You are on page 1of 41

DESI GNI NG AND EXPERI MENT TO

MEASURE THE FACTORS AFFECTI NG


THE FLI GHT OF A PAPER AI RPLANE
DESIGN AND ANALYSIS OF EXPERIMENTS MODULE
PROJECT


Presented To:
- Dr. Tamer Mohamed Adel
- - Eng. Malik Wagih
Presented By:
- Ahmed Khaled Etman
- Hazem Ali
- Karim Mohamed Khaled Hamouda 110253
- Mohamed Ahmed Bakr
110864
- Mohamed Khaled Hasseeb 110878
- Mohamed Mohamed Ashraf Salama 110573
ACKNOWLEDGEMENT:



























ABSTRACT:



























TABLE OF CONTENTS
Acknowledgement: ....................................................................................................... 2
Abstract: ......................................................................................................................... 3
List of Figures: ............................................................................................................... 5
List of Tables: ................................................................................................................ 6
1. Introduction: ........................................................................................................... 7
1.1. Project Aim: ................................................................................................................ 7
1.2. Problem Statement: .................................................................................................. 7
1.2.1. Choice of Factors: ......................................................................................................................... 7
1.2.2. Choice of ResponsE: .................................................................................................................... 8
2. Experiment Description: ........................................................................................ 9
2.1. Objectives: .................................................................................................................. 9
2.2. launching Device: ...................................................................................................... 9
2.3. Procedures: ............................................................................................................... 11
2.4. Sources of Errors: ..................................................................................................... 14
2.4.1. Improvement: .............................................................................................................................. 15
3. Data Analysis: ....................................................................................................... 16
3.1. Background: ............................................................................................................. 16
3.2. Hypothesis: ............................................................................................................... 19
3.3. Software Results: ..................................................................................................... 20
3.3.1. Machine 1: .................................................................................................................................... 21
3.3.2. Machine 2: .................................................................................................................................... 28
3.4. Data Verification: ..................................................................................................... 33
3.4.1. Machine 1: .................................................................................................................................... 33
3.4.2. Machine 2: .................................................................................................................................... 36
4. Conclusion: ............................................................................................................. 1
4.1. Data Interpretation: .................................................................................................. 1
4.2. Sources of Error: ........................................................................................................ 1
4.3. Recommendations: .................................................................................................... 2
Bibliography .................................................................................................................. 3









LIST OF FIGURES:
Figure 1 the making of the launching device ............................................................................................. 10
Figure 2 Devices tested ....................................................................................................................................... 10
Figure 3 the two sizes of the paper airplane with the centre of gravity at the tail ...................... 11
Figure 4 adjusting the angle on 15 degrees and 30 degrees respectively ..................................... 12
Figure 5 a combined small and large sizes trial with CG in the middle of the paper airplane 13
Figure 6 launching device attached to the square boxes for better stability ................................. 15
Figure 7 NORMAL PLOT OF RESIDUALS ...................................................................................................... 24
Figure 8 Residuals Vs Predicted Plot ............................................................................................................. 24
Figure 9 Residuals Vs size ................................................................................................................................. 25
Figure 10 Residuals Vs Angle............................................................................................................................ 26
Figure 11 Residuals Vs center of gravity ...................................................................................................... 26
Figure 12 Interaction AC ..................................................................................................................................... 27
Figure 13 Interaction AB ..................................................................................................................................... 27
Figure 14 Interaction BC ..................................................................................................................................... 28
Figure 15 Normal plot of residuals ................................................................................................................. 30
Figure 16 Residuals Vs. Predicted ................................................................................................................... 30
Figure 17 Interaction AB ..................................................................................................................................... 31
Figure 18 Interaction AC ..................................................................................................................................... 31
Figure 19 Interaction BC ..................................................................................................................................... 32








LIST OF TABLES:
Table 1 Small Sized Data .................................................................................................................................... 12
Table 2 Large Sized Data .................................................................................................................................... 13
Table 3 Test statistics outline ............................................................................................................................ 18
Table 4 Analysis of variance table ................................................................................................................... 21
Table 5 Analysis of variance table 2 ............................................................................................................... 22
Table 6 Analysis of variance table 3 ............................................................................................................... 28
Table 7: Machine 1 data output....................................................................................................................... 33
Table 8:The analysis of Variance table .......................................................................................................... 35
Table 9: The p values ........................................................................................................................................... 35
Table 10:Machine 2 data output ..................................................................................................................... 36
Table 11:The analysis of Variance table ........................................................................................................... 0
Table 12: The p values ............................................................................................................................................ 0
Table 13 Recommentadtions ............................................................................................................................... 2















1. INTRODUCTION:

1.1. PROJECT AIM:
The aim of this project is to design an experiment that can measure the effect of several
factors on the range of flight and stability of a paper airplane. The experiment should include
all the factors the affect the mentioned responses, and have statistical significance.


1.2. PROBLEM STATEMENT:
Paper air planes are easy and cheap to construct, it cannot be fixed because damaged parts
cannot be replaced, so if any part gets damaged it is easier to replace it than to fix it. Paper
airplanes can be used in numerous beneficial applications, to help do hard and relatively
precise work. It can be used in dangerous zones in which catastrophes had occurred, after
earthquakes, floods, fires, avalanches, and other disasters to search for injured people,
casualties, and survivors that is to keep firemen and lifeguards safe from death. Paper
airplanes can be used in spying and warfare surveillance as well, to discover enemy locations
and for field reconnaissance without being discovered due to their small sizes, and
manoeuvrability. But this sort of airplanes requires motors, batteries and control surfaces to
control the aviation and directing of the plane. In the studied models neither propellers nor
rotorcraft, it has a fixed wing design so it is affected by various factors which affect the flight
of the plane making it obligatory to study these factors.

1.2.1. CHOICE OF FACTORS:
The factors which affect the flight performance of the plane are air foil cross-section, aspect
ratio of the wing, and location of the centre of gravity. The selection of factors to study was a
hard decision to make due to the variety of factors. But it was decided to study size, launch
angle, and the location of centre of gravity. It would have been better to take aerodynamics as
a factor because it would give more precise and significant readings, but it was found to be
very hard to calculate so it was replaced by size of the plane, because it is stated that size is a
variable which plays an important role in the aerodynamic equations. Since the plane cannot
be considered a projectile because aerodynamic forces play an important role, tests for the
effect of the launch angle must be conducted to determine the effect of changing the lift
force. This is due to the fact that the lift force is directly proportional to the angle of attack
which in this case can be considered equal to the launch angle. The location of the centre of
gravity is a very important factor in aeronautics, and is very easy to locate and manipulate.

1.2.2. CHOICE OF RESPONSE:
Choice of response was based on the measurable effects that occurred during the preliminary
testing, the possible responses were first outlined which were the flight time, range, and flight
stability. The flight is considered to be the time measured between the moment the plane
departs from the device until the moment it touches the ground. However the flight time was
measured to be in the order of fractions of seconds and was hard to record. The range is one
of the most easily measurable and most meaningful responses. Any change in the above
selected factors should greatly affect the range. The range will be defined as the distance
between the airplane and the launcher on the horizontal axis. The final response which is the
stability could be a highly effective response. However no effective method for measuring the
stability exists. Therefore, it was decided to measure the stability as deviation or the distance
between the landing site and the horizontal axis.




















2. EXPERIMENT DESCRIPTI ON:

2.1. OBJECTIVES:
The main objective of this experiment is to test the effect of some controlled variable on the
desired parameters to be measured. In the case of the paper airplane, the set of these control
variables was as follows:
1) Size of the paper airplane.
2) Centre of gravity.
3) Launch angle.
These variables were set in order to experiment the effect on both the paper airplanes range
and stability as the main responses of these parameters.

In order to eliminate the human error or at least mitigate it as possible, a launching device
was needed. In addition to that, using such a machine will provide ease and accuracy when
controlling the main affecting variables.

2.2. LAUNCHING DEVICE:
A very simple device was made of available components at any stationery. Though
controlling the size and centre of gravity is possible by modification on the paper airplane
itself, the launching angle and launching force needed to be controlled via this device to
ensure that all trials undergo the same conditions of throwing force and the required angle.

FIGURE 1 THE MAKING OF THE LAUNCHING DEVICE

The device must be accurate as possible to achieve its purpose and ensure a smooth launching
force, that is why three devices were made, tested and then the one with the most accurate
result was chosen. The first device was made did not deliver satisfactory performance,
however , the second and third devices were used to collect data in order to reduce the device
error as much as possible and to make sure that the data of the two devices were in the same
range and finally the most accurate is used for the final experiment.

FIGURE 2 DEVICES TESTED

The device with the red stripes (Figure 2) showed the most accurate launching surface and
relatively powerful launching force with the minimum rotor unbalance which qualified this
device to be this experiments testing device.













2.3. PROCEDURES:
The experiment was made to test the chosen two paper airplane sizes with the change of
centre of gravity using three different angles.



FIGURE 3 THE TWO SIZES OF THE PAPER AIRPLANE WITH THE CENTRE OF GRAVITY AT THE TAIL

It was agreed upon to use the basic design of the paper airplane as three different flying
designs were tested but did not prove worthy as the basic design. Paper airplanes were made
using a regular notebook papers which have the same dimensions as an A4 paper (in case of
the large size). As for the small size, half of the mentioned paper was used (exactly 22.5 x 16
of size) (figure 3). The angles of launch were set to be 0, 15 and 30 respectively and the
centre of gravity was once at the middle of the paper airplane, once at the front and finally at
the tail of the paper airplane.


FIGURE 4 ADJUSTING THE ANGLE ON 15 DEGREES AND 30 DEGREES RESPECTIVELY




A 5 metres line was set exactly on collinear to the launching path in order to measure the
range (the distance covered by the paper airplane) and the stability (denoted by the deviation
from the horizontal axis of launch.

The 36 trials were divided as follows:
A) For the small sized paper airplane, the test was performed three times for each angle each
time with a different position of the centre of gravity. A total of 18 trails were recorded for
this size as follows:

TABLE 1 SMALL SIZED DATA
Size angle Centre of gravity
Size 1 (Small) 0 Middle (2 trials)
Front (2 trials)
Back (2 trials)
15 Middle (2 trials)
Front (2 trials)
Back (2 trials)
30 Middle (2 trials)
Front (2 trials)
Back (2 trials)
Total number of trials : 18

B) The same procedure was made for the large size.
TABLE 2 LARGE SIZED DATA
Size angle Centre of gravity
Size 2 (Large) 0 Middle (2 trials)
Front (2 trials)
Back (2 trials)
15 Middle (2 trials)
Front (2 trials)
Back (2 trials)
30 Middle (2 trials)
Front (2 trials)
Back (2 trials)
Total number of trials : 18

The responses were recorded in an excel sheet.

FIGURE 5 A COMBINED SMALL AND LARGE SIZES TRIAL WITH CG IN THE MIDDLE OF THE PAPER
AIRPLANE


















2.4. SOURCES OF ERRORS:
The main sources of error which affect the readings are:
Human error: mostly error in readings or slight error in placing the paper airplane at
runway.
Launching device: due to vibrations there might be a slight change in the launching
angle, fixation using 3 square boxes has been added to mitigate this error. Also two
other devices were preliminary used to provide data to check for the acceptable range
and eliminate the device error by any defects in a particular device.



FIGURE 6 LAUNCHING DEVICE ATTACHED TO THE SQUARE BOXES FOR BETTER STABILITY


Paper airplane: error in folding the paper or due to attaching the paper clip for the
centre of gravity position changing. Small paper clips have been attached to the neatly
folded paper airplane to reduce the possibility of this error to occur.
Wind: error due to wind, however, the test was performed in a closed place to
eliminate this error.




2.4.1. IMPROVEMENT:
As improvement, wind speed, different designs, elevation, more launching angle or more
parameters can be tested to get even more accurate results and to investigate the effect of
other parameters on the current responses or more responses such as duration of flight.

























3. DATA ANALYSIS:

3.1. BACKGROUND:
Statistical calculations throughout the experiment are done by the use of Stat Ease, which is a
common statistical tool designed especially for design of experiment problems. The
experiment was conducted as a multilevel general factorial design which measures the effect
of several factors on the range and stability of the paper airplane. The factors are:
1. Size (A)
2. Angle (B)
3. Centre of gravity (C)

So, the number of levels for A will be 2(a), the number of levels for B is 3 (b), and the
number of observations of C is 3(c). This means that there is a 2*3*3 general factorial
multilevel design. So, the total number of observations will be . So, the effect of
A, B, C and the interactions which are AB, AC, BC, and ABC will be tested. In order to get
the test statistic for the effects of A,B,C and the interactions, the mean square error will be
divided by the means square error. Equations will be solved using the following procedure:















Phase One:
The statistical model implies the use of several symbols and therefore a list of these symbols
and what they stand for is listed below:
a is number of levels for the factor A
b is number of levels for factor B
c is number of levels for factor C
n is number of observations

refers to the sum of the observations


Y
i
refers to the sum of observations of each level of factor A
Y
.j..
refers to the sum of observations of each level of factor B
Y
..k.
refers to the sum of observations of each level of factor C
Y
ijk.
refers to the sum of observations in each intersection between the three factors
Y
ij..
refers to the sum of observations common between levels of factors A and B
Y
i.k.
refers to the sum of observations common between levels of factors A and C
Y
.jk.
refers to the sum of observations between levels of B and C













Phase Two:
Test statistics outline


TABLE 3 TEST STATISTICS OUTLINE
Phase Three:
In which the symbols in the table are solved, when the values for SS
A,
SS
B,
SS
C,
SS
AB,
SS
AC,
SS
BC,
SS
ABC,
SS
E,
and SS
T
are achieved.
SS
T
=


SS
A
=


SS
B
=


SS
C
=


SS
AB
=


SS
AC
=

= ()
SS
BC
=

= ()
SS
ABC
=



= ()
SS
E
= ()

3.2. HYPOTHESIS:
The data can be analysed to a number of hypotheses:
- That each of the controllable factors affects the response factors (range and stability)
separately.
Sources of
Variations
Sum of Squares
Degrees of
Freedom
Mean Squares Fo
A SS
A
a-1 MS
A
Fo=MS
A
/MS
E

B SS
B
b-1 MS
B
Fo=MS
B
/MS
E

C SS
C
c-1 MS
C
Fo=MS
C
/M
SE

AB SS
AB
(a-1)(b-1) MS
AB
Fo=MS
AB
/MS
E

AC SS
AC
(a-1)(c-1) MS
AC
Fo=MS
AC
/MS
E

BC SS
BC
(b-1)(c-1) MS
BC
Fo=MS
BC
/MS
E

ABC SS
ABC
(a-1)(b-1)(c-1) MS
AB
C Fo=MS
ABC
/MS
E

Error SS
E
abc(n-1) MS
E
.....................
Total SS
T
abcn-1 .............................................
- That every pair possible of the controllable factors affects the response factors
separately.
- That all the controllable factors, measured at the same time, affect the response
factors.

















3.3. SOFTWARE RESULTS:
Stat Ease was employed in order to test and produce the ANOVA table, which will be
important for the calculation of the test significance.
0 ) ( one least at : . 0 ) ( :
0 ) ( one least at : . 0 ) ( :
0 ) ( one least at : . 0 ) ( :
0 ) ( one least at : . 0 ) ( :
0 one least at : . 0 :
0 one least at : . 0 :
0 one least at : . 0 :
1
1
1
1
1
1
1







ijk ijk o
jk jk o
ik ik o
ij ij o
k k o
j j o
i i o
H vs H
H vs H
H vs H
H vs H
H vs H
H vs H
H vs H








TABLE 4 ANALYSIS OF VARIANCE TABLE
Analysis of variance table [Classical sum of squares - Type II]

Sum of

Mean F p-value

Source Squares df Square Value Prob > F

Model 7603.92 17 447.29 1.27 0.2279 not significant
A-Size 210.56 1 210.56 0.60 0.4408

B-Angle 904.37 2 452.19 1.29 0.2810

C-CG 1481.53 2 740.77 2.11 0.1273

AB 1979.64 2 989.82 2.82 0.0650

AC 765.08 2 382.54 1.09 0.3408

BC 1200.08 4 300.02 0.85 0.4946

ABC 1062.66 4 265.66 0.76 0.5563


The test was conducted at 99% confidence level and therefore probability >F greater than
0.01 are considered to be insignificant. The test results from the ANOVA table above indicate
the selected factors are insignificant and do not affect the stability of the paper airplanes.












3.3.1. MACHINE 1:
Testing for range:

TABLE 5 ANALYSIS OF VARIANCE TABLE 2
Analysis of variance table [Classical sum of squares - Type II]

Sum of

Mean F p-value

Source Squares df Square Value Prob > F

Model 3.950E+005 17 23237.96 5.14 < 0.0001 significant
A-Size 1519.50 1 1519.50 0.34 0.5636

B-Angle 98371.06 2 49185.53 10.88 < 0.0001

C-CG 1.996E+005 2 99802.45 22.07 < 0.0001

AB 8411.87 2 4205.93 0.93 0.3983

AC 4624.44 2 2312.22 0.51 0.6014

BC 72840.29 4 18210.07 4.03 0.0047

ABC 9673.24 4 2418.31 0.53 0.7105


The test conducted at a 99% confidence level where Prob >F is larger than 0.01 is considered
to be insignificant. In our case the entire model is considered to be significant, however, some
of the model factors are insignificant. The size factor (A) is considered to be insignificant
since the Prob>F is 0.563 which is larger than the 0.01 level set. Factors (B) Angle and (C)
center of gravity which have a Prob>5 less than 0.0001 are considered to be highly
significant. Also interactions between factors are highly important to consider whether they
are significant or not. Interactions between A, AC and ABC are considered to be highly
insignificant. Although, Prob>F for interaction BC is higher than 0.01, it can still be
considered statistically significant. However, in our case it will be neglected.









Linear Regression:

Range = 192.89 + 27.68B[1] + 14.30B[2] + 8.43C[1] + 47.93C[2] - 4.89AB[1] - 7.50AB[2] -
6.12AC[1] + 9.07AC[2] - 44.87B[1]C[1] + 3.30B[2]C[1] + 33.72B[1]C[2] - 3.46B[2]C[2] +
8.64AB[2]C[1] - 5.71AB[2]C[1] - 5.71AB[2]C[1] - 6.89AB[1]C[2] + 16.51AB[2]C[2]

The linear regression mode is used to determine the relative size or impact of each factor. As
predicted, factors with the largest impact are found with factors B and C. Interactions
between factors, although were supposed to be insignificant when predicted with the
ANOVA table, have a comparatively large impact especially impact between factors B and
C.




















PLOTS:


FIGURE 7 NORMAL PLOT OF RESIDUALS

Normal Plot of residuals is used to test the normality of the data. The data appears to be
mostly normal although an outlier lies on the far right corner of the graph. The model can be
improved by repeating the data point which represents the outlier or neglecting it.




FIGURE 8 RESIDUALS VS PREDICTED PLOT
Design-Expert Software
Range
Color points by value of
Range:
420
45
Externally Studentized Residuals
N
o
r
m
a
l

%

P
r
o
b
a
b
i
l
i
t
y
Normal Plot of Residuals
-3.00 -2.00 -1.00 0.00 1.00 2.00 3.00 4.00
1
5
10
20
30
50
70
80
90
95
99
Design-Expert Software
Range
Color points by value of
Range:
420
45
Predicted
E
x
t
e
r
n
a
l
l
y

S
t
u
d
e
n
t
i
z
e
d

R
e
s
i
d
u
a
l
s
Residuals vs. Predicted
-4.00
-2.00
0.00
2.00
4.00
50.00 100.00 150.00 200.00 250.00 300.00 350.00

The residual vs predicted plot is also another normality test. It measures the range for each
cell and then compares with all cells by checking whether the difference between the
minimum and maximum for each cell is approximately equal or not. The outlier that appeared
on the normal plot of residuals also appears on this plot. The data may be considered normal
since most of ranges are approximately equal, however, some cells may be repeated to
improve the overall accuracy of the model.




FIGURE 9 RESIDUALS VS SIZE
Design-Expert Software
Range
Color points by value of
Range:
420
45
A:Size (Paper)
E
x
t
e
r
n
a
l
l
y

S
t
u
d
e
n
t
i
z
e
d

R
e
s
i
d
u
a
l
s
Residuals vs. A:Size (Paper)
-4.00
-2.00
0.00
2.00
4.00
1.00 1.20 1.40 1.60 1.80 2.00

FIGURE 10 RESIDUALS VS ANGLE

FIGURE 11 RESIDUALS VS CENTER OF GRAVITY

One outlier exists that needs to be removed or retested. Other than that, no sever violations
exist and model can be considered normal.

Design-Expert Software
Range
Color points by value of
Range:
420
45
B:Angle (degree)
E
x
t
e
r
n
a
l
l
y

S
t
u
d
e
n
t
i
z
e
d

R
e
s
i
d
u
a
l
s
Residuals vs. B:Angle (degree)
-4.00
-2.00
0.00
2.00
4.00
1.00 1.50 2.00 2.50 3.00
Design-Expert Software
Range
Color points by value of
Range:
420
45
C:CG (Pos)
E
x
t
e
r
n
a
l
l
y

S
t
u
d
e
n
t
i
z
e
d

R
e
s
i
d
u
a
l
s
Residuals vs. C:CG (Pos)
-4.00
-2.00
0.00
2.00
4.00
1.00 1.50 2.00 2.50 3.00

FIGURE 12 INTERACTION AC




FIGURE 13 INTERACTION AB
Design-Expert Software
Factor Coding: Actual
Range (cm)
Design Points
X1 = A: Size
X2 = C: CG
Actual Factor
B: Angle = 0
C1 Front
C2 Middle
C3 Back
A: Size (Paper)
C: CG (Pos)
Paper Half paper
R
a
n
g
e

(
c
m
)
0
100
200
300
400
500
Warning! Term involved in ABC interaction.
2
2
22
22
333
Interaction

FIGURE 14 INTERACTION BC

The interactions between factor A and other factors have deemed insignificant by previous
tests, however, interactions between Factor A and Factors Band C exist, although, it is very
little and can be neglected. This can be observed in the first two interaction plots above.
Interaction between factors B and C is significant and both factors are interact a lot as it can
be viewed from the third interaction plot by the many intersections between both factors and
their levels.


3.3.2. MACHINE 2:
Testing of Range:
In order to test whether a launching device error exists or not, the test was repeated using a
different launching device operating in similar conditions. Stat ease was also used to preview
the results.

TABLE 6 ANALYSIS OF VARIANCE TABLE 3
Analysis of variance table [Classical sum of squares - Type II]
Source Sum of Squares Df Mean Square F Value
p-value
Prob > F
Model 80167.70 23 3485.55 4.24 < 0.0001 significant
Design-Expert Software
Factor Coding: Actual
Range (cm)
Design Points
X1 = B: Angle
X2 = C: CG
Actual Factor
A: Size = Paper
C1 Front
C2 Middle
C3 Back
B: Angle (degree)
C: CG (Pos)
0 15 30
R
a
n
g
e

(
c
m
)
0
100
200
300
400
500
Warning! Term involved in ABC interaction.
2
3
33
22
22
33
22
22
Interaction
A-Size 3392.03 1 3392.03 4.12 0.0450

B-Angle 29734.57 3 9911.52 12.05 < 0.0001

C-CG 1598.55 2 799.27 0.97 0.3821

AB 2081.23 3 693.74 0.84 0.4733

AC 2877.72 2 1438.86 1.75 0.1793

BC 23018.78 6 3836.46 4.66 0.0003

ABC 17464.82 6 2910.80 3.54 0.0033


Both tests agree that factor A size is insignificant to our model, however, this test shows that
factor C center of gravity is insignificant. Moreover, this test was conducted on a 99%
confidence level where prob > F has to be less than 0.01 to be considered significant. Factor
A has Prob> F of 0.045 which compared to previous results shows that the size factor now
has a more significant effect since it is now less than 0.05. Also, the effect of the interaction
between factors BC and factors ABC appear to be significant. The normality was checked
through the residual plots shown below which shows that this test lacks any outliers or
abnormalities. The test needs to be repeated with designs that neglect size or center of
gravity as a factor in order to produce more accurate models which can help solve this
predicament. However, this device broke down due to unforeseen circumstances and the
experiment could not be repeated.






PLOTS:


FIGURE 15 NORMAL PLOT OF RESIDUALS

FIGURE 16 RESIDUALS VS. PREDICTED
Design-Expert Software
Range
Color points by value of
Range:
309
116
Externally Studentized Residuals
N
o
r
m
a
l

%

P
r
o
b
a
b
i
l
i
t
y
Normal Plot of Residuals
-3.00 -2.00 -1.00 0.00 1.00 2.00 3.00
1
5
10
20
30
50
70
80
90
95
99
Design-Expert Software
Range
Color points by value of
Range:
309
116
Predicted
E
x
t
e
r
n
a
l
l
y

S
t
u
d
e
n
t
i
z
e
d

R
e
s
i
d
u
a
l
s
Residuals vs. Predicted
-4.00
-2.00
0.00
2.00
4.00
160.00 180.00 200.00 220.00 240.00 260.00 280.00

FIGURE 17 INTERACTION AB

FIGURE 18 INTERACTION AC
Design-Expert Software
Factor Coding: Actual
Range (cm)
Design Points
X1 = A: Size
X2 = B: Angle
Actual Factor
C: CG = Front
B1 0
B2 10
B3 20
B4 45
A: Size (Paper)
B: Angle (Degree)
Large Small
R
a
n
g
e

(
c
m
)
100
150
200
250
300
350
Warning! Term involved in ABC interaction.
3
2
22
22
4
2
2
4
2
22
6
Interaction
Design-Expert Software
Factor Coding: Actual
Range (cm)
Design Points
X1 = A: Size
X2 = C: CG
Actual Factor
B: Angle = 0
C1 Front
C2 Center
C3 Back
A: Size (Paper)
C: CG (Position)
Large Small
R
a
n
g
e

(
c
m
)
100
150
200
250
300
350
Warning! Term involved in ABC interaction.
2
2
2
2
2
2
2
Interaction

FIGURE 19 INTERACTION BC

















Design-Expert Software
Factor Coding: Actual
Range (cm)
Design Points
X1 = B: Angle
X2 = C: CG
Actual Factor
A: Size = Large
C1 Front
C2 Center
C3 Back
B: Angle (Degree)
C: CG (Position)
0 10 20 45
R
a
n
g
e

(
c
m
)
100
150
200
250
300
350
Warning! Term involved in ABC interaction.
22
2
2
2
32
2
2
2
2
2
2
2
2
2
2
2
2 2
Interaction
3.4. DATA VERIFICATION:
This involves solving the equations in an algebraic way, which ensures the data resulted from
Stat ease analysis. The problem is solved using the known calculations, which were already
illustrated in the background section in an algebraic way as a kind of validating of the
answers given from Stat ease software.
The validation is done on the two machines as follows:

3.4.1. MACHINE 1:

TABLE 7: MACHINE 1 DATA OUTPUT







size
Angle

0 15 30

Center of gravity Center of gravity Center of gravity
l
a
r
g
e

front center back front center back front center back
153 312 211 183 207 74 104 124 95
290 284.1 140.5 227 323 116 210 171 76
128.5 407 138.5 303 237 293 162 180 45
132 222 239 155 191 156 178 143 74
222 303 213 311 249 157 279 165 53
216 324 190.2 273 217 260 275 180 47
s
m
a
l
l

171 375.6 228.3 88 303 116 160 227 63
311 196 143 148 139 163 86 170 178
111 180 151 189 248 113 170 122 73
96 404 142.8 178 258 108 98 206 150
183 221 142 233 396 120 269 196 85
196 398 165 339 252 136 420 139 60

Where:

= ( )


Yi...
2
=( )

(
)


( ) (
) (
)
( )

( )


( )

( )


( )

( )


( )

( )


( )

() () () () () ()

( ) ( ) (
) ( ) (
) ( )

(( ) ( ) (
)) (( ) (
) ( )) ((
) ( ) (
)) (( ) (
) ( )) ((
) ( ) ( ))
(( ) ( ) (
))
( ) ( ) (
) ( ) (
) ( ) (
) ( ) (
)
( ) ( )
( ) ( )
( ) ( )
( ) ( )
( ) ( )
( ) ( )
( ) ( )
( ) ( )
( ) ( )
Then the statistical analysis can be performed easily to get Fo as follow:
TABLE 8:THE ANALYSIS OF VARIANCE TABLE

From the tables of Fo, the following are the results:
TABLE 9: THE P VALUES
Fo

0.336029
0.5636
10.8771
Significant (<0.0001)
22.07074
Significant (<0.0001)
0.930118
0.3983
0.511334
0.6014
4.027053
0.0047
0.534795
0.7105


Sources of
Variations
Sum of Squares Degrees of
Freedom
Mean Squares Fo
Size (A) 1519.5
1 1519.5 0.336029
Angle (B) 98371.06
2 49185.53 10.8771
Centre of
gravity (C)
199604.8985 2 99802.45 22.07074
AB
8411.866852 2 4205.933 0.930118
AC
4624.436296 2 2312.218 0.511334
BC
72840.28593 4 18210.07 4.027053
ABC
9673.239259 4 2418.31 0.534795
Error
406974.1783 107 4521.935
.....................
Total
802019.4655 90
.............................................
In conclusion, the effect of factor B and C are significant, which are the centre of gravity and
the angle. No interaction between two factors gives significant effect. Moreover, the effect of
the







3.4.2. MACHINE 2:

TABLE 10:MACHINE 2 DATA OUTPUT


Using the same technique as the previous analysis we will get the following

= 332868740
167224646
221840754
size
Angle

0 10 20 45

Center of gravity Center of gravity Center of gravity Center of gravity
l
a
r
g
e

front center back front center back front center back front center back
207 260 272 162 176 171 206 194 193 189 186 193
175 262 213 165 198 177 248 239 215 194 200 197
245 269 184 166 199 191 273 245 220 195 213 211
237 252 199 169 235 267 309 257 225 208 215 213
186 279 259 192 258 180 309 274 239 237 234 280
s
m
a
l
l

216 223 161 173 191 223 154 159 229 179 178 221
227 224 171 181 207 199 179 177 240 180 180 221
242 230 173 207 214 116 211 186 251 180 199 223
251 255 227 219 217 163 291 189 255 181 203 233
254 275 234 220 221 154 295 223 256 195 208 234

570354
83694422
280148164
55987722
28122940


Then statistical analysis can be performed to get Fo as follow:
TABLE 11:THE ANALYSIS OF VARIANCE TABLE

From the tables of Fo, we get the following
TABLE 12: THE P VALUES
Fo

4.12447
0.045
12.0517
Significant (<0.0001)
0.971861
0.3821
0.843544
0.4733
1.749235
0.1793
4.664866
0.003
3.539433
0.0033


In conclusion, the effect of factor B is significant, which is the angle of launch. No
interaction between two factors gives significant effect. Moreover, the effect of the
size and the centre of gravity are not significant.






Sources of
Variations
Sum of
Squares
Degrees of
Freedom
Mean
Squares
Fo
Size (A)
3392.033333 1 3392.033 4.12447
Angle (B)
29734.56667 3 9911.522 12.0517
Centre of
gravity (C)
1598.55 2 799.275 0.971861
AB
2081.233333 3 693.7444 0.843544
AC
2877.2 2 1438.6 1.749235
BC
23018.78333 6 3836.464 4.664866
ABC
17465.33333 6 2910.889 3.539433
Error
159119.7 119 822.4167
.....................
Total
78952 96
.............................................

1
4. CONCLUSION:

4.1. DATA INTERPRETATION:
The tests conducted above were meant to amplify the factors affecting the flight
performance of paper airplanes under gliding conditions with no thrust or propelling
mechanisms. The chosen model measured from the first device has concluded that the
launch angle and centre of gravity play an important role in the performance of the
paper airplane. Moreover interactions between the three factors remain in the region
of indecisiveness, where they can be rejected based on a statistical anomaly or an
unforeseen nuisance factor. The second device has concluded that the centre of
gravity is insignificant, as well as, the size, yet the size and factor interactions remain
at a probability level very close to the probability threshold. In order to improve the
model, several tests need to be conducted that include two of the three factors where
the launch angle is always a constant. This will construct a more accurate model.

4.2. SOURCES OF ERROR:


Outliers:













2
4.3. RECOMMENDATIONS:
TABLE 13 RECOMMENTADTIONS
Observation Recommendation
The flight is slightly affected by the wind
speed variation
Perform the experiment in a wind free room
The planes are made by hand, which created
a variation in the design between them
Either use a machine to make the planes, or
apply standards for the design that can be
checked to see if there is any variation.
The launching device was not fully stable Build a better device























3
BIBLIOGRAPHY
Montgomery, D. C. (2012). Design and Analysis of Experiments. Wiley.
Roy, R. K. (2001). Design of Experiments Using The Taguchi Approach. Wiley-
Interscience.

You might also like