MECHANI

CALENGI
NEERI
NGPROJ
ECT
ACADEMI
CYEAR2
0
1
3

DEP
ARTMENTOFMECHANI
CALENGI
NEERI
NG
CHULALONGKORNUNI
VERSI
TY


2556

 2556

2556 13 .. 2557
2103499
2556
2103499 2556: .. ( ) ..
..
:
: http://www.me.eng.chula.ac.th/seniorproject

Mechanical Engineering Project, Academic Year 2013

Copyright Department of Mechanical Engineering, Chulalongkorn University
The Seminar on Mechanical Engineering Project, Academic Year 2013 on Thursday 13 March 2014 at the
Faculty of Engineering, Chulalongkorn University, is a part of the course 2103499 Mechanical Engineering Project,
Academic Year 2013 in the Bachelor of Engineering Programs in Mechanical Engineering, Automotive Engineering and
Naval Architecture and Marine Engineering.
2103499 Course Committee for 2013: Assc. Prof. Angkee Sripakagorn (Coordinator), Assc. Prof. Kuntinee Maneeratana
and Asst. Dr. Werayut Srituravanich
Cover Design: Mr. Chaiyasase Sangkhamaneenakorn
File Download: http://www.me.eng.chula.ac.th/seniorproject


Applied Mechanics
AM01
AM02
AM03 end milling rotary
AM04
AM05 Design a mechanical timer for AT-101 stove
AM06
AM07
AM08 specific damping capacity

AM09 Surface finishing of femoral head of prosthetic hip to meet ISO standard
AM10 The design of femoral component and guide component for knee prosthesis
AM11
AM12
AM13
AM14
AM15
Automotives
AU01
AU02 force feedback steer- by-wire
AU03 Forklift driving simulator
AU04

AU05
AU06
AU07
Thermo-Fluids
TF01 Effects of azimuthal control jets to main jet mass flowrate ratio on the entrainment of a jet in crossflow
TF02 ( .)
TF03

TF04

TF05 Taylor bubble

TF06
Master Program-Oriented Projects
MS01 Effects of crossflow Reynolds number on the entrainment of jet and controlled jets in crossflow
MS02
POD
MS03 A sorting of micro particle with microfluidics system
MS04
MS05
MS06
MS07 Ejector
MS08
MS09 A reconfigurable open architecture controller for a research robot
MS10 Design and control of a delta robot to work with human in real time
MS11
MS12
MS13
MS14
MS15
2555

2
8
14
18
26
28
34
40
46
48
54
60
62
68
70
76
82
84
90
96
98
100
106
112
114
120
122
124
126
132
138
144
150
156
162
164
166
172
178
184
190
196
202
208
210

2103499 2557 AM01

A Development of Closed-loop Controlled Micro-Gripper

5330226121 5330237821
..
( )

Arduino board

-


disturbance

Keywords: micro-gripper, displacement, controlling system

1.








2556 [1]









Arduino board UNO3

: , ,
Abstract
This project is to develop the thermal actuated microgripper, and its controlling system to manipulate the
displacement of micro-gripper. With this actuation technique,
the displacement is varied with the grippers temperature. In
addition to material properties, the material resistance is
usually varied to its temperature so that this project chooses
the material resistance as a controlling parameter instead of
grippers displacement or temperature which is more difficult to
deal with. This project is consisted of system design, input and
output function calibration, and software algorithm design by
using Arduino board. The calibration of input and output
functions showed that the results from experiments provide
almost the same with those from theories. Using closed-loop
feedback control with simple close-open algorithm for the
heater, we found that the micro-gripper will reach to a target
temperature more quickly and precisely than that using feedforward control especially when a disturbance was introduced
to the system.

2.


16190.08 mm3
0.07 1
(120.250.08 mm3 )



2

2103499 2557 AM01

(Motor driver circuit)

3 4

3
1 ( )

4
4.





()

3.
Arduino board

(Amplifier circuit)

4.1.
(0-200
)



ANSYS

2103499 2557 AM01

Thermal-Static mode



(DC Power Supply)

Thermocouple
Dino-lite
5
6



uncertainty
9-10

7
[2]

8


(50-100 m)


1.5A


9
0.25 m/C)

8

6
4.2.

2103499 2557 AM01

5
Thermocouple
12
13
Polynomial order 2nd 160C
8V ( )

4.3.

(0.01-0.1 V)


5-10 mV



Op-Amp UA741CN
6V non-inverting 10 R2=
4000 R1 = 100

11
0.042

y 0 offset
Op-amp

12

R2

R1

10 Op-Amp non-inverting

13

6.
-

( 8.5V )


()




- 10 50
-

14 x

11

5.







Ceramic power resistor 20W
5

2103499 2557 AM01

(), Vgref ,
Vinref Vgref Op-amp ,
Bref , B

15 60

7.2. disturbance

16 0.5
disturbance





16
disturbance

14
7.

feedback feed-forward 2


disturbance

7.1.
2


()

2



15
feedback

10
-

16 disturbance
8.

Arduino board

Ceramic power resistance

2103499 2557 AM01

-

STAR
Micro-Nano Fabrication
Technology Research (GSTAR 56-005-21-002)

..
.. (
)

[1] ,
,
, , ,
, , 2555
[2]

2103499 2556 AM02

Design and Manufacturing a Prosthetic Ankle Joint and Foot for Amputees
5330229021, 5330233521 5330464121
.. ( )

(Gait
Cycle) ISO 10328


0.13
ISO 10328

(Vacuum
infusion)
4.3 11.97

: , ,

80 8,000
1,000

ISO 10328

ISO 10328

Abstract
This project aimed for designing and manufacturing a
prosthetic ankle joint and foot in order to help amputees to
walk as normal people do. In designing methods, a gait cycle
and a ISO 10328 standard were employed with design
concepts; simplicity, stability, energy storage and safety. Using
finite element software to calculate, this foot can stored energy
up to 0.13 J/kg and passed separate cyclic test for ankle-foot
devices and foot units according to ISO 10328. A carbon fiber
material was formed as a prosthetic foot body by a vacuum
infusion method. A prosthetic foot prototype was produced and
able to swing with 4.3 degrees dorsiflexion and 11.97 degrees
plantarflexion which is similar to normal people's ankle.
Keywords: Prosthetic, Ankle, Foot, Carbon Fiber

2.
2.1

2 (Dorsiflexion)
(Plantarflexion) 1

1.

1,374,133
50,000 45%
25% 20%
20%

Dorsiflexion
Neutral Position
Plantarflexion
1

AM02
8

2103499 2556 AM02

(Ranjan Das)
3 5
12 -15


(Stability)

90%
80%
20%
10%

2.2 (Gait Cycle)

2
1) (Stance phase)
2) (Swing phase)

8

4 [2]
2.4


(Comparison of methods for the
calculation of energy storage and return in a dynamic elastic
response prosthesis) [3]

5
(Rotational
power and directly measured ankle kinetics)
0.09

Vertical force [N] Fore-aft force [N]

2.3

(Changes in walking
pattern caused by the possibility of a tripping reaction) [1]


10%

3
4

5:
2.5 (Failure analysis)
2.5.1 (Soderberg)

Ground Reaction Force

(1)
(2)
% of stride

(3)

3
AM02
9

2103499 2556 AM02

a
m
min
max
Se (Endurance Limit)
Sy (Yield
Strength)
n (factor of safety)
1
2.5.2 (Endurance Limit)
(Rita
Roy) [4]
71% 85% (Yield Strength)
3.
3.1

2

3.1.1





3.1.2





2.1





3.1.3

2.4

3.1.4



ISO 10328
3.2





-
(Light sport)
- 0 6
5 20
-

-

2.4
- 18-40
50-100
3.3




(Body) 2 (Twill
weave)
6

1) (Density) 1.356 g/cm3
2) (Youngs Modulus) 64.4484 GPa
3) 797.98 MPa

AM02
10

2103499 2556 AM02

3.4 (Detailed Design)

78

3.4.1 CAD (CATIA)


ISO 10328



(Stainless Steel) 7

7
3.4.2 CAE (ANSYS)
(Elastic region)

(Static Structural)

(Deformation) (Stress)
(Strain) (Strain Energy)
(Constraints)

(element) 4860
(frictionless contact)

3.3


10
10%
(point
load)
3.4.2.1

8 10.6 8a
4.9
8b

8a)
8b)
8
3.4.2.2


9

0.13 80 %

0.13 J/Kg

9
3.4.2.3


10
3.3

10
4 ISO 10328
ISO 10328

AM02
11

2103499 2556 AM02

4.1

2
11

5.1 (Mold)






13

1 1 ISO 10328

71% (797.98 MPa)

566.58 MPa
4.1.1 (proof strength test)

794.87 MPa

4.1.2 (fatigue strength test)


634.17 MPa
5.6409 x 10-6 MPa
(1)
1.045 12

13
5.2


(Vacuum Infusion) 14

2 (Twill weave)
200

14:

15
16
1 2
5


(Carbon fiber)

(Vacuum infusion)

15:

16:
AM02
12

2103499 2556 AM02

5.3


17

17:


18
(Motion Analysis Microscope) 125

18:

4.3 11.97
6.
(stiffness)

7.

CAE 10.6
4.9
11.97
4.3

0.13


ISO
10328

1.045

.. ..

Mr. Rommel Cruz

[1]: Changes in walking pattern caused by the possibility of a
tripping reaction []. :
http://www.sciencedirect.com/science/article/pii/S09666362
01001102 ( : 7 2556).
[2]: Biomechanikstudie_MRI_AKT-O_EN [].
:http://www.vacoped.com/media/files/pdfdateien/Biomechanikstudie_MRI_AKT-O_EN.pdf
( : 5 2556).
[3]: Comparison of methods for the calculation of energy
storage and return in a dynamic elastic response
prosthesis [].
http://www.sciencedirect.com/science/article/pii/S00219290
00001020 ( : 10 2556).
[4]: The influence of energy storage and return foot stiffness on
walking mechanics and muscle activity in below-knee
amputees []. :
http://www.me.utexas.edu/~neptune/Papers/clinbiomech2 6 (
10).pdf ( : 20 2556).

AM02
13

2103499 2556 AM03

end milling rotary

Improvement of end milling cutter by rotary technique
5330290221, 5330291921 5330392621
.. ( )

stationary


end milling

stationary

self-propelled
rotary (cutting mechanism)

CAD
ANSYS Workbench
Finite Element

ISO 8688-2: 1989(E)

rotary stationary

self-propelled rotary
stationary 2.86

propelled rotary tool holder compared to stationary tool

holder.
1.



(ISO 8688-2: 1989(E))

stationary

selfpropelled rotary 1.
2.

3.

4.

rotary


rotary end milling

end milling stationary
end
milling rotary

Abstract
Milling have significant role in manufacturing, especially
end milling. With standard stationary tool holder when its tool
life is reached, it cost large amount of money when need to
change an insert. This research concept is to extend the tool
life of an insert by using self-propelled rotary method, which
an insert will rotate around itself and propelled by it cutting
force. While the cutting mechanism is changed due to rotary
effect, tool life of an insert is increased.
The design and analysis of the self-propelled rotary tool
holder is done by CAD and finite element on ANSYS
Workbench program. The cutting force and cutting
mechanism are obtain from journal[1] and compared its tool
life with stationary tool holder by test cutting guided by ISO
8688-2:1989(E) standard in the same cutting environment.
The result shown a significant increase in tool life of self-

2.
2.1 face
milling self-propelled rotary[1]

14

2103499 2556 AM03

3
3.1
3.1.1

3.1.2

3.1.3
3.1.4
3.1.5

1 [1]
2.2
[2]
face milling self-propelled rotary
inclination angle 2

3.2
CAD ANSYS[4]

3

2 face miling self-propelled rotary

inclination angle [2]
inclination angle
45o

3
4

2.3
ISO 8688-2: 1989(E)[3]

- SteelC45

- end milling
feed per tooth, feed table, cutting speed, axial depth
radial depth
-
(flank wear)

Uniform wear : 0.3 ..
Localized wear : 0.5 ..

4 exploded view

15

2103499 2556 AM03

5.
-
CNC
CNC 3

7

1. M3
2.
taper

3.
4. fillet

5.

6.

inclination angle 45
5
4 bottom view

7
-


8

5 bottom view
4.
stress analysis finite element
ANSYS


Maximum stress fillet
1047.7 MPa

AISI4340 SNCM439

8
water jet 9

6 stress analysis
9 water jet

16

2103499 2556 AM03

water jet
10

steelC45
face milling

milling Lagun
spindle speed
Tachometer 14 feed per tooth, feed
table, axial depth radial depth milling
1

10
- Kyocera
11

11

14 Tachometer

- 12

1
parameter
specification
Table feed,Vf (mm/min)
825
Feed per tooth, fz (mm/tooth)
0.3
Cutting spindle revolution, n (rpm) 2750
Cutting speed, Vc (m/min)
276.46
axial depth (mm)
1
radial depth (mm)
5
cutting length (mm)
500

12

- stationary
Kyocera

motion
analysis microscope VW-Z1

frame rate 2000 15

13 stationary
6.
ISO 8688-2: 1989(E)
milling Lagun


cantilever beam

15 motion analysis microscope

17

2103499 2556 AM03

(flank wear) motion
analysis microscope VW-Z20R
16

uniform wear
localized wear stationary
2

18
2
stationary
cutting length(m) time(sec)
0.5
1
1.5
2
2.5
3
3.5

36.36364
72.72727
109.0909
145.4545
181.8182
218.1818
254.5455

tool deterioration
Unifom wear(micrometers) Localized wear(micrometers)
105.51
0
126.01
0
166.65
277.04
194.59
382.93
221.55
447.99
246.78
475.29
263.83
558.03

16 motion
analysis microscope
uniform wear
localized wear 17 ()

ISO 86882:1989(E) rotary
stationary

18
stationary

6.

self-propelled rotary
17 ()

uniform wear
localized wear stationary
3

19
3
self-propelled rotary
cutting length(m) time(sec)
0.5
1.5
3.5
5.5
7.5
9.5
10.5

17 () ()

18

36.36364
109.0909
254.5455
400
545.4545
690.9091
763.6364

tool deterioration
Unifom wear(micrometers) Localized wear(micrometers)
28.62
33.21
37.95
47.17
55.47
60.32
61.06
-

2103499 2556 AM03

self-propelled rotary









self-propelled rotary


stationary
self-propelled
rotary 5833
stationary
stationary 8

2036
self-propelled rotary
stationary 2.86

19
self-propelled rotary
5.



20
r = 10.35 ..
a = 6.75 ..
n = 2740 = 0.02189
20

10.35 10.35
6.75
69.645

0.004236
60

6.
.
.

14164.30
2360

7.
[1] Pradeep Kumar Baro, Suhas S. Joshi and S.G. Kapoor.
Modeling of inserts in a face-milling operation with selfpropelled round insert milling cutter. International Journal of
Machine Tools & Manufacture 45 (2005): 831839.
[2] Kaushikkumar M. P. and Suhas S. J. Mechanics of
machining of face-milling operation performed using a selfpropelled round insert milling cutter. Journal of Materials
Processing Technology 171 (2006): 6876.
[3] ISO 8688-2:1989(E), Tool life testing in milling part 2 :
End milling
[4] ANSYS,Inc. ANSYS Workbench Academic Research,
Release 12 [Computer Program]. 2010.

20 diagram

19

2103499 2556 AM04

Design and Development of 3D Printer by Delta Mechanism
5330242121 5330347421
5330508121
.. ( )

G-Code
2 1.


(Feedback
Control)


2.
3

Processor,
Field
Programmable Gate Array (FPGA) Microcontroller


G-Code

Abstract
This paper proposes the design and development of a 3D
printer with the Delta mechanism. The printer can be
commanded by G-code directly or from the 3D model. The
development is divided into two parts including 1. The structure
and mechanism that is driven by three permanent magnet DC
motors with gears and encoders in order to generate force to
control the position of the mechanism in the closed loop
fashion. This is different to a 3D printer in the market which is
normally driven by steppers and open loop controlled. The
structure is designed with fewer parts to reduce cost and time
to assemble. 2. The control system that is powered by three
processors to control the DC motor and process the encoder
signal at high speed. The processors consist of the Intel
CORE2 processor, FPGA, and Arduino MEGA 2560. The
experimental test demonstrates that proposed 3D printer

receives the G-code command and move in response to the

command accordingly
Parallel Manipulator, Permanent DC Motor
1.





(3D Printer)
(Material Addition Process)[1]




X-Y Z (Cartesian 3D
Printer)
[2]
Z





2.
[3] (Delta Mechanism)
(Parallel links)
3 (Closed
Kinematic Chain)
1

Linear Slider Guide

(Inverse Kinematic) 2

20

2103499 2552 AM04

2
Slider Linear Guide

3.
3.1.

2 Laser
Cutting
3
4

7
3
Laser Cutting

3.2. (Permanent DC
Motor)
(Closed Loop Control)
(Feedback Control)

4.2 (Inverse Kinematic)

(Inverse kinematic)[1]
Cartesian 3

Inverse kinematic
prismatic
8

5 Permanent DC Motor
3.3

6
4.
4.1


(Inverse Kinematic) 7

21

8
Origin J J (0,0,0)
A (x,y,z)

2103499 2556 AM04

- (JK) = 86.5
- (CF) = 250
-
(CB) = 53
- (AB) =
10

1. K L M (z = 0)
K : Kx = JK cos210 Ky = JK sin210
-(1)
L : Lx = JK cos330 Ly = JK sin330
-(2)
M : Mx = 0
My = JK
-(3)
Z x y F G
H K L M

9 3
2. C D E
C : Cx = X + CBcos210 Cy = Y + CB sin210 (4)
D : Dx = X + CBcos330 Dy = Y + CB sin330 (5)
E : Ex = X
Ey = Y + CB
(6)
3.
F C : XFC = Fx-Cx YFC = Fy-Cy
(7)
G D : XLD = Lx-Dx YLD = Ly-Dy
(8)
H E : XHE = Hx-Ex YHE = Hy-Ey
(9)
4. C D E Slider Linear Guide
C KF: CtoKF =
(10)
D LG: DtoLG =
(11)
E MH: EtoMH=
(12)
5. Z
C F: ZCF =
(13)
D G: ZDG=
(14)
E H: ZEH =
(15)
6.
Z
K F: KF = ZCF+AB+Z
(16)
L G: LG = ZDG+AB+Z
(17)
M H: MH = ZEH+AB+Z
(18)
A (x,y,z)

(Inverse Kinematic)

. Slider Linear Guide

4.3
4.3.1
(Permanent magnet DC
motor) 12
12 V Polulu Motor Driver


Polulu Motor Driver

4.3.2

MATLAB
Simulink

MATLAB
XPC (Real-time computer)

1 KHz XPC

FPGA XPC

Pulse width modulation(PWM)
PWM

(Encoder)
FPGA
XPC
MATLAB


5.

PID controller
(Encoder)

(overshoot )

G-code

X Y Z
MATLAB

22

2103499 2552 AM04

23

2103499 2556 AM04

5.



DC motor driver
motor driver



Pulse
width modulation (PWM)


high 11

11 Pulse width modulation

(voltage control)
PID (PID controller)


6.
(encoder)
(encoder)

12

12 encoder



High Low
Encoder
Encoder

Encoder Open collector

Encoder
Interface circuit
Encoder


Encoder Open collector Interface circuit

Encoder

7. (Hot End)
7.1

100k Thermistor



Thermistor
Voltage divider circuit 13

13 Voltage divider
(Vref)
(Vmeasured)

Ramp 1.4
Voltage Divider
Thermistor
Thermistor


thermistor
7.2 Controller Diagram

14 (Hot End)

24

2103499 2552 AM04

7.3 Control Law
U = PID(e)

(19)

8.

2
8.1 2
8.1.1

15 16

G-Code




17 18

G-Code

9.

G-Code
G-Code

15

..

8.1.2

16
Catia
8.2

8.2.1

[1] John Michael Brock. Fused Deposition Modeling (FDM)
Material Properties Characterization. ME 222 Final Project.
2000.
[2] A. Meyer. Delta vs. Gantry 3D Printers. Website:
https://groups.google.com/forum/#!msg/i3detroitpublic/3prXmFlv5BM/vyXPBD6sbCEJ. 1949
[3] J.-P. MERLET. Parallel Robots, G. M. L. GLADWELL, Ed.
Parallel Robots. ch. 2, pp. 19-59. Canada. 2006.
[4] Craig, J. J. Introduction to ROBOTICS: mechanics and
control. 2nd edition, Addison-Wesley Publishing Company,
1989.
[5] Wikipedia. Extruder lemio.svg.
http://reprap.org/wiki/File:Extruder_lemio.svg (Online). 2011

17
8.2.2

18

25

2103499 2556 AM05

Design a Mechanical Timer for AT-101 Stove

Jinjutha Khokasai 5330067721, Settawut Tanthanasarit 5330513221, Apiwit Puangsricharern 5330525821
Professor Tawan Paphapote (Senior Project Advisor)
Timer knob was added beside the Igniter knob as shown in
Figure 1. Timer knob has two working zone, Free zone for normal
uses and Timer zone for setting a time limit as shown in Figure 2.

Abstract
This report shows the background of the Mechanical Timer
project and designing processes. The process starts from make
the objective, and then decides on design criteria which are coagreement from the meeting with Lucky Flame Company. After
that is making a conceptual design to be a rough image of how
our product is going to works. Next process of the design is
making detail design which has to cover all of the detail from the
previous design, such as materials, dimensions, locations. And
lastly, use the data from detail design to make a prototype to test
in the real situation.
Introduction

Figure 2 Timer knob and Igniter knob

A How Free-zone works.

A1 Start from initial position
A2 Twist timer knob to Free-zone (Push-in and twist to the
left)
A3 Use Igniter knob to open-close the stove

Stove AT-101 from Lucky Flame Co,.Ltd has the highest sell
rate among other products. To add value and particularities to
this product, Lucky flame wants to develop a mechanical timer to
install in AT-101 because it has more stability than electrical one
and safer when using with gas system.

B How Timer-zone works.

B1 Start from initial position
B2 Twist Timer knob to Timer-zone (Push-in and twist to the
right)
B3 Use Igniter knob to open-close the stove
B4 When the time runs out, Timer knob will return to its initial
position

Objective
Design Mechanical Timer for AT-101 Stove.
Design Criteria
Material costs do not exceed 100 baht.

Life cycle more than 10000 times.

The Mechanical Timer installation will not affect gas

ignition system.

Simple design.

Have safety system that prevents gas leaking.

Timer range starts from 1 min to 1 hour.

Detail design
Originally, AT-101 stove only consists of Igniter valve, but our
design added an addition Timer valve to control the gas flow.
Timer valve is an extensional part which controlled directly from
the Timer knob.
Design1 Safety system is not only control Timer knob to use
after Igniter valve is turn off, but also make the user turn
Igniter off after the Timer valve cut out. Safety mechanism in
design 1 consists of 4 main parts: Locking Gear, Locking
Teeth, Bar and Non-Symmetric Ring as shown in Figure 3.

Conceptual design
Designs dimension is equal or less than igniter.

Timer valve works when the knob is push while turning.

Our design

Figure 1 Location of Igniter valve, Timer valve and gas flow

Figure 3 Mechanism design 1
26

2103499 2556 AM05

Conclusion
After the meeting with Lucky flame, they decide to choose
design 2 for production. Design 2 may not cover all of the human
possible error but in the term of marketing the production cost
and processes is very suitable and is in the design criterias
cover. Moreover, the possible case where both valve is open is
not because our product malfunction but from the user action. We
also made the safety manual for the user for Lucky Flame. CAD
files that we send for Lucky Flame approval are shown in Figure
5 and Figure 6.

In the design, after the timer cut off it cannot be open again if
Igniter valve is not at its initial position because the Locking
teeth will prevent it by Locking gear which is free wheel
mechanism, and Timer knob cannot switch between Freezone and Timer-zone neither. If user wants to switch zone
they have to push the switch button before changing it and to
do so, Igniter valve must be close.
Design2 Safety system focus mainly at Timer knob to
prevent an accident opens at timer valve from external force.
Safety mechanism consists of two main parts, locking key
and locking socket as shown in design 2.

Figure 5 Side view of the stove assembly after installed

Mechanical Timer

Figure 4 Mechanism design 2

Timer knob in this design requires user to push-in the knob

before twist to prevent an accidental Timer valve open from
user. When pushed, the knob along with Locking key will
move pass the Locking socket and cover Timer shaft. After
that, Timer knob and shaft will move together either to Freezone or Timer-zone to set the time.
Prototype
New and modified parts created by 3D printing. The parts
that use 3D printing cannot be used to test the Mechanical
Timers mechanism because the gas will leak out, but shows how
the assemble works and how the mechanism fit in the stove
without disturb Igniter system.

Figure 6 Top view of the stove assembly after installed Mechanical Timer

Discussion
Both design 1 and 2 succeed in our intent objective and can
successfully install in AT-101 and work correctly.
Design 1 cover all of the possible human errors but has
complex design and requires more new and modified parts,
furthermore the production cost and processes is high and
not worthy in term of marketing.
Design 2 has far more simple design. The only disadvantage
of design 2 is that there is possible accident which user open
Timer valve before closing the Igniter valve.

27

2103499 2556 AM06

Design and Construction of Automatic Control Unicycle Robot

5330402221
5330106521, 5330270721
.. ( )

(Unicycle Robot)


CATIA MATLAB

LABVIEW


(
)

, ,

Abstract
The objective of this project is to study how to design and
fabricate the unicycle robot. The early study initiate from
solving dynamic model via Lagranges equation, which allow us
to understand the systems behavior. After we obtained the
result, CATIA and MATLAB software are used to aid the
designing process before the fabrication. LABVIEW software is
used in designing the control system by controlling the
longitudinal and lateral axis. The functional objectives of the
Unicycle Robot are to build a robot that is structurally strong,
and able to balance itself at equilibrium position.
Keywords: Unicycle Robot, Lagranges equation, the
control system
1.




(Automatic Control)

2.

1. 2.

3.

1
2.1

(Dynamic
Behavior)

28

2103499 2556 AM06

(Differential Equation)






Unicycle Robot

d L L L T V
Qi ( )
dt qi

qi

T V Qi
generalized coordinate ( q i )
generalized coordinate
qi { , , , }

Qi

d L
L
(
)
dt q i
qi

I1 T1

2
( I 2 Mr2 ) (mD mF )r2 ( xcg r2 )(( sin ) (cos )) T2

I a ( m D m F ) g [( xcg r2 ) cos r2 ] sin mW gr2 sin T1

q
r (x r )
Ib (mD mF )r2 (xcg r2 )( ( sin ) (cos )) (mD mF )
2 cg
2
sin (mD mF )g (xcg r2 )sin cos T2

Dynamic model Linearize

Linearize 4
I1 T1 -----------------------------------------------------------------------1
I 2 Mr22 mD mF r2 xcm r2 T2 -------------------2
I a mD mF gxcg mW gr2 T1 -------------------------3
I b mD mF r2 xcg r2 mD mF g xcg r2 T2 --4
equation of motion 4 matrix

mF =
mW =

M =
r2 =

X cm =

X cg =
I1 = Fly Wheel I 2 =

I a =

Ib =
:
T

m D m F gxcg mW gr2

0
Ia

0
m
m
gx

D
F
cg mW gr2

T
1

0
1
mD mF r2 xcm r2 0
I2 Mr2

T2

0
m
m
g
x
r

mD mF r2 xcg r2
Ib

D
F
cg
2 1

Ib
mD mF r2 xcm r2

A
2

I 2 Mr2
mD mF r2 xcg r2

1
1
1
1
I1 2 I 2 2 I a 2 M ( r2 ) 2
2
2
2
2

A 1

1 2
r ( x r ) cos
I b ( mD mF )
2
2
cm
2

1
A

Ib
mD mF r2 xcm r2

I 2 Mr2 2
mD mF r2 xcg r2

I c m D m F r2 xcm r2

:
V ( mD mF ) g ( xcg r2 ) cos r2 cos
mW gr2 cos V0

1 1

I
Ia

1

T1

I
a

L T V

1
1
1
1
1
r
L I1 2 I2 2 Ia2 M(r2 )2 Ib2 (mD mF )
2
2
2
2
2
2
(xcm r2 )cos (mD mF )g (xcg r2 )cos r2 cos mW gr2 cos V0

29

Ib Ic
Ic

2
g ( I 2 Mr2 )
2

r2
I c I 2 Mr2

A
A

2

Ic g
r2

2103499 2556 AM06

matrix dynamic model
state space description

0


0


0

1

0
0
0

0 0 0 0 0
0 0 0 0 0
0 0 0 0 0

Mgxcg
Ia
0

Mgxcg
Ia

0 0 0 0 0

0 0 0 0 0

1 0 0 0 0
0 1 0 0 0
0 0 1 0 0

0
0
0

1 1

I
I

1 a

I g
Ib Ic

c
0

Ar2
A

0
0

I
T
a
1

2
2

I
I
Mr
gIc I2 Mr2
c
2
2 T2

A
r2 A

0
0
0

0
0

0
0
0

0
0

Dimensions

2.2

2.1 I

Encoder

planetary 12 vdc

planetary 12 vdc

250
rpm
100
rpm

60

500 p/r

20

500 p/r

simulate MATLAB

3
CATIA
MATLAB

4 simulate MATLAB

5 Simulate

3 CATIA

CATIA (Model)
- (try error)

4
7
5 simulate

30

2103499 2556 AM06

2.3
Controller
State variable feedback controller
multiple input multiple output (MIMO)

MATLAB

state
state
estimator state
state 8
encoder gyro sensor
roll pitch unicycle
0
controller
dynamic model State variable
feedback controller dynamic model state space
description

2.1

MATLAB SIMULINK response

roll pitch 5o initial
condition
X0[0 0 0 0 0.087 0.087]

6 MATLAB SIMULINK response

CATIA
unicycle dynamic model
state space description
0
0

0

0
1

0
0

7 Response Roll initial condition

0 0 0 0 0 27.8 0 113.1 0
0 0 0 0 0 0 171.4 0 87.5

0 0 0 0 0 27.8
0 2.0 0

0 0 0 0 0 0
74.9 0 25.0T1

0 0 0 0 0 0
0 0
0 T2

1 0 0 0 0 0
0 0
0
0 1 0 0 0 0
0 0
0

0 0 1 0 0 0
0 0
0

State variable feedback controller

closed loop poles
response State variable
feedback controller control law u Kx
K MATLAB
State variable feedback controller
unicycle
0.01 0 9 0 0 0 48 0
K

0 0.01 0 0.6 0 0 0 5.2

31

8 Response Pitch initial condition

Response Roll Pitch

unicycle

2103499 2556 AM06

9
3
3.1
2
Copley Compact RIO

2
Amplifier 220 Volt AC 24 Volt DC
Compact RIO Copley 2
3.2

2 1.

Compact RIO 2. Micro
Strain Gyro meter
Accelerometer 3
PC USB
9
PC

4.
2

1. (longitudinal) 2.

4.1 (longitudinal)


10 ()
()


11 ()
()



( )

10 longitudinal

32

2103499 2556 AM06

5.

11 longitudinal

4.2

12 () ()

Error
13 ()
()


Error

12 longitudinal

13 lateral

33

..

[1] . 2548. .
2000 . 2. :
.
[2] . 2554 LABVIEW
. 1.
:
()

2103499 2556 AM07

The Study of Shafts Critical Speed of Critical Speed Investigation Apparatus

5330156921, 5330177021 5330185021
.. ( )

(Displacement sensor)

Abstract
This article introduces the improvement of critical speed
investigation apparatus which diameter of shaft and mass can
be modified, also the mechanical experiments demonstrating
critical speed of shaft by using critical speed investigation
apparatus are performed to study the correlation of parameters
that influence the critical speed of shaft. The displacement
sensor is used to measure the amplitude of vibration in shaft
and evaluated the critical speed of the apparatus. In theory,
the amplitude of vibration is maximized when the critical speed
is equal to the speed of motors. From obtained results, it can
be concluded that the minimum critical speed of apparatus is
occurred when the position of mass is in the middle of two
bearings. The parametric study yields that the shafts critical
speed is directly proportional to the diameter of shaft and the
number of mass with the constant total weight of apparatus. In
contrast, the speed is inversely proportional to weight of mass
and distance between two bearings.

(Gas Turbine)

(Critical Speed)

2.

(First critical speed)
Stiffness

1.

1 1

34

2103499 2556 AM07

1 1
1 [1]

30

(1)

(rpm)
9.81 m/s2

(
)
(
) Static deflection (m)

1 2
Dunkerley [1]

2 1
Dunkerley Superposition

2
1

(2)

3
(3)
(Nm)
(Pa) (m4)

3
1 4 5

(4)

(5)
1 , ,

, ,

Roller

0
0
0

Pin

0
0
0

Fixed

0
0
0

x
y

2.1

Pin
Support, Fixed Support Roller Support
Elastic Curve 3

2.2 Stiffness support bearing

Stiffness
Stiffness
Critical Speed Map [1] 4

3 Elastic Curve

4 Lateral Critical Speed Map

35

2103499 2556 AM07

Critical Speed Map
Critical Speed Support Stiffness
Stiffness
Free Support
Stiffness
Fixed Support


Support Stiffness
3.
(Critical Speed Investigation
Apparatus) gunt TM620
(Critical Speed)

Fixed Support Pin Support

3.2.1
Fixed Support
2


1.

Fixed Support 6
Free Body Diagram 7

6
Fixed Support

5
3.1

1. 300 - 3000 rpm
2. 0.25 kW
3. (high-strength steel)
6 mm 8 mm 500 mm
4.
4 0.5 kg, 0.75 kg, 1 kg
1.25 kg
5. 50 mm
6. (
) 300 - 470 mm
3.2
1

Fixed
Support Clearance

7 Free Body Diagram

Fixed Support

,
0

3
4 5
Fixed Support

0
3

36

(6)

2103499 2556 AM07

(7)

8
Pin Support

2
Fixed Support

9 Free Body Diagram

Pin Support

192

2
3

Pin Support

2
3

(9)

2.
Distributed Load
Fixed Support
8

348

(8)

(N)

3.2.2
Pin Support
2


1.

Pin Support

Pin Support 8

37

(10)

3
Pin Support

48

93

2
93

2103499 2556 AM07

2.
Distributed Load
Pin Support
11

5
384

(11)

3.2.3

Critical Speed Map 4

Clearance
Fixed Support


Stiffness
Fixed Support Pin
Support

1.
2.
(Interpolation)
Pin Support Fixed Support 0 1

3.

4.


4.

10 2
4

1. (D)

6 8 mm

6 8 mm

2. (n)

1 2

1 ()

3. (x)

x1, x2, x3

x2 ()

4. (m)

0.5, 1.0 kg

0.5, 1.0 kg

5. (d)

45 cm ()

35 40 cm

4
28
400-3000 rpm

Displacement sensor
5.

11 (Amplitude)

11
= 45 cm
= 6 mm = 1
= 22.50 cm = 0.5 kg

1600 rpm

12, 13 14

10

38

2103499 2556 AM07

12
1

14
2

12

Static deflection

14

Static deflection
1 Static
deflection
1
6.

13
1
13


Static deflection
12 13

Static deflection

Static
deflection
5 14

1
2
3
4

11.25 cm 33.75 cm
0.5 kg
0.5 kg
0.5 kg
1.0 kg
1.0 kg
0.5 kg
1.0 kg
1.0 kg

39

..



[1] John, S., M. Introduction to Machinery Analysis and
Monitoring. Oklahoma: Pennwell Publishing Company,
1993.
[2] Hibbeler, R., C. Mechanics of Materials. Singapore:
Pearson Education South Asia Pte Ltd, 2005.

2103499 2556 AM08

Specific Damping Capacity

The determination of basic shear damping of materials

and the determination of Specific Damping Capacity of multilayer laminates by
using finite element analysis and verify by experiment
5330333621, 5330336521 5330372021
.. ( )
The objective of this project is to experimentally
determine the basic shear damping property of materials that
are components of aluminium laminate and laminated glass.
The basic shear damping property from this study will be
added up with extensional damping property conducted in the
last year project to determine specific damping capacity (SDC)
of multilayer laminates since the results of last year project had
a largely different between experimental method and finite
element method: last years experimentation is based on the
assumption that the shear damping property of viscoelastic
material is equal to its extensional damping property. In this
project, experimental method has been developed in order to
test free torsional vibration. Specimen support and exciting
device also have been designed for this experiment so that the
shear damping property of materials could be determined in a
more appropriate way. After experimentation, the results show
that shear damping propertyof viscoelastic materials, which are
polyethylene and glue, are approximately the same with their
extensional damping property. Therefore, the assumption in the
last year project can be used for determining SDC of multilayer
laminates.
Keywords: Specific damping capacity, multilayer laminates,
1.
(Vibration)

2

(Forced vibration)

(Free vibration)
Underdamped motion


(Damping)

(Logarithmic decrement),

(Damping coefficient), (Viscosity)

(Specific damping capacity)

(1)
3 ,
(2) PVB
Investigating on
damping property of laminated composite materials
Extensional damping

Viscoelastic material Extensional Damping
Shear Damping

26%-33%

Shear
damping
Extensional
Shear damping

40

2103499 2556 AM08

2555

(4)
(5)

= 0.0002 + 0.0038 + 0.003

= 0.0012 + 0.0213 + 0.1598
= 0.0024 + 0.0258 + 0.3183
= 0.0002 0.0033 + 0.0671

SDC

2.
2.1 (Free torsional
vibration)

1 Rod
Rod Ended mass
Rod

Ended mass
1

(1)

(2)

2
(

)+

(3)
(Fixed
support)
Boundary condition

2.4 (Specific Damping Capacity)

( ) SDC

(Strain) (Stress) SDC
(8)
=
(8)
(8)
(Dissipated energy Wd)
(Stored energy Ws)

SDC
(9)

(0, ) = 0

= Mass moment of inertia of ended mass

=
= Polar moment of inertia of the rod
2.2 (Damping) (Damping
capacity)
(Damping)

(Damping capacity)

2.3 (Logarithmic decrement)

( )
(Damping Ratio)
Ai
, Ai+n
n

=
(
)
(7)

Extensional Damping

1 Extensional Damping

( , )=

(6)

41

(9)

2103499 2556 AM08

2 Wd Ws
3.

ANSYS
ANSYS
( 5) (Support)
Mesh
( 1)

(Strain Energy Ws)
Ws

+. . . +

(10)

, =
,
SDC FEM

=2
(11)

(Aluminum sheet)

(Polyethylene PE)
2
(Glue)
4.2


(1)
(2)

2
(3105 H14), (MODIC M533),
(HDPE-N) 1 2


30 . 0.5 .,
30 . 3 ., 30 .
3 ., 30 . 4.2 .,
10 . 3 ., PVB 15 .
2 . 25 . 6.2 .


2 3

3 Ansys
Mesh

PE

PVB

(mm)
0.5
0.1
3
3
0.2

Mesh
(mmxmm)
0.5x0.125
0.5x0.05
0.5x0.5
0.5x0.5
0.5x0.05

1 Mesh Ansys
4.
4.1
2

(Layer)

,

PVB

42

(.)

()

1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
1

130
130
130
70
100
105
105
100
115
110
120
100
215
215
215
100

109.59
68.55
45.33
45.33
344.36
276.94
209.17
135.77
344.36
276.94
209.17
135.77
1260
1000
630
68.55

(Hz)
4.13
5.43
6.49
11.36
6.67
7.58
8.26
10.42
7.09
8.93
10.4
12.19
7.2
8.1
11.8
0.463

2103499 2556 AM08

4.3.4

(Laser displacement) LK-Navigator

2
(1) (2)

LK-Navigator

2 Free torsional vibration

(.)
()
(Hz)
1
165
1260
6.99
2
145
1260
8.1

3
135
1260
9.09

4
115
1260
10.99
5
125
630
12.5
1
230
1620
8.2

2
230
1260
9.09

3
230
1260
11.36
3 Free bending vibration

4.3
4
(1) , (2) , (3)
(4)
4.3.1
(Jig)

4.3.2
(Vacuum box),

(Vacuum pump) (Pressure

gauge)




25 kPa

4.3.3
(Function
generator) GW-INSTEK GFG-8216A,
(Bipolar operational power supply)
KEPCO BOP-36-6M
(Magnetic coil)

2


(Torsional vibration)

(Bending Vibration)

43

4.4





25 kPa

LK-Navigator

2-3

4
Function
Generator

Amplifier

LK-Navigator

Laser
Displacement

Magnetic Coil

Vibration of
Specimen

2103499 2556 AM08

4.5

4.4.1

(a)

(b)

5.
1.

1 Basic Shear Damping

1 Basic Shear
Damping
= 0.0099

(c)



4
4

(in-lb)

60

65

65


80 in-lb
4.4.2

(a)

+ 0.1922 + 0.5855

= 0.1598

+ 3.04 + 7.8118

= 0.2266

4.8688 + 58.168

2 Specific Damping Capacity

2.

(b)

(c)
(d)

3 Basic Shear Damping

3 Shear Damping

= 0.0523

+ 1.0872

PVB Damping

Basic Shear
Damping 0.93

44

2103499 2556 AM08

(Ws (J))
Aluminium

4 Specific Damping Capacity

6.

5 Extensional Damping Shear

Damping 165x30mm
Endmass 1260g 8.13 Hz
5 Extensional Damping Shear
Damping
Viscoelastic
Damping 30
% Shear Damping

2 SDC
( 1) Extensional Damping
Shear Damping ( 2) Extensional
Damping Shear damping

1 Extension damping Shear
Damping
SDC SDC
Extensional Damping Shear
Damping

45

Glue

Polyethylene

Normal

Shear

Normal

Shear

Normal

Shear

1148.527

0.058

0.269

24.677

9.800

19.124

1603.423

0.102

0.376

29.361

13.684

32.794

1922.213

0.139

0.450

32.294

16.405

43.989

1
4
SDC

PVB Viscoelastic

Shear Damping Extensional Damping
PVB
7.
Specific Damping Capacity

SDC

SDC











[1]
KritsadapornThongsawang,
NattaponKaewchalam,
NutthasanSrikwanma
and
WorapolPhraechinda,
Investigating on damping property of laminated composite
materials, Department of Mechanical Engineering,
Chulalongkorn University, 2012
[2] Mechanical Vibrations 5th, Singiresu S. Rao, Engineering
vibration 3rd, Daniel J. Inman
[3] Cornelis A. Van Eysden and John E. Sade, Resonant
frequencies of a rectangular cantilever beam immersed in
a fluid, Department of Mathematics and Statistics,
University of Melbourne, Victoria 3010, Australia, 2006
[4] Theory of Elastic ,Timoshenko

2103499 2557 AM09

Surface Finishing of Femoral Head of Prosthetic Hip to meet ISO Standard

Mr. Jirayoot Rajinda 5330072821, Mr. Samuel Chanhom 5330118021, Mr. Thanaphong Keawvichit 5330188021
Lect. Chanyaphan Virulsri (Advisor) & Lect. Pairat Tangpornprasert (Secondary Advisor)

In order to develop such a product, a number of qualities of

the prosthetic hip must be possessed such as sphericity, stem strength
material characteristics, etc. This project covers the issue on Surface
Roughness on the Femoral head (Figure 1). The prosthetic femoral
head needs to be smooth enough to not cause scratches to the
patients acetabulum cup.

ISO 7206-2
50 2


Drag
Finishing

Femoral Head

Acetabulum Cup

Abstract
To manufacture economically viable prosthetic hips in
Thailand, several requirements are needed. This project aims to
accomplish the requirements on Surface Roughness of the Femoral
Head since excessive roughness may damage the patients
acetabulum cup.
As stated from the International
Organization of Standardization (paper number ISO 7206-2) the
surface roughness of the femoral head at certain points (which will be
stated at article 2) must not exceed 50 nanometers. A method called
Drag Finishing is used in this project. It is used for exterminating
patterns caused by CNC milling machines after the machining process
which is called Lays. Therefore, an overall view of the project is to
present the best strategy in obtaining the lowest surface roughness for
the femoral head.
Keywords: Femoral Head, Surface Roughness, Dry Polishing,
Wet Grinding

Figure 1 showing parts of the prosthetic hip

2. Objective
To obtain average surface roughness (Ra) of 50 nanometers of
the femoral head at the pole position and the C-axis (measured at an
angle of 30 degrees below the pole) as described at Figure 2
Pole
C -Axis
30

Femoral Head Center

1. Introduction
Due to the lack of capability for manufacturing prosthetic
parts in Thailand, almost every hospital must import these parts from
international manufacturers thus increasing the cost to obtain the part.
Prosthetic hip surgery is an expensive operation but by manufacturing
the part domestically can significantly reduce the cost of operation and
provide better opportunities for patients with low income to receive the
treatment.

Figure 2 Position for measuring Ra

3. Experiment Design

46

2103499 2557 AM09

3.1 Material for experiment
For budget control two types of material will be selected for experiment
which is Steel S45C and Stainless Steel 304.
3.2 Method
The drag finishing process will be used since the finishing will deliver
constant abrasion and will not allow the workpieces to come in contact
which each other. (Figure 3) The drag finishing method uses a primary
drive unit to rotate the station and the holders have another motor to
rotate the workpiece during abrasion with the media.
3.3 Sequence for finishing & Abrasive Medias
Research using sequence of abrasive particle sizes [1] show that large
particles can efficiently decrease lays and then by using smaller
abrasive particles can decrease roughness. Similarly, two sequences
with different abrasive medias will conduct for the experiment.
3.3.1) Wet Grinding
Plastic grinding chips (Figure 4) will be used for this process.
During the process water will flow pass the workpieces and
down the drain to filter scrap from abrasion.
3.3.2) Dry Polishing
After wet grinding, the workpieces are to be polishing and
further eliminate surface roughness. Walnut Granule is used as
the media for this process. (Figure 5)
3.4 Measurement
In order to obtain surface roughness from the same position for each
and every trial of abrasion, laser marking would be used to mark
positions as indicated from article 2. (Figure 6) The value of Ra could
be obtained by using an Optical Inferometric Profiler.

From using different materials but same conditions in abrasions such

as speed and direction of the drag finishing machine we can consider
two scenarios
4.1 Result by using S45C as workpiece material
By plotting the result from wet grinding and dry polishing as shown in
figure 7 and 8 respectively, we can conclude that the process of wet
grinding for 75 minutes yields low Ra and also the Ra of each point on
the femoral head is less deviated. However porosity soon appears
continually at the dry polishing process (Figure 9) which limits the drag
finishing process to reduce Ra on the S45C material at 60 minutes of
dry polishing.

Figure 8 Result: Dry Polishing

Figure 7 Result: Wet Grinding
(S45C)
(S45C)
4.2 Result by using stainless steel 304
as workpiece material
Results of wet grinding and dry polishing
are shown in Figure 10 and Figure 11
respectively. Giving low Ra at 60
minutes of wet grinding and also
succeeding the objective of Ra lower Figure 9 Porosity of S45C at
than 50 nm from 60 of dry polishing at
microscopic scale
the specified positions.

Figure 10 Wet grinding result (SS304) Figure 11 Dry polishing result (SS304)
Figure 3 Drag Finishing

Figure 5 Walnut Granule

Figure 4 Plastic Grinding

Acknowledgements
The authors would like to thank their advisors, Dr.Chanyaphan Virulsri
and Dr.Pairat Tangpornprasert for advisement and support. The
authors would also like to thank the BDML and Mechanical
Engineering workshop staffs for encouragement and support.
Reference
[1]Pairat Tangpornprasert. Surface Roughness Improvement by a
Sequence of Abrasive Particle Size. Mechanical Engineering,
Chulalongkorn University, 2007

Figure 6 Laser Marking

4. Result
47

2103499 2556 AM10

The Design of Femoral component and Guide component for Knee Prosthesis
Tanachat Tanakornpipattanakul 5330186721 Matthew Pichitphan 5330367021 Rinrada jiravanstit 5330377221
Dr. Pairat Tangpornprasert (Advisor)
Preface
Osteoarthritis, also known as Arthrosis, is a chronic
disease that is frequently found in elderly people. Because of
the heavy usage of knee joints, along with the slow rate of
regeneration in elderly people, the only way to permanently get
rid of the disease is to have an operation, to substitute the joint
with the artificial one, which cost is slightly high since the
artificial one need to be imported from the other countries. This
project is created to serve the purpose of reducing cost of
importing the artificial joints and to further develop the artificial
joint within the country to save costs in treating the disease in
the future. The artificial joint that the project produce will be
Fix-bearing type which will be seperated into 4 parts, which are
Femoral Condyle, Guide surface, Stopper, and Cam. The
design will base on the theory of the differences in AnterorPosterior displacement on both side of the artificial joint in
order to make an Internal-External rotation. The angle of
rotation can be calculated by A-P displacement and take it to
design Guide surface in order to make it support the
moveability. After the Femoral Condyle and Guide surface was
produced, the CAD (CATIA), the 3D design program is in
action to calculate the size and the position of Stopper and
Cam. The product will be resulted in the artificial joint which
can be moved as one wish to. When we test the result of A-P
displacement, we found that the deviation from the ideal result
by 19.4 Percentage.
Keywords: Flexion Extention , Anterior - Posterior
displeacement, Internal - External rotation, Femoral
component, Guide component, Condyle, Cam, Stopper
1. Introduction
Nowadays, elderly people tends to have chronic diseases
more and more. The researchers found that among the elderly
people around 7.34 percent are likely to have Osteoarthritis,
which is a big problem in their living. Compare to the
population of people over the age of 65 years old, the research
found that over 463,939 are the elders with Osteoarthritis.
Because of the heavy usage of knee joints, along with the slow
rate of regeneration in elderly people, the only way to
permanently get rid of the disease is to have an operation, to

substitute the joint with the artificial one, which cost is slightly
high since the artificial one need to be imported from other
countries. The artificial joint is not available in Thailand yet. In
designing and developing the artificial joint as similar to the
movement of human joint as possible in Thailand would help
reduce its cost. The artificial joint are designed to have a
movement and flexibility with no muscle and tendon are
involved in the test.
2. Anatomy and Motion of knee

Figure 1 The components of knee

Components of the knee

1. Femur : the bone in the top part of leg, above your knee
2. Tibia:the bone at the front of your lower leg, between
knee and ankle.
3. Fibula : the outer narrower bone in the bottom part of leg
4. Patella : the bone at the front of knee main function of
Patella is improve moment arm of Quadriceps muscle.
5. Lateral Collateral Ligament (LCL) and Medial Collateral
Ligament (MCL)
6. Anterior Cruciate Ligament (ACL) and Posterior Cruciate
Ligament (PCL) : ACL is one of four major ligaments that
provide stability to the knee joint. PCL is one of four major
ligaments that connects the Posterior Intercondylar
area of the Tibia to the Medial Condyle of the Femur.

48

2103499 2552 AM10

Walking cycle in knee
Knee movement, in natural, can be explained into three
patterns as follows:
Flexion Extension: leg bending and stretching
Internal-External Rotation: knee joint twist
Anterior Posterior Displacement: front and back
movement of bone

Figure 4 contact position of Femur and Tibia in each

degree of flexion
Referring to the naturally knee movement, we can
conclude design specification as follows:
Figure 2: Three patterns of knee joint movement: Flexion-Extension,
Internal-External Rotation and Anterior-Posterior Displacement

Human walking cycle comprises with two phases of

process: Stance Phase (foot is touching on the ground) and
Swing Phase (leg and foot is moving).

Flexion-Extension
Internal-External Rotation
Anterior Posterior
Displacement

0-90 degree flexion

0-12 degree rotation
24-22 mm on Medial side
26-20 mm on Lateral side

3. Knee Prosthesis
Total Knee Arthroplasty is a surgery to cure severe knee
Arthrosis by removing degenerative knee Meniscus and
replacing with Knee Prosthesis.
Knee Prosthesis can be divided into two parts which are
Femoral and Tibial component.
Femoral component is a Femur-shape cover made from
chromium while Tibial components are artificial Meniscus and a
iron plate on Tibia.

Figure 3 walking step with flexion angle.

Figure 3, above, shows the walking step and flexion
degrees of knee joint. Each degree of flexion affects the
contact point between Femoral component and Meniscus, as
shown in Figure 4,

Figure 5 Component of a Prosthetic Knee

49

2103499 2556 AM10

The artificial Meniscus has Concave shape similarly with
knee joint surface. High quality of Polymer (hard texture, high
purity) is used to be the Meniscus materials. The iron plate
links between artificial Meniscus and Tibia. Metal core, on the
Tibia side, is implanted on Tibia.
4. Knee Prosthesis Design
While natural knee is bending, Internal-External Rotation
and Anterior-Posterior Displacement are implemented. Concept
of Knee Prosthesis design conducts by scale differences of
both Lateral and Media side of Femoral Component and of
Guide surface curve.
Femoral component is above Guide surface with the help
of Stopper to prevent split between Tibia and Femur.
Design process to create Internal-External rotation,
Anterior-Posterior displacement can be used to explain by
difference of two sides Anterior-Posterior displacement on
Femoral component. Fix-Bearing design pattern is implemented
in this study together with Anterior-Posterior displacement
values, in the table 1.

Developing from the equation, we will get radius value in

each section and can draw Lateral and Media of Femoral
Condyle in 3-D in Figure 6.

Figure 6 Femoral Condyle

As Femoral Condyle of Medial has very small size, it is

substituted by Lateral, bigger size, as shown in Figure 7.

Table 1 Anterior-Posterior displacement values on Medial and

Lateral side

Flexion Medial Lateral

20
26
0
19.8
20
10
19.6
18
20
19.4
16
30
19.2
15
40
19
14
50
18.8
13
60
18.6
12
70
18.4
12
80
18.2
11
90
18
10
100

Figure 7 Revised Femoral Condyle

4.2 Center of Cam and Stopper Positioning

Center of Cam and Stopper are motion controller of Knee
Prosthesis design. Their design correlate with rolling and
slipping concept.

Cam

4.1 Design of Femoral Component

Design concept of Femoral component is rolling without
slip. Pivot point is always on the ground. Equation 1 related to
the concept to find Femoral components curve radius while
rolling is on the flat is shown below.
(1)
that r
AP

: Radius of Femoral component each section

: Distance of contact position in each section
: Degree flexion in each section

Stopper

Figure 8 Stopper and Cam

Due to the fact that movement cannot be forced only by

curve of Guide surface, Stopper and Cam are used to support
position specification of Anterior-Posterior displacement of
Medial side of Femoral component as shown in the pictures
below.

50

2103499 2552 AM10

Center of Cam

Selecting new AP value which is a crossing point, red dot

in Picture below.
Stopper

Repeating step 1-4 for every 0-90 degree of flexion, we will

get the result as follows.

4.3 Surface Design of Guide surface

Guide surface perform Meniscus in knee to support three
patterns movement. The principles for the design Guide
surface is as follows
Draw a line connecting two points of contact for each
degree of flexion.
4.4 Appropriate Size of Cam and Position of Stopper
Calculate the radius of the Cam that the knee moves
according to value of AP with the following steps.

Drawing vertical line to define the Femoral Condyle and

Guide Component motion, as shown in red line in Picture

Drawing 3-D of Femoral Condyle and Guide surface

in CAD program (CATIA).

Moving position of AP along the red horizontal line. AP

value is fixed along the line.
Figure 9 Femoral Condyle

51

2103499 2556 AM10

Figure 10 Guide surface

Turning Femoral Condyle in every degree of flexion

and doing Internal-External rotation (twisting to
preferred degree)
Moving lowest point of Femoral Component to touch
on AP position on Meniscus at each flexion degree
Measuring distance from Stopper to center of Cam
perpendicularly to find the Cam radius

5. Testing
After design Femoral Condyle, Guide component, Cam and
Stopper next step, the knee is designed to measure Anterior Posterior Displacement and Internal - External Rotation.
Determination Anterior - Posterior Displacement each flexion
angle can be obtained as follows.
Draw 3-D of Femoral Condyle , Guide component ,
Cam and Stopper in program CAD (CATIA)

Figure 12 Smooth Guide components

Figure 11 find radius of Cam

Rotate Femoral Condyle by degrees flexion

Rotate Femoral Condyle by Internal - External

degrees

Repeating all steps from the beginning for 0 to 90

flexion degree.

After computing and measuring the Cam radius in each

degree, to create smoothly possible movement, trend is
created to be radius of new Cam. Values, resulting from the remeasurement are shown in table 2.
Table 2 Cam values on Medial and Lateral side

Flexion
0
10
20
30
40
50
60
70
80
90

R Medial
Cam

R Lateral
Cam

2.14
0.904
0.701
0.552
1.842
1.572
4.006
3.017
3.551
3.853

1.984
3.963
3.904
2.351
3.66
3.383
2.153
5.041
5.587
5.888

R new
Medial
Cam
1.3251
1.2411
1.2571
1.3731
1.5891
1.9051
2.3211
2.8371
3.4531
4.1691

R new
Lateral
Cam
3.1056
2.9416
2.8976
2.9736
3.1696
3.4856
3.9216
4.4776
5.1536
5.9496

52

2103499 2552 AM10

Cam on Femoral Condyle is exposed to the Stopper
Guide surface.

6.Conclusion
The Knee Prosthesis is designed to set 3 type movements
format. When Knee Prosthesis is moving in Flexion
Extension direction it affect movement in Internal External
Rotation and Anterior Posterior direction. From the
Simulation result, which in a muscular and tendon system in
not affect to the Knee Prosthesis movement condition, shown
that in 0 90 degree of Knee Prosthesis Flexion Extension
moving will affect to Internal External rotation range of 7 to -9
degree while In natural knee the range of moving is 3 to -1
degree while another movement Anterior Posterior direction
in Medial site will affect in the range of 8.7431 mm and Lateral
site in the range of 12.426 mm while natural knee in Medial
site is 2 mm and Lateral site is 16 mm. The Deviation
percentage of Anterior Posterior displacement in medial site
is 20.3 and Lateral site is 17.544.

Anterior-posterior displacement measured at the point of

contact between Femoral Condyle and Guide surface.

Repeating all steps from the beginning for 0 to 90 flexion

degree.
AP displacement values and error values are given in table 3.
Table 3 AP displacement values and error values on Medial
and Lateral side
Flexion

0
10
20
30
40
50
60
70
80
90

AP
Medial
(mm)
20
20
19
19
19
19
19
19
19
19

AP
Medial
(mm)
22.886
17.042
18.335
14.261
14.388
14.52
14.514
14.18
14.657
14.129

Error
(%)
14.43
14.79
3.5
24.942
24.274
23.579
23.611
25.368
22.858
25.637

AP
Lateral
(mm)
26
20
18
16
15
14
13
12
12
11

AP
Lateral
(mm)
25.693
19.902
15.383
17.126
17.068
16.888
16.481
15.654
14.731
13.267

Error
(%)
1.181
0.49
3.5
24.942
24.274
23.579
23.611
25.368
22.858
25.637

53

Acknowledgment
It would not have been successful to do this senior project
without the support from the kind people around me.
Above all, I would like to thank Dr. Pairat Tangpornprasert and
Dr. Chanyaphan Virulsri for their support and patience. Their
advices has been very worth to not only do my project but
also, sometimes, can be used in real life.
Likewise, without graduate students in lab BDML, there is
no way to complete this senior project smoothly.
Reference
[1] National Statistical Office
[2] P. Johal ; A. Williams ; P. Wragg ; D. Hunt ;W.Gedroyc .
2005. Tibio-Femoral movement in the living knee. A study of
weight bearing and non-weight bearing knee kinematics using
interventional MRI. Journal of Biomechanics 38 : 269276.
[3] Mow, Van C. ; Huiskes , Rik . Basic Orthopaedic
Biomechanics and Mechano-Biology . 3rd edition. USA :
Lippincott Williams & wilkins , 2005

21034992 2556 AM11

Design of the Buckling Tester Machine

5330041321 5330052221 5330137021
.. ( )

3





6-25%

1.

1
2 3
Mechanics of materials

Abstract
Buckling is a problem frequently occurs in engineering work.
Experimental analysis is one way to determine buckling which
demonstrates realistic bending behavior of the column and can
also lead to better understanding of the buckling behavior.
However, the existing tester used in the analysis cost a lot of
money. To reduce costs buying tester. Hence, this project has
been done to design and construct tester. The work of this
project can be divided into 3 step: studying the designation and
operation of existing tester, designing and production, evaluation
of the tester. It is found that the finished tester has a designed
strength and works as planned but the results got errors around
6 - 25 percent because of the imperfection of specimens.

54

21034992 2556 AM11

, ( ,

),
2.
Power screw, Pneumatic, Hydraulic,


Power screw, Pneumatic, Hydraulic
Pneumatic Hydraulic



Power screw


3.

4.
Load cell, Hydraulic load
sensor, Elastic ring deformation dial gauge

Load cell

Hydraulic load sensor Elastic ring deformation dial gauge


Load cell Strain
gauge

LabVIEW
5.

6.
Dial gauge,
, Laser distance meter Digital seismometer

Dial gauge


(parallax) Laser distance
meter dial gage Digital
Seismometer

2.

3
1.
2.

3.
3.
3.1

, ,
(Design specification)

, ,
, ,




3.2

- 2,000 N
- 20 N
- ( 2 mm,
10 mm, 200 300 mm)
- Fixed-Fixed supports, PinnedPinned supports, Fixed-Pinned supports
-
-
- 600 mm 300 mm
- 15 kg
- 50 N
- 40 m



1.

55

21034992 2556 AM11

7.

1 Linear shaft, manual slide & lock mechanism:
, , Slip,
Misalignment
2 Linear shaft,linearbearing,self-locking screw mechanism:
, Misalignment
1 1
3 Linear motion guide, manual slide & lock mechanism:
, Slip, Misalignment
2 2
4 Linear motion guide, self-locking screw mechanism:
, Misalignment
2 2
4
8.
Dial gauge

4

9.
3 Fixed-Fixed, PinnedPinned, Fixed-Pinned Fixed Builtin Pinned Knife edge

-

fixed-fixed

()
()
2
() Knife edge
()
4.
4.1

3

1

()
()
1
() Fixed
() Pinned
10.



2 Knife edge 2
2
- Knife edge ,

pinedpined

1.)

(2000N)

2.)
3.)

4.) Fixed-Fixed,
Pinned-Pinned, Fixed-Pinned
5.)

6.)

7.) 600.0 mm

(20N)

8.) 300.0 mm

(220mm)

9.) 15 kg
10.)
50 N
11.)
40 m

(2x10mm)
(*)
(0-260mm)

(450mm)

(5N)
(10m)

* Fixed support
Pinned support
Clearance

56

21034992 2556 AM11

gauge
Deflection Save
Plot
User interface 5

5 User interface LabVIEW

5.
5.1

Euler s Formula
5.1.1 20 cm Fixed-Fixed
6
1316
N 1030 N
25 %

3
4.2

precision 10 N

1. Load cell DAQ

2. Load cell

3.
calibration Load cell
4.

6 20 cm
Fixed-Fixed
5.1.2 20 cm Fixed-Pinned
7
671 N
603 N
10 %

4 Block diagram


Strain

57

21034992 2556 AM11

[ ]

(5.1)

(Initial bend)
(Critical load)

initial bend 9

7 20 cm
Fixed-Pinned
5.1.3 20 cm Pinned-Pinned
8
328 N
309 N
6 %
9
20 cm Pinned-Pinned

9

6.



,
Support Support

8 20 cm
Pinned-Pinned
5.2
(Critical load)

(initial bend),
Support Support
20 cm
Pinned-Pinned initial bend
Dial gauge 90 m

[4]

..

58

21034992 2556 AM11

[1] .. .Mechanics of Materials.
: -, 2548.
[2] . ,.. .
. 1. 10. :
, 2553.
[3] Russell C. Hibbeler. Mechanics of Materials (7 edition).
Singapore: Prentice Hall, 2008.
[4] Frederick A.Leckie, Dominic J. Dal Bello. Strength and
Stiffness of Engineering Systems. USA: Springer Science
Business Media LLC,2010

59

2103499 2555 AM12

Design a Tool for Helical Gear Parameter Measurement
5330132821, 5330359021
5330448121
.. ( )

100


overball, span Helix


Matlab

1 1

2
, ,
Abstract
This project aims to design and construct tools to
measure main parameters of helical gear of at most 100mm
diameter, included Module, Pressure Angle and Helix Angle of
a Helical Gear. Tools will be able to measure both left and
right handed Due to difficulties of direct measurement of each
parameter, we used indirect measurement of three parameters
.Overball distance, Span distance and helix angle at outside
diameter. These measured Parameters relate to main
parameters and will be calculated by Matlab. After measuring
with standard helical gear with designed tools and methods,
helix angle accuracy is 1, module accuracy is 1 mm but
pressure angle accuracy is 2.
Keywords: Module, Pressure angle, Helix angle

2.

(m), () ()


1.Overball 2.Span 3.
Helix
2.1 Overball Measurement
1()
M M

tan n

zmncos tan1
cos

Meven =
+ dp
cos cos

(1)

tan n

zmncos tan1

cos
90

Modd =
cos
+ dp
cos cos
z

...(2)

M - Overball Measurement
, dp , t
, mn - , z ,
, n
()
tan -

dp
tann -1 tann
- +
tan
(3)
180 zmncosn 2z cos 180
cos
=

2.2 Span Measurement

1()
W W

1.

1 Overball () span() [1]

60

2103499 2552 AM12

tann -1 tann
W = mncosn k - 0.5 + z
- tan
...(4)
cos
cos 180

W Span measurement, k
2.3 Helix Measurement

2
r

0.602 0.133
1.2265, 0.1456 1.8268
0.0014, 0.0302 0.0106

1

2 Helix Measurement


r d
(5)
= tan-1

h 180 do
d , , r
do , h

4 4

3.
2 3
Overball Span

mt

mt

30.16

K W

20

22

100.8

21.23

22.05

3.00

20

33.5

101.4

34.78

20.15

32.90

2.97

14.5

33.5

102.6

27.17

16.33

33.63

2.99

4.2
W,
M 3

3.5
14.5 30.17

3 Span & Overball Measurement() Helix

Measurement()

4.
2 1.
2.

4.1
3
W, M
Matlab
1
1

0.0518,

61

5.



Overball 2

6.

1 1
1

[1] KOHARA GEAR INDUSTRY CO.,LTD .Practical Information
on Gear.

2103499 2556 AM13

Development of a Paper-Based Pressure Sensor
5330235821 5330244421
.. ( )

pattern
water mask
Abstract
Nowadays, electronic devices have become one of the
significant human life necessities. Meanwhile, there are also
some disused electronic devices that have become electronic
waste which leading to some environmental issues. In these
past four to five years, there have been some developments of
devices using papers as the base of these devices. This
research project aims to develop pressure sensors by using
papers as the main material. From the experimental results,
the best way to fabricate the piezoresistive circuit is droping
carbon solution of a pattern created by water mask process.
1.
4-5




Piezoresistive Material

2.
2
PDMS 5cm*5cm
2cm*2cm 1b

1a Piezoresistive Material

(Deflection)

1 : (a) (b)

3.
3.1 Piezoresistor
Piezoresistor Piezoresistive
Material
Stress Colloidal
Graphite paste 2 Piezoresistive Material
Particle 3 m

2 : Colloidal Graphite Paste

3.2
PDMS
(Polydimethylsiloxane)

polymeric organosilicon

62

2103499 2556 AM13

PDMS Hydrophobic
PDMS

PDMS
1a
3.3

Graphite
Whatman Grade1 3
Cellulose
11 m Graphite

3 m 1

a
b
5 : Colloidal Graphite
(a) 1 (b) 5
2. Screen Printing
Screen Printing
screen block 6a
Colloidal Graphite Paste

Graphite Paste
6b
8 12
7
1

4 : Whatman Grade1
4.

(Mass Production)


4.1 Graphite

Graphite Colliodal
Graphite 1 : 200
Ethylene glycol 1 : 1




Colloidal Graphite

Graphite

63

a
b

6 : (a) Screen Block (b) Screen

Print

2103499 2556 AM13

1 :
*
(M)
(mm mm)
0.434
66
0.421
5 5
2.653
5 7
2.762
75
0.280
8 5
0.178
8 4
1.212
5 6
1.197
65

M

Carbon

Colloidal Graphite

4.3 Water Mask [1]

Water Mask

Polymer
Solvent Polymer Solvent
Solvent
Solvent
Polymer

Polymer Polystyrene Solvent Toluene
3% weight/volume
Pattern Water Mask
Pattern
2
4.3.1
Water Mask Graphite
Pattern Colloidal
Graphite
Pattern
Carbon

7 :

Water Mask (a) Whatman

Grade1 (b) (c)
Polymer Solvent (d)Polymer
(e) Solvent
Polymer
4.3.2 Colloidal Graphite Paste
Water Mask
Pattern


Colloidal

64

2103499 2556 AM13

Graphite Ethylene Glycol
Ethylene glycol
Pattern
: Ethylene Glycol 1:9
Colloidal Graphite

Pipet Chanel
Water Mask Hot
Plate
2

Colloidal Graphite Paste

Water Mask

9

2:

(C:Sol)
(l)

1:6

1:8

1:10

40
80
120

424.97
233.15
134.20

781.54
351.61
251.69

2298.01
657.30
320.97

9 : Graphite
Water Mask

5.
5.1



contact pad 8

5.4

Conductive carbon glue 10

10 : (a) Conductive Carbon Glue

(b) Conductive Carbon
Glue
8 :
5.2 Water Mask
Water Mask Pattern
4.3
5.3

65

5.5


Parafilm 11a
Hot Plate 60 C Parafilm
Parafilm 11b

2103499 2556 AM13

P ( N 2 )
m
a Plate
b Plate 12

b
(0,0)

11 : (a) Parafilm (b) Parafilm

Hot Plate 60 C

a
6.

Deflection
(Analytical Method) Deflection
(Experimental Method)

Deflection

6.1 (Experimental Method)

2cm 2cm
Deflection
5
7 3
3 : Deflection

(cm)
4
7
10
13
15

Deflection
(mm)
0.161
0.404
0.590
0.717
0.771

6.2 (Analytical Method)

Deflection

Plate Support
Fix Support 4 Deflection
Analytical method [2]
W

Pb 4
D

(1)

12: a b
D
D

Eh 3
12(1 )

(2)

Plate (m)

(Pa )

5
4
4 : Deflection

Deflection (mm)
(cm)
4
0.00102
7
0.00178
10
0.00254
13
0.00331
15
0.00382



PDMS

Deflection Deflection
Deflection
Deflection

5

66

2103499 2556 AM13

5 : Deflection

(cm)
4
7
10
13
15

Deflection (mm)
analytical
0.00102
0.00178
0.00254
0.00330
0.00382

experiment
0.08
0.13
0.16
0.18
0.19

Deflection PDMS



PDMS

Colliodal Graphite Paste Piezoresistive Material
Graphite Ethylene
Glycol 1:9 Water
Mask


Deflection



PDMS
Hydrophobic

Deflection

..

904

Shop

67

[1]
.

.

, 2555.
[2] C.E Imrak and I Gerdemeli. An Exact Solution for the
Deflection of a Clamped Rectangular Plateunder Uniform
Load. Faculty of Mechanical Engineering Istanbul Technical
University, 2007

2103499 2556 AM14

Change of Ship Hull to Reduce Resistance of Hull

5330140821, 5330432021
.. ( )

(Gross Tonnage) 30

(Length), (Beam) (Depth)
-(L-B)


-
(B-D)

Abstract
To reduce the total resistance of a ship will contribute
to the reduced fuel consumption, which is the main cost of the
sailing. So, to improve the ship in order to reduce the
resistance is an interesting thing in the design. Because of the
variety of Thai fishing vessels, there is a need to create the
prototype ship which is taken as an agent to change the hull
shape. The vessel has Gross Tonnage of 30 fixed and updated
shapes by changing three values which are length, width
(Beam) and depth. From the adjustment, we found that
adjusting the length-beam (L-B) affects resistance reduction the
most. When increasing the length, there is a linear function that
has change point through the first interval with the slope that is
greater than the end. Therefore, we are interested in the
change point range before adjusting the beam-depth (B-D).
The result shows that when being reduced the beam and
increased depth, the ship has less resistance and more
stability due to the prevention of water into the ship which
adjustment can be selected according to the requirements in

the design that can reduce resistance more than ten percent.
1.

2.

11

68

2103499 2552 AM04

Station

4.
-
(L-B)
()
20.884 (L=1.05)

2
Dimension Offset Table
DELFTship


Gross Tonnage 30
Gross Tonnage 30.24 Displaced
Volume 124.97 Length overall
19.889 , Maximum beam 4.904 Depth 4.36

4
-(B-D)

3
3.
3 Length(L),
Beam(B) Depth(D)

3 L-B, L-D B-D
Gross Tonnage
30 Displaced
Volume 124.97 m3 Gross Tonnage
30.235
DELFTship

Gross Tonnage 0.1025L 0.9757B
1.025, 1.05 1.225

69

5.
20..884
19.889
20.884
8 knot 3.856
15.185 11.329 25.4







[1] 2553
[2] J. Holtrop and G.G. J. Mennen,1982: An Approximate
Power Prediction Method, International Shipbuilding
Progress,Vol. 29, No. 335.

2103499 2556 AM15

Graphene synthesis by Chemical Vapor Deposition Method
5330219821 5330483021
.. ( )

(Graphene)
(Chemical Vapor Deposition)

Copper Foil
(CH4:H2 60:10) sccm
(Transfer) 2
PMMA

PMMA

2
(
)

Abstract
Graphene synthesis by chemical vapor deposition (CVD)
method can be made by gas mixture which flows in a quartz
tube with the most appropriate ratio between methane and
hydrogen. The best solution to synthetic graphene is growing
graphene on copper foil surface with flow gas mixture of
methane and hydrogen in a ratio of 60:10 sccm respectively.
Graphene can be transferred by 2 methods. The first method is
spin-coating with PMMA, the second method is Roll-to-roll
process. The best transfer method is spin-coating with PMMA.

CVD
(CVD)

1

SCCM
(Stand Cubic Meter Per Minute)
1000

2 2
20

Transistor, Solar cell, Electrochemical Sensor


2004

1 CV

70

2103499 2556 AM15

2
1.

(Solubility of Carbon)

[3]

2.

3
Nickel Film, Copper Film Copper Foil Ni/Cu
Silicondioxide (SiO2)
Ni/Cu Copper Foil
25-mm-thick Cu foil (Alfa Aesar, Item No.
13382)

Nickel Copper Film

Cleaned SiO2 Ni/Cu film
200-300 nm Sputtering Coater
NECTEC (National Electronics and Computer Technology
Center) Sputtering Coater
4

- Base pressure 7.6 x10-6 torr

- Plasma clean residue

- RF radio frequency Sputtering 100W
- Sputtering pressure 8.3 x10-3 torr

SiO2
SiO2 3
Diamond Tip-Head Pen 3
SiO2 1.5 x 1.5
3

(Substrate) 3 Nickel
Film, Copper Film Copper Foil CH4
H2 sccm (CH4:H2:40:10), (CH4:H2: 60:10),
(CH4:H2:80:10) 9 5

()

()

()

()

4 () Sputtering Coater ()
Ni ()
Ni Film
Ni/Cu 200300 nm

()

()

3 () SiO2
()
Diamond Tip-Head
SiO2 () SiO2 1.5 x 1.5

SiO2
SiO2
Acetone Sonication 20
IPA (Isopropyl Alcohol) DI Water
N2 Cleaned SiO2 Wafer

71

5 3
3 Copper Foil Glass Slide

Raman Spectroscopy



Raman Spectroscopic[1]
G Peak 1580
cm-1 2700 cm-1 6[2]

2103499 2556 AM15

Ni Film (CH4:H2:80:10)
3 Layers
Multilayer 9

6 Spectroscopic CVD

(Monolayer)
2D Peak
2 G Peak
2 G Peak 2D Peak
G Peak
2D Peak
9 Raman
Spectroscopy Raman

9 Raman Spectroscopy Ni Film (CH4:H2:80:10)

Copper Film (CH4:H2:40:10)
Bilayer 10

Nickel Film (CH4:H2:40:10)

Multilayer 7

10 Raman Spectroscopy Cu Film (CH4:H2:40:10)

Copper Film (CH4:H2:60:10)
Copper Film
G Peak 2D Peak 1580
2700 11
7 Raman Spectroscopy Ni Film (CH4:H2:40:10)
Nickel Film (CH4:H2:60:10)
Nickel Film
G Peak 2D Peak 1580
2700 8

11 Raman Spectroscopy Cu Film (CH4:H2:60:10)

8 Raman Spectroscopy Ni Film (CH4:H2:60:10)

72

2103499 2556 AM15

Copper Film (CH4:H2:80:10)
Bilayer 12

Light Microscope
(Uniformity)


Nickel Film, Copper Film, Copper Foil 16, 17,
18

12 Raman Spectroscopy Cu Film (CH4:H2:80:10)

Copper Foil (CH4:H2:40:10)
Bilayer 13

17 Nickel Film (CH4:H2:40:10),

(CH4:H2:60:10), (CH4:H2:80:10)
Objective Lens (X100)

13 Raman Spectroscopy Cu Film (CH4:H2:80:10)

16 Copper Film (CH4:H2:40:10),

(CH4:H2:60:10), (CH4:H2:80:10)
Objective Lens (X100)

Copper Foil (CH4:H2:60:10)

Monolayer 14

18 Copper Foil (CH4:H2:40:10),

(CH4:H2:60:10), (CH4:H2:80:10)
Objective Lens (X50)

14 Raman Spectroscopy Cu Foil (CH4:H2:60:10)

Copper Foil (CH4:H2:80:10)
Monolayer 15

15 Raman Spectroscopy Cu Foil (CH4:H2:80:10)

73

Raman Spectroscopy Light
Microscope

1
Copper Foil (CH4:H2:60:10)
Monolayer

(Transfer
Process)

2103499 2556 AM15

PDMS (Polydimethylsiloxan)
PDMS 5x5
300 500 micron
Silicone Elastomer Base Silicone Elastomer
Curing Agent 10: 1
19
2
19


70 C

PDMS 19

()

()

()

19 () Silicone Elastomer Base Curing Agent ()

5x5 300 micron ()
PDMS 5x5
300 micron

2
1. PMMA Film (Polymethyl
Methacrylate) Copper Foil PMMA
Spin Coating
PMMA
Copper Foil 2

(CH4:H2)

RIE

(Reactive Ion Etching) 2

Copper Foil
Ammonium Sulfate [(NH4)2SO4] Copper Foil
PMMA 20 PDMS
PMMA/Graphene
Deionized Water PMMA

74

2103499 2556 AM15

Acetone
PDMS 20

20 () Copper Foil ()
PDMS PMMA
2. PET
Film (Polyethelene Terephthalate), , PDMS, Copper
Foil 6
PET film (PETbottom /
Paper / PDMS / Copper foil / Paper / PETtop )
150 21
(NH4)2SO4
21 PDMS

[1] Yanwu Zhu , Shanthi Murali , Weiwei Cai , Xuesong Li , Ji
Won Suk , Jeffrey R. Potts and Rodney S. Ruoff *, Graphene
and Graphene Oxide: Synthesis, Properties,and Applications.
Adv. Mater. 2010, 22, 39063924
[2] Keun Soo Kim, Yue Zhao, Houk Jang, Sang Yoon Lee,
Jong Min Kim, Kwang S. Kim, Jong-Hyun Ahn, Philip Kim, JaeYoung Choi & Byung Hee Hong. Large-scale pattern growth of
graphene films for stretchable transparent electrodes.
doi:10.1038/nature07719, 2009
[3] YI ZHANG, LUYAO ZHANG, AND CHONGWU ZHOU,
Review of Chemical Vapor Deposition of
Graphene and Related Applications. ACCOUNTS OF
CHEMICAL RESEARCH, 23292339, Vol. 46, No. 10, 2013

..

Shop

22 () () Copper
Foil PDMS (NH4)2SO4
2
PMMA
Copper Foil
PMMA (NH4)2SO4
Copper Foil

Copper Foil
Copper Foil
Copper Foil
PDMS Copper
Foil

Copper Foil (CH4:H2:80:10)
sccm PMMA

75

2103499 2556 AU01

Design for a suitable type of small-scaled battery for the
Chulalongkorn Universitys Electric bus
5330180921, 5330205021 5330212321
.. ( )

7



3

3

( 3

)




20.84% 86.32%

()



13.70% 40.11%

addition, the Supercapacitor hybrid and Lithium are also present

their greatness in the available used energy (efficiency) of
13.70% and 40.11% more than Lead-Acid battery.
Keywords: Electric bus, Battery, Lead-Acid, Supercapacitor,
Lithium
1.

35
[1]
14.8% 5 [2]

33 65
[3]

4 1, 2 4

3

( 3)

3

3 3

3 1
3

Abstract
Chulalongkorn University electric buses have been
operating for about 7 years as shuttle buses for the students and
Universitys members. Due to the deficient of the bus service
during the peak hour, this project mainly finds out which battery
types are suitable to extend the bus operation time. The
experiment is setting up by using the daily buss load profile as a
load simulation on a test bench with the scale-down battery in
term of voltage. Three types of battery are used in the test which
are Lead-Acid the original battery used as a base in comparison,
Lead-Acid hybrid with Supercapacitor and Lithium. The result
shows that the Supercapacitor hybrid and the Lithium can extend
the operation time by 20.84% and 86.32% respectively. In

76

2103499 2556 AU01

3

1 3

8.00-10.00 10.00-12.00 12.00-14.00 14.00-16.00 16.00-18.00 18.00-20.00

1
2
3

Distance 3.0 km

9
1

1 3

(Supercapacitors)

[4]


[5]

1 3
9 2 (1)
(3)

3.0
2 3
( )

45 (20)

3.3 .
/ /
2.1 . / 6.75 . / 2.8 .

(24 , 307.2 V)

60 . / .
2.2 3
3
2





2

3


Hall effect sensor, Resistance
Circuit, National Instruments Data Acquisition (NI USB-6251)
Kayaba Drive recorder (DRE-400)
2.2.1

3
9 1 1) 2)
3) 4) 5)
() 6) () 7)

2.

3


Electronic load
(12.8 V) 3
(307.2 V)
3

2.1 3
3 4.0
45
(Energy for Environment Co.,Ltd)

3 1
2

77

2103499 2556 AU01

2.2.3 7

3
4250


7
3
4250
3 2

8) 9)
3 10
(3) 20

2.2.2


3
( 4250 ), 525
1050

3
2
3 3

2.3

2
3


(
3)
2.3.1
3
Electronic Load (Kikusui
PLZ1004W)
256
256 30
1

( 0.005%)
24 24
1
3

(Wh)
(km/h)
st
4250 (1 )
1004.5
9.85
nd
4250 (2 )
969.14
10.04
rd
4250 (3 )
1017.48
10.48
st
4775 (1 )
1067.12
12.03
4775 (2nd)
1158.53
11.60
st
5300 (1 )
1190.57
10.51
nd
5300 (2 )
1113.29
10.66
3

2 3 4250
3
2 3
5 Hz 21
6241
3 15.3 kW
0.6 kW

78

2103499 2556 AU01

3 30
30-60%
( 3)


( 3)
2.3.2

4250





3

2 4

4 57.7 [6]


3.
3 5

(YUASA EB130)
(Boostcap 1500F)
(CALB CA100) 5

6

4 3

10:40-11.01
11:05-11.24
11:30-11.52
11:55-12.15
12:20-12.40
12:45-13.06
13:10-13.30
13:35-13.56
14:00-14.20

( 8 )
27.6
39.9
43.0
50.4
71.5
41.3
35.3
20.3
15.7

( )
15
20
20
25
35
20
20
10
10

5
5

(V)
12.8 V
12.8 V

(Ah)
130 @ 6.5 A
130 @ 6.5 A

12.8 V

100 @ 30A

5
(Nominal Voltage) 12.8 V

( 6
(Circuit
2.7 V)
(Connector)
Breaker)






( 0.7 Ah)



LiFePO4
3.2 V 4

3
4

79

2103499 2556 AU01

7

4.
2

Peukerts Law

Peukert
1.33
1.02

6.5 A (Ah)
100.77
98.05

4.1 (Capacity
Test)

(Cut Off
Voltage)
1
(1)
C (Ah), I
(Constant Current)
(A) t (h)
6

7 Peukert
2 Peukert 1



1 2
3
(3)

(C) (I) 6.5 A

1 A (E)
5.
3

8

(A)
20
50
65
30
50
70
(Ah)
69.6 51.0 47.4 94.9 94.5 93.1

st

(1 )
nd

(2 )

st
(1 )

nd
(2 )
st
(1 )
nd
(2 )

30%
20 A 65 A
2%
4.2 Peukerts Law

Peukerts Law 2

13
13
16

4 hr. 45 m
4 hr. 45 m
5 hr. 47 m

(Ah)
58.07
57.46
71.80

(%)
57.63
57.02
71.25

16

5 hr. 41 m

71.34

70.79

25
25

8 hr. 49 m
8 hr. 53 m

95.08
95.99

96.97
97.90

8
3


13


16 25
20.84%
86.32%

Peukert

(2)

I t 1 E
1 A k Peukert
1
[7]



6.5 A

80

2103499 2556 AU01

13.70%
40.11%

3


Electronic Load


3
20.84%
86.32%

13.70% 40.11%

..




.

[1] ,
2555, http://www.m-society.go.th/article_attach/9831/14069.pdf
25 2557
[2] ,
,
http://vigportal.mot.go.th/portal/site/PortalMOT/stat/ index20URL/
25 2557
[3] . 2556.
ED95, http://www.thaipost.net/xcite/180413/72332 25 2557
[4] Bogdan, Vulturescu., et al. Implementation and test of a
hybrid storage system on an electric urban bus. ELSEVIER
No.30 (2013): 55-66.
[5] Nansi, Xue., et al. Design of a lithium-ion battery pack for
PHEV using a hybrid optimization method. ELSEVIER No.115
(2014): 591-602.
[6] Sarah Catherine Walpole., et al. The weight of nations: an
estimation of adult human biomass. BMC Public Health (June
2012). doi: 1186/1471-2458-12-43910.
[7] .
. 1 . :
, 2554.

7






3


3

6.

3


3
,

81

2103499 2556 AU02

Force feedback Steer- by-Wire

Driving Simulator for Evaluating Force feedback impact on Driving precision with Steer- by-Wire System
5330433721 5330223221
.. ( ) . ,
( )

Force feedback
Steer-by-Wire
Deadband 0 5 Force
feedback model Force feedback torque
0
Deadband
Deadband 4
, , Force
feedback, Steer-by-Wire (SBW), Deadband, Force feedback
model

Force feedback
SbW
0

[1] Deadband
Steering input Force feedback Deadband
SbW
Driving Simulator
Deadband
Deadband
Steer-by-Wire
Steer-by-Wire

Abstract
This paper presents an Evaluation of Force feedback
impact on Driving precision with Steer-by-Wire System by
Driving simulator. Experimental was designed to focus on
Deadband size from 0 to 5 degrees of Force feedback model
which is a zero zone of a relation between Force feedback
torque and hand wheel angle. The result has shown that the
driver has the best driving precision with Deadband size of 4
degrees.
Keywords: Driving simulator, Driving precision, Force
feedback, Steer-by-Wire (SBW), Deadband, Force feedback
model

2.

2 Software design
Hardware design
2.1 Software design

1

Interface LabVIEW
Vehicle Model Unity 3D
Display Force feedback model LabVIEW
Interface Force
feedback

1.
Drive-by-Wire (DbW, by-Wire or x-by-Wire)



Steer-by-Wire (SbW)
SbW Force feed back
SbW
Force feedback
Force feedback

82

2103499 2556 AU04

2.2 Hardware design
Hardware design angle
sensor potentiometer Inertia mass
Inertia

2.2.1 Inertia mass
1:1
Inertia mass


Slot

Inertia mass
SolidWorks 3

Error
Average, Error SD, Error Max
Force feedback
Torque Steering angle
Dead band Torque
5 0,2 , 3,4 5
Dead band
0.026, 0.044 N.m/deg [2] Dead
band
4.
10
5 Dead band
Error average

5 Dead band 5
Error average
Dead band = 4, 0.44 N.m/deg
Error average

3 Inertia mass
2.2.3 Potentiometer
Potentiometer
Potentiometer multi-turn Potentiometer
10
Potentiometer
Inertia mass 1:2
2.5
Potentiometer
Potentiometer
L
Potentiometer
Potentiometer
Potentiometer
SolidWorks
4

4 Potentiometer
3.

10
Force feedback

83

5 Error Dead band

..
. ,

[1] Susan G. Hill Jason S. Metcalfe, and Kaleb McDowell, The
Use of a Steering Shaping Function to Improve Human
Prerformance in By-Wire Vehicle, Army Research Laboratory,
2008
[2] D. Toffin, G. Reymond, A.Kemeny, J. Droulez, Influence of
Steering Wheel Torque Feedback in a Dynamic Driving
Simulator, DSC North America 2003 (October 2003).
[3] Ronald R. Mourant and Praveen Sadhu, EVALUATION OF
FORCE FEEDBACK STEERING IN A FIXED BASED DRIVING
SIMULATOR, Proceedings of the HUMAN FACTORS AND
ERGONOMICS SOCIETY 46th ANNUAL MEETING --200

2103499 2556 AU03

Forklift Driving Simulator

Mr.Natthapong Angsupasirikul 5330143721, Mr.Kawin Koonvisal 5330021821
Asst. Dr. Sunhapos Chantranuwathana (Project-consultant)
Abstract
This paper present design and production of forklift driving
simulator for beginner forklift driver training. The forklift
simulator system consists of simulator programming,
production of driver console and parameter adjustment. The
simulator must be able to simulate driving, turning, tilting and
lifting a pallet while able to check if the forklift crash with
environment. The result demonstrates that the simulator can be
used to train forklift driver at beginner level.
1. Introduction
Nowadays, forklifts are used to transport stuff in many
industries. To use a forklift, a skilled driver is needed. From
OSHAs forklift-related accident statistic [1] demonstrates that
every year there are nearly 34,900 times of accident, and 85 of
these have dead casualties. In the past, test tracks are used to
train the drivers, but these training led to injuries and damage
costs. Training in a simulator before driving real forklift can
eliminates those problems.

Figure 1: Component of forklift [2]

There are mainly 4 things that forklift need to be controlled
to transport stuff.
1. Straight driving. Some forklifts are drove by an internal
combustion engine, mostly diesel engines, and some
are drove by electric motors. An internal combustion
engine forklift is chosen to be forklift model for this

simulator.
2. Turning or steering. Chosen forklift model has 2 front
wheel which cannot steer, and 2 rear wheel used to
steering.
3. Fork lifting. Forklifts carriage can be lifted up and
down along a mast to transport a pallet.
4. Mast angle rising. Because sometimes a driver needs
to lift a pallet with an angle to the ground, so the mast
must be able to tilt up and down.
Forklift driving simulator creating has three main tasks.
First is simulator programming in a program named Unity3D.
Second is driver console design and production. And the last is
forklift parameter adjustment
2. Simulator Programming
There are two alternatives to create forklift driving simulator
by Unity3D
First is to change position of forklift graphic model and
other graphic model from physics equation script. To do this,
we need to create a formula to calculate where the forklift and
other object should be according to human input. Programming
this way is very difficult and simulation of forklift driving on non
planar plane is nearly impossible.
Second alternative is to use physics engine that comes
with Unity3D. To do this, we need to build a physical collider of
every object. The position of forklift and other object is not
controlled by script but by changing variable. To drive a forklift,
we can change the torque of forklift wheel, and steering by
rotate rear wheels. To lift a pallet, we change the appearance
of forklift model (lift carriage upward) and a pallet that place
above will be lifted. Pallet position is controlled by physics
engine that calculate every force on it which include gravity
and friction. This alternative is easier than the first one but
sometimes the simulator becomes unstable because physics
engine calculated in discrete time, and sometimes physics
engine isnt work as it should be like forklift driving acceleration
is not related to its whole mass and wheel torque but related to
wheel mass and wheel torque instead. So, to use physics
engine, we must check that physics are work as planned.

84

2103499 2556 AU03

Alternative two is used to build the simulator
2.1 Design simulator with Unity3Ds physics engine.
Unity3D is programmed by an OOP, Object-Oriented
Programming, concept. To create a forklift object in simulator,
first, create object with forklift graphic model, which can be
downloaded for free.

Figure 2: Forklift graphic model in Unity3D

Then, place physical collider around the forklift which
includes WheelCollider. Place WheelColliders according to
graphic model. All size parameter of the forklift will be this
graphic models instead of the real one.
The WheelCollider component has many variables that can
be controlled via scripting. Only three are controlled to simulate
driving and turning.
Table 1: list of used WheelColliders variable
Variable
Function
Type of use.
name
motorTorque Torque generated by wheel.
Controlled
brakeTorque Braking torque generated by
Constant
wheel
(vary if brake
is applied)
steerAngle
Angle of wheel (used to
Controlled
turning)
spring
Suspension spring for wheel
Constant
damper
Damper of wheels suspension Constant
To have driver feel like driving on a real forklift, the
simulator has 3 windows and use head tracking system to
change position of camera according to driver head position.
This simulator also uses 3D stereoscopic to give driver an
ability to know distance of object he sees.
3. Driver Console
The driver console must be similar as much as possible to the
real forklift control panel in term of consoles location, motion,
size and output signal. There are four steps in process of
creating the console.

85

3.1 Forklift control panel

We use FG20, forklift from department of Mechanical
Engineering Laboratory as the model. The forklift control panel
has eleven controllers with many difference motion types.

Figure 3: Forklift control panel

Table 2: Forklifts controller and its function & motion
Forklift Controller
Function
Motion
Steering wheel
Direction control
Rotation around
axis
Horn
Warning sound
Push button
Hand brake
Brake control
Lever
Headlight
Create light
Pull button
Turn signal
Create turn signal
Lever
Direction gear
Direction control
Lever
Speed gear
Speed control
Lever
Fork lifting gear
Reach fork control
Spring lever
Fork tilting gear
Reach fork control
Spring lever
Accelerator
Speed control
Pedal
Brake
Brake control
Pedal
3.2 Alternative controller
From the forklift control panel, we can use alternative
controllers to create similar motion as the original one.

Figure 4: Example of alternative controllers

2103499 2556 AU03

Table 3: Alternative controllers
Forklift Controller
Alternative
controller
Steering wheel
Encoder

Motion
Rotation around
axis
Push button
Push button
Push button
Lever
Lever
Lever
Spring lever
Spring lever
Pedal
Pedal

Horn
Push button
Hand brake
Step push button
Headlight
Step push button
Turn signal
Step Lever
Direction gear
Step Lever
Speed gear
Step Lever
Fork lifting gear
Spring lever
Fork tilting gear
Spring lever
Accelerator
Accelerator pedal
Brake
Brake pedal
3.3 Building the console
Driver console can be divided into three parts.
3.3.1 Steering wheel controller
Using coupling to directly connect between steering wheel
shaft and encoder and grip them with the frame.

Figure 5: Steering wheel controller

After the design process, we build a prototype of the
steering wheel controller in mechanical engineering laboratory.

Figure 6: Prototype of the steering wheel controller

3.3.2 Lever Controller
Design the location for levers to stay at the same place
with the forklift control panel.

Figure 7: Lever Controller

After that, we created a prototype of lever controller using
alternative controller from table 3

Figure 8: Prototype of lever controller

3.3.3 Pedal Controller
Using the ready-made pedal (accelerator and brake) and
place it to the ground level under the lever controller.

Figure 9: Pedal Controller

3.4 Computers connection
Normal wires can be used to connect between the console
(Steering wheel controller, Lever Controller and Pedal
Controller) and NI USB-6251, the signal receiver/generator.
Using LabVIEW, computer program to generate 5 V. output
signal through NI and send to the console, and then use NI to
receive the input signal that pass through the console and
send it back to LabVIEW to be the input data for simulation.
4. Forklift parameter adjustment
Size of forklift can be obtained from graphic model.
However, parameters related to 4 abilities of simulator must be
obtained from real forklift. For this simulator, we use FG20, a
20-ton size forklift of department of Mechanical Engineering, as
an operation model.

86

2103499 2556 AU03

change in to engine curve of forklift. (Assume that maximum
speed of the engine is 6000 RPM and gear ratio is 1 for
Forward-Low.)

Figure 10: Real FG20 forklift

4.1 Driving ability adjustment.
Parameters related to driving ability are acceleration and
speed, which take throttle as an input variable. Before using
physics engine to simulate driving ability, we need to confirm if
relation between motorTorque variable and acceleration is
linear. To confirm this, we vary motorTorque variable and
calculate acceleration from position changed each frames in
simulators. Result is that relation of motorTorque and
acceleration is linear within 1% error.
Driving ability was adjusted by obtain relation of
acceleration vary by speed and throttle. But forklift testing had
many obstacle, GPS cannot be used, only device available for
driving testing is accelerometer which give relation of
acceleration and time. And it is very difficult to perform many
trials.
4.1.1 Driving ability testing
The testing track has three points. Distance from first to
second point is 15 meters while distance from second and third
point is 10 meters.
Test was performed by maximizing throttle with ForwardLow gear. Velocity of forklift was stable before reaching the
second point. So, the maximum velocity of Forward-Low gear
can be obtained from time of last 10 meters.
4.1.2 Driving ability data processing

Figure 11: Graph of acceleration and time

After 4th polynomial graph of acceleration was created,
integrated into velocity and found multiplier that will make final
velocity equal to velocity obtained from last 10 meters, and
then multiply this multiplier into acceleration graph. Lastly, plot
acceleration (vertical line) against speed (horizontal line) and

87

Figure 12: Engine curve (wheel torque) of maximum throttle.

Before torque from engine curve can be used, resistance
force, from brake torque constant, must be added so the
torque applied can be used to create net acceleration.
brakeTorque constant can be obtained from time that the
forklift used to stop from max velocity.
Acceleration of forklift is effected by throttle ratio, but given
that throttle-vary test was unable to do. So, assumption of
throttle and torque relationship has to be done.
To assume relation of torque and throttle, Maximum-Power
concept was used. With this concept, maximum power supplied
to engine varied by throttle ratio. If torque from engine curve is
less than maximum power supplied, torque of engine curve is
chosen. But if torque from engine curve is higher, torque will
be limited to torque form maximum power supplied. Maximum
power supplied at maximum throttle is maximum power of
engine curve. Origin of this concept is that the throttle used to
vary amount of fuel flow, but sometimes fuel flow is more than
engine can use so torque is limited by engine curve.

Figure 13: Torque from maximum power supplied

2103499 2556 AU03

4.2 Turning ability adjustment.
Forklifts are turning by steering their rear wheels. Forklift
turning ability can be simplified into bicycle model, which there
are only one front wheel and one rear wheel. Turning radius of
bicycle model can be calculated by steering angle of rear
wheel and distance between front and rear wheel, however,
distance from graphic model was used instead of FG20.

Figure 14: Size parameter for bicycle model calculation.

Radius of turning(R), steer angle of bicycle model () and
steer angle of real wheel () are calculated by these equations
below
tan( ) L/ R
(1)
tan( ) L / R D / 2
(2)
Because turning ability use WheelCollider which is physics
engine calculation, we need to test if turning radius calculation
of physics engine is correct compare with real radius of turning
calculated by position of each frame. The result is that bicycle
model can be used with below 1% error.
So, in this adjustment, information needed is relationship of
steering wheel on driver console and steer angle of rear wheel
in bicycle model.
4.2.1 Turning ability testing
Steer angle of bicycle model rear wheel can be calculated
by steer angle of one of rear wheel. Steer angle was obtain by
stick the wheel with laser pointer and then spin steering wheel,
measure distance created by steer angle change. Laser
distance was measured each 180 degrees of steering wheel.

Figure 15: Setting for measure steer angle with laser.

4.2.2 Turning ability data processing

From relation between steer angle of bicycle model rear
wheel and steering wheel, 6th polynomial relationship was
created to use within simulator.

Figure 16: Steer angle vs. Angle of steering wheel

4.3 Lifting ability adjustment
Lifting range and lifting speed are important parameters to
use in the simulation system. Lifting range can be easily
measured by mark the location when the reach fork is in the
lowest position and when it is in the highest position then
measured the different range. As for lifting speed, the value
can be calculated by the basic equation below
lifting speed = (lifting range)/(lifting time)
(3)
4.3.1 Lifting ability testing
Maximum lifting range can be measured by sticking a laser
pointer to the reach fork and make it point to the meter rule
that perpendicular to the ground, then mark the position where
reach fork is in lowest and highest position.
As for lifting speed, it can be found by measured the time
that the reach fork use to move from its lowest to highest
position and vice versa.

Figure 17: Lifting range method

4.3.2 Lifting ability data processing
From the calculation on finding the different range between
the highest and the lowest position of reach fork. The
maximum lifting range is 3 meters. Using the equation (3),
lifting speed for upward is 0.222 meters per second and lifting
speed for downward is 0.3 meters per second.

88

2103499 2556 AU03

4.4 Mast tilting ability adjustment
Mast is a part of forklift that the carriage can lift up and
down along with. Because tilting speed is determined by a
controller. So, maximum tilting speed and tilting range must be
known.
4.4.1 Mast tilting ability testing
Maximum tilting speed can be obtained by pushing the
controller to maximum and measure time needed for tilting
lowest to highest. Laser pointer was used to measure the
angle of tilting.

Acknowledgement
Advice given by Assist. Prof. Sunhapos Chantranuwathana,
Assist. Prof. Angkee Sripakagorn and Lect. Nuksit Noomwongs
has been a great help in the project.
Assistance provided by teachers from department of
mechanical engineering laboratory, Hans building, was greatly
appreciated.
Advice given by graduate students from Smart Mobility
Research Center has been a great help in the project.
References
[1] TOOLBOXTOPICS, Forklift Fatalities, 2012. Available:
http://www.toolboxtopics.com/Gen%20Industry/Forklift%20F
atalities.htm.
[2] WIKIPEDIA, Forklift Truck, 2006. Available:
http://en.wikipedia.org/wiki/File:Forklift_Truck.jpg.

Figure 18: Diagram for angle measurement with laser pointer.

Because center of tilting is at the axle, the carriage must
be set align with the axle. Then, stick a laser pointer at the end
of fork and perpendicular to ground. Tilt the mast and then
record distance the red dot moved.
4.4.2 Mast tilting ability data processing
Upward tilting angle can be calculated by equations.
x r r cos
(4)
(5)
y h r sin
(6)
r y tan x
And downward tilting angle can be calculated by these
equations below, notes that angle is always positive.
x r r cos
(7)
(8)
y h r sin
5. Conclusion
Forklift Driving Simulator Project uses the driver console
that each of them has similar motion and location to the real
forklift vehicle. Combining with the simulation system that has
the dynamic model and other important parameters as same
as the real model. The project also provide the large-scale
environmental simulation such as roads, hallways, large rooms
to drive in, objects that can be picked up or wall and
obstruction to avoid. With these functions, when drivers use the
simulator, they will feel like they are driving the real forklift. In
addition, the other function of the project is Training Program
that provides simulation of events which can easily cause
accidents such as driving up or down the slope. Trainees can
use the forklift simulator for their first-time driving before they
drive a real one to reduced practice time and accident rate.

89

2103499 2556 AU04

Comparison and prediction of fuel economy between Hybrid electric vehicles and Internal combustion engine vehicles in
Bangkok Traffic
5330060221, 5330083721, 5330413121
.. ( )

(Hybrid Electric Vehicles : HEVs)
(Internal Combustion Engine Vehicles :
ICEVs)
Vehicle Specific Power (VSP)


Engine Control Unit (ECU)
(OBD)

HEVs ICEVs 24.4%, 48.4%
53.9% ,

HEVs
10.9%,-18.8%
0.3% ICEVs

-7.5%, -21.9% 12.9%
, ,

Abstract
This research focuses on developing a simple method to
predict fuel economy and compare the fuel economy of Hybrid
electric vehicle (HEVs) and Internal combustion engine vehicle
(ICEVs) fuel economy characteristic in the different traffic
patterns by using Vehicle Specific Power Methodology. Vehicle
Specific Power (VSP) represents the vehicle power demand
from speed, acceleration and friction which were obtained
through the OBD interface. By comparing the fuel economy
between ICEVs and HEVs in Highway, Suburban and City
traffic, HEV fuel economy were 24.4%, 48.4% and 53.9%
respectively lower than ICEV. By using Vehicle specific power
model to predict fuel economy in different pattern of traffic HEVs
fuel economy were 10.9% ,-18.8% and 0.3% different from
actual in City, Suburban and Highway traffic respectively. The
prediction of ICEVs fuel economy was -7.5%, -21.9% and

12.9% different from actual fuel used respectively in City,

Suburban and Highway traffic.
Keywords Hybrid electric vehicles, Vehicle specific power, Fuel
economy , On road measurement.
1.




35.7%
[1]











HEVs ICEVs

ICEVs
HEVs
(Vehicle Specific Power: VSP)

2.
(On road measurement)
3

2 HEVs ICEVs

90

2103499 2556 AU04

1,000
2

2.1
Global Tech Stream (GTS)
Electric control unit (ECU)
(On board diagnostic
system :OBD-II) Speed,
Engine speed ,Fuel injection volume ,Engine on/off condition
10 Hz 1
Global
TechStream
Data Logging
Software

Vehicle
Interface
Module

On Board
Diagnostic
(OBD) Port

Data Logger

Electric
Controlled
Unit
(ECU)

Data
logging
1

2.3

(Injection Volume)
(Mass
air flow)

(Exhaust Gas Recirculation System : EGR)







1.86%
2

Software

2.2
(Vehicle Specific Power : VSP)


[3] VSP

( 1) kW/ton
(

(1)

v : (m/s)
a : (m/s2)
: (m/s2)
:
o
: ( )
(
= 0 o)
= 0.132 m/s2 = 0.000302 [3]
(1) VSP
VSP 14
1
1 VSP
Mode VSP(kW/ton) Mode VSP(kW/ton)
1
VSP-2
8
13<VSP16
2
-2<VSP0
9
16<VSP19
3
0<VSP1
10
19<VSP22
4
1<VSP4
11
22<VSP25
5
4<VSP7
12
25<VSP28
6
7<VSP10
13
28<VSP31
7
10<VSP13
14
VSP>31

(2)
Injection Volume : (mL)
N : (rpm)
: Gasohol 91= 737 kg/m3
n : = 4
K : 1
2.4

[2]
2

-
2

-
3

91

2103499 2556 AU04

500m

2.5

3
1. (City Traffic)

10 km/h

2. (Suburban Traffic)

10-60 km/h

3. (Highway Traffic)
60 m/h

3
Type
of
vehicle
ICEV
ICEV
ICEV
HEV
HEV
HEV
ICEV
ICEV
ICEV
HEV
HEV
HEV
ICEV
ICEV
ICEV
ICEV
HEV
HEV
HEV
HEV

3 km

Type
of
route
HWY1
HWY2
HWY3
HWY4
HWY5
HWY6
SUB1
SUB2
SUB3
SUB4
SUB5
SUB6
CTY1
CTY2
CTY3
CTY4
CTY5
CTY6
CTY7
CTY8

Average
speed
(km/h)
68.3
75.4
78.2
69.0
70.5
72.0
15.4
35.7
34.0
15.9
36.4
37.7
4.0
7.1
3.1
6.5
6.3
2.7
3.3
5.3

Distance
(km)

Stop/km

37.7
37.2
36.7
37.8
39.4
37.3
11.7
11.4
11.7
11.5
12.4
11.6
4.0
7.4
3.0
7.9
6.2
2.7
3.3
3.6

0.03
0.05
0.00
0.08
0.05
0.03
2.49
0.44
0.26
1.22
0.16
0.26
9.42
6.46
14.43
8.11
5.98
19.10
8.73
8.68

%
Engine
off
0
0
0
29.2
23.8
15.6
0
0
0
67.2
71.3
62.8
0
0
0
0
81.1
83.0
79.6
85.0

%
Cycle
Idling
3.5
0.6
0.0
1.4
1.1
4.8
6.8
5.6
15.0
4.0
1.4
4.3
72.2
56.6
75.2
56.4
59.2
76.1
73.5
65.9

* HWY : ,SUB : CTY :

4 km

2.6
Toyota Prius
3rd Generation
Toyota
Corolla Altis
2

4

92

Fuel
Economy
(L/100km)
7.0
6.1
5.4
4.7
4.3
4.5
10.6
6.8
6.0
5.4
2.8
3.8
31.3
17.5
37.0
18.5
9.0
15.9
11.8
9.1

2103499 2556 AU04

VSP VSP

Engine
Dimension
(Length x Width x Height)
Displacement volume(cc)
Compression ratio
Engine Power (kW/RPM)
ICE torque (Nm/RPM)
Vehicle gross mass (kg)
Electric motor (kW)
Electric motor torque (Nm)
Battery type
Battery capacity (Ah)
Battery nominal voltage (V)
Combined max. power (kW)

Toyata Prius
3rd Gen.
2ZR-FXE
4480 x 1745
x 1505
1798
13:1
73/5200
142/4000
1725
60
207
Ni-MH
6.5
201.6
100

Toyota
Corolla Altis
2ZR-FBE
4540 x 1760
x 1465
1,798
10:1
104/6,000
177/4000
1245
104

3.
3.1
ICEVs HEVs

HEVs
ICEVs 53.9% HEVs






80% 3 HEVs
ICEVs
(Idling) ICEVs
HEVs

HEVs ICEVs 48.4%

HEVs

HEVs ICEV 24.4%
HEVs

2.6

ICEVs HEVs 16.00-19.00 .

(3)

HEVs

ICEVs

25.0

Fuel Economy (L/100km)

2.7

VSP

VSP
VSP 7
3

20.0
15.0
10.0
5.0

10.86 23.53

3.75 7.27

4.53 5.98

City

Suburban

Highway

0.0

F
Di
fi
T
i

:
:
:
:
:
:

( L )
( % )
(g/s)
(s)
( g/L )
VSP

5 ICEVs HEVs
3.2 VSP-Based Fuel Consumption Model

VSP ICEVs
HEVs 3 6 7

93

2103499 2556 AU04

Fuel Consumption (g/s)

VSP ICEVs HEVs

20% (Normal Distribution)
VSP-Based Fuel Consumption Model 8
ICEVs HEVs VSP
3.00
2.75
2.50
2.25
2.00
1.75
1.50
1.25
1.00
0.75
0.50
0.25
0.00
1

City

Suburban

7
8
VSP Mode

Highway

10

11

12

13

14

12

13

14

13

14

6 ICEVs

Suburban

Highway

VSP Mode

10

11

7 HEVs

Fuel consumption (g/s)

Fuel Consumption (g/s)

City
3.00
2.75
2.50
2.25
2.00
1.75
1.50
1.25
1.00
0.75
0.50
0.25
0.00

ICEVs AVG Fuel consumption

3.00
2.75
2.50
2.25
2.00
1.75
1.50
1.25
1.00
0.75
0.50
0.25
0.00
1

HEVs AVG Fuel consumption

VSP Mode

10

11

12

8 ICEVs HEVs
94

2103499 2556 AU04

3.3

3
4
4 ICEVs HEVs

HEVs

ICEV
s

Type
of
Route
CTY

Average
Speed
(km/h)
8.78

8.07

Predict
Fuel
(L)
0.56

Actual
Fuel
(L)
0.63

Stop/
km

Distance
(km)

4.71

Different
(%)
10.85

SUB

39.60

0.75

26.56

1.50

1.27

-18.81

HWY

68.95

0.08

37.76

1.77

1.78

0.26

CTY
SUB
HWY

3.38
34.01
68.25

9.47
0.26
0.03

5.38
11.67
37.72

1.80
0.84
2.31

1.68
0.69
2.65

-7.47
-21.89
12.85

4.

HEVs ICEVs HEVs ICEVs
53.9%
48.4% 24.4% HEVs

HEVs
ICEVs

VSP

VSPBased Fuel Consumption Model
HEVs ICEVs
HEVs

-10.9% -18.8% 0.3%
ICEVs

-7.5%, -21.9%
12.9%

..


[1] ,
- 2556.
.
http://www.dede.go.th/dede/images/stories/stat_dede/sit_56/sit_dec.pdf
13 2557

[2] ,
, .
http://cpd.bangkok.go.th:90/web2/new_cpd/PostCityPlan/bkkp
lan_01_hires.pdf
14 2557

[3] Haibo Zhai , H. Christopher Frey, Nagui M. Rouphail.

Development of a modal emissions model for a hybrid electric
vehicle. Transportation Research Part D 16 (2011) : 444-450.

95

2103499 2556 AU05

Safe Driving Evaluation based on Longitudinal and Lateral Acceleration Measurement

5330200821, 5330202021

5330216921
.. ( )

4

Microsoft Visual Basic 6.5
MATLAB 2012a





Abstract
This project aims to create a database of driving behavior
and rating driver program. We bring data, recorded by the
driving recorder, verified by the reference driving recorder, to
analyze and classify accident risk to four events, rapid
acceleration, harsh braking, rapid cornering, and exceed speed
limit. We use Microsoft Visual Basic 6.5 for creating database
and use MATLAB 2012a for rating drivers. The results of this
project are able to create a database application and rating
driver program. However, those programs are inconvenient for
practical use because there are many steps to use. They need
to develop a model of the program to be easy to use.
Keywords: Driving behavior, Driving recorder, Acceleration
1.



,
,

2.

2

2.1
KAYABA
5 Hz
1 Hz

1 KAYABA DRE-400
2.2
VBOX

100 Hz

2 VBOX 3iSL

96

2103499 2556 AU05

VBOX

KAYABA

Acceleration (g)

0.6
0.4
0.2
0
-0.2

10

20

30

40

50

Time (s)

-0.4

R2012a Microsoft Visual Basic 6.5

5

MATLAB
R2012a

(
) 2

-0.6

4.

MATLAB R2012a
1 2

3 (Acceleration)

3.

(Threshold)
4
(Rapid Acceleration) (Harsh
Braking) (Rapid Cornering)
(Exceed Speed Limit)
Listing of all files in: C:\Users\SONY\Desktop\Data m file
File Name
Date Modified File Size (Kb) Event
Max Acceleration (g)
5/2/2014 21:12
0.1 Rapid Acceleration
0.37
AC100205021402.m
AC200205021402.m
7/2/2014 14:14
0.1 Rapid Acceleration
0.25
7/2/2014 14:14
0.1 Rapid Acceleration
0.4
AC300205021402.m
AC400205021402.m
7/2/2014 14:16
0.1 Rapid Acceleration
0.2
7/2/2014 15:30
0.1 Harsh Braking
0.89
BR100205021402.m
7/2/2014 13:42
0.1 Rapid Cornering
0.53
CU100205021402.m
CU200205021402.m
7/2/2014 15:24
0.1 Rapid Cornering
0.59
6/2/2014 14:34
2.3 N/A
importfile.m
importfileSpeed.m
7/2/2014 13:47
2.3 N/A
7/2/2014 13:44
0.1 Exceed speed limit
0.19
SP100205021402.m
7/2/2014 13:48
0.1 Exceed speed limit
0.2
SP200205021402.m
SP300205021402.m
7/2/2014 15:19
0.1 Exceed speed limit
0.21
7/2/2014
15:20
0.1
Exceed
speed
limit
0.15
SP400205021402.m
SP500205021402.m
7/2/2014 15:21
0.1 Exceed speed limit
0.18

Choose Event
Rapid Acceleration
Find Data
File Name
Date Modified File Size (Kb) Event
Max Acceleration (g)
5/2/2014 21:12
0.1 Rapid Acceleration
0.37
AC100205021402.m
7/2/2014 14:14
0.1 Rapid Acceleration
0.25
AC200205021402.m
7/2/2014 14:14
0.1 Rapid Acceleration
0.4
AC300205021402.m
AC400205021402.m
7/2/2014 14:16
0.1 Rapid Acceleration
0.2

4 Microsoft Visual Basic 6.5

R=

Er

Er =

Pi

(1)
(2)

R (Driver Rating) Er
(Event Rating) Pi
(Point) i, n 1
0 1 6

6

5.





4

ISUZU All New D-max X-series
Smart Mobility Research Center

5
MATLAB R2012a
4

MATLAB

97

[1] Chalermpol Saiprasert, Wasan Pattara-Atikom, Computer
of Safety Rating Using In-Vehicle Data Recorder: Case
Study of Bangkok Public Transportation, National
Electronics and Computer Technology Center, 2012.

2103499 2556 AU06

Hardware in the loop testing for student formula car suspension
5330437221 5330540121 5330390321
.. ( )

Skidpad






: , ,

2.


Skidpad

[1]



VBOX sensor
Stroke sensor
VBOX sensor

Abstract
The objective of this experiment is to find the difference
between the car suspension simulator and the student formula
car to prove the simulators reliability. Test constructed using
Skidpad method, by driving in a constant radius circle
driveway. After the test, the collected data in terms of graph,
the steering angle and the lateral acceleration, is compared
between the data acquired from the experiment and the data
from the simulator. The trend of the results both from the
experiment and from the simulator are very similar and have
an oversteer behavior of the car. Therefore it proves that the
simulator has a level of precision enough to substitute the need
of designing and manufacturing the prototype car.
Keyword: Steering angle, lateral acceleration, over steer
turning

3.
3.1

3
1. (Understeer)
2. (Oversteer)
3. (Neutral steer)



( )
(wheelbase , L) (R) (1)[2]

1.

(HILs)

(1)


1

1 [3]

98

2103499 2556 AU06

3.2

Stroke sensor
VBOX
sensor
Stroke sensor



5 -25 25
0
2

(I),
(W), (), ( )
(V) Labview 8.5
2555


1

5.

2
2

(2)

Steer Angle [degree] = -427.38(Voltage [V])

+ 200.31(Voltage [V]) 1.2691

3
HILs,

(2)

4.
2



VBOX sensor
Skidpad
8
15

5

30


Stroke sensor
VBOX sensor
(2)


1

99

6.

(HILs)


.. ..

VBOX sensor

[1] Yee Man Chan, Chi Yung Ng. Road Resistance. Wikipedia,
http://en.wikipedia.org/wiki/Skidpad 5
2556
[2] Neha Ravi Dixit. Evaluation of Vehicle Understeer Gradient
Definitions. Masters Thesis, , Department of Mechanical
Engineering, The Ohio State University, 2009.
[3] Gillespie T. D., Fundamentals of Vehicle Dynamics, SAE,
Warrendale, PA, 1992

2103499 2556 AU07

EVALUATION RATING ECO-DRIVING

5330069021, 5330172921 5330369221
..
( )

(Driving
Simulator)
(EcoDriving Index)

(Velocity Model)
(Fuel Consumption Model)
ISUZU DMAX X-SERIES
(ECO-Rating)
5




0

Abstract
The purpose of this project is to build Driving Simulator,
which evaluates drivers driving behavior by using Eco-Rating
Index as an indicator. To ensure the reliability and accuracy of
Driving Simulator, the velocity model and the fuel consumption
model have been produced by collecting data from test drives
with real car, ISUZU D-MAX X-SERIES. The Eco-Rating is
created by collecting average fuel consumption at various
average velocity from five test drivers in the same traffic
condition via Driving Simulator. The fuel consumption rating of
each test driver will be compared and the Eco-Rating is sorted
by group. The test drivers who have the lowest fuel
consumption from driving simulator in each group will get full
score and those who have the highest fuel consumption will
get 0 score. Others will get the score sorted by order.
1.

Driver Eco Rating

Driving Simulator

2.
Innovate
Sport : LM-2 Digital Air/Fuel Ratio Meter

100

Motor

2103499 2556 AU07

(1)

(Time Constant)

(Sensor) (Actuator)
(Step input)

2 Port

() =

OBD 1.
(RPM) 2.(Km/h) 3.
(%) 4. (%) 5.Mass Air
Flow (g/s) 6.Air Fuel Ratio
3.

(Dynamic model)


(Differential Equation)

2
(Close form solution)
(Numerical Method)

3.1

Isuzu D-max X-series Simulator
Acceleration Paddle ()

(2)

(Steady
state Velocity) f[(t)] f[(t)]

DC
gain (t)
time constant transient b
First Order 3
() = f[(t)]

(3)

(
) () step function
5 Dynamic Model

Paddle Shift 1 1,2 -
1 (Step 1) (
40 ) (Step 2)
1 6 Step
1 3,4,5 - 3,4,5
1,2 3,4,5
0 km/h 3
paddle shift 3

Step 1,2,3 1
4 Step 2

3



First Order with no zero (DC gain)
[1]

101

4 1 Step

2103499 2556 AU07

(Steady state
velocity) ()

Fit curve Matlab 3rd order
polynomial

5 Fit curve
Time constant b

Innovate data logger

( Step function)

f(x)=
Fit curve a
f((t))
b step input b
Time constant

8 3

1 2 3

(RPM) (%)
(Km/h) 4
1>2 2>3 3>4
4>5



7


9

6 Fit Curve Time constant

Input Matlab
7

9
4
Fit Curve Matlab 10

7 (Input : Pedal, Output :Velocity)

102

2103499 2556 AU07

11 %Engine load x MAF Fuel
consumption(L/s)

10 Fit Curve


3.2


[2]
Fuel consumption = K (%Engine load MAF) + c

(4)

Fuel consumption %
Engine Load
MAF (Mass air flow)

1 5

Average fuel consumption

km/L Average velocity Innovate data
logger km/h Average velocity
Average fuel consumption
L/s

%Engine Load MAF Innovate
data logger
%Engine Load MAF

Fit curve
Matlab K
2
%Engine Load MAF 11

103

%Engine Load

%Engine
load 2



%Engine Load 12

12 %Load

MAF


MAF

3rd. order polynomial Fit curve

MAF
1 5

2103499 2556 AU07

4.1.2 Calibrate
Fish Eye
Calibrate
Pixel Centroid
Centroid

13 MAF MAF
4.
Driving Simulator
(Ghost Car)

Driving
Simulator Innovate
Data Logger 2
Matlab
4.1 2
Webcam 2 Code
Matlab Webcam
Centroid
14 L 2 L
Centroid R
Centroid

15 Matlab

16 Pixel Centroid
Centroid
2
4.2
4 1
6 3 2 1.5
10 Km/h 20 Km/h 30 Km/h 40Km/h

14
D Sin
(5)

104

2103499 2556 AU07

17



(ghost car)
Driving Simulator
V1 = V2 + V

(6)

V1 =
V2 =
V = s / t =

5. Driving Simulator
Driving Simulator
Unity 3d Unity3D



Driving Simulator
Driving Simulator
1. Fuel
Consumption
2. (Ghost Car)
3. Eco-Rating



0

20 Driving Simulator
4.
Dynamic Model Isuzu D-max Xseries 5 2

Driving Simulator
Driving Simulator
Eco-Rating


.
.
..
.


Smart mobility research center

Isuzu All D-max
x-series

[1] Dr. Sunhapos Chatranuwathana, Automotive Controls
Class Note , Present at Faculty of Engineering,
Chulalongkorn University, 2005
[2] A. Alessandrini, F. Filippi and F. Ortenzi, "Consumption
Calculation of Vehicles using OBD Data", 20th International
Emission Inventory Conference. Florida USA, 2012

19 Driving Simulator

105

2103499 2556 TF01

Effects of azimuthal control jets to main jet mass flowrate ratio on

the entrainment of a jet in crossflow
Settasit Chaikasetsin 5330460621 Teerapat Sushewakhul 5330220321 Piyapon Panusittikorn 5330279421
Assc. Prof. Asi Bunyajitradulya (Advisor)
Abstract
The effects of azimuthal control jets to main jet mass
flowrate ratio (rm) on the structure and entrainment of a jet in
crossflow (JICF) are investigated. The effective velocity ratio of
main jet to crossflow (r) is about 4 with initial jet velocity profile
being a fully developed turbulent pipe flow and the crossflow
Reynolds number (Recf) of 5,900. For the two controlled JICF
cases with two different rm, control jets are injected at the
same azimuthal positions = 135 (I135) but with two
different rm = 2% and 4% respectively. The experiment
employs Stereoscopic Particle Image Velocimetry (SPIV) to
measure three components of velocity on four cross planes,
x/rd = 0.5, 0.75, 1.0 and 1.5. The results are compared with
those of Zaman and Foss [1], Witayaprapakorn [2], and they
show good agreement. The results show that rm have effects
on JICF structures and volumetric entrainment. Specifically,
I135 at rm=4% causes the jet structures to penetrate much
deeper into the crossflow, causes change in vorticity structure,
and gives much higher entrainment than JICF, up to 57%
increase at x/rd = 1.5.
Keywords: Jet in crossflow, Entrainment, control jets, SPIV
1. Introduction
Jet in crossflow (JICF) is the flow in which a jet is injected
normally into a crossflow. JICF has wide range of applications,
e.g., mixing of fuel and air in combustors, film cooling on gas
turbine blades, dispersion of pollutants from smoke stacks. The
governing jet characteristics that influence the effectiveness of
the use of JICF in these applications are jet trajectory,
entrainment, and mixing. In this regard, the ability to
manipulate and control these governing characteristics,
especially entrainment and mixing in the case of combustors, is
desired since it can improve the effectiveness and efficiency of
these equipments.
Past researches on JICF have shed lights on JICF in many
aspects including the jet structures and characteristics. In
regard to the jet structures, Fric and Roshko [3] identified four
main structures in JICF, which are jet shear layer, horseshoe

vortices, wake structures, and counter-rotating vortex pair

(CVP) as shown in Fig. 1. Smith and Mungal [4] found that,
although CVP is the main mechanism for entrainment of JICF
in the far field, it does not result in entrainment enhancement
over a free jet. Instead, it is the formation of the CVP in the
near field that results in entrainment enhancement over that of
a free jet. In regard to the formation of the CVP, Yuan et al. [5]
proposed that the CVP is formed from hanging vortices, which
in turn formed in the skewed mixing layers at the lateral edges
of the jet. Yuan and Street [6] found that the jet entrainment is
related to the jet trajectory by power law in the far field (x/rd >
0.8) where r is effective velocity ratio and d is jet diameter.

Fig. 1. Vortical structures of the jet in crossflow, Fric and Roshko [3].

On the other hand, numerous techniques have been

proposed to manipulate and control JICF. Examples are tab
(Zaman and Foss [1]; Bunyajitradulya and Sathapornnanon
[7]), swirling jet (Wangjiraniran and Bunyajitradulya [8];
Bunyajitradulya and Sathapornnanon [7]), and pulsing
(MCloskey et al. [9]). Following the ideas of the entrainment
enhancement by the formation of the CVP and how the CVP is
formed described in the past works above, our Fluid Mechanics
Research Laboratory (FMRL) has developed the azimuthal
control jets technique, which has been proven to be energy
efficient and effective in controlling the trajectory and
entrainment of JICF (Kornsri et al. [10]; Witayaprapakorn [2]).
Specifically, Kornsri et al. [10] found that the azimuthal control
jets technique can be used to control the jet trajectory while
Witayaprapakorn [2] showed that the use of a pairs of
azimuthal control jets at the azimuthal positions = 135 can
be used to promote entrainment significantly over the baseline
uncontrolled JICF. These discoveries lead us to expand the

106

2103499 2556 TF01

investigation on the use of azimuthal control jets to control
JICF and to our present objective.
The objective of this experiment is to investigate the effects
of azimuthal control jets to main jet mass flowrate ratio (rm) on
the volumetric entrainment (E) of a jet in crossflow. In this
respect, it should be noted that, upto present, the
determination of the volumetric entrainment (E) of a jet by
experimental technique is still relatively rare and the
development of such technique is still challenging. This stems
from the difficulty in identifying the jet edge and extent
excluding the pure crossflow region, hence the jet volume
flowrate and entrainment, experimentally. In this work, we also
therefore develop a technique to circumvent this issue and can
determine the jet volume flowrate and entrainment effectively.
The details will be described in the next section.

measurement of velocity. As a marker for the jet extent, the

tracer particles clearly mark the jet fluid, jet extent, and jet
edge and clearly differentiate the jet extent from the pure
crossflow region (in which there are no tracer particles). As a
marker for the measurement of velocity, SPIV will therefore
register non zero velocity only in the jet region (in which there
are tracer particles) and will register absolute zero velocity in
the pure crossflow region (in which there are no tracer
particles). With this seeding technique, the jet edge and extent
can then be clearly identified, and the jet volume flowrate
determined. In effect, we then define and identify the jet extent
in this work as the region in which there are some jet fluids as
marked by the tracer particles in that region. Naturally, this jet
region contains some entrained crossflow fluid, but excludes
the pure crossflow region in which there are no jet fluid at all.

2. Principle of the experiment

2.2 Definition and evaluation of entrainment

2.1 Technique to identify boundary of the jet in surrounding

crossflow

The reference coordinate used in this experiment is shown

in Fig. 2: the origin is at the jet exit center, the x, y and z axes
are in the streamwise, transverse and spanwise directions,
respectively. The volumetric entrainment are defined by

In order to determine the volumetric entrainment (E) of

a jet, the jet volume flowrate (Qj) at any downstream cross
plane must be determined. In turn, in order to determine the jet
volume flowrate (Qj), the jet edge and extent excluding the
pure crossflow region - must be clearly identified. Past works
identified the jet edge and extent by, for example, setting the
threshold level for quantities such as turbulence. This results in
some arbitrariness regarding the threshold value to be used to
indicate the edge of the jet, hence some arbitrariness in the jet
extent. In addition, the region so identified naturally also
includes, and is modulated by, some part of the pure crossflow
region in some instants. This renders, at least conceptually,
some inaccuracy in the determination of the jet volume flowrate
due to these arbitrariness and modulation.
In this work, we use Stereoscopic Particle Image
Velocimetry (SPIV) to measure the velocity field. In this regard,
the use of SPIV in past works employs the SPIV in which both
the jet fluid and the crossflow fluid are seeded with tracer
particles, resulting in the inability to identify and differentiate
the jet edge and extent from the pure crossflow region clearly.
As a result, some threshold value in some mean quantities,
e.g., turbulence quantities, must be chosen to identify the jet
edge and extent, which in turn results in some arbitrariness
and modulation as mentioned above. To circumvent this, in this
work we use SPIV in which only the jet fluid and not the
crossflow fluid is seeded. As a result, the tracer particles act
both as the marker for the jet extent and the marker for the
107

Qj

(1)

Q0

where Q0 is the initial jet flowrate at jet exit plane and Qj is the
volume flowrate of the jet at any cross plane, defined by

(2)
Q j ( x, t ) Vx ( x , t )dA
A j ( x ,t )

Vx ( x , t )

where
is streamwise velocity field and the domain of
integration is limited to over the jet cross sectional area Aj(x,t)
only excluding the pure crossflow region - as depicted in Fig.
2. Due to the jet-seeding only scheme, if we denote the jet
streamwise velocity as registered by the SPIV as Vx , j ( x , t ) ,

Vx , j ( x , t ) then equals to Vx ( x, t ) when x is in the jet region

and equals to absolute zero when x is in the pure crossflow

region (no tracer particles to register a non-zero velocity).
Therefore, Eq. (2) can be simplified to

(3)
Q j ( x, t ) Vx ( x , t )dA Vx , j ( x , t )dA
A j ( x ,t )

A( x )

where A(x,t) is the total area (as captured by the SPIV).

Subsequently, we consider time-averaged volume flowrate
of the jet. Since the velocity fields data obtained from SPIV are
finite, we can further simplified Eq. (3) into summation form,

Q j ( x, t ) V x , j ( x, t )dA V x , j ,ij A
(4)
A( x )

ij

where V x, j ,ij is the time-averaged streamwise velocity of the

jet obtained from SPIV at location (i,j) and A is area of each
velocity element, which is constant in this experiment.
Finally, the effectiveness of the use of the azimuthal

2103499 2556 TF01

control jets in controlling JICF is defined by
E
cJICF
E JICF

(5)

where EcJICF and EJICF are the entrainments of the controlled

JICF and the baseline uncontrolled JICF, respectively.

Fig. 2. Schematic diagram of the experiment.

3. Experiment
3.1 Experimental setup
The experiment is conducted in the Fluid Mechanic
Research Laboratory (FMRL), Department of Mechanical
Engineering, Faculty of Engineering, Chulalongkorn University.
The setup is shown in Fig. 2. Briefly, the crossflow is
generated in the wind tunnel with 50x50 cm2 cross sectional
area and 240 cm long. The main jet and control jets
configuration is shown in Fig. 3. The main jet with diameter (d)
of 22.5 cm is surrounded by the azimuthal control jets with
diameter (dcj) 1 mm, which are spaced circumferentially by 15
degrees apart. To ensure the fully-developed turbulent pipe
flow velocity profile at the jet exit, the length of straight pipe
leading to the jet exit is
44d. The six-jet atomizer is
used to seed tracer particles
which are 5% by volume of
glycerin/water solution.
Stereoscopic Particle
Image Velocimetry (SPIV)
system from TSITM is used
to measure all three
components of velocity field
(Vx, Vx, Vz) in the cross
planes. The SPIV system is
composed of laser, light arm
and light sheet optics,
Fig. 3. The main jet and control
camera
and
laser
jets configuration, Kornsri et al. [10].

synchronizer. The laser is New Wave Research Nd:YAG laser

(Solo 200XT), which generates nominal energy of 200 mJ at
532 nm wavelength. The light arm and light sheet optics are
used to guide the laser beam and deliver the laser sheet to the
test section. Two identical CCD cameras (PowerView
Plus11MP) are used to capture the images of the flow field.
For the imaging lens, Tokina macro 100 mm f/2.8D lenses are
used in all cases, except the case I135 rm=4% at x/rd = 1.5
where Nikkor 50 mm f/1.8D lenses are used instead because
of the larger field of view. The laser, the cameras, and the data
acquisition are synchronized by the synchronizer. Six-jet
atomizer is used to seed the tracer particles into the main jet.
For each case and plane, 4,000 consecutive image pairs are
captured by each camera. The images are processed by TSITM
Insight4G program to to give 4,000 instants of the threecomponent velocity field. The velocity data are further analyzed
with in-house Matlab programs.
3.2 Experimental condition
The experiment is conducted at the main jet area-averaged
velocity Vj,0 of 17.4 0.2 m/s with the fully-developed turbulent
pipe flow velocity profile at the jet exit. The crossflow velocity
Vcf is 4.3 0.2 m/s. The effective velocity ratio r , defined by
2
2
r j V j ,0 cf Vcf is about 4.0 0.3, where j and cf
are jet and crossflow density, respectively, which are equal in
this experiment. In the case of controlled JICF (cJICF), a pair
of azimuthal control jets located at the azimuthal positions ()
of 135 (I135) are deployed to inject the control jets into the
main jet. The control jet to main jet mass flowrate ratio (rm)
which is defined by
control jets m
main jet
(6)
rm m
where m control jets is the total volume flowrate of the two control
jets and m main jet is the volume flowrate of the main jet. In this
experiment, two cases of rm are used, 2% and 4%, hereafter
referred to as case I135 rm=2% and case I135 rm=4%
respectively. Crossflow Reynolds number Recf Vcf d is
5,900, where d is the jet diameter. Velocity fields in cross plane
(yz plane) are then acquired by SPIV at four downstream
locations: x/rd = 0.5, 0.75, 1.0 and 1.5.
4. Results
4.1 Probability of finding the jet fluid at a point
Probability of finding the jet fluid at a point is defined by
ij Nv,ij N

where Nv,ij is the total time of finding the jet fluid (non-zero

108

(7)

2103499 2556 TF01

velocity) at point (i,j) and N is the total data acquisition time,
including those periods that jet fluid is not found. Figure (4)
shows that in all cases (JICF, I135 rm=2% and I135 rm=4%) ij
has high value in the center of the jet and decreases towards
zero at the jet boundary, reflecting the decreasing of the
probability in finding the jet fluid towards zero at the far edge of
the jet. Note that the white spaces in the center of the jet ij in
Fig. 4 correspond to ij > 0.95

another word, our time-mean jet velocity is modulated by the

probability of finding the jet fluid at that point. Hence, to
compare the results of the two works quantitatively, this
probability must be taken into account.
Figure (5) shows the distribution of time-mean jet velocity
Vx /Vcf and ij of the present result in comparison to ZF timemean field velocity Vx /Vcf , at x/rd = 0.5 and 1.0. The present
results reveal the kidney shaped structure similar to ZF. We
also find similar upper local peak as in ZF. Quantitatively,
overall our magnitude of velocity already taking the
probability into account - at x/rd = 0.5 and 1.0 are lower than
ZF about 20% and 3%, respectively. While quantitatively there
is some difference between the present result and ZF, the
overall structures are considered to be in good agreement.

Fig. 4. Distribution of probability of finding the jet fluid at a point.

4.2 Effect of rm on the jet structure

4.2.1 Verification of results
Zaman and Foss [1] (hereafter referred to as ZF)
conducted experiment under effective velocity ratio (r) of 4.58
and a jet with top-hat initial velocity profile. Before we compare
our results to ZF, an important point regarding the
interpretation of the two results must be made. Specifically, ZF
used hot-wire anemometer to measure the velocity and
presented the time-mean field velocity. As a result, their mean
velocity at any one point (i,j) is the time-mean velocity of the
field, regardless of whether at the instant the point (i,j) is
occupied by the jet extent or by the pure crossflow extent. As a
result, as we move towards the jet edge into the crossflow,
their mean velocity approaches the crossflow velocity. On the
other hand, the present result presents the time-mean jet
velocity the instantaneous jet velocity at point (i,j) is
considered zero when the point is occupied by the pure
crossflow extent. Consequently, towards the edge of the jet,
even though the instantaneous jet velocity can be as high as,
or even higher than, the crossflow velocity, the time-mean jet
velocity will be low and approach zero since the total time
period of finding the jet fluid near the jet edge to the total data
acquisition time approaches zero. In other words, even though
the jet velocity can be high near the jet edge, the event of
finding the jet extent near the jet edge is relatively rare when
compared to the event of finding pure crossflow fluid. Yet in

109

Fig. 5. Comparison of distribution of Vx /Vcf at x/rd = 0.5 and 1.0 between

Zaman and Foss [1] and present work.

To collaborate the present results further, they are

compared to those of Witayaprapakorn [2]. The two works
have similar experimental setup and conditions, and employ
the same SPIV technique in which only the jet fluid is seeded.
Figure (6) shows the structures of the distribution of Vx /Vcf in
both works, and the comparison shows good agreement. While
not shown, we find that the distributions of mean speed, mean
streamwise vorticity, and turbulent kinetic energy in our work
are also in good agreement with those of Witayaprapakorn [2]
both in JICF and cJICF cases.
We also repeat our experiment for JICF at x/rd = 0.75.
Both runs show good agreement as given as an example for
the structure of Vx /Vcf in Fig. 7.

2103499 2556 TF01

structures, which resemble a pair of comma-shaped structures.
On the other hand, those of I135 rm =4% are both elevated
and stretched in vertical position. However, both the control
jets and rm have no effect on the sense of rotation of vorticity
since the sense of rotation of all cases of cJICF remain the
same as the CVP in JICF.

Fig. 6. Comparison of distribution of Vx /Vcf at x/rd = 0.5 and 1.0

between Witayaprapakorn [2] and present work.

Fig. 9. Distribution of xd/Vcf .

Fig. 7. Comparison of distribution of Vx /Vcf at x/rd = 0.75 in case JICF
between (left) present experiment and (right) repeated experiment.

4.2.2 Effect of rm on the structure of streamwise velocity

Figure (8) shows the effects of rm on the structure of
streamwise velocity, Vx /Vcf. The results show that the
structures of I135 rm =2% is similar to JICF while those of
I135 rm =4% penetrates much deeper into the crossflow and
the velocity becomes more uniform when compared to JICF. In
all cases, the location of the peak value of Vx /Vcf is above the
center of the circulation of V yz / Vcf .

4.3 Entrainment corresponding to SPIV calibration

SPIV is calibrated against reference pitot tube. The
calibration curve, not shown here, has an equation of VPitot =
1.1VSPIV + 0.34. In corrected case, streamwise velocity fields
are corrected by this equation. Figure (10) show that value of E
from corrected Vx case, denoted as (1), is higher than value of
E from uncorrected Vx case, denoted as (2). However, there
are large amount of percentage of errors at low velocity due to
the constant term in extrapolated calibration curve equation
which cause the increase in Vx and in turn cause significant
increase in entrainment. Due to the uncertain of calibration
process mentioned above, all jet structures shown here are
from uncorrected Vx case i.e. calibration curve is not used.
Therefore, we choose to report our value of E with uncorrected
case which we have more confidence as a lower limit and with
the corrected case as an upper limit as shown on the Fig. 10.

Fig. 8. Structure of Vx /Vcf and in-plane vector Vyz /Vcf .

4.2.3 Effect of rm on the structure of vorticity

Figure (9) shows the normalized streamwise vorticity,
xd/Vcf. The results show that the magnitude of the normalized
streamwise vorticity decreases as the flow develops
downstream. This is consistent with the effect of friction in the
flow. The xd/Vcf structures of I135 rm =2% is similar to JICF

Fig. 10. Evolution of E in JICF , I135 rm =2% and 4% cases;

solid line: uncorrected (1), dashed line: corrected (2).

110

2103499 2556 TF01

4.4 Effect of rm on entrainment and effectiveness

Acknowledgement

The present results are compared with those of

Witayaprapakorn [2] as mentioned earlier in section 4.2.1. In
regard to the entrainment, present entrainment results differ
from Witayaprapakorn no more than 6% at x/rd = 0.5, 1.0 and
1.5, and 14% at x/rd = 0.75 for both JICF and I135 rm = 2%
cases. Therefore, two works show good agreement. In
addition, the entrainment was compared with the repeated
experiment. And both runs show good agreement as the
difference is less than 5%.
Fig. 10 show that in all cases the jet entrainment from
uncorrected Vx case (solid line) increases as the jet develops
downstream. In case I135 rm=2%, the jet entrainment slightly
increases from JICF, but not larger than 7%. On the other
hand, I135 rm=4% gives considerably high entrainment
compared to JICF, 57% increases at x/rd =1.5. Similarly, the
results in Fig. 11 show that the effectiveness of case I135
rm=2% is about 1, while that of I135 rm=4% is higher than 1 at
every x/rd and reaches 1.5 at x/rd =1.5.

We take this opportunity to express our profound gratitude

and deep regards to our advisor, Assc.Prof. Asi Bunyajitradulya
for his guidance, constant monitoring and encouragement
throughout this project. Assistance provided by Mr. Apichet
Srimekharat and Mr. Taned Witayaprapakorn in teaching and
guiding are appreciated. Finally, we wish to thank Mr. Suphak
Dawyok and Mr. Kittikun Wongthongsiri for their helpfulness
and friendliness for all times.

Fig. 11. Evolution of .

5. Conclusion
In this research, the effects of control jets to main jet mass
flowrate ratio (rm) on the structure and entrainment of a jet in
crossflow are investigated. The Stereoscopic Particle Image
Velocimetry (SPIV) is used to measure velocity field on four
cross plane, x/rd = 0.5, 0.75, 1.0 and 1.5. In order to clearly
identify the jet extent and to differentiate it from the pure
crossflow fluid, the jet-fluid seeding only scheme is used for the
SPIV. The results show that the control jets to main jet mass
flowrate ratio does have effects on JICF structure and
volumetric entrainment. Specifically, it is found that the case
I135 at rm=4% causes the jet structures to penetrate much
deeper into the crossflow and cause significant change to the
vorticity structure. In addition, the case I135 at rm=4% gives
much higher entrainment than JICF, up to 57% increase at
x/rd = 1.5.

111

Reference
[1] Zaman, K. B. M. Q. and Foss, J. K. (1997). The effect of
vortex generators on a jet in a crossflow. Phys. Fluids, Vol.
9, pp. 106-114.
[2] Witayaprapakorn, T. (2013). Effects of azimuthal control
jets on the entrainment of a jet in crossflow. Master thesis.
Department of Mechanical Engineering, Faculty of
Engineering, Chulalongkorn University.
[3] Fric, T.F. & Roshko, A. (1994). Vortical structure in the
wake of a transverse jet. J. Fluid Mech. 279, 1-47.
[4] Smith, S. H. and Mungal, M. G. (1998). Mixing, structure
and scaling of the jet in crossflow, J. Fluid Mech., Vol. 357,
pp. 83-122.
[5] Yuan, L. L., Street, R. L., and Ferziger, J. H. (1999).
Large-eddy simulation of a round jet in crossflow, J. Fluid
Mech., Vol. 379, pp. 71-104.
[6] Yuan, L. L. and Street, R. L. (1998). Trajectory and
entrainment of a round jet in crossflow, Phys. Fluids, Vol.
10, No. 9, pp. 2323-2335.
[7] Bunyajitradulya, A. and Sathapornnanon, S. (2005).
Sensitivity to tab disturbance of the mean flow structure of
nonswirling jet and swirling jet in crossflow, Phys. Fluids
17, 045102.
[8] Wangjiraniran, W. and Bunyajitradulya, A. (2001).
Temperature distribution in non-zero circulation swirling jet
in crossflow, Proceedings of The Fifteenth Conference of
The Mechanical Engineering Network of Thailand,
November 28-30, 2001, Bangkok, Thailand, Vol. 1, pp.
TF104-TF116.
[9] M'Closkey, R. T., King, J. M., Cortelezzi, L., and
Karagozian, A. R. (2002). The actively controlled jet in
Crossflow, J. Fluid Mech., Vol.452, pp. 325-335.
[10] Kornsri, P., Pimpin, A., and Bunyajitradulya, A. (2009). A
scheme for the manipulation and control of a jet in
crossflow: The use of azimuthal control jets. The TwentyThird Conference of the Mechanical Engineering Network
of Thailand, November 4 7, 2009, Chiang Mai.

2103499 2556 TF02

(.)

Study and Design the Air Conditioning System of Shuttle Bus by using Ice Storage
5330038521 5330473821 5330475021
.. ()
1.
.

(condenser)
(compressor)

.

17
6.65



(Indirect System)
.


20
215 200
1

2.
.

Abstract
The objectives of this project are to study the air conditioning
system and examine the root causes of inefficient systems in Shuttle
buses. After the study and examination, we found out that the heat
load in the passenger room is about 17 kW but air conditioning
systems nowadays can only eliminate 6.65 kW of the heat. We then
designed a new system by using the Air Conditioning Ice Storage
System. Our group chose the system which uses water as a
refrigerant to convect the coolness from the ice and set the size of
the system based on the Shuttle bus. However, due to the limitation
of the budget and time, we decided to use a prototype system that
has a smaller ratio compared to the designed one. In conclusion,
using 215 Kilograms of ice which cost about 200 baht (for one hour
of driving), our prototype is able to eliminate the heat loads of about
20 Kw.

1 .
1

[1]

159 W

(%)
0.90

[1]

343 W

2.00

444 W
2887 W
7190 W
5961 W
16.98 kW

2.60
17.0
42.4
35.1
100

[1]
[2]
[2]
[3]

.
Vapor Compression Refrigeration
6.65kW

112

2.

2 Evaporator
10 kW Specification Heat Exchanger
Evaporator 3
1 kW 10
3 Heat Exchanger Evaporator [4]

2 3

KU/C3-780

1/2-1/4

Copper Rows
NO. Inches
2
3/8

Q(kW)
L (in.) H (in.)
11
11
1

215 L 21.5 L

18 m 1.8 m
3.3.

40
2


3.4
4
4

2
2

(Q)
20 kW*
(L)
334.5 kJ/kg
(t)
1
. (WxLxH) 0.47x2.2x5.6 m 3

ICE STORAGE

QE
6.65 kW
20 kW

288 V
215 kg

37 /1
200 /1

.
6.65 kW
.

17 kW

20 kW
200
.1 1
215 kg

HP.

3.2.
3.1 10

3 3

MODEL

(1)

(1)

215 kg

10 L 260 L

0.4 m
1.9 m 0.35 m

[1] Incropera, Introduction to Heat Transfer Edition 5, Wiley, 2007
[2] ASHRAE Handbook of Fundamentals SI Edition, 1989
[3] Wilbert F. Stoecker, Industrial Refrigeration Handbook,
McGraw-Hill, 1998
[4] URL: http://en.wikipedia.org/wiki/Latent_heat, access on 25/2/14
[5] URL: http://www.kulthorn.co.th/pdf/22.pdf, access on 24/2/14

3.


3.1.

113

2103499 2556 TF03

The analysis for thermal comfortand thermal performance of a room with the glass
window and a venetian blind
5330190121, 5330211721 5330511921
.. ( ) ..

4
0 , 45 -45
45
PMV

-45
0
, , , PMV,

Abstract
This report presents the analysis of the experimental work
for the thermal comfort and thermal performance of the glass
window and a venetian blind system.The purpose of the study
is to find the guidelinefor choosing the proper blind which can
reduce heat transmission into space and maintain a proper
thermal comfort condition.The study is based on clear glass
window and clear glass window and a venetian blind with
different slat angle(0, 45, -45 degrees). The result shows that
the glass window with blind of -45 degree slat anglegives the
lowest value of PMV. However, the thermal performance of 45
degrees blind system and 0 degree blind system are almost
the same
Keyword Glass Window,Venetian Blind, Thermal Comfort,
PMV, Thermal Performance
1.

60%

1.

2.


2.
(Predicted Mean Vote, PMV)

[-3,3]
PMV
Operative(

)
1,2 3
0.036M

PMV (0.303e
0.028) [(M W) 3.05 10 (5733 6.99(M W) - Pa )
5
0.42((M W) 58.15) 1.7 10
M(5867 Pa) 0.0014M(34 Ta)
(1)
- 4
4
4
- 3.96 10
fcl (Tcl Tmrt ) fcl hcl (Tcl Ta)

Tsmrto

114

To 4

hcg
p

(T o T a)

0.25

(2)

2103499 2556 TF03

18 va 0.55 Forced Convection (3)

3 ( To Ta ) 0.25 Free Convection

M = ,W/m2
W = , W/m2
Pa = , kPa
fcl =
Tmrt = , K
Ta = , K
Tcl = , K
hc =
, W/m2

=
Tsmrto=
operative temperature, K
To = , K
va = , m/s

p =
( = 0.97)
= S
tefan-Boltzmann,W/m2k4
Tmrt = 6 , K
hcg max of

(Glass
Heat gain)


q U (Tout Tin ) (SHGC) I total
(4)

3.

3.1
4

2.8 3.15
4.32
18
Chilled Water System

0.8 1 6
optic = 0.80,
= 0.08,
= 0.12, =
0.84
1
25.2 . 2 20.0 .
74 . optic
( ) = 0.71, = 0.83

q =
,W/m2
U =
,W/m2K
Tout = , K
Tin = , K
SHGC = Solar Heat Gain Coefficient
Itotal = , W/m2
Heat Balance
Room Heat Gain Wall convection and conduction
Heat Gain Infiltrati on load Power load
Glass Heat Gain

1
(5)

Room Heat Gain = , W

Wall Convection and Conduction Heat Gain =
6
, W
Infiltration load =

, W
Power load = , W
Glass Heat Gain=
, W

115

2103499 2556 TF03

3.2

1. Air Velocity Transducer :

Innova
MM0034 [2]
2. Air Humidity Transducer :
Innova MM0037 [2]
3. Air Temperature Transducer :
Innova
MM0038 [2]
4. Operative Temperature Transducer :

54 . 160 .
[2]
PMV
Thermal- comfort meter Innova 1221[2] software

()

()()

()
()
3
() Air Velocity Transducer
() Air Humidity Transducer
() Air Temperature Transducer
() Operative Temperature Transducer
() Thermal Comfort meter Innova 1221

(5)
Thermocouple Type J
0.5 C 4

Thermocouple Type J flow meter
- Room Heat

Gain (Data logger)

NI 9211

Pyranometer
Kipp&Zonen CM6B
4
(shadow ring) 1
FRONT
WALL

FLOOR

CEILING
RIGHT
WALL

LEFT
WALL

BACK
WALL

4 Thermocouple

5
-

6 Pyranometer
Kipp&Zonen CM6B
4.

4
0 , 45
-45

7

116

2103499 2556 TF03

7 Pyranometer

5.

Pyranometer2
()
Horizontal Pyranometer
Vertical
8() 8()

10.21%

16:30
45
9

18

()

()

()

()

()
8
() Horizontal Plane () Vertical Plane

()

-45

0

117

9()
() 0 () -45 () 45

2103499 2556 TF03

13 PMV 4
10

Operative 4 ( 11)

( 10)
Operative
Operative

45
Operative Operative transducer

PMV ( 13)

4 ( 12)
PMV
PMV
-45
PMV
13


Operative PMV
45
Operative
PMV

-

+
6

11 Operative 4
0.01 C

()

12 4
0.2C
118

2103499 2556 TF03

()

15 4
()

16
4

()

14

() 30 .. 2557 ( )
15:36 16:36
() 5 .. 2557 (
0 ) 16:13 17:13
() 13 .. 2557 (
45 ) 15:40 16:40
() 12 .. 2557 (
-45 ) 15:11 16:11

4

-45

15 16

119

6.

PMV
45

PMV



18

..
..

[1] .


.

[2] InnovaAirtech Instruments. 1997. Thermal Comfort

2103499 2556TF04

Design and Construction of Heat Pump Water Heater
5330489921, 5330492721, 5330476721
..
( )

4 [1]

2.5
45
25 35

(Double-piped Heat Exchanger)

3.5 kW 7

2.5
45 2 5
31

3.03

Abstract
Presently, production of hot water by heat pump is widely
popular, especially in hotel industry, because of its saving four
times more than other methods.
The objective of this engineering project is to design and
build heat pump to produce 2.5 liter per minute of hot water, at
45 degree Celsius, when the entering water temperature is 25
degree Celsius, and the ambient temperature is 35 degree
Celsius.
Condenser, which is the most important component in the
system, was selected to be a double-piped heat exchanger,

because of its simple in construction and good heat transfer.

From the calculation, it is found that the condenser has to be 7
meter long for 3.5 kW heating.
From the result of testing, it was found that at the water flow
rate of 2.5 liter per minute, the machine can deliver hot water
at 45 degree Celsius approximately, while the entering water
temperature is 25 degree Celsius, and the ambient
temperature is 35 degree Celsius. This result shows that the
performance of the system is a little bit higher than expected.
COP of the system was found to be about 3.03, which is quite
low. This low COP may come from the low COP of the
compressor and the low efficiency of the evaporator.
2.
2.1

2.1.1 ( Condenser)

2.1.2 ( Compressor)

2.1.3 ( Evaporator)

2.1.4 ( Expansion Valve)

5
55

2.834 kW
3.486 kW
0.652 kW
(pressure ratio)
3.725
2.2

120

2103499 2555TF04


[2]

[2] 1

1
Dobson Chato [2]

[2] 2

3.2

28
31
5

(COP) 6

6 (COP)

2

7


3.

3
4

3.1

3.1.1
3.1.2 2.5

3.1.3 1.2 1.8 3

121

5.

2.5 45
25
31

3.03

6.
[1]
,URL:http://www2.dede.go.th/km_berc/downloads/menu4/

//04%20

.pdf, access on 12/07/2013

[2] Weerawoot Arunwattana (2010). A Model for Improvement
of Water Heating Heat Exchanger Designs for Residential
Heat Pump Water Heaters. Unpublished Master's Thesis for
Master's Degree, Master of Engineering at Massey University,
Palmerston North, New Zealand
7.
...

2103499 2556 TF05

Taylor bubble
The investigation of Taylor bubble shape in various diameter of vertical pipe
5330243821, 5330257621 5330420521
..
( )

Slug flow 3

k model

( D p )


3 Froude Number (Fr), Reynolds Number
(Re) Etvs Number (Eo)
FrD

0.0089
0.0725

1
ReD

1 0.11Re

ReD

EoD

Abstract
Slug flows consist of three main components; i.e. Taylor
bubbles, liquid slugs between two bubbles, and falling films
around the bubble surface. Generally, characteristics of a slug
flow are affected by shapes of Taylor bubbles. The objective of
this work is to study the effect of pipe diameters on the shapes
of Taylor bubbles by using a numerical method. This work uses
k model as governing equations. The results show that
larger pipe diameters yield thicker bubble shapes due to
turbulence intensity.
1.

3
Slug flow
(Taylor bubble)
(Slug)

([1] [2])
Slug flow
Dumitrescu

Dumitrescus model potential
flow

0.33
D

l wt D p
l

41
1 1.96
EoD

4.63

(1)
(2)

l g gD 2p

(3)

(4)

wt FrD gD p

(1) (4)
FrD

1.96

0.0089

1 41

2
0.33

gD

g
p
Fr gD 3

p
l D
0.0725
1 0.11

3
l

l FrD gD p

4.63

(5)
FrD g , l , l
D p FrD
0.351
Dumitrecus model
2
25C 100 kPa =71.99x10-3 N/m, g =
1.18 kg/m3, l = 997 kg/m3 l = 0.891 10-3 Pa s.
(5) Dp 0.03 - 0.1 m
FrD 0.351
5
0.0300, 0.0475, 0.0650, 0.0825 0.1000 m
Dp 5 ReD


k model

CAFFA.EXE

Ferziger, J.H. and Peric [3]

122

2103499 2556 TF05

2.

(1/2)Dp 1
(6) (7)
1

wb2
l

Rb R p 1
2
2 gz
nose wb

l a1 1 ea2 z / D

(6)

(7)

Rb Rp
z znose z z
wb
wt (6) (7)
a1 a2

15Dp

2 RMSDP Grid search

3.

5Dp

Dp

Rb
3

1



(8)
1
V
(8)
p g pl
2 l

Rbs

n l

a1 a2


(Uniform)
Root-mean-square derivation

( RMSDP ) (9)
RMSDP
)2
nj 1 ( p p
j
nose
RMSD
p
n 1

( 2) Resolution
1 2

(9)

i
p nose
n

Grid search
( 2) RMSDP
(6)
1 2 RMSDP
1 2
Grid search RMSDP
pi

123

3


4.
FrD 0.351

ReD
Dp Dp ReD



Rb/Dp


[1] Sotiriadis, A.A. and Thorpe, R.B., Liquid Re-Circulation in
Turbulent Vertical Pipe Flow Behind a Cylindrical Bluff
Body and a Ventilated Cavity Attached to a Sparger,
Chemical Engineering Science, 60, 981 994 (2005)
[2] Van Baten, J. M. and Krishna, R. CFD Simulation of Mass
Transfer from Taylor Bubbles Rising in Circular Capillaries,
Chemical Engineering Science, 59 (2004), 2535-2545.
[3] Ferziger, J.H. and Peric, M. Computational Methods for
Fluid Dynamics 3rd edition. Springer. Germany (2002)
p.249

2103499 2556 TF06

Ice Maker Design and Production
5330427021, 5330506921 5330526421
..
( )

13.6

Dual Fluid


Evaporator
Heat Load

Dual Fluid, Evaporator, Heat Load
Abstract
The objective of this project is to design, create, and test
an ice maker of which the production rate is at 13.6 kilograms
per hour. For the design process, the project has started from
studying ice production cooling system, after that Dual Fluid
cooling system was chosen to apply in the project. According
to the experiment, the results have shown that the efficiency of
the Ice maker is lower than Design Specification due to an
error on the Evaporator plate as well as an error on the Heat
Load calculation. In conclusion, the Ice maker can produce ice,
but the quantity of the ice is lower than expected.
Keywords: Dual Fluid, Evaporator plate, Heat Load
1.

4
Evaporator


Evaporator
0 C
Evaporator
Dual Fluid [1]
Secondary fluid reservoir Refrigerant
Condenser Evaporator

3.
3.1 [2]
(1)

Qe = UA(LMTD)

Evaporator
Overall Heat Transfer Coefficient
A

LMTD
Evaporating
3.2 Wattelet convective boiling correlation [3]
Qe
U

Nu
Re
Xtt

Nusselt number [4]

Pr
Reynolds number [4]
F
Lockhart-Martinelli parameter [3]

(2)
Prandlt number [4]
F parameter [3]

4.
4.1
Design Specification P-h diagram 1
Condensing 50 C Evaporating -15 C
Superheat
5 C
Subcooling
10 C
QSGHX : 0.14 [kW]
5

T4 : 39.7 [C]

QC : 1.87 [kW]

T2 : 86.3 [C]

TC : 50.0 [C]

3
T3 : 86.3 [C]

T5 : 31.4 [C]

2.
3

W : 0.66 [kW]
m : 0.011 [kg/s]
QE : 1.26 [kW]

T7 : -10.0 [C]

TE : -15.0 [C]

X6 : 0.38 [kg/kg]

8
T8 : -9.0 [C]

1 P-h diagram

124

T1 : 5.6 [C]

2103499 2556 TF06

4.2 Evaporator
(1) (2) Evaporator 50 cm x 50 cm
Drawing 2

2 Drawing Evaporator
4.3 Expansion valve
Expansion valve Capillary tube
DanCap Capillary tube
0.064 2.06 m
4.4
1.Suction line ( Compressor )
Suction line
saturation
0.5-2 [2] DIRcalc 3/8
0.818 C 3
2.Liquid line ( Condenser )
Liquid line
1-2.5 m/s [2] DIRcalc
1/4 0.997 m/s

3 Suction line
5.
5.1
4a
5.2 Evaporator
Drawing
4b
5.3
4c

a. b. Evaporator c.
4

125

6.
1 Design Specification
Design
Parameter
Results
Specifications
Condensing

50 C

39 C

Evaporator

-15 C

-14 C

0 C

10 C

Power Input

680 W

790 W

2 mm

3 mm
(
)

1 min

45 sec

13.6 kg/hr.

1.5 kg/hr.

7.

Heat Load
Evaporator Cooling Capacity
Refrigeration Cycle
Refrigeration Cycle
Cooling Capacity Power
Input Compressor Design Specification

..



.. 2556

[1] Oak Ridge National Laboratory. Ice-Maker Heat Pump
Development : Final Report. Department of Energy: Union
Carbide Corporation, 1980.
[2] Wilbert F. Stoecker, Industrial Refrigeration Handbook,
McGraw-Hill, United states of America: R.R.Donnelly &
Sons Company, 1998
[3] M.K. Dobson, J.P. Wattelet, and J.C. Chato. Optimal Sizing
of Two-Phase Heat Exchangers, Mechanical & Industrial
Engineering Department, University of Illinois, 1993.
[4] .. .
.
, 2554

2103499 2556 MS01

Effects of crossflow Reynolds number on the entrainment of jet

and controlled jets in crossflow.
Kittikun Wongthongsiri 5330037921
Assoc. Prof. Asi Bunyajitradulya (Advisor)
Abstract
The effects of crossflow Reynolds number on the
entrainment of jet and controlled jets in crossflow are
investigated by Stereoscopic Particle Image Velocimetry
(SPIV). The experiment is conducted for a baseline jet in
crossflow (JICF) with effective velocity ratio (r ) of 4 0.3 . In
the cases of controlled jet in crossflow (cJICF), a pair of
azimuthal control jets at the azimuthal positions
135 (I135) is deployed with the control jet to main jet
mass flow rate ratio (rm ) of 2%. The crossflow Reynolds
numbers (Re cf ) of 5,900 (L) and 12,500 (H) are experimented
and the measurements are made on the cross planes located
at x/rd = 0.5, 0.75, 1.0, 1.5. In order to clearly identify the
extent of the jet excluding the pure crossflow region - only
the main jet fluid is seeded with tracer particles. The result
shows that Reynolds number has little effect on peak velocity
magnitude in near field and has almost no effect in the far
field. As for the effect of Re cf on entrainment (E ) , it is found
that the increase in Re cf results in the increase in entrainment
in the near field but the decrease in entrainment in the far field,
for both JICF and cJICF cases.
Keywords: Jet in crossflow, azimuthal control jet, Crossflow
Reynolds number, Entrainment
1. Introduction
Jet in crossflow (JICF) is the jet that exits normally into a
stream of crossflow. As the jet interacts with the crossflow,
various flow structures result. This in turn affects the
characteristics of the jet such as jet trajectory, entrainment, and
mixing. JICF has a wide range of applications such as in
mixing of air and fuel in a combustion chamber, chemical
reactor, aquarium oxygen machine, trajectory and dispersion of
smoke from smoke stacks, film cooling on turbine blades, and
in V/STOL airplanes, etc. Thus, knowledge and understanding
of characteristics of JICF such as its structures are important to
the development and improvement of these engineering
equipments.

Past works on JICF can be roughly grouped into two

aspects: on jet characteristics and structures, and on how to
manipulate and control JICF. In regard to structure, Smith and
Mungal [1] found that, although counter-rotating vortex pairs
(CVP) is the main mechanism for entrainment of JICF in the far
field, it does not result in entrainment enhancement over a free
jet. Instead, it is the formation of the CVP in the near field that
results in entrainment enhancement over that of a free jet.
Yuan and Street [2] found that the jet entrainment has close
relation with jet trajectory, especially in the far field where x >
0.8rd; namely, as the jet entrainment increases, its trajectory
becomes lower. Furthermore the bending mechanism of JICF
is very complex as it depends on both pressure drag and
Reynolds number. Muppidi and Mahesh [3] found that the
deformation of jet and formation of CVP are resulted from jet
acceleration from crossflow in streamwise axis. Moreover
development of jet depends on both Reynolds number and
effective velocity ratio. Yuan et al. [4] found that the CVP is
formed from hanging vortices, which in turn has its origin in the
skewed mixing layers at the lateral edges of the jet.
As for the manipulation and control of JICF, Zaman and
Foss [5], and Bunyajitradulya and Sathapornnanon [6] used
delta tab. The use of tab is simple, uses small amount of
energy, but cannot be adapted for active control easily.
Niederhaus et al. [7], Wangjiraniran and Bunyajitradulya [8],
Bunyajitradulya and Sathapornnanon [6], and Denev et al. [9]
used swirling jet. The use of swirling jet has significant effect
on structure causing it to be asymmetric, but has a little effect
on entrainment. Eroglu and Breidenthal [10], MCloskey et al.,
[11], Narayanan et al., [12] used pulsing jet. It is found that the
pulsing jet causes JICF to penetrate more deeply into the
crossflow. However, both the uses of swirling and pulsing jets
require high amount of energy. Kornsri et al. [13] and
Witayaprapakorn et al. [14] used azimuthal control jets. The
use of azimuthal control jets is considered here to be more
suitable technique for controlling JICF so far since it consumes
less amount of driving energy and can be adapted for active
control easily.

126

2103499 2556 MS01

In thus regard, there are few researches that focus on the
effect of Reynolds number on entrainment of JICF directly and
none on the effect of Reynolds number on entrainment of
cJICF. Therefore this research will focus on these issues.
2. Experimental setup
This laboratory is located at Fluid Mechanics Research
Laboratory (FMRL), Department of Mechanical Engineering,
Faculty of Engineering, Chulalongkorn University. This
experiment setup consists of 2 parts. One is research rig.
Second is SPIV. The experiment setup and SPIV configuration
has shown in fig 1.

Fig. 1 The experiment setup and SPIV configuration.

2.1. Research rig
Wind tunnel has been use to provide crossflow. It has
square section 50 x 50 cm and 240 cm long.
Main jet pipe has 22.5 mm in diameters and 44D straight
long pipes to ensure fully developed pipe flow.

Azimuthal control jets configuration has shown in fig 2

which is a disk with holes inside that act as tube around
circumference. Each hole has diameter 3 mm which are
spaced circumferentially by 15 apart, install perpendiculars
to main jet pipe.
Tracer particle inject to main jet pipe by Six-jet Atomizer
(TSITm model 9306A). The tracer particles are 5% glycerol
solution.
2.2. SPIV
This experiment use Stereoscopic Particle Image
Velocimetry (SPIV) from TSI. The coordinated is shown in fig
2, where origin of coordinate is at center of main jet pipe, x is
streamwise axis, z is spanwise, y is tranverse axis. Laser is
New Wave Research Nd:YAG laser (model Solo 200XT), which
generates nominal energy of 200 mJ at 532 nm. Light arm
(model 610015) and light sheet optics (model 610021-SIL, -25
mm cylindrical and +500 mm spherical). Two CCD cameras
(PowerView Plus 11MP, model 630062). CCD resolution is
4008 x 2672 pixels, Pixel size 9 m x 9 m , intensity
dynamic range 12 bits. Each camera has Tokina Macro 100
mm f2.8D as image lens. Laser, cameras, and computer are
synchronized with a synchronizer (model 610035). The
acquisition is made through TSITm Insight 4G software, with the
acquisition rate of 2.07 Hz and the total of 2,000 velocity fields
(2 x 2,000 particle images for each camera). In processing for
velocity data from pairs of particle images, the initial
interrogation area of 64 pixels x 64 pixels is used with overlap
of 50%, resulting in the final interrogation area of 32 pixels x
32 pixels. This parameter will lead to different spatial resolution
in each case.

a)
2.3 Experiment conditions
The experiment conditions are shown in table 1. The
experiment has baseline as JICF-L case, Re cf of 5,900 with
no control jets. Whereas high Reynolds number case is
Re cf of 12,500. In case of azimuthal control jets are deployed,
they deploy at 135 (I135) of both Re cf of 5,900 (I135L) and 12,500 (I135-H) with azimuthal control jets to main jet
mass flow rate ratio of 2%. In all case, the jet initial velocity
profile is fully-developed turbulent. The experiments is conduct
at 4 planes that are x/rd = 0.5, 0.75, 1.0, 1.5.

b)

Fig. 2 Configuration of azimuthally control jets

a) Top view b) Side view

127

2103499 2556 MS01

Table 1 Experiment condition
Condition
T
P
R
rm
Seeding
For Recf
Vj
Vcf
P(Seeding)
No. nozzle
For Recf
Vj
Vcf
P(Seeding)
No. nozzle
Spatial resolution
x/rd = 0.5
x/rd = 0.75
x/rd = 1.0
x/rd = 1.5

From section 2, SPIV will generate 3D-velocity field by

detect tracer particle so this will lead to 2 options. One is
seeding only main jet and the other is seeding both main jet
and crossflow. From fig 3 (a) show that in case of seeding only
main jet, SPIV will detect only jet which clearly isolate jet
structure from crossflow. While fig 3 (b) shows that in case of
seeding both main jet and crossflow SPIV will detect both
velocity which we cant separate between jet and crossflow.
Since this research considers entrainment and structure of jet,
seeding only main jet was chosen.
When only jet fluid is seeding, SPIV will generate velocity
field from region of jet only, but in crossflow region SPIV will
get 0 velocities. This function can be written as

V ( x , t ) 0 when x is in the jet region

(3)

V ( x, t ) V ( x, t )

Values
28 C
101.3 kPa
4 0.3
2%
Glycerol 5%
5900
17.4 0.2
4.3 0.2
10
1
12,500
37.0 0.2
8.9 0.2
15
2

3. Entrainment
Entrainment ( E : Volumetric entrainment at any single
plane) can be define as
Qj
E
(1)
Q0
Where Q 0 is volume flow rate of initial main jet and Q j is
volume flow rate of jet in any cross plane that define as

Q j ( x, t ) Vx ( x , t ) dA
(2)
A ( x,t )

Where Vx ( x, t ) is streamwise velocity field and A j ( x, t ) is

area of jet at corresponding plane.

b)

when x is in the crossflow region

Q j ( x, t )

1.068 x 1.068 mm
1.198 x 1.198 mm2
1.32 x 1.32 mm2
1.35 x 1.35 mm2

a)

Hence equation 2 can substitution from area of jet

entire area of SPIV Ax , PIV as

SPIV

( x , t ) dA

(4)
Considering entrainment in time mean (Q j ) , equation 4
turn out to be
T
1
x, j

A( x )

Q j ( x)

Q j ( x)

Q (t ) dt
j

T
1
T

V
0

A( x )

x, j

( x , t ) dA dt

1T

T V x , j ( x , t ) dt dA
A( x )
0

Vx, j ( x ) dA

(5)
Subsequently, we consider time-averaged volume flow rate
of the jet. Since the velocity fields data obtained from SPIV are
finite, so integral form in equation 5 is inappropriate. Therefore
equation 5 can be further simplified into summation form.

Q j ( x) V x , j ( x ) dA Vij Aij
(6)
ij
A( x )
Where Vij is velocity at pixel i, j and Aij is area of resolution
at corresponding with Vij . Since areas Aij are equal in entire
plane so we will substitution Aij with A . Thus equation 6 can
be rewrite as
Q j ( x) Vij A
(7)
ij
From equation 7, where V x, j ,ij is the time-averaged
streamwise velocity of the jet obtained from SPIV at location
i, j and A is area of each velocity element, which is
constant in this experiment.
Q
(8)
E
Q
Effectiveness can be defines as
E

(9)
E
Where E Other case is entrainment on other case (JICF-H, I135-L,
I135-H). So this effectiveness describes how much modification
cases affect entrainment compare to baseline case.
A( x )

Other case

Fig. 3 Sketch picture of result from seeding

b) Seeding only main jet
c) Seeding both main jet and crossflow

A j ( x, t ) to

JICF L

128

2103499 2556 MS01

4. Result
4.1 Verification result with Zaman and Foss [5]
Zaman and Foss [5] (hereafter referred to as ZF)
experiment has some distinctive from this research. First, they
have r 4.58 while this research has r 4 . Second, they
measured time mean field velocity field that has velocity in
every point and every time in the interval of experiment
because they measured both jet and crossflow velocity. While
this research, by only jet fluid is seeded, SPIV will detect only
jet velocity result in at instantaneous time some point will has 0
velocities. Thus jet probability of finding the jet fluid at the point
is presented. To apply this feature, multiply velocity from ZF
with probability will result in this research domain velocity
called time mean jet velocity.

agreement.

Fig. 5 Distribution of velocity between present research and

Witayaprapakorn [14]

Fig. 6 Distribution of streamwise velocity between present

research and Witayaprapakorn [14]

Fig. 4 Distribution of streamwise velocity between present

research and Zaman and Foss [5]
Consider structure, Fig 4 shows that this research has
resembled the same shape as ZF. Consider magnitude of
velocity, at the upper local peak when multiply ZF velocity with
probability, this research has magnitude lower than ZF at x/rd
= 0.5 and x/rd = 1.0 by 20% and 3% respectively. At bottom
local peak, both researches have magnitude of velocity
converge to 0.
4.2 Verification result with Witayaprapakorn [14]
This research and Witayaprapakorn [14] are conducted at
the same laboratory and parameter, so this will check whether
experiment are consistent.
Fig. 5 and 6, show that in case of JICF-L and I135-L both
distributions and streamwise velocity have almost the same
shape and magnitude. Therefore both researches have good

129

4.3 Probability of finding the jet fluid at the point.

Since this experiment is turbulent jet, so jet will have
different shape at any instantaneous. Thus, it have to find
probability of finding the jet fluid at the point to identify where is
high or low chance to find a jet. Probability of finding the jet
fluid at the point can be define as
N v ,ij
ij
(10)
N
Where N v,ij is number of pictures that has vector velocity at
point i, j . N is total number of pictures that taken.
Fig, 7 shows that all case JICF-L, JICF-H, I135-L and I135H have the same result. At plane x/rd = 0.5 area of high
probability of finding the jet has found at the middle of jet and
decrease in magnitude when expand from center of jet until
have nearly 0% probability at boundary of jet. When jet
developed to downstream, area of high probability still occurs
in the middle of jet but have a expand area with lower
probability. Moreover its shape will be more round, wider and
higher as jet developed.

2103499 2556 MS01

4.4 Effect of azimuthal control jets and Re cf on the jet
structure
Figure 8 shows the structures of JICF and cJICF for all
cases: JICF-L, JICF-H, I135-L and I135-H. Overall, the
structures are similar in all cases, signifying that both the
azimuthal control jets and Re cf has little effects on the jet
structure. Nonetheless, increase in Re cf at x/rd = 0.5 causes
increase in the peak of streamwise velocity. When the jet
develops further downstream into the far field, however, the
effect of Re cf disappears; all cases have similar structures and
peak.
4.5 Effect of azimuthal control jets and
entrainment

Re cf

on

Fig. 7 Distribution of Probability of finding

the jet fluid at the point.
Fig. 9 Effect of

and azimuthal control jets

on Entrainment
Re cf

Fig. 10 Effect of Re cf and azimuthal control jets

on Effectiveness

Fig. 8 Distribution of normalized streamwise

velocity Vx Vx / Vcf

Figures 9 and 10 show the jet entrainment and the

effectiveness of the use of azimuthal control jets at low and
high Re cf . At x/rd = 0.5, increase in Re cf result in increase in
entrainment and effectiveness. However, as the jets develop
further downstream, increase in Re cf result in decrease in
entrainment and effectiveness, for both cases of JICF and
cJICF. Nonetheless, as is clearly seen, the use of the

130

2103499 2556 MS01

azimuthal control jets in case I135 result in more entrainment
and effectiveness than JICF at their corresponding Re cf .
5. Conclusion
SPIV is used to study effects of crossflow Reynolds
number on the entrainment of jet and controlled jets in
crossflow. The experiment is conducted for JICF with the
effective velocity ratio (r ) of 4 0.3 . In the case of controlled
JICF, a pairs of azimuthal control jets are deployed at
135 (I135) with the control jets to main jet mass flow
rate ratio (rm ) of 2%. The crossflow Reynolds numbers
(Recf ) experimented are 5,900 (L) and 12,500 (H). In order to
identify the jet extent clearly, and hence determine the jet
volume flow rate and entrainment more effectively, only the jet
fluid and not the crossflow fluid is seeded with tracer
particles.
The result shows that Reynolds number has little effect on
the jet structure and the peak magnitude of the streamwise
velocity at x/rd = 0.5 but has almost no effect in the far field,
for both cases of JICF and cJICF. For entrainment (E) and
effectiveness ( ) of the use of azimuthal control jets, it is
found that the increase in Re cf results in the increase in
entrainment and effectiveness only in the near field at x/rd =
0.5, which is the result from the increase in the peak of the
magnitude of streamwise velocity. However, as the jet
develops further in the far field, the increase in Re cf results in
the decreases in entrainment and effectiveness.
6. Acknowledgement
I would like to thank you Assoc. Prof. Asi Bunyajitradulya
(Advisor), for his valuable time to give me advice, knowledge,
analyses and problem solving skill. Thank you to Mr. Apichet
Srimekharat, Mr. Taned Witayaprapakorn graduate researchers
for their dedicated time to guide me. Also thank you Mr.Settasit
Chaikasetsin, Mr. Teerapat Sushewakhul, Mr. Piyapol
Panusittikorn, Mr. Suphak dawyok senior researchers in my
laboratory for their friendship and cooperated for solving
problem in laboratory. Finally I would like to thank you my
family for their understanding and encouraging me.
References
[1] Smith, S. H. and Mungal, M. G. (1998). Mixing, structure
and scaling of the jet in crossflow, J. Fluid Mech., Vol. 357,
pp. 83 - 122.
[2] Yuan, L. L. and Street, R. L. (1998). Trajectory and
entrainment of a round jet in crossflow, Phys. Fluids, Vol.
10, No. 9, pp. 2323-2335.

131

[3] Suman Muppidi and Krishnan Mahesh (2006). Twodimensional model problem to explain counter-rotating
vortex pair formation in a transverse jet, Phys. Fluids 18,
085103 (2006).
[4] Yuan, L. L., Street, R. L., and Ferziger, J. H. (1999).
Large-eddy simulation of a round jet in crossflow, J. Fluid
Mech., Vol. 379, pp. 71 - 104.
[5] Zaman, K. B. M. Q. and Foss, J. K. (1997). The effect of
vortex generators on a jet in a crossflow. Phys. Fluids, Vol.
9, pp. 106-114.
[6] Bunyajitradulya, A. and Sathapornnanon, S. (2005).
Sensitivity to tab disturbance of the mean flow structure of
nonswirling jet and swirling jet in crossflow, Phys. Fluids
17, 045102.
[7] Niederhaus, C. E., Champagne, F. H., and Jacobs, J. W.
(1997). Scalar transport in a swirling transverse jet. AIAA
J., Vol. 35, No. 11, pp. 1697-1704.
[8] Wangjiraniran, W. and Bunyajitradulya, A. (2001).
Temperature distribution in non-zero circulation swirling jet
in crossflow. Proceedings of The Fifteenth Conference of
The Mechanical Engineering Network of Thailand,
November 28-30, 2001, Bangkok, Thailand, Vol. 1, pp.
TF104-TF116.
[9] Denev, J. A., Frhlich, J., and Bockhorn, H. (2009). Large
eddy simulation of a swirling transverse jet into a crossflow
with investigation of scalar transport. Phys. Fluids 21,
015101.
[10] Eroglu, A. and Breidenthal, R. E. (2001). Structure,
penetration, and mixing of pulsed jets in crossflow. AIAA
J., Vol. 39, No. 3, pp. 417-423.
[11] MCloskey, R. T., King, J. M., Cortelezzi, L., and
Karagozian, A. R. (2002). The actively controlled jet in
crossflow. J. Fluid Mech., Vol. 452, pp. 325-335.
[12] Narayanan, S., Barooah, P., and Cohen, J.M. Fluid
dynamics of highly pitched and yawed Jets in Crosssflow.
AIAA Journal Vol. 41 (December 2003) : 2316- 2330.
[13] Kornsri, P., Pimpin, A., and Bunyajitradulya A. (2009). A
scheme for the manipulation and control of a jet in
crossflow: The use of azimuthal control jets. The TwentyThird Conference of the Mechanical Engineering Network
of Thailand, November 4 7, 2009, Chiang Mai.
[14] Witayaprapakorn, T. (2013). Effects of azimuthal control
jets on the entrainment of a jet in crossflow, Master thesis,
Department of Mechanical Engineering, Faculty of
Engineering, Chulalongkorn University

2103499 2557 MS02

POD
Effects of the azimuthal control jets to main jet mass flow rate ratio
on the flow structures of jet in crossflow via POD
5330505221
.

POD
Stereoscopic Particle Image
Velocimetry (SPIV)


40.3
5,900
fully-developed turbulent pipe flow
135o rm = 2% 4%


near field far field
CVP

rm

downstream
rm = 2%
far field
, ,
proper orthogonal decomposition (POD),
Abstract
The effects of the azimuthal control jets to main jet mass
flow rate ratio on the most energetic flow structures of jet in
crossflow are investigated by proper orthogonal decomposition
(POD) analysis. The velocity field in the cross planes is
measured by Stereoscopic Particle Image Velocimetry (SPIV).

In order to identify the jet structures only and clearly, excluding

the pure crossflow fluid, only the main jet fluid and not the
crossflow fluid - is seeded by tracer particles. The experiment
is conducted at the effective velocity ratio r of 40.3, crossflow
Reynolds number of 5,900, and the jet initial velocity profile of
fully-developed turbulent pipe flow. The results show that the
effect of using the azimuthal control jets at position = 135o
and the control jets to main jet mass flowrate ratio rm of 2%
and 4% changes the most energetic turbulent structure in both
the near and far fields to CVP structure and makes dominant
structure more stable. In addition, the azimuthal control jets
has an effect in promoting the most dominant structures as can
be seen from the transformation of the energy distribution
among modes from broad and low-profile in case of JICF to
narrow and high-profile in case of cJICF. Furthermore, as rm
increases, this promotion effect lasts longer in the downstream
direction as can be seen from the change in the energy
distribution profile of case rm of 2% approaching that of JICF,
while that of case rm of 4% remains relatively narrow and high
profile, as the jet develops downstream.
Keywords jet in crossflow, azimuthal control jets, proper
orthogonal decomposition (POD), dominant structures
1.
(Jet in Crossflow : JICF)


(Entrainment)

132

2103499 2552 MS02

Laser sheet
Crossflow
y

Crossflow

Control jets

Jet

1 [8]
film cooling

(Trajectory)




S.H. Smith and M.G. Mungal [1]
CVP (Counter rotating vortex pair)
far field JICF Free jet
CVP near field JICF
Free jet
rd Yaun et al. [2] LargeEddy Simulation (LES) CVP Hanging vortices
Skewed mixing layer

1

Delta Tab
Vortex (Zaman and Foss [3]
Bunyajitradulya and Sathapornnanon [4]) pulsing
(MCloskey et al. [5]) swirling jet (Bunyajitradulya and
Sathapornnanon [4]) Delta tab
active control pulsing
swirling jet Kornsri et al.
[6] (Azimuthal
control jets) active control

15o

Witayaprapakorn [7]

135o

15o

133

x
ccd
Cameras

Rotameter
Seeding particle
Six-jet Atomizer,
glycerol water solution 5% by volume

Main jet

(most energetic turbulent structures, turbulent kinetic energy)
POD (Proper Orthogonal Decomposition)

Meyer et al. [8] POD JICF
(effective velocity ratio, r) 3
Wake vortices
r = 1.3 Jetshear layer vortices
Kornsri et al. [6]

(rm)

(most energetic turbulent structures, turbulent kinetic energy)

POD

2.




2

50x50 240

2103499 2557 MS02

(a) Top view

Insight 4G
Spatial resolution
1.07x1.07 x/rd = 0.5 2.07x2.07
x/rd = 1.5
POD
(Vj)
17.40.2 / (Vcf)
4.30.2 /
( r V / V ) 40.3
(Recf) 5,900
fully-developed turbulent
pipe flow

2 x
streamwise ( ) y
transverse () z spanwise
() SPIV
( yz) x/rd = 0.5, 0.75,
1.0 1.5

2
(x/rd)
(JICF) 135o rm
= 2% (I135-rm2) 135o rm = 4%
(I135-rm4) rm
j

(b) Side view

3 (Witayaprapakorn [7])
(d)
22.5
44d
fully-developed turbulent pipe flow (
Witayaprapakorn [7] 3)
1
135o
rotameter Cole-Parmer (model TW03217-30)
Stereoscopic Particle Image Velocimetry (SPIV)
Six-jet Atomizer (TSITM model 9306A)

glycerol water solution 5%

New WaveTM (model Solo 200XT) 200 /
532
(laser sheet) 2 glycerol water solution

CCD (Powerview Plus 11 MP, model
630062) Tokina 100
mm f2.8D Macro
135o rm = 4% x/rd = 1.5
Nikor 50 mm f1.8D Synchronizer (model
610035) ,

2.07 Hz
4,000
4,000 TSITM

2
j

cf

rm

2
cf

m cj
m j

m cj

m j

3. POD

POD

(Mode)
(Energy) Mode


POD SPIV
mode 0
POD
(fluctuation)
(u Nj , v Nj , w Nj ) u, v w
j N

U

134

2103499 2552 MS02

u1N

N
uj

v1N

v Nj

w1N

w Nj

u12

u 2j
v12

2
j
2
1

w2j

(1)

JICF

I135-rm2

I135-rm4

4.43 %

6.42 %

6.49 %

Vx/Vcf

u11

u1j
1
v1
1 2
N
U [u u u ]

v1j
1
w1

1
w j
~
C UTU

(2)

eigenvalue eigenvector
~
CAi i Ai

(3)
eigenvalue
mode

Mode 1

autocovariance

Mode 2

eigenvector (3) POD

Mode
N

n 1

i
n

3.75 %

Au
n 1

i
n

3.36 %

JICF

I135-rm2

I135-rm4

4.22 %

4.90 %

5.96 %

4.16 %

4.23 %
() x/rd = 1.5

2.95 %

Vx/Vcf

POD Modes eigenvalue
eigenvalue eigenvector
Mode
Mode Mode

3.53 %
() x/rd = 0.75

4.

SPIV

POD

4.1
(POD mode 0)
4
135o rm = 2% (I135-rm2) x/rd = 0.75

spanwise
(JICF) rm = 4% (I135-rm4)
Wake

downstream x/rd = 1.5

135

Mode 1

Au

Mode 2

4 POD Mode 1 2
Contour

Vx / Vcf V yz / Vcf
() x/rd = 0.75 () x/rd = 1.5

2103499 2557 MS02

() x/rd = 0.75

() x/rd = 1.5
5 mode
4.2

x/rd = 0.75 mode 1
lobe 2 lobe lobe
lobe mode 2
CVP
downstream x/rd = 1.5 mode 1
Jet-shear layer mode 2 4
lobe
135o rm = 2% mode 1
CVP
mode 2 lobe 2 lobe
lobe
lobe downstream
x/rd = 1.5 CVP
lobe lobe
mode 2
Jet-shear layer
135o rm = 4% mode 1
rm = 2%
CVP lobe 2 lobe

6 mode
Wake mode 2
Jet-shear layer downstream
x/rd = 1.5 mode 1
CVP Wake mode 2
Jet-shear layer Wake


x/rd (x/rd = 0.5 x/rd = 1.0

) 135o rm 2%
4%
downstream

4.3

5 mode
x/rd
135o
rm = 2% 4%
mode
far field
x/rd = 1.5
135o rm = 2%

far field rm
mode
near field
6
mode
() mode
50% mode
5%

136

2103499 2552 MS02

5.


POD

Stereoscopic Particle Image Velocimetry
(SPIV)

40.3
5,900
fully-developed turbulent pipe flow

135o rm = 2% 4%
135o

rm = 2% 4%

(POD mode 1)
near field far field CVP

mode

far field rm

near field
6.

.
, , ,
, ,
,

137

7.
[1] Smith, S. H. and Mungal, M. G. (1998). Mixing, structure
and scaling of the jet in crossflow, J. Fluid Mech., Vol.
357, pp. 83-122.
[2] Yuan, L. L., Street, R. L., and Ferziger, J. H. (1999).
Large-eddy simulation of a round jet in crossflow, J. Fluid
Mech., Vol. 379, pp. 71-104.
[3] Zaman, K. B. M. Q. and Foss, J. K. (1997). The effect of
vortex generators on a jet in a crossflow. Phys. Fluids,
Vol. 9, pp. 106-114.
[4] MCloskey, R. T., King, J. M., Cortelezzi, L., and
Karagozian, A. R. (2002). The actively controlled jet in
crossflow. J. Fluid Mech., Vol. 452, pp. 325-335.
[5] Bunyajitradulya, A. and Sathapornnanon, S. (2005).
Sensitivity to tab disturbance of the mean flow structure of
nonswirling jet and swirling jet in crossflow, Phys. Fluids
17, 045102.
[6] Kornsri, P., Pimpin, A., and Bunyajitradulya A. (2009). A
scheme for the manipulation and control of a jet in
crossflow: The use of azimuthal control jets, The 23rd
Conference of the Mechanical Engineering Network of
Thailand, November 47, 2009, Chiang Mai.
[7] Witayaprapakorn, T. (2013). Effects of azimuthal control
jets on the entrainment of a jet in crossflow, Master thesis.
Department of Mechanical Engineering, Faculty of
Engineering, Chulalongkorn University.
[8] Meyer, K.E., Pedersen, J. M., and zcan, O. (2007). A
turbulent jet in crossflow analysed with proper orthogonal
decomposition. J. Fluid. Mech., vol. 583, pp. 199-227.

2103499 2556 MS03

A sorting of micro particle with microfluidics system

5330507521
..
( )

1-3 [1]




2 1
Pillar Structure
5 10
2




Separation technique, Pillar Structure, Magnetic
Separation, Microfluidics, Malaria
Abstract
Nowadays the spreading of malaria disease is still a
problem of the people who lives in tropical climate countries.
This disease can kill the malaria-infected patients about 1-3
millions of people each year around the world. Sorting of
malaria disease at present needs use the specialist to decide
that owner of sample is infected or not, which is the cause of
delaying in separation the people infected from the community
for preventing the spreading of disease. In this senior project,
we study and design the microfluidics system device. This
project is divided into 2 parts. The first part is studying and
designing the sorting device based on size with pillar structures
and doing the experiment sorting the particles with size of 5
and 10 m. The other part is studying and designing the
separating device using the magnetic forces induced by with
magnetic field. There is the calculation of magnetic force
exerting on the particle and controlling the direction of particle
motion path for designing the device
Keywords: Separation technique, Pillar Structure, Magnetic
Separation, Microfluidics, Malaria

1. Introduction
According to many sources of articles indicating the
technique used microfluidics system for sorting the particle, we
have reviewed each different methods and summarized in
Table 1 that show the main characteristic of each technique.
Table 1 showing the concept of each separation technique. [2]
Method

Separation
Markers
FACS
Fluorescence
Fluorescent labels
Magnetic
Magnetic field
Magnetic
Susceptibility
Dielectrophoresis Inhomogenous electric Polarizability
field
Electrophoresis Homogenous electric Charge density
field
Optical
Optical force
Refractive index
and polarizability
Acoustic
Ultrasonic wave
Density
and
compressibility
Pillar Structure
Laminar flow
Size
PFF
Hydrodynamic
force Size
(parabolic
velocity
profile)
Hydrodynamic
Hydrodynamic force
Size
filtration
Inertial
Shear-induced and wall- Size and shape
induced lift
Surface affinity
Specific binding to Surface
surface markers
biomarkers
Biomimetic
Hydrodynamic force / Size
Fahraeus effect
This project aims to design a sorting system of malariainfected red blood cell using microfluidics system. The whole
system is shown in Fig. 1. It is consisted of pillar structure
which is used to sort based on size. The other part is that can
separate the magnetic particles from non-magnetic particles
using magnetic force.

138

Mechanism

2103499 2556 MS03

Figure 2 displaying the motion of the both different size of

particles.
microposts lie systematically in microchannel and using the
nature of laminar flow that is the characteristic of flow in the
microchannel. Each row of micropost is shifted laterally and the
distance between 2 pillars in row is as equally as in the pillar
between 2 rows as in Fig. 2. Because of fluid flow which
moves pass micropost, The particles which flow along the fluid
are driven by drag force as a result of the particles are moving
along streamline. The green particles that are small flow in
streamline that can pass through. However, the red particles
that are bigger than green particles cannot pass through the
Figure 1 Schematic of separation system in this project. Sorting the array of microposts and then they will move laterally that have
particles based on size and then separate with the distance called lateral displacement which is the key
parameter of sorting by size. The motion pattern of the green
magnetophoresis
particles is called zigzag mode and the motion pattern of the
In principle, we might able to increase the efficiency of
red particles is called bump mode.
malaria separation by sorting blood cell based on size first as
the parasite known as the cause of disease live in red blood
2.1.2 Design
cell of human by eating hemoglobin in blood cell and change it
The cut-off size at which the transition from zigzag mode to
into the magnetic substance called Hemozoin. It is possible to
bump mode occurs is critical diameter, Dc, and corresponds to
separate infected blood cells using magnetic force. The size of
twice the width of the first streamline. The relationship can
Hemozoin is proportional to the maturity of parasite living in
written as
.
red blood cell. In the first part, we will test the pillar structure
= .
(1)
for it with polystyrene particles mimicking size of red blood cell
Where Dc is critical diameter, G is the gap between the
is 5, 10 m.
surface of two pillars and N is shift row. Equation 1 shows the
In the second part, the study of separation with magnetic
critical diameter is a function of the period of the array, as has
field induced by current passing through 2 electrodes is
been shown, and the distance between posts, the gap (G in
performed. The scope of this second part is to design
Fig. 3). In the experiment condition, we would like to separate
dimensions of the system and show the result by analytical
the particle size 5 m and 10 m; therefore, we have to find
method.
the appropriate gap in order to get the critical diameter that the
parameter indicating motion of particles. There are four pillar
sections. Gap is shown in Table 2 and post dimensions. From
2.1 Pillar Structure
the designed device in Fig. 4, we would like to manipulate the
2.1.1 Literature review
Another name of this method is Deterministic Lateral
sample solution to flow into the expected position in the
Displacement (DLD), which consists of a number of symmetric
system. Therefore, the system is designed with 3 inlet ports.
The sample solution flows into middle port and the buffer
Table 2 showing calculation of the Equation 1 for designing device
solution flows into others on the right and left. Moreover, in the
Gap between pillar (G) (m)
Critical Diameter (m)
design, one of the inlet port has the width is 5 times of other
15
6.95
ports so that the flow rate in this bigger port is 5 times of the
20
9.27
other for maintaining fluid velocity passing through the system
25
11.59
as desired. The channel height is around 50 m.
30
13.91

139

2103499 2556 MS03

Figure 3 Calculation data on the transition between different types

of motion of particle for different normalized particle sizes
as a function of the inverse period of array of the device.
The lines represent the theoretical values for empirically
deduced value used for constructing the devices used
herein.[3]

Instrument:
1. Pump: Before pumping, we have to input 2 values in
order to pump accurately that are inner diameter of syringe
and the flow rate in whatever unit that each pump can supply.
2. Syringe: The inner diameter is measured with vernier
caliper which has resolution 0.02 mm.
3. Camera by Dino-Lite AM331 which connect with
computer that can record video or capture image while running
system.
4. Sample that is well-prepared from experiment will be
mentioned in next section.
5. Microfluidics device are fabricated by TMEC (Thai
Microelectronic Center)
6. Computer: Record the video while doing the experiment.
Sample Preparation is done with the following procedure
1. Buffer solution: The ratio of surfactant (Tween20) over
DI water 1:2 by volume
2. Sample: Dilute particles solution with buffer solution by
1:100 by volume

Figure 4 Left the designed chip sorting particles Right the

expanded image of device for checking the structural size
of pillar by microscope.
2.1.3 Experiment
Our setup is shown in Fig. 5. We have calibrated our
pumps and the results are shown in Table 3. Details of each
instrument are explained as follows.
Table 3 Data of measurement and working range of pump.
Inner
Flow rate
Flow rate
No. Diameter
boundary
boundary
(Cole-Parmer Model 100) (NE-1600/NE-1800)
(mm)
0.1 uL/m
0.007 uL/m
1
4.37
31.6 uL/m
605.5 uL/m
0.1 uL/m
0.0319 uL/m
3
9.38
146.1 uL/m
2.553 mL/m
0.001 mL/m 0.0552 uL/m
5
12.64
0.264 mL/m
4.422 mL/m
0.001 mL/m 0.0845 uL/m
10
15.20
0.383 mL/m
6.777 mL/m
0.001 mL/m 0.1377 uL/m
20
19.80
0.650 mL/m
11.02 mL/m

Experiment is done with the following procedure

1. Use the syringe No.3 (Maximum volume 3 mL) 3 items
pull the prepared solution. Two syringes are buffer solution and
the other syringe is mixed solution.
2. Connect all the syringes with connector and then
connect with the device. In the outlet, micron diameter line
connects to microcentrifuge tube for counting later.
3. Begin flushing by driving all syringes for 20 minutes
4. Recording the video in the inlet area, the middle channel
and the outlet area.
5. After flushing, we will store the sample in each outlet to
microcentrifuge tube for 20 minutes.
6. Pouring collected sample onto glass slide and then
inspect with microscope and then count the particle in the
area.

Buffer
Sample
Buffer
Figure 5 Installation of all instruments.
2.1.4 Result
According to the design, the device has 5 outlets;
therefore, we drain about 1 L from each outlet and drop it on

140

2103499 2556 MS03

Figure 6 Distribution of the number of particles 5 m to

different outlet.

Figure 9 Example of measuring size of pillar in middle area of

device. Expected size is 125 m but average size is
122.487 m.

Bubble
Figure 7 Distribution of the number of particles 10 m to
different outlet.

Figure 10 Bubbles obstruct particles moving through the device.

Figure 8 Distribution of the number of mixed particles 5 m

and 10 m to different outlet.
the glass slide. We count the number of the particle with using
the VIS software in different area of glass slide and then
average them. The results are shown in Figs. 6-8.
In the first condition, feeding only 5 m, most 5 m
particles flow out at outlet 2 about 71.59% and in second
condition, feeding only 10 m, most 10 m particles flow out
at outlet 2 about 56.86% as well as feeding mixed 2 sizes of
particles most particles also flow out at outlet 2 with 5 m of
53.05% and 10 m of 47.37%.
2.1.5 Discussion
The results are unexpected and the efficiency of separation
is not good enough from experiments. We have considered all
problems while doing experiment.
1. Size of pillar is not exist like design size. By measuring
size of all pillars with microscope, we find out that the
measured size is smaller than designed size as shown in Fig.
9. It causes the gap between pillars which is the function of
critical diameter that is the parameter separate type of motion
of particles and also all particles have a same type of motion
that is zigzag motion. Therefore, we cannot separate the
particle between 5 and 10 m from each other.
2. In microchannel device, there are a lot of pillars. It

141

Group of formed particles

Figure 11 The image captured by microscope displaying the

formation of particle.
causes the device to have so much pressure drop and then
bubbles occurring while installing instrument in experiment
cannot be pushed away from device. Furthermore, the
experiment condition use very low flow rate (5 L/min) that
does not have enough force to push the bubbles away and
then result in blocking the particle flowing along streamline as
shown in Fig. 10. Thereby, the particles will not move following
expectation.
3. The problem of formation of group of particles because
of the attractive force results in the bigger size of particle
flowing through the device. It causes the flowing particles to
clog in device or to change behavior into bigger particles. From
observation while doing experiment, when we release the
mixed solution at room temperature, it can be formed a group
of particles between near particles as shown in Fig. 11.

2103499 2556 MS03

2.2 Magnetic Separation
2.2.1 Literature Review
Magnetic Force
Magnetic force occurs on paramagnetic object which is in
inhomogeneous magnetic field. The force can be found the
magnitude of magnetic force as
=2

Figure 13 Left cross-section of device magnetic sorting

indicates dimension of both electrodes and
microchannel. Right schematic shown separation of
each electrode for computing total magnetic field
any expected points.

(2)

Where Fm is magnetic force (N), medium is permeability of

medium (N/A2), particle is permeability of particle (N/A2) and a
is diameter of object and H is magnetic field (T). Considering
the Equation (2), the magnitude of force is proportional to size
of object and gradient magnetic field. Magnetic field used in
this senior project is created by current, which describes it
later.
Magnetic field can be found by calculating from current,
which is explained with Biot-Savart law. The equation can be
written as
=

Figure 14 Left Distribution of magnetic field in x-axis varying

height Right Distribution of magnetic force in x-axis
varying height at z=0, z=25 and z=50 um

(3)
where dB is magnetic field , 0 is permeability of medium =
8.85 10-12 , I is current, dl is the direction vector of current, lr
is the direction vector from point of current to the interested
point and R is the displacement from point of current to the
interested point.
We can define the conductor wire is infinite length which is
the length of wire is greater than cross-section area of wire
that cause the magnetic field in the parallel direction of wire is
eliminated. Therefore, in this case, we consider the wire into 2
dimensions as shown in Fig. 12.

Figure 15 Right Free body diagram of particle flowing in

microchannel. Left Prediction of motion that
magnetic force exerting on the particle.
Fig. 14 and use the range of suitable amount of force for
exerting on the particles.

Figure 12 In case infinite length, we can consider magnetic

field in 2-Dimension for calculating easier.
This project finds the magnetic field occurring in any areas
is come from 2 electrodes which each electrode is 50
micrometers in width, 80 micrometers in height and 25
millimeter in length. Calculating magnetic field can be more
apparent by separating the electrode into sub unit cell and then
cumulating the magnetic field from each sub unit cell that can
give a better solution. We choose the Microsoft Excel to help in
calculating where we have to cope with a lot of information. In
calculation, the electrode is divided into 40 pieces and then
vary distance both 2 dimensions for calculating magnetic field
and magnetic force later as shown in Fig. 13. We can get the
distribution of magnetic field and magnetic force as shown in

2.2.2 Design
Equation of motion and consider the force acting on the
particles while flowing in channel as Fig. 15, all forces in z-axis
are neglected due to their small magnitude comparing to the
lateral magnetic force in y-axis, so we can define the 2nd law of
motion in y-axis as
6

(4)

From the calculation data of magnetic force, we can find

the magnetic force equation that is the function of z-axis and yaxis at any points of microchannel but for simplicity we ignore
the result of z-axis. We choose the lowest magnetic force for

142

2103499 2556 MS03

design at z = 50 m which is the further height from electrode
and x from 250 m to 350 m by using the MATLAB with
cftool
We will get the second order differential equation and it
apparently cannot solve with direct method. We can use the
numerical method with fourth-order Runge-Kutta method in
order to solve this equation for exposing time of particles in
magnetic field. The solution can indicate the position of
particles at each time step. We can gather the data to predict
motion in Fig. 16.
We define the lateral distance (in y-axis) 50 m, the
experiment particle size of 15 m (Expected range 5-15 m),
mass of particles calculating from density and volume of
particle, the velocity of fluid is 100 m/s and fluid flowing in
channel is water (=8.9410-4 Pa.s).
According to computing with MATLAB, we will get time
calculating in y-axis which the particle move laterally
approximately 50 m so that we can approximate the time in
order to get the minimum length in x-axis of device
We choose the safety factor is 2 for flow uncertainty and
the initial distance for fully develop velocity profiles. Therefore
we select the minimum distance for separating the magnetic
particles from the main stream is 20 mm and also indicate the
dimension of device in Fig. 17.

Figure 16 The motion path of 15 m particles solved by fourthorder Runge-Kutta method by MATLAB.
3. Conclusion
We review the separation technique using microfluidic
system and then we choose two techniques that fit in this
project. In the pillar structure section, we succeed in designing
the device and fabricating the device. However, the device
cannot separate the particles size 5 and 10 m. Dimensions of
channel change from the parameter defined as designed
equation of separation might be a problem. In magnetic
separation section, magnetic separation using the program for
helping find magnetic force and solving differential equation for
particle motion which can solve with numerical method by
fourth-order Runge-Kutta method. The analysis provides the
minimum length of the device.

143

Figure 17 Schematic of magnetic separation design.

Acknowledgement
This project is financially supported by STAR research fund
of Chulalongkorn University under Micro-Nano Fabrication
Technology Research (GSTAR 56-005-21-002), and research
fund of Department of Mechanical Engineering, Chulalongkorn
University. In addition, the author would like to thank Dr.
Werayut Srituravanich and Dr. Angkee Sripakagorn (Faculty of
Engineering, Chulalongkorn University), Dr.Wutthinan
Jeamsaksiri and staffs of TMEC, Dr. Achariya Sailasuta,
Dr.Prapruddee Piyawiriyakul, Dr. Dettachai Ketpun and
Mr.Anusorn Kanungthong (Faculty of Veterinary Science,
Chulalongkorn University for their great assistance.
References
[1] , , ,
,

. 13,25 (
2552): 1-13.
2] Ali Asgar S. Bhagat, Hansen Bow, Han Wei Hou, Swee Jin
Tan, Joogyoon Han, Chwee Teck Lim, Microfluidics for
cell separation Med Biol Eng Comput 48,(2010): 9991014.
[3] Stefan H.Holm. Microfluidic Cell Sorting and Analysis.
Masters Thesis, Department of Physics, Faculty of
Science, Lund University, 2012.

2103499 2556 MS04

Gear Design for Reducing Power loss
5330048821
.. ( )

1.

Abstract
This paper presents the design of gear for reducing power
loss while using but the gear can support torque load as well
as normal gear. In design step, we set the conditions of torque
and speed, and then design the gear sets with difference
geometric parameters for support the load conditions. By
analyzing effect of geometric parameters to sliding loss and
windage loss, we get guidelines of choosing parameters for
reducing loss gear design. We found when decrease the size
of module and increase the length of face width, the sliding
loss will decrease. In addition, in range of low to medium
speed windage loss is lower than sliding loss greatly. For
examining which gear sets make the lowest loss, we use
mathematic-model-program developed by masters in
laboratory. By these step above, it possible to design gear with
low power loss.

2.

144

2103499 2556 MS04

3.

Kv, Kf Ns
1. Kv


Kv

3.1
3.1.1

Lewis
W

W = SF Y m

3+V
V

Kv =

(1)

5.56 + V
(4)
5.56
V = (m/s)
2. Kf (actual stress
concentration factor) 1.2-1.7
1.4-2.0

Kf =
1.5

Kf
3. Ns

Ns
1.0-2.0
Ns = 1

2

SF Y m
T=
r
(5)
K v K f Ns
r
Kv =

S = (1/3 )
F = ()
Y = Lewis ( 1)
m = ()

S45C (tensile strength) 569 MPa

1 Y
Lewis
1 Y Lewis

3.2
3.2.1
[4] [5]

3.1.2
[1]

Lewis

W=

SF Y m
K v K f Ns

(3)

(2)

145

2103499 2556 MS04

4

3.2.2

Handschuh Dudleys Gear

Book Design [2]

2
2 O1 O2 v1
O1
O1 v2 O2
O2 v1 v2

v1 v2
v2



3

Pwindage

15 N D 5L D
=

+
0.746 1000 100 100 100

(6)

Pwindage =

15 N D

0.746 1000 100

5L
Rf
D
+ (7)

tan 100
100

P = (HP)
N = ()
D = ()
L = ()
=
Rf 1

3 1

146

2103499 2556 MS04

1 Transverse diametral
pitch transverse module Rf
Transverse diametral pitch
4
6
10
16
24

Transverse module
6.35
4.23
2.54
1.59
1.06

3
4

Rf
7.2
6.7
6.1
5
3.8

3.3

(-)

()

120

900

80

1200

150

1500

90
1:1 5

25

11-12

3

25
Y
11-12

147

2103499 2556 MS04

5

(-)

()

( )

(.)

()

(.)

25

36

2.50

27.50

121.67

25

40

2.25

15

24.75

118.23

25

50

1.80

30

21.60

117.20

25

45

2.00

45

24.00

123.75

25

45

2.00

24.00

78.00

25

50

1.80

15

21.60

83.32

25

60

1.50

30

17.25

81.66

25

55

1.64

45

18.86

83.15

25

30

3.00

36.00

149.24

25

33

2.70

15

29.70

153.42

25

35

2.57

30

30.84

148.18

25

32

2.80

45

33.60

153.21

()

()

190.84

0.0015

159.30

0.0102

136.73

0.0056

146.75

0.0048

126.32

0.0033

105.31

0.0188

92.61

0.0105

98.36

0.0091

347.79

0.0085

294.61

0.0586

281.58

0.0417

300.00

0.0355

148

2103499 2556 MS04

5000

3

6 6


1 2 4
1
3 5
3

()
250
200
150

100

..

50

0
0 1

2 13

6


7
5000

30000
2 15
1

[1]
.. ;
2, 2523
[2] Dennis P. Townsend, Dudleys gear handbook, 2535
[3] C. Yenti, S. Phongsupasamit and C. Ratanasumawong,
Analytical and Experimental Investigation of Parameters
Affecting Sliding Loss in a Spur Gear Pair, Engineering
Journal, Vol.17, No.1, (2013), p.79-93.

()
400

300

[4]
C. Ratanasumawong, P. Asawapichayachot, S.
Phongsupasamit, H. Houjoh, and S. Matsumura, Estimation of
Sliding Loss in a Parallel-Axis Gear Pair, Journal of Advanced
Mechanical Design, Systems, and Manufacturing, Vol.6,
No.1, (2012), 88-103.

200

4
100

0
0

20000

()

7
1

149

2103499 2556 MS05

The Study of Load Behavior of Spur Gear by

Using Mathematical Models and Finite Elements Method

5330217521

.. ( )

2
1. 2.

2

ANSYS

Von-mises

,
,

Abstract
This research aims to study load behavior of spur gear
teeth in any meshing position. This paper offers two different
ways of study which are creating a simple mathematical model
and analyzing by Finite Element Method. For creating a simple
mathematical model , we utilize Hertzian Contact Stress
Theory which explains the stresses on the surfaces of the
touching areas of two curvature objects which were pressed
together by the external forces. The study is conducted by
simulating the meshing together of surfaces of gears in each
position which is similar to the compression behavior of two
different sizes of cylinders. In analyzing by Finite Element
Method , we apply ANSYS software by creating spur gear

model and select the material properties and appropriate

boundary conditions. The results inform us that the calculated
Von-Mises stress from both methods has a similar pattern and
its values will change according to the position of the meshing
gear from the start till the end. The meshing position of two
pair teeth has less stress than the meshing position of just one
pair. However, the magnitude of the stresses from two
methods still has a significant difference which may be the
result of the division of the element is not enough.
Keywords: Load behavior of spur gear , analyzing by simple
mathematical model , analyzing by Finite Element Method
1.

1.

2. 1

2





1

1.

2.
Finite Element

150

2103499 2556 MS05

2.



2
,
, ,
2 P
1 R1 R1
R2 R2 2
2

Hertzian Contact Stress

Theory
tooth
profile involute
(face width)
crowning 4

4 Hertzian Contact Stress Theory

R1 R2


5 R1 R2 Crowning

1 2

2 2
2 P


2
3

5 R1 R2
3

151

2103499 2556 MS05

Crowning Hertzian Contact
Stress Theory
2

3
b
k

(1)

b cb 3 P

(2)

1 v12 1 v 22

E1
E2

a
b
P
cb
k

E1
1
E2
2
v1
1
v 2 2

A B

(3)

7 2


Crowning Crowning R1 R2
1/R1 1/R2 0

= 0 (4) (5)
B

A , B B/A
2

1
1 1
1
1
1 1 1
1 1
1
1 1
1 2

4
A

sin
4 R1 R2 R '1 R'2 4 R1 R'1 R2 R'2
R1 R'1 R2 R'2

1
1 1
1
1
1 1 1
1 1
1
1 1
1 2

sin

4 R1 R2 R '1 R' 2 4 R1 R'1 R2 R ' 2
R1 R'1 R2 R' 2

R1
R2
R1 Crowning
R2 Crowning

R1 R2
cb k 6

(4)
(5)

1 1
1

2 R1 R2

(9)

A0

B/A
k = 0 (1) k = 0 a
2a

(2) k = 0
b

2w

(10)

w
(3)

xx

yy 2v

zz

(11)

Von Mises Stress

0.5 x y 2 y z 2 z x 2 3 xy2 yz2 zx2

(12)
(11) (12)
0 Von Mises Stress

0.5 xx yy 2 yy zx 2 zz xx 2 (13)

6 cb k

152

2103499 2556 MS05

3.
MATLAB

Crowning Hertzian
Contact Stress Theory

9 MATLAB

1 2 2
pitch 1
pitch 2


2
1 2 (mc)

,, ,
,,
R1 ,
R2 Von
Mises Stress

Von Mises Stress
10

mc

AB
Pb

(14)

AB
Pb
1:1

1
()
30

3
()
20
(mm)
20
(N-m)
200

0.3
(GPa)
200

10 Von Mises Stress

MATLAB
Von Mises Stress
9

153

2103499 2556 MS05

4. Finite Element
Method
Finite Element Method
ANSYS
quasi-static

GearTrax2008

Rough

Frictional -


No

Seperation

15
11 Model GearTrax2008
(Assembly)
1-2
CATIA

4.2
ANSYS mesh
Sizing - Element
Contact Sizing - Element

Refinement - Element

Contact Sizing Element
( ) 1 mm
28.0740.2623.24 mm3

14 Model

Model Finite Element Method

ANSYS Static Structural


4.1
ANSYS

Bonded

No Separation

Frictionless

16 Element

4.3
ANSYS Supports
Fixed Support -

Displacement -

154

2103499 2556 MS05

Frictionless Support -

Frictionless Support

4.4
ANSYS Loads
Pressure -
Force -
Moment-
Moment 200 N-m
Solution
Deformation
Strain , Stress Energy
Equivalent (Von-Mises) Stress run

Stress Stress
45
17

19 Von Mises Stress

Finite Element Method
5.
Von Mises



20

17 Von Mises Stress 45

18


20 Von Mises

...


MATLAB

18
Von Mises Stress


19

155

[1] Arthur P. Boresi , Richard J. Schmidt Advanced mechanics
of materials. 6th ed. United stated of America: John Wiley &
Sons, Inc. 2003

2103499 2556 MS06

Parametric Study on Fabric Duct System Design
5330171221
. ( )

3 ,















, , ,

Abstract
The purposes of this project are studying and making the
mathematical model of fabric duct for using results to design
and predict parameter distributions in the system. The
experiments are divided into 3 major parts. The first part is
testing flow through fabric duct. The second part is tensile test
on small piece of fabric duct material and the third is
calculation flow from mathematical model. Static Pressure and
velocity of open hole and angle of open hole trend to increase
along the length of duct. The factors that influence on pressure
drop in fabric duct are the ratio of the sum areas of all lateral
holes to duct area and flow velocity in the duct. Found that the
pressure drop significantly increases with increasing velocity
and aperture ratio. The material of fabric duct must have a
minimum of stress to make cylinder shape duct. Results show

that stress direct variation with static pressure at each position

on the duct. So, the method to design fabric duct has to
consider static pressure distribution along the duct or predict
flow distribution by mathematical model that has a result
calculation as like as the result of experiment. But the accuracy
of mathematical model is affected from the discharge
coefficient.
Keywords: Fabric Duct, Static pressure, Discharge Coefficient,
Mathematical model
1.

2.

1

()
( )
1 2

156

2103499 2556 MS06

3.3
3.3.1




(Discharge Coefficient ,Cd)

1 :
(1)

= =1 + =1
1

1 + 1 2 = 2 + 2 2 + =1 2 + =1 2
2

(2)

3.

2

DRD 3

Cd

3.3.2




5 Velocity Potential (Vp)

cos() =

3.1
-
-
- 2
-

-
-
-
-

- 2

3.2

= +
(3)

157

(5)

1 +2
2

= 2

2 :

(4)

()

[ =

(6)

3.4

3.4.1

m =
(7)
2
2

f
Moody Chart
Miller f
= 0.25

3.7

5.74

0.9

(8)

2103499 2556 MS06

3.4.2

K Local loss coefficient

ml =

(9)

3.4.3
1
3 (tee)
3

3
(Dividing Flow) 3

3 :
Gardel 3

(K)

(10)

32 = 0.03(1 )2 + 0.352 0.2(1 )

31 = 0.95(1 )2 + 2 1.3 0.3 +

0.9

1
2

0.4(

= ,

1+

1
3

0.40.1
2

(11)

1
3

4 :

(Elastic Deformation) (Hookes
law) E (Youngs modulus)
(modulus of elasticity)
(y)

stress-strain curve


0.2% plastic
strain stress-strain diagram
stress-strain diagram yield strength
3.5.2
(strain gage)


3.5.3

5

3.5



3.5.1


Stress-Strain Diagram
(tension test) 4

5 : Cylindrical Pressure Vessel

Hoop Stress

=

158

(12)

2103499 2556 MS06

3.5.4




3
- :

- :

- (Strain Gauge) :


Strain Gauge
3.6

(K)

= (, , , , , , 0 )

1
2
2

= , , , 0

(K)

= ,

Fix

6 : DRD
4.

3

( 7)
(parachute)

7 :
4.1



254 (10 )
5.3 [Aperture Ratio (A.R.)
]
4-1

(13)

3.7 (Mathematical model)

3.1 3.5

DRD


(Constraint)

159

1
2
3

[]
10
10
10

[]
5.3
5.3
5.3

[]
2.5
2.5
2.5

AR.

10
20
40

0.625
1.25
2.5

4.2

2103499 2556 MS06

3 1 3
7 20, 30, 40,50 ,60
,70 80 21

1.5,2.5 3.5

5.
5.1
8


3

Moody Chart

9 :
1 3 60 Hz

10 :
1 3 60 Hz
8 :

1 3

Moody Chart

5.2

9 10

5.3

0.2%Offset

172 MPa
20 MPa

11
10 28 Pa
0.23 MPa

(20MPa)
0.8 MPa
(Safety Factor) 25

11 :

160

2103499 2556 MS06

12 : () ()
5.4

2.5
Gardel

9 10

1.5 2.5
3.5
Cd
1.5 2.5
Cd
3.5
Cd

13


1.5 2.5 3.5
Cd

13 :
1.5,2.5 3.5

161

6.




1.




2.





3.

Discharge Coefficient

[1] , ,
2552.
[2] Donald S. Miller, Internal Flow Systems, BHRA Fluid
Engineering 1978.
[3] Ductsox.com, DuctSox Corporation, Available from:
www.ductsox.com
[4] Gardel, A., ChambresdEquilibre, Lausanne, Switzerland,
1956.
[5] Newman B G, A Hydrographic Solution for Flow leaving a
Manifold through a slit, Canadian Aeronautics and Space
Journal, 1989, 35:205-210
[6] Roger Gordon Kinsman, Outlet Discharge Coefficients of
Ventilation Ducts, Master thesis, Macdonald College of
Mcgill University, Montreal, 1990.

2103499 2556 MS07

Ejector
Design of Ejector Nozzle for Refrigeration Use
5330318221
. ( )

Ejector

Ejector

Evaporator 2
Ejector

Motive nozzle 0.4 mm Ejector
0.6 mm Ejector

Ejector
Expansion valve

Ejector

2.
Ejector
1

Abstract
Ejector nozzle is equipment that can improve a
refrigeration system performance, but refrigeration system
needed to be modified before use properly with ejector
nozzle. The modified refrigeration system has two
evaporators. After that, conservation laws are applied to
design ejector nozzle with ideal process assumption. Finally,
a motive nozzle with end size 0.4 mm and a ejector throat
size 0.6 mm making the ejector operates on the favorable
conditions.

1 Ejector
Ejector

1. Evaporator
2.

-
2

1.







Expansion valve

3. Ejector

- condensing temperature 50OC evaporating
temperature 5OC
- ejector 15 OC

162

2103499 2556 MS07

Ejector
15 OC
3
Motive nozzle, Suction chamber Diffuser

(
Isentropic )

2
--------- (4)
4. Ejector

Ejector
4

2 -
Ejector
- R-22
- 0.02
kg/s
Ejector Conservation of
mass, First law of thermodynamics Second law of
thermodynamics
Ejector
Ejector
--------- (1)
--------- (2)
--------- (3)
Second law of thermodynamics
- 3

Ejector

4 Ejector

.

163

2103499 2556 MS08

(Stress analysis of human teeth subjected to mechanical loads using finite element method )
5330011521
.. ( )

-
250 N
Fixed Support

(Tetraheral Element) Node
31386 - 15399
18.6 MPa

-

, ,
Abstract
The objective of this project is to determine the stress
analysis of human tooth with mechanical load simulation as
chewing activity, with finite element method. This study focus
on stress in regular-shape tooth react by 250 N force acting on
the top surface, which the tooth was fixed-supported at the
root.Analyze by using tetrahedral mode of finite element
method, containing 31,386 nodes and 15,399 elements. The
maximum stress is 18.9 MPa. The study not only provide the
dentist an important information but also helps them to be able
to understand and evaluate the tooth structure correctly.
Keywords: Finite element , Tooth , Stress

3
1.analytical methods


2. experimental techniques
[2]



3. computational methods
[3, 4] -

2.
1.
ANSYS


2. ANSYS

1 .

1.
-

1.

164

2103499 2556 MS08

3. ANSYS

ANSYS

Engineering data
Youngs modulus 200 GPa Mesh
(Rectangular element) 12,000
Node 60,447 Node 4.

cantilever

0.1 x 0.15 m2 6 m Youngs Modulus
200 GPa P1 P2 2,400 N
5,000 N
Fixed Support 2 .

4 . Mesh
Boundary condition
5,000N Y Line Pressure
50,000 N/m
2,400 N Z
Pressure 160 kPa
Fixed support 5 .

2 .

bending = My / I1
= (5000 N x 6 m)(0.075 m) / 0.000028125 m4
= 80 MPa
axial = P1 / A
= (2,400 N / 0.015 m2 )
= 0.16 MPa
X
80.16 MPa 79.84 MPa

3.

5 . Boundary
Solve
Z - X
2.
Exact Solution Fixed Support
Stress Z 114.37 MPa
113.79 MPa 6.

6 . Z
Fixed Support

3 . Geometry

165

2103499 2556 MS08

Fixed Support
- Fixed Support 3 m

bending = My / I1
= (5000 N x 3 m)(0.075 m) / 0.000028125 m4
= 40 MPa
axial = P1 / A
= (2,400 N / 0.015 m2 )
= 0.16 MPa
X
40.16 MPa 39.84 MPa
Z
40.014 MPa 39.22 MPa
7 .

7 .
Fixed Support 3 m
exact
solution numerical solution
40.16 MPa 40.014 MPa
0.36
4.
4.1
1 .
(Crown):
-


(Root):

(Enamel):

(Dentine):



(Pulp):


1. [5]
1.
Material

Youngs
modulus
(GPa)

Poissons
ratio

Enamel

80

0.3

Dentine

18.6

0.31

Pulp

0.0021

0.45

Periodontal ligament

0.0689

0.45

Bone tissue

12

0.3

4.2 Modeling
CT
scan
CT
scan .dcm Dicom file (Digital
Imaging and Communications in Medicine) 8.

8.
CT scans .dcm (DICOM)

2103499 2556 MS08

.stl STL
(STereoLithography)
9.

1 2. Mesh Creation
9. (.stl)


CATIA Digitized Shape
editor Generative Shape Design split
cloud Root Crown
10. 11.

Merge Cloud
Mesh Creation
13.

1 3. Mesh Creation
Solid
Close Surface 14.

10. Spilt Cloud Root

14. Solid

11. Spilt Cloud Crown

Mesh creation
12.

4.3


Crown
3.1 250 N [1]
-
4.4.
Mesh
Default 15.
15,399 Node 31,386

2103499 2556 MS08

Dentine 1. Youngs modulus
18.6 GPa Poissons ratio 0.31

15. Mesh
Boundary Condition 250 N

16.

18. Stress
Stress
17.
z
z 21.3 Mpa
x y 6.6 Mpa 7 Mpa

16. 250 N

Fixed Support
17.

17. Fixed support

5. Result
(Von-Mises Stress)
Crown Root 18.6
MPa 1 8.

19. X Y Z

2103499 2556 MS08

Z (Von-Mises stress)

20. Z

21.

..

20.

21. Z

6.

250 N 18.6 MPa
Crown
Root




7.

1 .

[1] Josipa Bori, Robert Antoni, Miranda Muhvi Urek, Nikola
Petrievi,Petra Nola-Fuchs, Amir ati and Ivica Smojver. 3-D
Stress Analysis in First Maxillary Premolar. Mechanical
Engineering and Naval Architecture, University of Zagreb
[2] Palamara, J.E; Palamara, D; Messer, H.H.: Strains in the
marginal ridge during occlusal loading,Australian Dental
[3] Toparli, M; Sasaki, S.: Finite element analysis of the
temperature and thermal stress in a postrestored tooth
[4] Roberto Sorrentino, Raffaella Aversa, Valeria Ferroc,
Tommaso Auriemma,Fernando Zarone, Marco Ferrari , Antonio
Apicella .Three-dimensional finite element analysis of strain
andstress distributions in endodontically treated maxillary
central incisors restored with diferent post, core and
crown materials. 2006
[5] Josipa Bori, Robert Antoni, Miranda Muhvi Urek, Nikola
Petrievi,Petra Nola-Fuchs, Amir ati and Ivica Smojver. 3-D
Stress Analysis in First Maxillary Premolar. Department of Oral
and Maxillofacial Surgery, School of Dental Medicine,
University of Rijeka, Rijeka, Croatia

2103499 2556 MS09

A Reconfigurable Open Architecture Controller for a Research Robot

Kitithorn Phanatamporn 5330043621
Assoc. Prof. Ratchatin Chancharoen, Ph.D. (Advisor)
Abstract
Industrial Robot plays important role in modern industry
where environment and cost are concerned. Its role becomes
increasing important towards the future where robots are
working closely with human in our living space. This presented
controller is Reconfigurable Open Architecture robot controller
that combines 3D printer software, i.e., Repetier and Slic3r with
the powerful Matlab to develop a full function robot controller.
To demonstrate the performance of the controller, it is used to
control a Delta robot. The results proof that this controller is
ready to operate in any new situation.
Keywords: Open Architecture, Reconfigurable
1. Introduction
An Industrial robot [1] plays an important role in modern
manufacturing facility since 1970 to present [2]. It is designed
with precision motion control capability that is effectively used
for transferring, inspection, and processing parts in industrial
plant. The controllers of these robots are closed architecture in
order to gain stability which work very well in industry.
Although the capability of an industrial robot is very interesting,
its working performance is only achieved in the environment
that suits the robot. This robot should not be used in nonindustrial domain such as in a hospital or an office where
working space contains unstructured dynamic environment and
sometimes human.
The design of a non-industrial robot is very challenging
since this robot requires sensors to sense object and/or
environment in real time and human-friendly interface devices.
This robot must be 1) safe for the environment, human and the
robot itself and 2) intelligent such that it can still work
effectively in dynamic environment. The commercial robot
controller is normally closed architecture which its capability is
limited [3] [4]. Moreover, the control technique should be
considered not only the motion but also force. If we have to
control both motion and force simultaneously, the controller is
going to be very complex. If we use the closed architecture
controller, it is going to be very expensive and use a lot of
effort to design and build.

This project is to develop a Reconfigurable Open

Architecture Controller [5] [6], the next generation controller, to
be the solution of mentioned problems [4] [7] [8] such that we
can utilize the devices and source code from several domains
to use with that robot to lower the cost of the controller.
Another thing, the product life time is very short, maybe only 2
years. So, the robot cant work at the same thing and same
place during the whole life of it. The robot should adapt to the
new product new work space and new facility. Therefore if we
design the reconfigured controller, once the product market is
already end we can reconfigure the robot to be ready to do the
new thing. Thats why we need reconfigurable open
architecture controller.
2. Features of the Reconfigurable Open Architecture
Controller
There are many robotics issues have been considered to
design this system such as processor, kinematics,
performance, control device, Mathematics, logic level, safety,
timing communication and community. The purposed controller
should handle these kinds of properties effectively.
2.1 Processor
This is one of the most important thing to control the robot.
It integrates and processes all devices to work together. In
robotics field, there are many involved processors such as
Window PC, Real-time PC (xPC) [9], Microcontroller and FPGA
[10]. The great suitable processors are needed in the
controller.
2.2 Control devices
One preferred property of this controller is that it should be
able to integrate with many modules and devices to improve
flexibility, reliability, reconfigurability [11] and quality of work as
well. If we design the open controller system, we can use the
input devices from any company from several domains. We
can find the best and reasonable cost one to fit with our tasks
and the market will be bigger, the whole system cost will be
lower.

166

2103499 2556 MS09

2.3 Robotic Mathematics
The robotic control has much involving Mathematics, so
required property is reconfigurable Mathematics. Normally, the
robot controller must be able to handle the robot Mathematics,
including forward and inverse kinematic, Jacobian [1] and the
low level PID controller which must be solved in real-time as a
minimum.
2.4 Logic level [12]
The use of 5-V Vcc has long been the standard for both
core and memory logic. However, with the increase in
complexity and the functionality of application-specific
integrated circuits (ASICs), central processing units (CPUs),
microprocessors, and digital signal processors (DSPs), it has
become necessary to reduce the structure size of these
elements. Making electronic devices more complex, without
enlarging the overall size of the chip area, requires reducing
the structure size, which also requires reducing Vcc. Moreover,
power consumption always is a concern for new system
designs. A reduction in supply voltage produces an exponential
decrease in power consumption.
2.5 Safety
Nowadays, the robot becomes close to human more and
more. Safety has been firstly considered. In the other way,
after the code has been written, it has to test with the robot
many times for debugging. To avoid the serious damage to the
robot and optimize the hardware, safety technique will be
proposed in the next section.
2.6 Timing
The accepted speed to control the robot is 1 kHz or higher
[13] [14] for great performance and safe. Additionally, there are
many devices working together which each device has different
characteristics. Digital camera works at 30 Hz, the force sensor
might work at 100 Hz but the system has to run at 1 kHz.
Therefore, the controller should handle these servoing rates.
2.7 Community
Due to the complex system and sometimes the time is
limited. Solving the accidentally met problem alone wastes
much time and could make the project delay or even fail. Open
source codes from communities can also be, either directly or
with slightly modification, integrated to the top level code. This
is a very comprehensive technique to develop a real-time
advanced code for the controller.

167

3. Design of the Reconfigurable Open Architecture

Controller
3.1 Specification
From the features in previous section, this project has 3
main concepts. This presented controller must 1) control and
do the robotic full function tasks 2) has lower price 3) has
smaller size, which can specify to
1. Handle much complex Mathematics involved in robotic
field.
2. Safe for human environment and robot itself.
3. Receive the encoder signal and drive 3 independent
DC motors at the same time.
4. Control position and force of the robot simultaneously.
5. Automatic drive via G-code [15].
6. Receive the various input and feedback devices.
3.2 Hardware
We propose that the controller should have at least two
processors such as Window PC and Real-time PC (xPC) which
can communicate and exchange data between two computers
via TCP/UDP. The external devices can connect to the
Windows based computer via USB and most of software need
to run on it as well. However, this processor works precisely at
the speed of 100 Hz which is not enough to control the robot
steadily. So, this project chooses Real-time PC (xPC) to work
with Window PC as a dedicated processor for high speed lowlevel control at 1 kHz. This is a key to handle both real-time
and non-real-time codes at the same time.
This project also appreciate FPGAs performance such its
flexibility. Unlike working with chips and wires, the design,
prototyping and debugging phases of a project using an FPGA
are very fast. In the past, designing and building complex
projects required the use of graph paper, discrete logic chips,
breadboards and jumper wires, making it a slow and tedious
project. However, in the world of FPGAs, after updating design,
all we have to do is pressing the "implement" button. In
addition, when finishing with one project, it can be easily
reused the development board for the next. This project uses
this advantage of FPGA integrated with two powerful
processors by RS232 communication to get the signal from
several devices and also command to drive the robot as shown
in Fig. 1. In this way, most devices in the market can interface
with the system especially the newly one.

2103499 2556 MS09

Visual signal

E.stop button

Motor Driver

GamePAD

Phantom
Omni
Haptics

Encoder signal

RS232

TCP/UDP

USB

Motor
power

Motor
signal

USB

FPGA

Window PC

Opto signal

Real-time PC
G-code
Pulse/direction signal
Microcontroller
Pulse signal

Manual
Pulse
Generator

Fig. 1 Hardware diagram

For safety, this project chooses soft circuit approach for
emergency stop button. Normally, emergency stop button is
used to discontinue the power going to the system directly or
using relay. It would improve more flexibility and function if we
get and process its signal by FPGA. Three opto interrupts,
photoemitter-detector device, are chosen for setting the robots
home position because it is non-contact device that has longer
used-time than mechanical contact device such as limit switch.
3.3 Software
One feature of our controller is that it is developed based
on software with strong community for developer. In the
community, we can share ideas, problems, techniques and
even codes. The proposed controller should benefit from the
newly proposed techniques and codes from the community.
The Matlab seems ideal to our case since it can, 1) handle
heavy mathematics through a number of toolboxes, 2) manage
codes that run at different servoing rates, 3) solve ODE in
runtime and 4) implement on real-time computer that run at a
high servoing rate. However, there is no G-code interpreter in
Matlab and its community. Matlab does not effectively handle
sequential logic and loop, even with its Stateflow, code that run
at varying time, and ASCII text that are all required for G-code
programming. In this project, Repetier [16] and Slic3r are
chosen to handle G-code and STL file respectively. Both
Repetier and Slic3r are very good open source programs that

are designed for DIY 3D printer. Roughly, Slic3r is used to

convert STL model into the G-code and Repetier is used to
generate the command signal at runtime from the resulting Gcode.
In the proposed controller, Repetier interfaces with Matlab
through Arduino MEGA 2560 connected to Window PC. FPGA
receives the command position from various input
simultaneously including Repetier, Manual pulse generator and
Matlab which also can switch and manage the position
references from various devices.
3.4 Control technique
The proposed control diagram (Fig. 2) is designed such
that it can get the signal from various ways such as Manual
pulse generator (MPG), gamepad, Phantom omni haptics and
Repetier. Then, send the signal to inverse kinematic building
block and control by PID method. For feedback, encoder is
used for getting the position and digital camera is used to
observe the end tip of the robot. This system is also
reconfigurable to develop and test the new ideas. For example,
the proposed controller can be easily reconfigured such that
the reference trajectory is constructed from two input devices,
i.e., Repetier and MPG (Fig. 3). The X, Y positions from
Repetier are merged with the Z position from MPG such that
robot goes automatically in plane while an operator controls
the depth manually. Fig. 4 shows the real overall system.

168

2103499 2556 MS09

Delta Robot
Gain

Inverse
Kinematics

PID
controller

1 kHz

30 Hz

To monitor

Visual signal

Repetier

Fig. 2 A graphical control diagram

xREP

Delta Robot

yREP
zREP
Repetier

Gain
Inverse
Kinematics

PID
controller

xMPG

1 kHz

yMPG
zMGP

To monitor

30 Hz

Visual signal

Fig. 3 A reconfigured control diagram such that the XY trajectory is driven by G-code while the Z trajectory is manually controlled

Fig. 4 The whole system

169

2103499 2556 MS09

4. Preliminary test and results
The proposed reconfigurable open architecture controller
is used to control the Delta robot (Fig. 5) because it is a
parallel mechanism which is one of the most complex
mechanical systems and has much difficulty to control. The
pen and digital camera are set up to draw complex trajectory
in plane. The height can be controlled by manual pulse
generator (MPG) with the visual feedback from the digital
camera. The robot is firstly commanded by manual pulse
generator to the required print position. The controller then
receives the signal to draw the trajectory.

Opto
interupt

Linear
slide
joint

Timing
belt

Camera
Pen
Base

Fig. 6 Device level result of each motor

PM DC Motor
& Encoder

Fig. 5 Delta robot

The result is separated in two parts, device level and
machine level. Device level part shows the performance of
controlling each motor (Fig. 6), blue line is actual output
while green dot line is the sinusoidal-function reference
commanded via Matlab/Simulink.
For the machine level results, the robot is commanded to
draw a circle (Fig. 7, http://youtu.be/s0eecpoHAgU) ten times
via Matlab/Simulink where the green line is the actual
printing and the red dot line is the reference command.
Moreover, it is G-code commanded to draw the Batmans
symbol via Repetier (Fig. 8, http://youtu.be/J3x56qZQ4Lk)
The controller is successfully implemented to control the
robot as mentioned, the robot go along the defined trajectory
very well as shown in the circle and batmans symbol.
However, there are some errors seen from the device level
result (Fig. 6) that could make some mistake while operating.

Fig. 7 Machine level result via Matlab/Simulink

Fig. 8 Machine level result via G-code

170

2103499 2556 MS09

7. Conclusion
These preliminary results demonstrate the capabilities of
the controller as it can interface with many tools from several
domains such as manual pulse generator, Logitech C920
camera and Repetier. Furthermore, it can handle heavy
Mathematics, kinematics, timing with good accuracy and
repeatability. And from the result, we can realize that this
controller is open and reconfigurable for any software and
hardware.
Although, this project tries with only few software and
hardware but from whole systems characteristics we got, it
ensure that this system is ready for the new robotic tasks
with new devices, facilities or places. Besides, this controller
can apply with 3D printer application as well.

[6] Phanatamporn, K. and Chancharoen, R.

A Reconfigurable Open Architecture Controller for a
Research Robot. in 4th TSME-ICoME, Pattaya, 2013.
[7] Hong, K. S., Kim, J. G., Huh, C. D., Choi, K. H. and
Lee, S., A PC-based Open Robot Control System: PCORC, in IEEE International Symposium on Industrial
Electronics (ISIE), 2001.
[8] Santini, D. C. and Lages, W. F. An Architecture for
Robot Control Based on the OROCOS Framework, in
Robocontrol, 2010.
[9] Wikipedia, The Free Encyclopedia, access 20/2/57.
Available: http://en.wikipedia.org/wiki/XPC_Target.
[10] Field, M. Introducing the Spartan 3E FPGA and VHDL,
2012.
[11] Liandong, P. and Xinhan, H., Implementation of a PCbased Robot Controller with Open Architecture, in IEEE
International Conference on Robotics and Biometics,
2004.
[12] Huchzermeier, J. Benefit and Issues on Migration of 5-V
and 3.3-V Logic to Lower-Voltage Supplies, Texas
Instruments Incorporated, 1999.
[13] Gurdan, D., Stumpf, J., Achtelik, M., Doth, K. M.,
Hirzinger, G. and Rus, D. Energy-efficient Autonomous
Four-rotor Flying Robot Controlled at 1 kHz, in IEEE
International Conference on Robotics and Automation,
Roma, Italy, 2007.
[14] Senoo, T., Yamakawa, Y., Mizusawa, S., Namiki, A.,
Ishikawa, M. and Shimojo, M. Skillful Manipulation
Based on High-speed Sensory-Motor Fusion, in IEEE
International Conference on Robotics and Automation,
Kobe, Japan, 2009.
[15] Wikipedia, The free encyclopedia, access 20/2/57.
Available: http://en.wikipedia.org/wiki/G-code.
[16] Hot-World GmbH & Co. KG, Repetier, The software
driving your 3D printer, access 20/2/57. Available:
http://www.repetier.com/.

8. Acknowledgement
I would like to thank you Assoc. Prof. Ratchatin
Chancharoen, my advisor, for his priceless time to teach and
encourage me closely. Thank you to Mr. Jaruboot Kananai,
senior researcher in my laboratory for his valuable advices.
Thanks to Ms. Sureerat Luepongpattana, Mr. Nithiwit
Wattanawonganan and Mr. Paak Nakpeerayuth for their
Delta robot which is used to demonstrate my controller. The
author also thank you to my family, friends and supporters
who understand and walk along me consistently.
9. Reference
[1] Craig, J. J. Introduction to ROBOTICS: mechanics and
control, 2nd ed., Addison-Wesley Publishing Company,
1989.
[2] Blomdell, A, Bolmsjo, G., Brogardh, T., Cederberg, P.,
Isaksson, M., Johansson, R., Haage, M., Nilsson, K.,
Olsson, M., Olsson, T., Robertsson, A. and Wang, J.
Extending an Industrial Robot Controller :
Implementation and Applications of a Fast Open Sensor
Interface. in IEEE Robotics & Automation Magazine, pp.
85-94, 2005.
[3] Santini, C. D. and Lages, F. W. An Open Control
System for Manipulator Robots. in 20th International
Congress on Mechanical Engineering, 2009.
[4] Oliveira, A., De Pieri, E. and Moreno, U. An Openarchitecture Robot Controller applied to Interaction
Tasks. in Advances in Robot Manipulators, 2010.
[5] Ford, W. What is an Open Architecture Robot
Controller?. in IEEE International Symposium on
Intelligent Control, 1994.

171

2103499 2556 MS10

Design and Control of a Delta Robot to work with human in real time
Patharawut Suphama 5330344521
Assoc. Prof. Ratchatin Chancharoen, Ph.D. (Advisor)
Abstract
The Delta robot is designed and built for tasks where an
operator controls its motion via human machine interface (HMI)
to work with human in real time. Three HMI devices, including
manual pulse generator, Kinect camera, and PHANToM OMNI
Haptic Device, are benchmarked on task performance when
used to control the Delta robot. The Fitts Law is used as a
framework to evaluate the task performance. The result
demonstrates that PHANToM OMNI Haptic Device is the best
HMI since it gives a better response to human operator and
mobility to the task and thus the shortest time to completion.
Keywords: Robotics, Human machine interface.
1. Introduction
A telemanipulator is a device for transmitting hand and/or
finger movements to a remote robotic device, allowing the
manipulation of objects that are too heavy, dangerous, small,
or otherwise difficult to handle directly [1]. In a more usual
context, it is a remote control of a (slave) robotics system
where human partially or fully control the operation. In the
beginning, the mechanical linkage is used as a link between a
control device and a slave manipulator as used in nuclear
industry 40 years ago. A telemanipulator is now designed with
more advanced technology.
A modern telemanipulator uses Haptic as a human
machine interface, which can sense a human movement and
then generate the control signal to control a slave manipulator,
while generating the feedback force, which the manipulator
senses the environment, back to the operator. A mechanical
linkage is now replaced with electronics wires which are used
as a link between Haptic and a slave manipulator. The modern
telemanipulator is a very complex system involving human
factor, a complex robotic system, and an advanced
communication.
In this project, a telepresence Delta robot is designed for
tasks to improve overall working performance. The Delta robot
is designed and built for three dimensional precision tasks

along with electronics and software which are designed such

that it can be reconfigured. To improve our telepresence Delta
robot, a performance evaluation framework is designed based
on Fitts Law [2] where similar works can be found in [3] and
[4]. Roughly, the task is to control the robot to go three
dimensional point to point (PTP) in image space while the time
to completion is the performance index (PI). Design and
Control of a Delta Robot to work with human in real time need
to study about effects of Three HMI devices on completing
tasks.
Three input devices are investigated including (1)
Manual Pulse Generator (MPG) (2) Kinect and (3) PHANToM
OMNI Haptic Device and used to control the slave Delta robot
to complete the task. The paper is divided into eight sections
where section 1 is introduction. Section 2, 3, and 4 describes
the Delta robot, the Human Machine Interfaces, and the control
diagram, respectively. Section 5, 6 and 7 describes the
Assisting system, the evaluation task and the results
respectively. Section 8 is the conclusion.
2. The Delta Robot
A Delta robot consists of three parallel mechanisms to
constrain the end effector such that it parallel to the base at all
time. Three motors sit on the base where their joints, called
Hip joints, connect to the rigid link called upper link. The upper
link connects to the parallelogram mechanism called lower link
using two spherical joints where the other side of lower link
connects to the end plate via the other two spherical joints. In
this way, the rotation of the three motors is transferred to three
translations of the end plate. The torques from the motors can
be transmitted to the end effector via low inertia rigid link.
Thus, the robot is very rigid with very low inertia. In total, there
are eleven links, twelve spherical joints, and three active
rotational joints that are driven by three motors. In this project,
two delta robots are adapted to complete task because the
design of a delta robot to work with human must be a model
that can be used with any delta robots.

172

2103499 2556 MS10

2.1 Kinematics of the delta robot
Compared to serial link manipulator, the kinematics
solution of the parallel Delta robot is more complicated since it
involves twelve passive spherical joints which their positions
must be eliminated prior to map the encoders positions to the
end effectors position in the closed form kinematics. Several
techniques to solve kinematics are proposed including [5-9]. In
this project, the solution proposed by [6], is implemented to
solve inverse kinematics in real time. Let set the reference
axes at point O, then the parameters and direction of x, y and
z will be as in Fig. 1.

respectively. Three 24V PMDC motors, with 500 PPR encoders

and 1:600 planetary gears are used to drive the robot. Thus,
the resolution of the joint position is 300000 PPR. Logitech
C920 HD camera is installed at the end plate.
Table.1 Specification of the Delta robot
Motor and Gear
PMDC with 1:600 planetary gear.
Encoder
500 PPR with 1:600 multiplier.
Camera
Logitech C920
Motor driver
Copley Controls 4122P
Design and Control of a Delta Robot are developed into a
model that is a good optional control system to work with
human in real time effectively. Thus, every delta robots should
be adapted to this model to complete task so other delta robot
is tested and it made in satisfaction as in Fig. 2.

a) The kinematic model.

Fig. 2 The second Delta Robot.

3. Human Interface Devices
Three Human Interface Devices (HMI) are investigated and
used to command the Delta robot to complete a task, including
(1) Manual Pulse Generator (MPG) that is widely used to jog
the CNC machine, (2) Kinect sensor that is used as HMI for
Microsoft XBOX game console, and (3) PHANToM OMNI
Haptic Device that is directly designed to tele-operate a robot.
b) The built Delta robot.
Fig. 1 The Delta Robot.
2.2 The design and built delta robot
The built Delta robot is three translational degrees of
freedom robot (Fig. 1). The lengths of the upper and lower link
are 152 and 296 mm respectively while the side lengths of the
base triangle and the end effector are 207.8 and 92.6 mm

173

3.1 Manual Pulse Generator

A manual pulse generator (Fig. 3) is an interface device
designed for CNC machine to jog its position. It consists of one
rotating knob with an encoder attached. The signal from the
encoder can be directly wired to the Delta robots controller
which is processed at the same speed as the motors encoder.

2103499 2556 MS10

There are switches to select the axis to jog and the
amplification.

the Cartesian position and command force to the device. Its

interface is IEEE-1394 Firewire. In this project, The
Prok.Phantom [12] is used as an interface to Matlab and the
Matlabs Instrument Control Toolbox is used to send the
Cartesian position to the Delta robots controller.

Fig. 3 Manual Pulse Generator.

3.2 Kinect sensor
Kinect sensor is an interface device for the Microsoft's
Xbox 360 video game console. It consists of the RGB camera,
depth sensor, multi-array microphone, and motorized tilt (Fig.
4). The camera and the depth sensor is able to output the 8-bit
RGB and 11-bit depth images with 640x480 resolutions at the
speed up to 30 fps respectively. Both images can be combined
and treated as a single image with four dimensions called
RGBD (Red, Blue, Green, and Depth) image. The depth image
significantly simplifies the segmentation tasks, especially when
an objects color is close to an environment. Using Kinect, the
object and background with the same color can be easily
distinguished by their depths and the RGB image can be
processed further to extract the useful information of an object
such as the centroid of a hand that is used in this project. The
hand tracking program from Prime Sensor [10] is modified to
print the position of the hand through TCP/UDP to the Delta
robots controller.

Fig. 5 The PHANToM OMNI Haptic Device.

4. The Control Diagram
The signals from the three input devices: MPG,
PHANToM OMNI and Kinect are sampled at 1 kHz, 300 Hz,
and 30 Hz respectively. The signal from MPG is printing to the
robot controller directly while the signals from The PHANToM
OMNI and Kinect are printing to the Windows based computer
which are then sent to the robot controller via 100 MB/s
TCP/UDP. There are software switched to manipulate these
signals. These signals are the position reference that the robot
will follow. The inverse kinematics block is then used to
convert the position reference into the three joint references
which are then compared to the actual joint positions sensed
by encoders. The PID controller block is then process the
difference signals in parallel and generate the control effort
signals to the amplifiers which drive the current into the motor.
The PID gains are tuned for the best performance. The
Logitech C920 camera transmits the image signal back to the
Windows computer. The live video is shown on screen along
with useful information from image processing.

Fig. 4 Kinect.
3.3 The PHANToM OMNI Haptic Device
The PHANToM OMNI [11] (Fig. 5) is a commercial six
degrees of freedom haptic that is used to generate the
Cartesian (x, y, z) position and orientation and, in the same
time, give the force feedback to an operator. This device is
widely used as an interface device to control a slave
manipulator in tele-operation task. The device comes with the
driver on Windows and low level C library (HDAPI) to receive

Fig. 6 The Control diagram of the Delta Robot.

174

2103499 2556 MS10

5. The Assisting system
The one important part of Design of a Delta Robot to work
with human in real time is the Assisting system because
human need tools to help him complete task easier. Then,
Matlab is software to build assisting display screen on video
from Logitech C920 camera attached at the hand of robot. The
camera capture picture of tested plate as in Fig. 7 and the
assisting display screen report status of working and warn
users to move the delta robot to the right position in 3D
following desired constraint as in Fig. 8.

easily move the target in image plane such that the centroid
and bounding box of the target match the assisting marks.
Once the position of the Delta robot is on the target, there is
a beep to inform the operator and the next target symbol is
shown on screen. The next target will be in different color. In
case that the target, as seen by the camera, is too close or too
far, the text ZOOM OUT or ZOOM IN is displayed on
screen respectively.
The elapsed time from one target to the next is recorded
and analyzed. This time is then used as the performance
index. The images captured at each of the target positions are
also recorded to check that the robot is really at the target
positions within the giving accuracy.

Fig. 7 Tested plate.

Fig. 9 Connect the Dot task.

Fig. 8 Assisting display screen.
6. The Evaluation Task
The evaluation task is the three dimensional point to point
in image space, i.e. to complete connect the dot as in Fig. 7
and Fig. 9 but only Fig. 9 will be described. An operator sees
the target objects only in a screen and then makes decision
and commands the Delta robot through HMI device. The target
objects are the circle marks in three colors: red, green, and
blue and in three slightly different sizes. There are 21 targets
consists of seven- red, green, and blue targets. For each color
targets, there are letter mark, A, B, C, D, E, F, and G. The
operator is to go from red A, then green A, , until blue G,
and come back to red A to complete the task. The marks are
in a zigzag pattern as seen in Fig. 9.
The position of the Delta robot with eye in hand camera is
to be controlled such that the target is at the center of the
image while its size is also as specified. The positioning error
must be within 10%. There is an assisting mark and box at the
center of the image to aid the operator such that he/she can

175

To study on effect of delay time on operating time under

the condition of different human machine interface devices
between Phantom Omni Haptic device and Manual Pulse
Generator, a Delay box in Simulink of control diagram is
required (blue in Fig.10). Then the operator completes the
evaluation task and repeat with different delay time. Including,
to study on effect of resolution on operating time under the
condition of different human machine interface devices
between Phantom Omni Haptic device and Manual Pulse
Generator, a Quantizer box in Simulink of control diagram is
required (red in Fig.10). Then the operator completes the
evaluation task and repeat with different delay time.

Fig. 10 Delay time and resolution in control diagram.

2103499 2556 MS10

7. Experimental Results
The result indicates that PHANToM OMNI gives the best
performance, followed by Kinect and MPG respectively (Fig.
10) . The PHANToM OMNI reduces the time to completion by
4 times compared to MPG and about 50% compared to Kinect.
This is because MPG can only move the robot only one axis at
a time and thus take a much longer time for a three
dimensional PTP task. The signal from PHANToM OMNI is
about 10 times higher sampling than from Kinect and the
resolution of the signal is much better. Furthermore, the
attractive force is activated to assist the operator to know the
absolute position of the device which is unknown when using
Kinect. This feature is really help since there is a noticeable
delay in the system. The operator has to face both slow
sampling and delay when using Kinect to control the robot.
On table 2, It is noted that Kinect cannot command the
robot to the position (round) 21 where the robot is near its
workspace boundary. This is because the signal from Kinect is
quite slow and fluctuating. This signal results in a shaking
motion of the robot and also the camera, especially near the
boundary of the robots workspace. In our test, an operator
cannot go to position 21 and skip this position after trying in
250 second period. The MPG, which is directly connected to
the robots controller and is sampling in the same rate as the
servo algorithm gave the smoothest motion of the robot even
at the position 21. In comparison, PHANToM OMNI gives a
shorter time but little shaking motion is noticed. However,
PHANToM OMNI still gives the best overall performance in
term of time and working experience.
35

MPG

30

Kinect

Haptic

Table 2 Overall comparative performances (time) of the three

input devices: MPG, PHANToM OMNI, and Kinect.

Round
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Average

MPG
Start
12.5
13.1
21.3
10.4
20.2
23.2
13.1
30.0
20.3
25.9
31.9
13.1
12.3
21.6
16.9
8.2
16.5
12.8
9.7
22.5
14.5
17.6

Time (s)
Haptic
Kinect
Start
Start
3.5
3.1
1.3
5.6
3.4
7.0
7.0
6.6
1.7
2.7
5.7
5.0
1.7
2.0
0.1
2.5
10.4
18.2
3.9
3.8
2.3
3.6
2.7
3.2
3.9
1.7
7.2
18.4
5.2
6.2
2.4
13.3
1.9
13.6
6.6
2.4
1.1
2.8
5.2
NA
6.7
2.5
4.0
6.2

The result of delay time indicates that the effect of delay

time on operating time between PHANToM OMNI and MPG,
there are 2 trend linear graphs. In Fig. 12, points of
Intersection are accepted delay time at 100 ms on PHANToM
OMNI and 300 ms on MPG. Delay time which is less than the
points do not affect the operating time significantly.

Time (s)

25
20
15
10
5
0
1 3 5 7 9 11 13 15 17 19 21
Round

Fig. 11 Comparative performance of the three input devices:

MPG, PHANToM OMNI, and Kinect.
Fig. 12 Effect of delay time on operating time under the condition
of different human machine interface devices between Phantom
Omni Haptic device and Manual Pulse Generator
176

2103499 2556 MS10

The result of resolution indicates that the effect of
resolution on operating time between PHANToM OMNI and
MPG, there are 2 trend exponential graphs. In Fig. 13, points
5% of maximum are accepted resolution at 1.78 mm on
PHANToM OMNI and 1.35 mm on MPG. Resolutions which
are more than the points affect the operating time significantly.

Fig. 13 Effect of resolution on operating time under the condition

of different human machine interface devices between Phantom
Omni Haptic device and Manual Pulse Generator
8. Summary
The PHANToM OMNI Delta robot is designed and built for
tasks where performance evaluation framework is designed to
improve the working performance. Based on the proposed
evaluation framework, PHANToM OMNI gives the best
performance to the telepresence system compared to MPG
and Kinect, i.e., shortest time to completion. Accepted delay
time and Accepted resolution of PHANToM OMNI are 100 ms
and 1.78 mm respectively.
Acknowledgement
Thank you to Assoc. Prof. Ratchatin Chancharoen, my
advisor, for teaching and giving me suggestions. Gratitude to
Prof. Viboon Sangveraphunsiri and Dr. Phongsaen
Pitakwatchara for devices and valuable supports to complete
the work and also thanks Mr.Jaruboot Kananai and
Mr.Nopasorn Kowathanakul, senior researchers, for the Delta
robots.
References
[1] The American Heritage Science Dictionary.
[2] P. M. Fitts, The information capacity of the human motor
system in controlling the amplitude of movement, Journal
177

of Experimental Psychology, vol. 47, no. 6, pp. 381-391,

1954.
[3] McLean, G.F., Prescott, B. and Podhorodeski, R (1994).
Teleoperated System Performance Evaluation, IEEE
Transactions on Systems, Man, and Cybernetics, vol. 24,
no. 5, May 1994, pp. 796-804.
[4] Mihelj, M., Kamnik, R. and Bajd, T. Assessment of
Teleoperator Performance, Electrotechnical Review, vol.
65, pp. 7-13, 1998.
[5] M.A. Laribi, L. Romdhane and S. Zeghloul: Analysis and
dimensional synthesis of the DELTA robot for a
prescribed workspace, Mechanism and Machine Theory
vol. 42, Elsevier, pp. 859870. (2007).
[6] P.J. Zsombor-Murray, Descriptive Geometric Kinematic
Analysis of Clavels Delta Robot, Centre of Intelligent
Machines, McGill University, USA. (2004).
[7] M Lopez, E Castillo, G Garca, and A Bashir, Delta robot:
inverse, direct, and intermediate Jacobians, Proceedings
of the Institution of Mechanical Engineers, Part C: Journal
of Mechanical Engineering Science, Vol. 220, No. 1. , pp.
103-109. (2006).
[8] K. Miller, Experimental verification of modeling of DELTA
robot dynamics by direct application of Hamilton's
principle, Proceedings of IEEE International Conference
on Robotics and Automation (ICRA 1995), Nagoya, Aichi,
Japan, Vol. 1, pp. 532 537. (1995).
[9] S.T. Staicu, Dynamic analysis of Clavel's Delta parallel
robot, Robotics and Automation, Proceedings of IEEE
International Conference on Robotics and Automation
(ICRA03), Vol. 3, pp. 4116- 4121. (2003).
[10] PrimeSense Limited Company. Prime Sensor NITE 1.3
Algorithms notes, URL: http://www.primesense.com
[11] Jarillo-Silva, A., Domnguez-Ramrez, O.A., Perra-Vega,
V. and Ordaz-Oliver, J.P. (2009). PHANTOM OMNI Haptic
Device: Kinematic and Manipulability, Electronics,
Robotrics, and Automotive Mechanics Conference 2009,
September 2009, pp. 193-198.
[12] http://prok-phantom.googlecode.com

2103499 2556 MS11

A study of the thermal comfort effect on people who sit near a glass window with
a venetian blind installed
5330230621
.. ( )

(Shortwave radiation PPD)

( Longwave
radiation PPD)
(Total radiation
PPD)










EnergyPlus

:
Predicted Mean Vote (PMV) Predicted Percentage of
Dissatisfied (PPD)

Abstract
This paper is about a study of the thermal comfort effect
on people who sit near a glass window with a venetian blind
installed. The effect of the related parameters of glass windows
and the venetian blind on human thermal comfort are studied.
It is found that at a constant solar profile angle, the thermal
comfort is dependent on optical properties of the glass window
and the slat angle. If the glass window has a higher solar
transmittance and a lower solar absorbtance, it will give the
shortwave radiation PPD higher than the glass window having
a lower solar transmittance and a higher solar absorbtance.
It is also gives a lower value of the longwave radiation PPD.
Nevertheless, the value of the total radiation PPD of the glass
window with a higher value of solar transmittance is still higher.
In addition, when the slat angle is set in the position that
obstructs the solar radiation, the value of the total radiation
PPD becomes smaller. On the contrary, when the slat angle is
set in the position that allow more radiation passing into the
room ,the value of the total radiation PPD becomes larger
(which is mainly resulted from the shortwave radiation PPD).
Finally, comparing the results between the predictions from the
developed mathematical model and from the EnergyPlus
mathematical model, it is found that the results from the two
models give the same trend of thermal comfort condition but
they also have a certain difference in magnitude of the related
parameters due to the different approaches in the
mathematical models.
Keyword: Thermal comfort Mean radiant temperature (Tmrt)
Predicted Mean Vote (PMV) Predicted Percentage of
Dissatisfied (PPD)
1.

178

2103499 2556 MS11

(

) Tmrt

3
(3)
Tumrt = Unirradiant Mean Radiant Temperature
4
(4)

2 .

Fanger [1]
2 Predicted Mean Vote (PMV)
6


Predicted Percentage of Dissatisfied (PPD) (
)
1 2

=
Bn = Radiosity n, W/m2
Fp-n = Angle Factor n
fp
= Project Area Factor

ak =
( = 0.6)
p =
( = 0.97)
I
=
, W/m2
(5)
I = Idirv T() + Idiffv Ther
Idirv =
(W/m2)
Idiffv =

(W/m2)
T() =

Ther =

(1)

(2)

Pa

f cl

Tmrt

Ta

Tcl
hc

=
, W/m2
=
, W/m2
= , kPa
=

= (Mean radiant
temperature),C
= ,C
=
=

,C

2
, W/m K
1 (Tmrt)

1
(Tmrt) ( 1)
black enclosure

3.

[2] [3] [4]
2

179

2103499 2556 MS11

Matrix layer calculation




3 1 6 . 1
5 . 2 6 .

6 . 3
40 .
1 2
1

Thickness

6 .

5 .

Conductivity

0.917
W/m-K
0.840
0.840
0.801
0.069
0.069

0.917
W/m-K
0.840
0.840
0.477
0.053
0.053

Front emissivity
Back emissivity
Transmittance
Front reflectance
Back reflectance

Slat thickness
Slat separation
Slat width
Slat conductivity
Transmittance
Front reflectance
Back reflectance
Front emissivity
Back emissivity

6 .
/
0.917
W/m-K
0.840
0.840
0.650
0.130
0.080

0.3 .
20 .
25.4 .
120 W/m-K
0
0.6
0.6
0.87
0.87

3



EnergyPlus
1.
1.
2.
Effective layer
2
3.

diathermanous (

)

2.
3
3.

3 EnergyPlus

EnergyPlus [5]



3

180

4.
3

1 2
(b) 3 0 ,45 - 45
Solar Profile (s) = 46.497 b
s 4

2103499 2556 MS11

557.38 W/m2
221.49 W/m2


1

4 b s



1. Blind temperature
2. Mean Radiant Temperature (Tmrt)
3. Predicted Mean Vote (PMV)
4. Longwave radiation PPD
5. Shortwave radiation PPD
7. Total radiation PPD
4

EnergyPlus 5
4.1


4
-45

Tmrt PMV
Total radiation PPD
PPD 58% PPD Shortwave 53%
PPD Longwave 5%


45

Tmrt PMV
Total radiation PPD Longwave
radiation PPD Shortwave radiation PPD

Tmrt PMV Total radiation PPD

Tmrt PMV Total radiation PPD








[2]




4.2
EnergyPlus
5
EnergyPlus









4.3


EnergyPlus

EnergyPlus




diathermanous


EnergyPlus


0 45
Tmrt PMV Total
radiation PPD

181

2103499 2556 MS11

4

45

- 45

Blind temperature (K)

Tmrt (K)
PMV
Longwave radiation PPD
Shortwave radiation PPD
Total radiation PPD

49.097
35.293
1.421
20%
26%
46%

46.112
32.758
1.042
16%
12%
28%

30.742
37.282
1.641
5%
53%
58%

Blind temperature (K)

Tmrt (K)
PMV
Longwave radiation PPD
Shortwave radiation PPD
Total radiation PPD

44.071
32.206
0.96
14%
11%
25%

42.73
30.828
0.758
12%
5%
17%

33.849
33.345
1.105
6%
25%
31%

47.463
34.071
1.238
18%
19%
37%

45.619
32.221
0.963
15%
9%
24%

31.924
35.222
1.361
5%
38%
43%

6 .

Blind temperature (K)

Tmrt (K)
PMV
Longwave radiation PPD
Shortwave radiation PPD
Total radiation PPD

5 .

2 6 .
6 .

5 EnergyPlus

45

- 45

44.009
34.044
1.229
13%
24%
37%

43.323
31.577
0.868
12%
9%
21%

34.146
39.024
1.906
6%
66%
72%

39.655
31.033
0.787
9%
9%
18%

39.297
29.629
0.585
9%
3%
12%

33.303
33.764
1.163
6%
27%
33%

43.135
32.915
1.063
12%
17%
29%

42.907
31.106
0.799
12%
7%
19%

33.275
36.272
1.515
6%
46%
52%

6 .

Blind temperature (K)

Tmrt (K)
PMV
Longwave radiation PPD
Shortwave radiation PPD
Total radiation PPD
Blind temperature (K)
Tmrt (K)
PMV
Longwave radiation PPD
Shortwave radiation PPD
Total radiation PPD

Blind temperature (K)

Tmrt (K)
PMV
Longwave radiation PPD
Shortwave radiation PPD
Total radiation PPD

5 .

2 6 .
6 .

182

2103499 2556 MS11

EnergyPlus





EnergyPlus


-45

5.


Predicted Mean Vote (PMV)
Predicted Percentage of Dissatisfied (PPD)


EnergyPlus
solar profile (s) = 46.497

2

45 PPD
0 -45 45

-45













diathermanous

diathermanous

183

..
.


[1] Fanger, P.O. Thermal comfort analysis and application
in environmental engineering. Kansas State university:
McGraw-Hill, 1970.
[2] .



.
,2551.
[3] Chaiyapinunt, S. and Worasinchai, S. Development of a
mathematical model for a curved slat venetian blind with
thickness. Solar Energy 83 (2009) pp.1093-1113.
[4] Chaiyapinunt , S. and Worasinchai, S. Development of a
model for calculating the longwave optical properties and
surface temperature of a curved venetian blind. Solar
Energy 83 (2009) pp.817-831.
[5] EnergyPlus, EnergyPlus Engineering Reference:
The Reference to EnergyPlus Calculation, October
2013.

2103499 2556 MS12

A Thermal Performance Study of a Glass Window with a Venetian Blind Installed
5330238721
. . ( )

3

0 45 -45


Solar Heat Gain Coefficient (SHGC)


45

EnergyPlus

EnergyPlus


Solar
Heat Gain Coefficient
Abstract
This article is about the study of thermal performance of a
glass window with a venetian blind installed. Different types of
glass window; clear glass, tinted glass, and double pane glass,
are chosen for the study. Blind with different slat angles (0, 45,
and -45 degree) are also chosen for the study. The results
from the analysis show that at the specified solar profile angle
in this study the clear glass window and setting blind slat angle
to -45 degree gives the largest value of the direct solar heat
gain coefficient and the room heat gain. EnergyPlus
mathematical model for calculating the thermal performance of
the glass window with blind is also considered. The analysis
indicates the agreement between the results of the solar heat
gain in the shortwave part predicted from the developed
mathematical model and EnergyPlus mathematical model is
quite good. There are certain disagreements of the solar heat
gain in the longwave part from both mathematical models
found in this study.

Keywords: thermal performance, glass window, venetian

blind, solar heat gain coefficient
1.

(solar energy)
(thermal energy)
SHGC (Solar heat Gain
Coefficient)
SHGC



(overall heat transfer coefficient, U)

184

2103499 2556 MS12

SHGC

2.


1




(1)
q U T T SHGC I cos SHGC I
q = , W/m2
U = , W/m2oC
Tout = , oC
Tin = , oC
= , o
SHGCd SHGC
SHGCD SHGC
Idir , W/m2
Idif , W/m2
out

in

dir

diff

inside

outside

2 azimuth

optic
optic
optic




inward
flowing fraction SHGC

1
SHGC

(2)
SHGC , T , N A ,
fH
T = directional-hemispherical transmittance
Ni = inward-flowing fraction
Afi = directional-hemispherical front absorptance
layer i
= azimuth , o
SHGC
, azimuth
, 2
fH

solar profile
-1
(3)
tan sin tan
= solar profile, o
SHGC 2
solar profile
SHGC
diffuse SHGC


TfH SHGC
NiAif SHGC

i 1

185

3.


[1-3]

optic

effective layer
3
optic

optic specular (
)

2103499 2556 MS12

optic non specular (
)


optic optic

optic
Matrix Layer Calculation
optic
Furlur [4]
Finlason [5]

optic
Chaiyapinunt Worasinchai [2]

optic
radiosity
6
inwardflowing fraction


1 3

diathermanous layer (
) Wright
Collins [6]
4

inward-flowing fraction


[6]

3 effective layer (layer 3)

(layer 1 2) [2]

EnergyPlus [7]


1. optic

4a
EnergyPlus
4 b

(a)

(b)

4
2.
diathermanous
layer EnergyPlus
diathermanous layer
3.



effective layer EnergyPlus

5

5
EnergyPlus
4.



=
557.38 W/m2
= 221.49 W/m2 = 32.56oC
= 1.2 m/s =46.5

6 5
6 6

186

2103499 2556 MS12

0 45 -45
optic 1
=
25.4 = 20.0
= 300 =0.6
= 0.87
=
40

1 optic

Trn
0.80
0.48
0.65

Solar Energy
Emissivity
Reff Refb
Ab
Ef Eb
0.07 0.07
0.13
0.84 0.84
0.05 0.05
0.47
0.84 0.84
2
0.12 0.08 0.84 0.84

1 : Trn = transmittance, Reff = front reflectance, Refb = back

reflectance,Ab = absorptance,Ef = front emittance Eb = back emititance
2 : absorptance 1= 0.137
absorptance 2 = 0.098

fH

NiAif
SHGC d

q(W/m2)

-45o

0o

45 o

-45 o

0o

45 o

-45 o

0o

45 o

0.714
0.059
0.773
436.2

0.147
0.357
0.504
349.0

0.065
0.315
0.380
280.1

0.396
0.122
0.518
309.7

0.081
0.301
0.382
267.7

0.034
0.277
0.311
226.9

0.554
0.096
0.650
372.2

0.118
0.369
0.487
319.8

0.055
0.339
0.394
269.5

: TfH = directional-hemispherical transmittance , NiAif =

, SHGCd SHGC q =

2



, TfH





,
-45
optic

187

0 45 (
)

SHGC
TfH




cooling load

-45

0
45


0 46.5


45
-45
45
-45

SHGC
-45 SHGC
0
45
-45
0 45
(
)

3
EnergyPlus

fH

N A

f
i i

SHGC

q(W/m2)

-45o

0o

45 o

-45 o

0o

45 o

-45 o

0o

45 o

0.736
0.128
0.864
476.1

0.152
0.254
0.406
411.6

0.069
0.261
0.330
324.3

0.410
0.153
0.563
373.4

0.083
0.255
0.338
330.4

0.036
0.239
0.275
288.6

0.570
0.131
0.701
413.9

0.122
0.275
0.397
362.8

0.059
0.281
0.340
302.5

2103499 2556 MS12

EnergyPlus
3 Tfh

EnergyPlus
NiAif

inward-flowing fraction

diathermanous layer

SHGC



optic
6 8


SHGC
solar profile 0 45
-45

6 SHGC

6 0 45 -45

7 SHGC

5 0 45 -45

8 SHGC

0 45 -45
6 8 SHGC
solar
profile
SHGC


0
45 solar profile
SHGC

-45 solar profile

SHGC
solar profile 45

SHGC
45
SHGC
solar profile


45

5.
1. optic

2.

188

2103499 2556 MS12

..
.

[1] ; .

.
, 2551
[2] Chaiyapinunt, S. and Worasinchai, S. Development of a
mathematical model for a curved slat venetian blind with
thickness Solar Energy 83 (2009) pp. 1093 1113.
[3] Chaiyapinunt, S. and Worasinchai, S. Development of a
model for calculating the longwave optical properties and
surface temperature of a curved slat venetian blind Solar
Energy 83 (2009) pp. 817 831.
[4] Furler, R.A. Angular dependence of optical properties of
homogeneous glasses. ASHRAE Transactions: Symposia, IN91-19-1, 1991.
[5] Finlayson, E.U., Arasteh, D.K. Huizenga, C. Rubin, M.D.,
and Reilly, M.S. WINDOW 4.0: Documentation of calculation
procedures. Publication LBL-33943/UC-350, Lawrence
Berkeley Laboratory, Energy & Environment Division, Berkeley,
CA, 1993.
[6] Collins, M.R. and Wright, J.L., Calculating center-glass
performance indices of windows with a diathermanous layer,
ASHARE Transactions (2006) 112 Part 2, pp. 22-29.
[7] EnergyPlus Engineering Reference, U.S.Department of
Energy, 2013.

189

2103499 2556 MS13

Design and Manufacturing Ocean Wave Energy Converter Prototype
5330442321
..
( )

1
, 2

Abstract
This article present the design of an ocean wave energy
converter prototype that can convert a wave power into 1
kilo-watt of electricity. This machine is designed using
pneumatics circuit which located on a base buoy. The
dimensions of base buoy are 2 meters long, wide and tall. The
machine is capable to converts an unpredictable form and
frequency of ocean wave in to a steadier form of energy. The
machine starts to work when wave reaches the float buoys
which float around the base buoy and connected to pneumatic
cylinders by the force-increasing mechanism. The float
buoys are lifted and the cylinders are compressed while wave
passes through buoys. After cylinders are compressed, the air
will flow through check valves then in the air reservoir. The
pressure inside will slightly rise until it reaches the designed
pressure, as soon as the pressure hit the limit the air inside will
flow out through pneumatics logic circuit and air motor to
atmosphere with a constant flow rate. Air motor is able to
produce almost a steady power in a short period.
Keywords: 3-5 words, 12 point size

(Greenhouse
effect)



71 3 4








4

1.








PowerBuoy,
Pelamis

1.

190

2103499 2556 MS13

2.




VOITH
Oceanlinx

3.

,




Oyster

4.

Wave Dragon, CETO Wave Energy Converter,

= 0.365
= 6

191

2103499 2556 MS13

2. (Ocean Wave Converter Design)
2.1















xxx

Air reservoir

2.2

2.2.1
1

GAST 6AM-FRV-5A

90

2 7

2.2.1

8 0.2
2.2.2

F Buoyancy = pfluid V g
= pfluid A X g
F Buoyancy:
pfluid:

V :

A :

X :


F = P1A1 P2A2

F = mg

h
x

192

2103499 2556 MS13

2.2.4




FBbase = moverall g
Xbase = ((moverall g) - pwater)/Abase

X = ((Fcylinder * Ratio)/Afloat g) + ((pfloat hfloat)/pwater)

F = P1A1 P2A2
F = mg

h
x

Fbase :
Moverall :
Xbase :


Abase :


Xbase A

X = - ((Fcylinder * Ratio)/Afloat g) + ((pfloat hfloat)/pwater)

Fcylinder :
Ratio :

pfloat :

pwater :
hfloat :

g:

X1

Y
2.2.5

2.8

X2
Y = hfloat X1 + X2
X1 :
X2 :
Y:

2.2.3

Vcylinder = (Y) (Acylinder)

t = VTank / ((Scylinder) (fwave) (Acylinder) (ncylinder))

Scylinder :
fwave :

Acylinder :
ncylinder :
t:

193

JK flip-flop

2103499 2556 MS13

7 J
flip-flop 2 K NOT
Output AND

3.

8
16
1 0.2
FLR unit 1 1 1
JK flip-flop NOT AND 1

4.

..

..
2556

5.
[1] ,,
http://www.electron.rmutphysics.com/sciTech/pdf/c07_2554_ene
rgy.pdf, 6 2556
[2] GAST Air motor,
http://www.gastmfg.com/catalogs/Airmotor_D10_Catalog_Final_Feb09-2012_lo-res.pdf, 9
2552
[3] PowerBuoy, Ocean Power Technologies,
http://www.oceanpowertechnologies.com/ 23
2556
[4] Pelamis , www.pelamiswave.com 23
2556
[5] Voith, http://voith.com/en/index.html 23

194

2103499 2556 MS13

195

2103499 2556 MS14

Design of Compact Underactuated Mechanical Gripper with Simplicity

5330364021
.. ( )

(self-adaptibility)

-




(SARAH

M1) [1]


0.43 1.72

[2]

Abstract

[3]
The objective of this paper is to design and develop

the geometry of compact underactuated mechanical gripper

which has generality and simplicity. The gripper consists of 2

2-DOF fingers underactuated by pulley-tendon system and

movable pulley for underactuation between fingers. A parallel

linkage is used to improve fingertip grasp. This works
(ejection phenomenon)
implement stability and force isotropy to optimize the design.
(stability frontier)
The result gripper. The prototype has 0.43 of pulley ratio and
(particular
1.72 of link-length ratio. From primitive-shape grasping test, the
design)
gripper is able to achieve the stable configuration.
Keywords: underactuated, gripper, stability, grasp
2.
2.1
1.

( 1 )

2

4

-

1

(closing process) 1

2 2
(underactuation)

196

2103499 2556 MS14

3 1

2 3

(opening process)
1 2

r1

1
2.2
(
)

4 r1 - 1 r2 - 2 l1 1 l2 - 2 4



RP r2 / r1 RL l2 / l1

R p RL

1


(
(Fingertip Stability Frontier))

1 - 1 k2 2 2

: 1 / 2 3 / 4 k2
0.25l2 0.75l2 3


[4]

Is

* (k1 , 1 )dk1d1

dk1d1

197

(1)

2103499 2556 MS14

1
0

* (k1 ,1 )

; stable
; unstable

(2)

(force
isotropic surface) 3 2 -
2 1 k1 - 1
1 k2 - 2 2


2 / 8 3 / 8 , k1 0.25l1
:
0.75l1 k2 0.25l2 0.75l2
4

I s - I f

IG I s I f

(4)

- RP - RL

RP RL IG
5

5 RP RL IG

[4]

If

* ( 2 , k2 , k1 )

f1 (2 , k2 , k1 ) f2 (2 , k2 , k1 ) d dk dk
2
2
1
2 max f1 ( 2 , k2 , k1 ), f 2 ( 2 , k2 , k1 )

* ( 2 , k2 , k1 )d 2 dk2 dk1

(3)

2.3

-1 ( 6) - RP 0.43
- RL 1.72


[5] ; l1 l2 92.31mm

10mm

65.3mm 1

198

2103499 2556 MS14

3.
2



( 7 )

( 7 )

3.1.
3.1.1

8 ()



8 ( )

8 ()

6 -1

Gripper

Object
8 ()
()

7 ()
()

199

3.1.2

9 ()
9 () 10

2103499 2556 MS14

3.2

( 1)


3-4

1-2




2 1 2

-1
2 1

-1

10

GRIPPER STABILITY TEST

9 ()
()
4.


-

TEST OBJECT
GRASP TYPE CONTACT POINT STABILITY
OBJECT-SQUARE 10mm x 10mm enveloping
no
OBJECT-SQUARE 10mm x 10mm
fingertip
2
yes
OBJECT-SQUARE 30mm x 30mm enveloping
4
yes
OBJECT-SQUARE 30mm x 30mm
fingertip
2
yes
OBJECT-SQUARE 50mm x 50mm enveloping
3
yes
OBJECT-SQUARE 50mm x 50mm
fingertip
2
yes
OBJECT-CIRCLE 10mm
enveloping
no
OBJECT-CIRCLE 10mm
fingertip
2
yes
OBJECT-CIRCLE 30mm
enveloping
4
yes
OBJECT-CIRCLE 30mm
fingertip
2
yes
OBJECT-CIRCLE 50mm
enveloping
3
yes
OBJECT-CIRCLE 50mm
fingertip
2
yes

..

200

2103499 2556 MS14

Lionel Birglen,
Thierry Lalibert Clment Gosselin





[1] Thierry Laliberte, Clement M. Gosselin, Underactuation in
Space Robotic Hands, 2001
[2] Thierry Laliberte, Clement M. Gosselin, Simulation and
Design of Underactuated Mechanical Hand, 1998
[3] Lionel Birglen, Clement M. Gosselin, On the Force
Capability of Underactuated Fingers, 2003
[4] Lionel Birglen, Thierry Lalibert,Clment Gosselin,
Underactuated Robotic Hands, 2007
[5] B. Alexander, K. Viktor, Proportions of Hand Segments,
2010.

201

2103499 2556 MS15

Passive Mechanism with Self Adaptive Capability

5330553321
. ()

3
()

2.62

, ,
Abstract
This research proposes a design and development of a
passive mechanism with self-adaptive capability. This passive
mechanism has 3 degree of freedom and driven by a single
hydraulic actuator. The design is based on force isotropic and
the relation between linear & angle movements. According to
the design, the passive mechanism uses ratio of profile areas
of hydraulic cylinders equal to 2.62. From experiment the
mechanism has self-adaptive capability and can capture
objects with various figures and sizes.
Keywords: Passive mechanism, Self-adaptive capability,
hydraulic
1.

(Degree of
freedom, DOF)
(Active mechanism)




" " (Passive
mechanism)
(Actuator)
(Underactuated
Mechanism) [1]


Super under-actuated multi-fingered
mechanical hand [2]

Underactuation in robotic grasping hands [3]
(Linkage)
10 Design of a Fully compliant
Under-Actuates Finger with a Monolithic Structure and
Distributed Compliance [4]

- (Pulley Tendon) [5-7] 1
- (Pulley-belt) [8-9]

202

2103499 2552 MS15

1 2
( )
1
1
2
(Mechanical Limit) 1 2
1

1 . [3]
. - [6]
. - [2]

(Self-adaptive capability)
2

() (Hydraulic)

5
8

2 [11]

(Link) 2
2

2.2
3
1

a1 a1
1 1

203

11

2.

2.1

(Differential
Mechanism) (T-Pipe) [10]

10

F
5

1 Base, 2 Link 1, 3 Link 2, 4 Hydraulic, 5

Hydraulic Lock, 6 Link a1, 7 Link a2, 8 return spring, 9
and 10 the first and second rotation
3
2.3

1 2 1
2

1
2

1 2
2
2

2103499 2556 MS15

3, 4, 5,
2.4

(Contact force)
(Force isotropic)

0 180
a
40 .
2.4.2 (Contact force)

2.4.1
-
(Slider-crack mechanism)

(Cos law)
a Ls h 2(Ls )(h )cos
(1)
a A d O
S h
h Ls
2

4
(1)

a2
)
2(Ls )(h )

(2)
a
Ls 20 .
cos1(Ls 2 h 2

(Degree)
a=3

6
O1O2
1, O2Z 2 O1 O2
1 2
O1O2 = L1 O2Z = L2
f1 - 1
1 (N)
f2 - 2
2 (N)
F1 - a1
a1 a1
2 force member (N)
F2 - a2
a2 (N)
- 1 a1 ()
- 2 a2 ()
- 1 ()
- 1 2 ()
d1 - O1 1
(.)
d2 - O2 2
(.)
1

Ls (cm)

5 (2)
5 a 30 .

204

2103499 2552 MS15

6 O1
()
f d f (d L cos ) F Ls sin 0
(3)
1 2
O2
f d F Ls sin 0
(4)

cos ( (1 sin )(a Ls sin ) Ls sin

(5)

a
n 1, 2
(3) (4)
1 1

1 2

2
n

2
n

2
2

F1Ls1 sin1
f1
f2

F2Ls2 sin 2
(d 2 L1 cos2 )
d2
2
d1

(6)
(7)

F2Ls2 sin 2
d2

(6)
(7)
(Force isotropic)
f f
(8)
d1 = 0.5L1 d2 = 0.5L2 (
)
L1=L2
F
sin
(9)
(2 2cos )

F
sin
,

F
(10)
2 cos
F
(0,90) cos
[0,1]

F PA
(11)
(10) (11)

A
(12)
2
4
1

3.

(DOF) 2

3.1

20 CATIA
(3D Printer)
(Acrylonitrile-butadiene-styrene)

(Syringe) 10 ml 20
ml (Pin)
5 .
100 200
/ 1 1.5 2
3

A2

8
3.2

15

7 CATIA

205

2103499 2556 MS15

9

(25 .)
(50 .) (75 .) 10 .

.
.
9 . 10 .
. 4 .

10

4.




3 1

10

1 2
3 9
10

206

2103499 2552 MS15

[1] G. Carbone, Grasping in Robotics, Mechanism and

Machine Science 10, DOI: 10.1007/978-1-4471-4664-3,
London : Springer-Verlag, 2013
[2] Wenzeng Zhang - Demeng Che - Hongbin Liu - Xiande Ma
- Qiang Chen - Dong Du and Zhenguo, Super underactuated multi-fingered mechanical hand with modular selfadaptive gear-rack mechanism, Department of Mechanical
Engineering, Tsinghua University, Beijing, China, 2008
[3] Thierry Laliberte Lional Birglen and Clement M.Gosselin,
Underactuation in robotic grasping hands, Department of
Mechanical Engineering, Laval University, Canada, 2003
[4] Peter Steutel Gert A.Kragten and Just L.Herder , Design
of a Fully compliant Under-Actuates Finger with a
Monolithic Structure and Distributed Compliance,
Biomechanical Engineering, Delft University of Technology,
The Netherlands, 2010
[5] R.Rizk S. Krut and E. Dombre, Grasp-stability analysis of
a two-phalanx isotropic underactuated finger, LIRMM,
University Montpellier 2, France, 2007
[6] Clement Gosselin Frederic Pelletier and Thierry Laliberte,
An Anthropomorphic Underactuated Robotic Hand with 15
Dofs and a Single Actuator, IEEE International Conference
on Robotics and Automation, Pasadena, USA, 2008
[7] Aaron M.Dollar, Simple, Robust Autonomous Grasping in
Unstructured Environments, Havard University, Cambridge,
USA, 2007
[8] M.R. Cutkosky, Development of a Self-Adaptive Prosthesis
for Restoring Natural Grasping, Design Division,
Department of Mechanical Engineering, Stanford
University, 2004
[9] Guoxuan Li Bowen Li Jie Sun Wenzeng Zhang
Zhenguo Sun and Qiang Chen, Devolopment of a Directly
Self-adaptive Robot Hand with Pulley-belt Mechanism,
Mechanical Engineering, Tsinghua University, Beijing,
China, 2013

207

[10] Hebin Liu Wen Zhen and Lv Hai Feng, A survey on

Underactuated Mechanism, Shanghai Key Laboratory of
Mechanical Automation and Robotics, Shanghai, China,
2010
[11] Springer Verlag Berlin Heidelberg, Underactuated Robotic
hand, ISBN 978-3-540-77458-7, 2008

 2555
2555 6 .. 2556
08:30 12:10 . 3 3 12:00 13:30 . 3
08:30 08:55
08:55 09:20
09:20 09:45
09:45 10:10
10:10 10:30
10:30 10:55
10:55 11:20
11:20 11:45
11:45 12:10

Room 301
AM11
AM07
AM03
AU02
AU06
TF03
TF07

Room 304
Room 305
Room 306
AM10
AM09
AU05
AM06
AM05
AM04
AM02
AM01
AU01
AU03
AU04
AM08
Coffee Break @ Room 309
AU07
TF01
TF02
TF04
TF05
TF06
TF08
MS09
MS08

Applied Mechanics
AM01

AM02 The Study and Experimental Investigation of Blade

Vibration Effect of LVAD Model
AM03 Investigating on Damping Property of Laminated
Composite Materials
AM04
AM05
AM06 Design and Construction of a Test Rig for the CablePulley Power Transmission System
AM07 Improving Material Removal Rate by "Waterjet Milling"
AM08

AM09

AM10 Design and Manufacture Argumentative Mobility Device

for Osteoarthritis Patients in Thailand
AM11

Automotives
AU01
AU02 HCCI DME
AU03
Driver Eco Rating
AU04


AU05 The Possibility of Installing Extra Battery to Improve
Hybrid to EV-like Vehicle in Bangkok Traffic

208

Room 308
MS03
MS02
MS01
MS04
MS05
MS06
MS07
MS10

AU06 Hardware-in-the-loop simulator

,
Student Formula
AU07 Design, Analysis and Testing of Semi-Monocoque
Frame for Student Formula 40
Student Formula

Thermo-Fluids
TF01
TF02

TF03
19-25
TF04

TF05
ASHRAE .. 2556
TF06

TF07

TF08

Master Program-Oriented Projects

MS01

MS02

MS03 Steer-By-Wire System Evaluation using Fitts Law

MS04
GPS

MS05

MS06
MS07

MS08
POD

MS09

MS10

CU Engineering - Isuzu Research Fund

..
3,500

The Study and Experimental Investigation of Blade Vibration Effect of LVAD Model

..

2,500

,
..
..
750
Steer-By-Wire System Evaluation using Fitts Law

..
..
750

,
..
1,500

,
..

1,000
209



..
TF02
.. AM11
..
AM12, MS04, MS05
..
AM02, AM09
..
AM14, MS06, MS07
.
AM05
..

TF04, TF06, MS13
..
AM08
..
AM07
..
AU02, AU05, AU06
..

TF05
..
AM03, AM09, AM10
..
MS08
..
AM04, MS09, MS10
..
MS14, MS15
..
AM06
..
AM13, AM15
..
TF03, MS11, MS12
..

AU02, AU03, AU07
..

AM01, MS03
..
TF01, MS01, MS02
.. AU01, AU04

MS08
AU03
MS01
TF02
AM11
MS09

..

..









..







..

..


..








..




..

210

MS04
AM11
AU07
AM09
AM06
AM09
AM12
AM11
AU03
AM07
MS06
AU07
AM07
AM07
AM10
AM09
TF03
AU05
TF03
AU05
AU02
AM01
AM01
AM02
MS11
AM02
MS12
AM04
TF05
AM13
AM06
AM03
AM03
MS07
AM08
AM08
MS10
AM04
AM12
MS14
AM10
AU07
AM08
AM10
AU06

..



..

..


..


..

AM03
AM06
TF03
TF06
AU02
AU06
MS13
AM12
TF01
AM02
TF02
TF02
TF04
AM15
TF04
TF04
MS02
TF06
MS03
AM04
AM05
AM05
TF06
AU06
MS15
AU04
AM05
AU04
AU01
AU05
AU01
AU01
MS05
AM15
TF01
AM13
TF05
TF01
AU04
TF05
AM14
AM14