pistons are used in the RIICE engine. Two pistons are bolted to each crank
180 degrees apart. The purpose of the pistons is to transfer the energy supplied by the
compressed air to the cranks. Each piston is made of two components: The piston and its
face seal. These components are shown in Figure 32.
Figure 32: Piston Assembly.
37
The piston is machined from aluminum and it is this part that bolts to the crank. On the
trailing end of the piston is a face seal. This seal is screwed into the piston and prevent the
air from passing between the piston and the surrounding walls. This assembly is shown in
Figure 32.
The actual aluminum piston never comes in contact with any rotating face thus reducing
friction. This is achieved by inserting an o‐ring between the crank and piston.
5.2.2 Crank Assembly
Each crank is used to convert the expansion force applied to the piston to torque. Each
crank forms half of the bottom toroidal piston chamber. Keyed into the innermost surface
of the crank is a sleeve that transmits the torque of the crank to a set of indexing one‐way
bearings that are pressed into the sleeve. These indexing one‐way bearings in turn transmit
the applied torque to the driveshaft of the RIICE engine. On the outermost section of the
sleeve is a hardened steel collar that provides a rolling surface for the backstopping bearing.
The crank assembly is illustrated in Figure 33.
38
Figure 33: Crank Assembly.
Each crank applies torque to the driveshaft of the RIICE engine through the sleeve and
indexing one‐way bearing assembly at intermittent intervals. As one crank assembly
rotates, thus rotating the driveshaft, the other crank assembly remains stationary. Once one
crank has completed a stroke, its remaining energy is transmitted to the other crank
assembly using impact rings located on the innermost face of the crank assembly. The
indexing one‐way bearing allows the shaft to rotate freely when torque is no applied. This
enables the momentum of the engine to be conserved.
Located between the crank assemblies is an inner seal that prevents compressed air to
escape in this location. The inner seal is design to be stationary on a crank while allowing
the other crank to rotate with respect to itself. Another seal, the outer seal, is located on the
outermost section of the crank, between he crank and casing. This seal prevents air leakage
39
between the rotting crank and the fixed casing. In addition a thrust bearing provides a
smooth rolling surface between the two innermost crank faces.
5.2.3 Casing
Each casing is mounted over a crank assembly closing the toroidal piston chamber. The
casing provides mounting points for the engine mount, pillow block assembly and shaft
encoders. Furthermore, the casing is used to clamp the crank assemblies together. A total
of twelve bolts located thirty degrees apart are used to provide the required clamping force.
Figure 34 shows the two halves of the engine’s casing enclosing the crank assembly
discussed in the previous section.
Figure 34: Crank and Crank Assembly.
From Figure 34, one can observe that the casing incorporates three ports: The injection
port, the exhaust port and the intake port. The injection port is used to allow the
40
compressed air to enter the engine. Simulating the four strokes of conventional internal
combustion engines, the exhaust port is used to remove air from the engine after it has
produced useful torque. The intake port is used to draw fresh air into the engine in
preparation for the compression and power stroke of the engine.
5.2.4 Pillow Block Assembly
The Pillow Block is attached to the casing using four bolts. As shown in Figure 34 the pillow
block is mounted over the protruding section of the crank sleeve. Pressed into the
outermost surface of the pillow block is a radial ball bearing used to take any radial load
applied to the engine. These bearings are also used to keep the shaft true with respect to all
the other engine components. Also pressed into the pillow block is the backstopping one‐
way bearing. The pillow Block Assembly can be seen in Figure 35.
Figure 35: Pillow Block Assembly.
41
The backstopping bearings are to use to prevent the crank that does not transmit power to
the driveshaft from rotating backwards. This is achievable since the backstopping one‐way
bearings are pressed into the pillow block that in turn is grounded using the casing. As the
compressed air enters the toroidal piston chamber, the expansion force is equally applied to
the fore piston and the back piston. The fore piston will move and produce useful torque
via the crank and indexing one‐way bearings. The expansion force prevents the back piston
from rotating forward with the fore piston. This piston is also restricted from moving in the
opposite direction by the backstopping bearing. Therefore, the back piston and its crank
assembly will remain stationary until an energy transfer occurs between the moving crank
and the stationary crank.
5.2.5 Engine Mount
The engine mounts are used to secure the engine while it is in operation. An engine mount
is shown in Figure 36. The engine mounts also provide a study base for the engine. Each
mount bolts into the two half of the engine casing at opposite side of the casing itself. A slot
at the outermost section of the casing allows for the installation of each engine mount.
42
Figure 36 Engine Mount.
6 Testing
6.1 Open Casing ‐ Model
6.1.1 Manual Test
Objective:
Test the arrangement of the drive train assembly to verify:
• The backs piston stays stationary when force is applied. • The forward piston rotates forward and transmits torque to the driveshaft. • Cranks can be rotated intermittently.
Method:
• The drive train was assembled using the model of the crank and piston assembly. Team member manual applies force to the back piston to engage the backstopping
43
piston. Team member manually applies force to the forward piston to engage indexing bearings and transmit torque from piston to the shaft via the crank.
• High‐pressure air was used to apply force to the pistons in a similar method as described above. The air was applied at different orientations to determine how assembly of the pistons, cranks and bearing would react.
Figure 37 Crank piston assembly
Results:
1. The manual operation of the bearings successfully proved that when a force is applied to
the back piston, the backstopping piston engages immediately with no backlash. This shows
then piston crank assemblies will not rotate in the wrong direction during operation of the
engine.
When a manually applied force was applied to the forward piston, the indexing bears
successfully grabbed the shaft and torque was transmitted from the pistons to the shaft. It
was also proven that when one crank assembly rotated forward, the other crank assembly
did not move. This proved that when the indexing bearings were not engaged the shaft is
free to rotate in the same forward direction.
2. When a pulse of compressed air at 50psi was shot at the piston from a vertical position, as
shown in Figure 37, the back piston stayed in the same position, while the forward piston
44
rotated forwards. This proved the backstopping bearing can hold the piston in place against
the force of air and the torque can be transmitted from the forward piston to the shaft.
6.1.2 Compressed Air Test
Objective:
Use compressed air to continuously turn the model to evaluate the energy transfer and
dynamic characteristics between the two piston crank assemblies.
Method:
Compressed air at 75psi was shot at the model with the full drive train in place. The
solenoid valve was held horizontally with the pistons positioned 90° apart as in Figure 38A.
The air was turned off and on and the angle was adjusted to see the different effects.
Figure 38 Dynamic testing of the RIICE model
45
Results:
1. When the compressed air was turn on, the pistons in Figure 38A started to turn
clockwise, Figure 38B. When the forward moving crank impacted the stationary crank in
Figure 38C, the kinetic energy of the forward piston set was transferred to the back piston
set. This energy transfer pushed the back piston set into the path of compressed air.
2. We see that piston set with no bolts visible, in Figure 38D, is in the same back piston
location. This replacement of the back pistons occurred every rotation, and is a key factor in
giving continuous rotation.
3. There is continuous shaft rotation as the two cranks rotate intermittently.
4. A rotational speed of 620rpm was achieved.
5. The compressed air pressure of 75psi allowed for a potential torque of 175in·lb. Combing
this with the 60psi it took to induce rotation from test 1, there was and estimated 35in·lb of
torque applied to the drive shaft. When rotating at 620rpm, this torque led to an estimated
power of 0.35hp.
Backstopping Bearings
Objective:
Evaluate the performance of the new set of backstopping bearings.
Method:
The full assembly of the engine was used minus the casing. The pillow blocks were mounted
to a temporary stand that was clamped to the tabletop. Air was shot at the pistons at
various pressures, starting low and increasing to a maximum of 80psi. This was a similar
test set up as test 2, save the prototype was used in place of the model.
Results:
1. The new backstopping bearings have far less friction than the old set.
46
2. Rotation was induced in the setup at a pressure 50psi.
3. When the air was applied on piston would start to rotate and the other would stay
stationary. When the impact between the two pistons occurred, the two pistons would hit
and stick and start to rotate together about the shaft. The cause of this was believed to be
the lack of friction in the backstopping bearings. The lower friction allowed the crank
assembly to continue to rotate when energy was transfer between the two cranks.
6.1.3 Open Casing ‐ Prototype
Objective:
Test the crank and bearing assembly to verify:
• Achieving proper piston and crank motion as described in the previous section with the prototype.
Method:
• The prototype crank, pillow block, and bearing assembly were assembled. • Bracket made to support assembled components and to allow for secure grounding
of the assembly. • High‐pressure air was applied horizontally to a vertical piston face using the
solenoid valve.
Results:
During this test, compressed air ranging from a pressure of 60 psi to 120 psi was applied to
the crank and bearing assembly. This force supplied by the compressed air in this test was
not sufficient to cause the assembly to rotate indicating a large amount of friction. The
excessive friction within the assembly was found to be generated by the steel sleeves and
backstopping bearings.
1. As previously mentioned, the steel sleeves house the indexing bearings. In addition, each
sleeve is stepped down to allow for a hardened steel inner ring to be pressed onto the
sleeve. The inner ring provides the rolling surface for the backstopping bearing. This
assembly is illustrated in Figure 39.
47
Figure 39: Sleeve Assembly.
Similar to the inner ring, the indexing bearings are pressed into the sleeves. After the
fabrication of the sleeves, the pressing of the inner rings and indexing bearings was
performed. This task required a pressing force greater than expected for the inner rings,
indicating that conservative tolerances during the manufacturing of the sleeves were
employed.
Using a lathe chuck and a dial it was determined that each collar was out of round by 0.008
in and 0.011 in respectively. These imperfections were cause by pressing the inner rings
onto sleeves that were slightly larger in diameter than required. Being stretched and out of
round, the inner rings did not rotate freely in the backstopping bearings thus introducing
friction into the system.
48
Although not confirmed, it is believed that that the sleeves’ inner diameter and outer
diameter were slightly eccentric, generating an angular error. Similar to the inner ring
imperfection, additional friction would be generated into the system.
From this test, it was concluded that new sleeves were required in order to move the
project forward. Two options were then explored. First, the defective sleeves could be
turned down in order to potentially salvage the indexing bearings and inner rings. Second,
new indexing bearings and inner rings could be purchased and mounted onto the newly
fabricated steel sleeves.
The second option was chosen since it was determined that the first option required a large
time commitment and the risk of damaging the indexing bearings and inner rings was very
high. Furthermore, it was not possible to determine if the salvaged components would be of
any use. To resolve this issue, new steel sleeves were fabricated using small tolerances, new
indexing bearings and inner rings were purchased. The purchased indexing bearings and
inner rings were then installed onto the newly manufactured sleeves.
2. As previously mentioned the backstopping bearings roll onto the inner rings located on
each sleeves. These bearings are pressed into the pillow block that acts as a ground since it
is bolted to the casing. Prior to pressing the backstopping bearings into the pillow blocks,
small imperfections on these bearings were noticed.
It appeared as if the plastic cage of the backstopping bearings might have been poorly fitted
into the bearing itself. The backstopping bearings had been deemed acceptable from a
previous test. However, these bearings may have been further damaged by their continued
use or by the out of round inner rings.
A stationary lathe was used instead of a conventional press to press the backstopping
bearings into the pillow blocks. In turn, this may have contributed to the damaging the cage
of these bearings.
During this test, it was noticed that there was a large amount of friction in the backstopping
bearings. Furthermore, it was determined that the cage rubbed onto the rollers since while
rotating the sleeve inside the backstopping bearings by hand, a rubbing sound could be
heard.
49
The radial bearings and backstopping bearings were pressed out their pillow blocks using a
conventional press. Each component was inspected for damages and it was found that
indeed the cages of the backstopping bearings were damaged. The damaged cage can be
seen in Figure 40.
Figure 40: Damaged Backstopping OneWay Bearing.
One of the radial bearings was found to have some play between its inner race and outer
race. This may have been caused by the misalignments of the sleeves into the pillow block
or simply by a defective bearing. In order to resolve this issue new, more robust,
backstopping bearings were installed into the pillow blocks along with new radial bearings.
Figure 41 shows an illustration of one of the new backstopping bearings.
50
Figure 41: High Quality Backstopping OneWay Bearing.
One can observe that the edge of the cage is fully restrained within the outer shell of the
bearing. The construction of these bearings is such that both sides of the shell are
manufactured in this fashion. This is not the case for the previously used back stopping
bearings, since the outer shell only enclosed a single side of the roller cage.
6.2 Closed Casing ‐ Prototype
6.2.1 All Piston Seals – Old Seals
Objective:
Test the arrangement of the initial design of the piston seal to verify:
• The size of the pistons is correct and fit inside the casing. • The cranks will rotate easily by hand inside the casing. • How much pressure will make it turn.
51
Method:
• Piston seals were assembled on to all piston faces. • One piston of each crank was arranged at the top dead center of the engine at the
compressed air inlet. • Followed the test procedure outlined in Appendix A.
Results:
1. Cranks rotated by hand in the casing requiring some effort initially. 2. The cranks did not rotate with the air; there was too much friction due to the piston
seals. The piston seals were too thick, which added more friction than needed. Air pressure was at 130 psi.
3. Most of the air was escaping from the intake and exhaust ports. This was because the piston seals were undersized, which meant air just went around the pistons instead of pushing the pistons.
6.2.2 Half Piston Seals – Initial Design
Objective:
Test the arrangement of half the number of piston seals to verify:
• The cranks will turn inside the casing with compressed air. • How much pressure will make it turn.
Method:
• Only the leading piston seals were assembled on to the pistons. • One piston of each crank was arranged at the top dead center of the engine at the
compressed air inlet. • Followed the test procedure outlined in Appendix A.
Results:
1. The shaft rotated approximately 90 degrees with 100 psi; there was still too much friction
due to the piston seals.
2. Most of the air was escaping from the intake and exhaust ports. This was because the
piston seals were undersized, which meant air just went around the pistons instead of
pushing the pistons.
52
6.2.3 Piston Seals – New Design
Objective:
Test the design of the new piston seal to verify:
• The quality of seal. • The cranks will turn inside the casing with lower pressures. • Number of seals versus pressure that will make it turn. • Arrangements of seals versus pressure that will make it turn.
6.2.4 Sub – Test A: Quality of Seal
Method:
• Attach one new piston seal on the back of the leading piston and one old piston seal on the front of the back stopping piston only.
• Those pistons of each crank were arranged at the top dead center of the engine at the compressed air inlet.
Results:
1. The shaft rotated approximately 110 degrees with 80 to 85 psi; there is less friction due
to decreased number of piston seals as well as better sealing.
2. There was less air leaks through the ports. Most of the air leaked through the exhaust
port once the leading piston reached that position. Some air leaked through the back
stopping piston because of the undersized seal.
6.3 Sub – Test B: Eliminating Back Stopping Seal Leaks
Method:
• Attach one new piston seal on the back of the leading piston and one old piston seal on the front of the back stopping piston as well as one old piston seal on the back of the back stopping piston only.
• Those pistons of each crank were arranged at the top dead center of the engine at the compressed air inlet.
Results:
1. The shaft rotated approximately 180 degrees with 80 to 85 psi; this meant that the old
seal was had some air leaking through.
53
2. There was less air leaks through the ports. Most of the air leaked through the exhaust
port once the leading piston reached that position.
3. The leading piston did not have enough momentum for the crank to impact the second
crank.
6.3.1 Sub – Test C: Transmitting momentum from one crank to another using higher
pressure
Method:
Seal arrangement was same as in Sub – Test B.
Results:
1. The shaft rotated approximately 190 degrees with 95 psi; this confirmed that increasing
the pressure will successfully allow for a complete revolution of the shaft.
2. There was less air leaks through the ports. Most of the air leaked through the exhaust
port once the leading piston reached that position.
3. The leading piston had enough momentum to shear off the four #2 brass screws that held
the impact ring in place. This was calculated to be a force of 6.2 kN and a torque of 122 Nm.
Consequently, the second crank did not turn.
6.3.2 Sub – Test D: Compression cycle – No impact ring
Method:
Seal arrangement was same as in Sub – Test B, only attaching one old piston seal onto the
front of the compression piston.
Results:
1. The same results as in Sub – Test C were achieved. This was due to the air leaking around
the old piston seals; therefore no compression.
2. There was not enough momentum to overcome the friction produced by a fourth piston
seal.
54
6.4 High Pressure Argon
Objective:
• Determine if higher pressure (greater than 130psi) will produce continuous rotation.
Method
• Argon was used at various input pressure (150psi, 200psi, 260psi).
• Argon was chosen because it was readily available. The compressibility of argon is very similar air, which made these tests comparable.
• Different arrangements of piston faces were used to determine
Results:
1. The high‐pressure tests using argon and a high‐pressure regulator did not produce
continuous rotation of the pistons. Over an increasing range of pressure, the angular
displacement of the pistons was between 90° and 120° for multiple tests. The different
piston arrangement of piston faces did not have an affect on the on the angular
displacement. These results show that the engine’s inability to continuously rotate was not
caused by the lack of high enough pressure. One could speculate that a much greater
pressure may produce continuous rotation, but the team is confident that the problems lie
elsewhere. The ability to rotate and engine a full revolution in earlier tests at lower
pressures shows that the problem is most likely caused by the friction and improper sealing
of the engine.
7 Sealing
Benchmark Testing
Objective:
• The objective of this testing is to establish a basis of comparison for the seal testing in following sections. In addition, this test provides qualitative results as to the amount of work the engine is able to produce without sealing.
Method:
55
• The engine was assembled without seals installed. This includes the outer, inner, and piston seals. This leaves large gaps between the pistons and the casing (0.0500 in.), and minimal gaps between the cranks and the casing (0.0015 in).
• Compressed air was induced in the engine at 60 psi, and observations were made as to: how many degrees the engine turned; the amount of air escaping from the intake and exhaust ports, and flow pressure.
• Following these observations, the intake and exhaust ports were plugged, and observations were made as to: the amount of air escaping from the outer and inner seals; and flow pressure. This flow pressure is the benchmark to determine the resistance from sealing methods tested in the following sub‐sections.
Results:
• From these tests it was found that the benchmark resistance pressure with no seals installed is 40 psi with a static input pressure of 60 psi. From this the additional resistance of any seals installed can be measured.
Outer Seals
Objective:
• The objecive of the outer seal testing is to refine the seal geometry to minimize friction, and maximize sealing.
Method:
• This testing was done as an iterative process by testing sequential seal designs, until an effective seal, with low friction was manufactured. Following each iteration, recommendations were made to improve the design.
• For each iteration, the frictional torque required to rotate the crank in the casing was measured. From this measurement, it was determined whether the friction was too high. If the friction was not too high, the sealing effectiveness was measured by plugging the intake and exhaust holes, and injecting compressed air. The air was injected at a stagnant pressure of 60 psi, and the flow pressure was recorded. The difference between the flow pressure recorded and the benchmark flow pressure, is the effective sealing resistance of the seal.
Results:
1. The first seal tested was press fit into the casing, with an o‐ring located on the exterior
face, between the seal and a shoulder on the casing. This made an effective seal, but could
not be turned in the crank by hand. The anticipated reason for this is that the seal was
constrained radially and axially by the crank/ casing assembly. This caused the seal to rub
excessively on the outside diameter and the face of the shoulder of the crank. In order to
eliminate one of these constraints, it was recommended to undersize the seal axially, and
56
press fit the seal on the crank, with an o‐ring located on the interior face between the
shoulder on the casing and the seal. This eliminated the axial friction from the seal, but may
reduce the sealing effectiveness.
2. The second seal tested was designed from recommendations made in the previous test
results. In addition, in order to reduce the contact area between the seal and the casing
bore, two groves were added on the outer surface of the seal. The outer diameter of the seal
was originally cut to 3.500 in., and the inner diameter to 3.200 in. It was observed that the
seal could rotate easily in the casing or crank when inserted independently, and the seal add
an additional flow resistance of 5 psi. However, when the seal was fit between the casing
and crank, the torque required to turn the crank was measured as 10.1 ft‐lbs. Therefore, it
was recommended to undersize the outer diameter of the seal by 0.030 in. to 3.470 in.
3. The third seal tested was designed from recommendations made in the previous test
results. The torque required to turn the crank in the casing was measured as 3.1 ft‐lbs. This
method does not fully seal the casing, but when 60 psi is applied to the engine, the seal adds
an additional 1 psi of resistance. This is a low sealability, but the seal has very low friction
torque of 2.3 ft‐lbs. From these results, it is recommended that these seals be used in the
prototype testing, as we want to reduce the amount of friction in the engine.
Inner Seals
Objective:
• The objective of the inner seal testing is to refine the seal geometry to minimize friction, and maximize sealing.
Method:
• The inner seal geometry was sized based on results from the outer seal dimensions. However, since the dynamic seal only contacts the crank on the outer diameter, the seal is less constrained than the outer seal. Therefore, it was chosen to size the outside diameter of the seal exactly to the crank outer diameter, and allow a clearance between the inner diameter of the seal and crank.
• From the design the frictional torque required to rotate the cranks independently was qualified. In addition, the sealing was measured by plugging the intake and exhaust holes, and injecting compressed air. The air was injected at a stagnant pressure of 60 psi, and the flow pressure was recorded. The difference between the flow pressure recorded and the benchmark flow pressure, is the effective sealing resistance of the seal.
57
Results:
1. It was found that the seal did not have a considerable amount of friction torque,
generated a resistance flow of 1psi. It is recommended that this resistance should be
improved upon. However, due to time constraints, it is recommended that this seal be used
to test the prototype.
8 Control System
8.1.1 Solenoid Valve – LabView
Objective:
Prove the solenoid valve be fired from LabView. Test the on/off capabilities of the valve for
different pressures.
Method:
Connected the relay switch to the DAQ card of a PC. Attached the high voltage lines from the
relay to the 12V power supply. Run the LabView VI entitled “riicecontroller.vi”. The air
tanks were opened and the regulator was adjusted to the correct pressure. The ball valve
was then opened in the airline.
Results:
1. It was found that there has to be a back pressure on the solenoid valve in order for the
pilot valve to close the airflow.
2. The minimum pressure require to turn on and off the valve with the required ¼” NPT
fitting was 40psi.
3. The valve open and closed on demand for all pressures ranging from 40 to 130psi.
8.2 Rotary Encoders ‐ LabView
Objective:
Read the rotary encoders in LabView and be able to manipulate the data to trigger the
solenoid valve.
58
Method:
The encoders were attached to the DAQ device of a PC and brought into LabView. A
LabView program entitled “RIICE Control Encoders.vi” was used to read the encoder counts,
manipulate the data and output the number of counts of each encoder, the angular position
of each encoder relative to the starting location and a sinusoidal plot of the angular position.
Results:
1. The program successfully read, manipulated and output useful data about the angular
position of the data.
2. The encoders count start at zero and counts up while the encoders are rotated, no matter
what the direction or rotation.
3. The index location is a fixed location on the encoder and increments at the same spot
regardless of starting position.
8.2.1 Solenoid Valve and Rotary Encoder ‐ LabView
Objective:
Fire the solenoid valve independently twice per revolution of each rotary encoder.
Method:
Both the solenoid valve and each rotary encoder was attach to the DAQ device of a PC and
read into a LabView program entitled “RIICE Control Valve and Encoders.vi” The rotary
encoders were setup on a Lego platform and were turned independently by two electric
motors.
Results:
1. Each encoder fired the solenoid valve twice per revolution.
2. This system can pulse the air much faster than the design goal of 120rpm (corresponding
to 240 pulse per minute).
3. The system could continuously pulse even at high pressures 130psi.
59
4. It was determined that the optimal design of the controller should be allow of dynamic
control of the amount of air input into the system.
5. It was decided that in order to avoid and cumulative error in the encoder counts, the total
number of counts should be reset every time the index pulse increase.
8.2.2 Pulsating Air ‐ LabView
Objective:
Determine whether or not pulsating the input of air into the engine would aide in the
rotation of the pistons.
Method:
With the rotary encoders and the solenoid connected to the DAQ device and imported into
LabView. The encoders were mounted to Lego test setup and turned using electric motors,
similar to Test 3. The solenoid valve was attached to the engine and connected to a
compressed air source at 95psi. The electric motors were turned on and the program was
initiated. The ball valve on the air hose was opened and the air began to pulsate. All four
pistons had one face seal attached to the trailing edge.
Results:
1. The initial movement caused by the first pulse of the pistons was approximately 100°.
The piston slowed and began to slowly rotate forward until the impact rings between the
two cranks came into contact.
2. The pistons moved together for an unknown angular displacement. This was not
measured because the rotary encoders were being use to pulsate the flow, not to measure
the angular position.
3. Eventually the crank piston assemblies ceased to move as not enough force was
generated by the air to overcome the friction of the piston faces.
60
4. It was noticed that shaft continued to rotate very slowly even after the cranks had
stopped. It was noticed that cranks were expanding axially along the shaft with each pulse
of air. It was determined that this axial expansion was causing the shaft to “walk” in the one‐
way bearings.
5. It was found that when the cranks stopped, if the intake port was blocked than the cranks
would start again. This rotation continued until the cranks reached the location of the
exhaust port. Plugging the exhaust port did not have much of an effect on the rotation of the
cranks.
It was deemed that plugging the ports to aide in the rotation of the cranks while pulsing the
air was not beneficial since in order to have a positive effect the exhaust port would have to
be plugged and then the intake port. This was not a practical design change to incorporate
and an unreliable method of turning the engine.
6. Different air pressures were tested in an effort to gauge the effect or pressure on the
pulsating flow. The best results occurred with 50psi, which resulted in a complete shaft
rotation, but that was before the team discovered when the pistons had stopped moving. It
was determine that pistons traveled about 180 degrees.
9 Budget
The development, manufacture and fabrication of the RIICE conceptual design is presented
in this section under two sections. First the expenses of the project are presented.
Following are the man‐hours spent by Dalhousie University machine shop technicians as
well as the man hours spend by the group. The entities responsible for the fabrication of
each component of the project are also presented.
The allotted budget for the RIICE project was $1500. This amount was to be expensed as
follows: $200 during the fall term and $1300 during the winter term. The itemized cost of
the project over both terms is tabulated in Table 5.
Table 5: RIICE Itemized Cost.
Item Cost
Drive Train $456.52
61
Model $155.68
Materials $241.82
Air System / Sealing $453.58
Hardware $27.74
Miscellaneous $42.28
Total $1375
Adding the cost of Table 5 yields a total project cost of $1375, which is under the assigned
budget of $125. Machining expenses were not added to Table 5 since the group was not
responsible for this cost. The machining and fabricating time however is presented in Table
6.
Table 6: Machining and Fabrication ManHours.
Fabricator ManHours CNCHours
Advanced Manufacturing
Group
4 24
Dalhousie University 140 0
RIICE Design Group 350 0
Pricing Dalhousie University’s and the Advanced Manufacturing Group machining time at
$60/hour, a machining cost of $8640 is estimated. Note that this cost is excluding the CNC
machining time since CNC rates are specific to the work being performed. A summary of the
division of the fabricating work supplements the information presented in Table 6. This
information is tabulated in Table 7.
Table 7: Fabrication Work Division.
Component/Feature Fabricator Comment
Casing Dalhousie University Excluding Fillet
Casing Fillet Advanced Manufacturing
Group
Fillet Only
Control System RIICE Design Team
62
Crank Dalhousie University Excluding Fillet
Crank Fillet Advanced Manufacturing
Group
Engine Mount Dalhousie University
Inner Seal RIICE Design Team
Outer Seal RIICE Design Team
Pillow Block Dalhousie University
Piston RIICE Design
Team/Dalhousie University
Piston Seal RIICE Design Team
Sleeve RIICE Design Team First Iteration
Sleeve Dalhousie University Second Iteration
10 Design Evaluation
The inability of the RIICE to continuously rotate can be attributed to two main factors:
friction and sealing. These two factors were discovered through multiple tests with
different sealing and piston configurations. There were subcomponents in the design that
functioned quite well. The drive train consisting of the unique one‐way bearing
arrangement performed well, with open case testing achieving speeds of 620rpm. These
results gave the team confidence in the RIICE concept and the design. The main source
friction came from the piston faces. The tests conducted with few piston faces generally
achieved great angular displacement, with the best results coming from tests with only two
seals. There exists a very delicate relationship between sealing and friction. In order to
properly seal the pistons with face seals require precision machining and iterations in the
design to find the optimal shape, material and configuration on the pistons. The
manufactured and tested two different piston seal designs with improvements made
between iterations. In order to have successful operation of the engine, more investigation
into relationship of friction versus sealing is required.
63
11 Future Considerations In further development and testing of the rice engine it is recommended to investigate a
method for starting the rotation of the crank assemblies before sustaining the rotation of
the engine with compressed air. With some modifications to the current design it would be
possible to implement a pull cord or an electric starter motor the start the engine to the
desired engine rpm. The addition of a flywheel to the shaft can help smooth the intermittent
motion of the cranks. A modification to the current cranks is needed to reposition the
pistons to the desired location between tests without disassembling the engine. This would
allow for significantly quicker engine startup time. A larger solenoid valve is needed to
permit a greater range of input pressures.
The difficult balance of sealing and friction will necessitate testing different seal designs
with different materials. The current use of Acetal seals can be replaced with better
materials such as Teflon or Ultra High Molecular Weight Polyethylene (UHMWPE). Teflon is
non reacting and extremely low friction and UHMWPE is abrasion resistant, impact
resistant, non‐sticking and self‐lubricating. These mechanical properties make these
materials ideal for sealing.
12 Conclusion
The RIICE is an innovative concept that holds a promising future. The team believed in the
RIICE concept and the design, and developed a strong passion for the project. The team
achieved 12 of the 16 design requirements. Successful testing of the open casing proved the
design that the one‐way bearing drive train. Sealing and friction are a major issue of the
design and prevented continuous operation of the engine. The team proposed future
recommendations to render the operation of the engine successful. These recommendations
include methods of starting the engine, such as a pull cord or a starter motor and a flywheel,
and investigation of different materials for sealing.
64
APPENDIX A – Detailed Control System
65
User Manual for Engine Testing
1. Open the engine casing and align pistons to the correct starting location. 2. Bolt the two halves of the engine casing together. Bolts should be snuggly tightened
using a wrench. 3. Fasten the engine casing to a bench top and insure it will not move or vibrate off the
table. 4. Wire the solenoid valve to the 12V power supply. Insure power supply is turned off.
The flagged wire lead is connected to the 12V terminal, the other is connect to ground.
5. Connect the rotary encoder leads and the relay switch board leads to the DAQ Card pin‐out block. Follow the pin‐outs in the table below:
Solenoid Valve Encoder 1 Encoder 2
Relay Ground
(1)
Port 33
(Digital ground)
5V Port 35
Digital 5V
5V Port 34
(Digital 5V)
Relay Switch
(2)
Port 27
(Digital line 0)
Ground Port 33
(Digital ground)
Ground Port 24
(Digital
Ground)
Relay Ground
(3)
Port 33
(Digital ground)
Quadrature
A
Port 47
(Counter 0)
Quadrature
A
Port 41
(Counter 1)
Relay Switch
(4)
Port 25
(Digital line 1)
Index Port 29
(Digital line 2)
Index Port 26
(Digital line 4)
6. Start LabView program “Master Control_Rev 3.vi” 7. Run Program. The following graphical user interface (GUI) should appear:
66
8. Turn on the 12V power supply 9. Insure the green LED on the GUI is dark, this means the valve is energized and
closed. This can be done by adjusting the slider on the right side of the GUI. SAFETY NOTE: The current solenoid valve is a NORMALLY OPENED valve, meaning
that unless energized the valve will be open and air will flow.
10. Turn on the compressed air supply. 11. Adjust the pressure regulator to the desired input pressure. 12. Test the solenoid valve before the attaching to the engine.
a. Attached the air hose to the solenoid valve at port 1, marked on the valve. b. Insure port 3 of the valve is plugged. c. Port 2 is the valve outlet. Hold the valve close to a table to restrict the air
flow d. Insuring the green LED on the GUI is dark, open the ball vale in the air line,
all the while keeping the valve held against the table. The solenoid has a pilot valve that uses backpressure to close.
e. Open the ball valve. f. The solenoid valve, being energized, should close g. If the valve does not close, close the ball valve and check the wiring of the
solenoid to the DAQ pin‐out. h. When the valve is closed, test the on/off capabilities, which is done by
adjusting the slider value on the GUI. The slider can be either increased or decreased, turning on and off the green LED, which corresponds the opening and closing of the valve.
i. Close the ball valve. j. Deplete the air in the hose and solenoid valve.
13. Attach the female quick connect fitting to port 2 of the solenoid valve. 14. Attach this quick connection to the male quick connection on the engine. 15. Stop the LabView program 16. Adjust the slider bar to the desired triggering values. 17. Open the ball valve. 18. Press the white arrow in the top left of the screen to run the component.