Vous êtes sur la page 1sur 11

Engineering 9516

Golf Tournament Report

Group F
Amanda Ryan
Pragyandeep
Sahoo
Younes Salami
Vahit Saydam
Peter Seifert

December 3, 2015

200903466
201299955
201498359
201581030
200930246

ENGI 9516 Golf Tournament Report Group F

Executive Summary
To be written after report is completed.

ENGI 9516 Golf Tournament Report Group F

Table of Contents
Executive Summary...............................................................i
1.0 Overview.......................................................................1
2.0 Design of Experiment.....................................................1

2.1 Long Distance Model............................................................1


2.2 Short Distance Model...........................................................1

3.0 Results..........................................................................1

3.1 Long Distance Model............................................................1


3.2 Short Distance Model...........................................................1

4.0 Recommendations..........................................................1
Appendices..........................................................................3

List of Figures and Tables


Figure 1..................................................................................................2
Table 1....................................................................................................2

ENGI 9516 Golf Tournament Report Group F

Overview

PRAGY
Review what the project was (explanation of project, goals, etc.)
List materials we used.

Design of Experiment

AMANDA
Trial runs to decide on models to go forward with

1.1 Long Distance Model


PETER
Review details of design.
Using face-centred CCD for LDM.
Discuss what factors we used, levels, etc. Check # of centre points
Refer to appendix for run sheets of model.

1.2 Short Distance Model


AMANDA
Review details of design.
2-level factorial with 2 factors. Angle and Direction.
# centre points, replicates.
Refer to appendix for run sheets of model.

2.0

Results

AMANDA
Results of tournament.
How well our model worked, etc.

2.1 Long Distance Model


YOUNES
Design Expert results.
Include screenshots of all figures, tables, etc as output from Design Expert.

ENGI 9516 Golf Tournament Report Group F

2.2 Short Distance Model


East-West Model
Anova table for the short distance model in the east-west direction is formed in Design
Expert as follows;

Table 1: Anova Table


Anova table shows that angle is the only significant effect. Direction does not have
significant effect on the distance. It is seen that lack of fit is significant. This stems from
the fact that the contours of the floor has enormous influence on the distance.
Particularly, uneven parts of the floor affected the short shots even more. It was clearly
observed during the collection of the data.
The R-squared values seem to be in reasonable agreement. Adequate precision
value,15.415, is greater than 4 which is desired.
Now, Anova assumptions defined below are to be studied through the plots taken from
Design Expert.

Normality assumption: The experimental data should follow the normal distribution.
Constant variance: The experimental data has constant variance. There should be no
enormous deviations.
Independence: The experimental data should be collected independently and randomly.
There should not be any patterns.

Figure1: Normal plot

Figure 2: Residuals vs Predicted

Figure 1 indicates that the data obtained follow the normality line verifying
normal distribution. In figure 2, we see that the data points are within the
margin which confirms the equal variance assumption.

Figure 3: Residuals vs Run

Figure 4: Predicted vs Actual

Figure 3 presents that the data are randomly scattered without any pattern
while figure 4 demonstrates the data points that are evenly split by 45 degree
line.
Box-Cox plot shown on the left
indicates that we need log
transformation
in
order
to
eliminate the significant lack of fit.
However, due to the reasons
mentioned in the interpretation of
the
Anova
table,
log
transformation could not rule out
significant lack of fit.
Figure 5: Box-Cox plot

Figure 6: Interaction plot, Angle


The interaction plot shows that as we increase the angle, the distance also
increases exponentially for both direction. However, the distance of shots
made in the west direction seem to be more than the shots made in the east
direction.

Page 1 of X

Table 2: Equations in terms of coded and actual terms


As can be seen from the table 2, we have 2 equations for actual factors in
both direction. The coefficients of both equations are close to each other
owing to the insignificant effect of direction.
North-South Model
Anova table for the short distance model in the north-south direction is formed in
Design Expert as follows;

Table 3: Anova table


Anova table shows that our model is significant. Angle is the only significant
factor since its p-value is less than 0.05. It is seen that lack of fit is significant.
This is caused by the same reasons that were explained in the east-west model.
R-squared values are in accordance with each other while adequate precision
value is greater than 4.
Now, Anova assumptions defined below are to be studied through the plots taken
from Design Expert.

Normality assumption: The experimental data should follow the normal


distribution.
Constant variance: The experimental data has constant variance. There should
be no enormous deviations.
Independence: The experimental data should be collected independently and
randomly. There should not be any patterns.

Page 2 of X

Figure 7: Normal plot


Predicted

Figure 8: Residuals vs

Figure 7 indicates there are no significant deviations from the normality line while
in figure 8 all the data points are within the acceptable margin without outliers.

Figure 9: Residuals vs Run


Actual

Figure 10: Predicted vs

Figure 9 indicates the randomly scattered data points. Predicted vs actual plot
shows the evenly split data points by 45-degree line.

Figure 11: Box-Cox plot

Page 3 of X

Box-Cox plot indicates the need for square root transformation due to the
significant lack of fit.
One factor plot of angle shown on
the left presents that the increase in
angle leads to exponential rise in the
distance.

Figure 12: One factor, angle

Page 4 of X

Table 4: Equations in terms of coded and actual factors


Table 4 above gives the model equations. The coefficients for both directions are
quite alike.
VAHIT
Design Expert results
Include screenshots of all figures, tables, etc as output from Design
Expert.

3.0

Recommendations

PRAGY
On what we would have done differently looking back.
Table 1
Figure 1

Appendices
Run sheets

Vous aimerez peut-être aussi