CATAPULT PROJECT
Group Members:
• Abdullah Amini
•
•
•
•
•
Kia Vakili
Pedram Karam Beigi
Ramit Shrivastav
Riyanka Daga
Roozbeh Zad
Professor: Jay Hamade
May 8th, 2012
D
M
A
I
C
DEFINE
MEASURE
Problem
statement
Problem
objective
sipoc
metrics
ANALYSE
IMPROVE
CONTROL
PROBLEM STATEMENT
• 35% of our Middle East customers that are currently using the latest
Catapult-Forza, are returning the product launched in January 2012
for their military training purposes because the distance travelled by
the ball doesn’t meet their requested specification range of 200 +- 2
inches. Resulting in a negative profit impact of $10M and reducing
market share around 20%.
PROJECT OBJECTIVE
• Reducing the product returns from our middle east customers from
35% to 15% by the end of May 2012 to save $5.7M, by modifying
and revising the hardware and functionality of the Catapult-Forza ,
such that it can meet customers shooting specification range (200 +-
2 inches) in every attempt.
SUPPLIER
INPUT
Ohio Willow
Wood
Wood
Hercules Rubbers
Rubber Bands
Bolt Depot
Screws & Bolts
Woodcraft
Blueprints
Archbold Co
Labor
Dewalt
Tools
1. Cut The Wood
As Per Dimensions
2. Drill Holes In The
Wood As Per
Dimensions
3. Join The Sides and
The Arm To The Base
4. Punch In The Screws
and The Bolts
5. Fit The Rubber
Bands
6. Fix The Ball Holding
Shell
OUTPUT
CUSTOMER
Catapult with
Desired
Specifications
Military
Training
Academy
PRIMARY AND SECONDARY METRICS
•
Primary Metric is used to measure process performance and is the
gage used to measure success.
• In this case distance travelled by the ball is our primary metric
•
Secondary Metrics is the vertical distance from the location of our
catapult to the floor
DEFINE
MEASURE
ANALYSE
Gage R&R
Normality test
Capability test
metrics
IMPROVE
CONTROL
GAUGE R&R ANALYSIS REQUIREMENTS
 5 Different Parts (Shoot by Catapult)
1. Black Stone
2. Marble Ball
3. White Stone
4. Paper Clip
5. Gray Stone
 3 Different Operators (Measuring the Distance)
 Randomized Reading
GAGE R&R GRAPHICAL OUTPUT
The following charts are the result of running Gage R&R study for the collected data
(measurements) by operators.
Gage R&R (ANOVA) for Results
G age name:
Date of study :
Reported by :
Tolerance:
M isc:
03/06/12
Components of Variation
Results by Parts
100
% Contribution
240
Percent
% Study Var
220
50
200
0
Gage R&R
Repeat
Reprod
Black-stone
Part-to-Part
paper-clip
R Chart by Operators
Sample Range
Zad
Abdulla
Pedram
white-stone
Parts
grey-stone
marble-ball
Results by Operators
UCL=12.20
240
10
5
_
R=4.74
0
LCL=0
220
200
Zad
Xbar Chart by Operators
Zad
Abdulla
Pedram
Abdulla
Operators
Pedram
Operators * Parts Interaction
240
240
220
UCL=215.53
_
X=210.68
200
LCL=205.83
Average
Sample Mean

Operators
Abdulla
Pedram
220
Zad
200
Black-stone
paper-clip
white-stone grey-stone
Parts
marble-ball
GAGE R&R ANALYSIS

Components of Variation
“Part-to-Part“ variation is 98.32% .
Repeatability and Reproducibility together have a total of 1.67% of variation.
This is an ideal result which shows the accuracy and consistency of operators’ measurements.

R-Chart by Operators
It shows all the measurements performed by different operators.
Most measurements that were recorded were very close to the average.

X-Bar Chart by Operators
The above X-Bar chart shows that some points are inside the control limits. This means these parts
variations (third and fourth parts) are not easy to detect. This chart shows our measurement system is
making it difficult to measure part to part differences for part three and four for operator one and two but
for operator three just part three is inside the limit and difficult to measure .
GAGE R&R ANALYSIS

Results by Parts
The measurements that were taken should vary little from each other.
Most measurements that were recorded were very close to the average.
The Marble ball readings were the most accurate measurements recorded compared to other parts.

Results by Operators
The above chart shows the measurement of each part by each operator. In this case the total number of
measurement is 15 (5 Parts x 3 Times).
The variations between the measurements of each operator is different. The averages are varying for all 3
Operators. Ideally, the variation in measurement of each operator must be the Same.
Reasons are Human Errors, Reduction in Elasticity Of The Rubber Band, Instrument Related Errors, Setup
For Measurements

Operators / Parts Interaction
Average measurement taken by each operator on each part
The variation in the measurement is very low
NORMALITY TEST

Select one part for the Normality Test – Marble Ball

Shoot the ball 30 times from the catapult
Probability Plot of Result
Normal
99
Mean
StDev
N
AD
P-Value
95
90
Percent
80
70
60
50
40
30
20
10
5
1
232
234
236
238
Result
240
242
244
237.7
2.150
30
0.958
0.013
PROCESS CAPABILITY ANALYSIS
Process Capability of Result
Calculations Based on Weibull Distribution Model
LSL USL
P rocess Data
LS L
198
Target
*
USL
202
S ample M ean 237.677
S ample N
30
S hape
117.598
S cale
238.751
O v erall C apability
Pp
0.24
PPL
3.25
PPU
-7.82
P pk
-7.82
E xp. O v erall P erformance
P P M < LS L
0.00
P P M > U S L 1000000.00
P P M Total
1000000.00
O bserv ed P erformance
P P M < LS L
0.00
P P M > U S L 1000000.00
P P M Total
1000000.00
198
204
210
216
222
228
234
240
METRICS
250
225
200
175
150
125
100
75
50
25
0
1
2
3
4
5
6
7
8
9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Results
LSL
USL
DEFINE
MEASURE
ANALYSE
IMPROVE
Process Map
Fishbone diagram
C&E Analysis
FMEA
CONTROL
DETAILED PROCESS MAP
Arm Holder
Base
Arms
Object Holder
Cut the Wood
as per
Dimension
Input
Arm Holder
Base
Arms
Drill Holes
Input
• Wood C
• Base C
• Tools C
• Arms C
• Blue Prints S
• Arm Holder C
• Operator N
• Object Holder C
• Supplier S
• Tools C
• Blue Prints S
• Operator N
Partially
Assembled
Catapult
Assembly of
Arms and Base
Input
• Arms C
• Base C
• Arm Holder C
• Position of Pin
on Fixed Arm C
• Position of Pin
on Moving Arm C
• Tools C
• Blue Prints S
Catapult-Forza
Fix the Rubber
band & Install
Angel
Measurement
Input
• Assembly of
Catapult C
• Rubber band N
• Nuts S
• Bolts S
• Operator N
• Angel of moving
Arm C
FISHBONE DIAGRAM
TOP 3 CAUSES
1. Angel of Moving Arm
2. Position of Pin on Fixed Arm
3. Position of Pin on Moving Arm
C & E MATRIX
1
10
1
2
3
4
8
9
10
11
12
13
14
15
catapult
7
116
1
10
10
10
121
blueprint
tools
partially assembled catapult
1
1
1
1
5
5
5
1
57
17
1
1
10
5
62
1
1
1
1
5
10
10
5
5
10
1
5
57
112
22
57
61
580
19
20
Fix rubberband and angular measurment
rubber band
angle of moving arm
nuts and bolts
operators
Total
Lower Spec
Target
Upper Spec
0
10
0
5
0
10
assemble arms and base
0
1
Drill holes
0
53
13
21
26
22
22
62
21
21
17
21
17
57
17
17
57
0
5
1
1
1
1
1
5
1
1
1
1
1
5
1
1
5
0
1
1
1
1
1
1
1
1
1
1
5
1
5
5
5
5
0
1
1
5
5
1
10
1
5
5
1
1
1
1
1
1
1
0
1
1
5
10
10
1
10
5
5
5
5
5
1
1
1
1
0
wood
tools
blueprint
operators
suppliers
Base
arm
arm holder
tools
operators
blueprint
shell
arm
base
arm holder
operators
position of pin in stationary
arm
position of pin on moving arm
0
cut the wood as per dimension
63
Total
Process Input
18
22
23
24
25
6
Process Step
17
21
5
71 1 1 1 1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
partially
assembled
catapult
1
unassembled
arm,base,shell
Process Step
1
Arms base and
sides with holes
Rating of Importance to
Customer
FMEA ANALYSIS
Process Step/Input
What is the process step/input under
investigation?
GROUP 3
Potential Failure Mode
Potential Failure Effects
In what ways does the input What is the impact on the
go wrong?
Output Variables
(Customer Requirements)
or internal requirements?
S
E
V
O
C
C
Potential Causes
What causes the input to go
wrong?
How often does cause of FM occur?
Responsible:
Prepared by:
GROUP 3
FMEA Date:
(Orig.) 04/06/2012
(Rev.)
Current Controls
Prevent
Detect
What are the existing controls and procedures (inspection and test) that
prevent/detect either the Cause or Failure Mode?
Should include an SOP number.
D
E
T
R
P
N
How well can you detect cause or
FM?
CATAPULT
How sever is the effect to the
customer?
Process or Product Name:
Actions Recommended
What are the actions for reducing the
occurrence of the Cause, or improving
detection? Should have actions only on
high RPN’s or easy fixes.
Position of the pin on stationary arm
Inappropriate Position of
the pin
Distance wanted not
achieved
10
Wrong hole selected for the
pin
5
Inspecting to avoid errors in
specifications
Test and find the location of
the pin
3
150
Testing and Finding appropriate pin
position
Position of the pin on moving arm
Inappropriate Position of
the pin
Distance wanted not
achieved
10
Wrong hole selected for the
pin
10
Inspecting to avoid errors in
specifications
Test and find the location of
the pin
3
300
Testing and Finding appropriate pin
position
Angle of the moving arm
Improper selection of angle
Distance wanted not
achieved
10
Wrong angle selected
10
Inspecting to determine the angle of
operation
Test and find the appropriate
angle
5
500
Testing and Finding appropriate angle
Page
Responsible
Who is responsible for the recommended
action?
Actions Taken
What are the completed actions taken with the recalculated RPN? Be sure to
include completion month/year.
Testing and Trouble Shooting Team
Testing and Trouble Shooting Team
Testing and Trouble Shooting Team
Fixed position of the pin is selected by testing on 04/08/2012
Fixed position of the pin is selected by testing on 04/08/2012
Fixed angle of operation is selected by testing on 04/08/2012
1
of
1
S E V O C C D E T R P N
10
10
10
1
1
5
1
1
1
10
10
50
DEFINE
MEASURE
ANALYSE
IMPROVE
DOE
Interaction Plot
Pareto Chart
Equation
CONTROL
PARETO CHART OF THE EFFECTS
•
•
•
•
Minitab displays the absolute value of the
Effects on the Pareto Chart
The Chart shows which Effects are active
meaning which effects are affecting the
distance
The plot shows that Position of the Pin on
Stationary Arm is active
Chart also shows the interaction
between other factors
INTERACTION PLOT FOR RESULTS
• This Graph helps us look at the
Significant Interaction between the 3
sources of error
• It tells us how big each effect is
• Here, In order to get highest yield from
our experiment, angle should be set to
point 4, position of the stationary pin
should be set to 3 and position of the
pin on the moving arm to 3
NORMAL PLOT OF THE EFFECTS
•
•
The Normal Plot and Pareto Chart shows
which effects influence the yield
The graphs shows all the points are outside
the fitted line hence active
MAIN EFFECTS PLOT FOR RESULTS
• The Plot shows the effects of changing angle
and the positions of the pins on the stationary
and moving arm
• Here we can see that the Positions of the Pins
on the Stationary Arm has the major effect on
achieving the target spec and then the
Position of the Pin on Moving Arm
Factors
Size Of Interpretation
Effect
Angle
+15.17 Runs at 4 had better yields
than runs at 2
Pin
+57.29 Runs at 3 had better yields
Stationary
than runs at 1
Pin
Moving
+33.45 Runs at 3 had better yields
than runs at 1
CUBE PLOT FOR RESULT
• From The Cube Plot, In order To Get The
Desired Specification Of The Distance i.e.
78.74 inches,
 The Angle should be set to point 4
 The Position of the Pin on the Fixed
Arm should be between 1 & 3
 The Position of the Pin on the
Stationary Arm should also be
between 1 & 3
OPTIMIZATION PLOT
• As the name suggest, the plot gives the
combination of effects for optimum
efficiency, i.e. to meet the desired
specifications
• In our case,
 The Angles should be set to point 4
 The Position of the Pin on the Fixed
Arm should be at point 1.7822
 The Position of the Pin on the
Stationary Arm should be at point
1.4021
FORMULA
Y = F(X)
Target Value = F (Position of Pin On Stationary
Arm) + F (Position of Pin On Moving Arm) +
F (Angle Of Moving Arm)
78.74 = F(1.7822) + F(1.4021) + F(4.0)
DEFINE
MEASURE
ANALYSE
IMPROVE
CONTROL
IMR Chart
X Bar – R chart
Normality Plot
Conclusion
BLUE PRINT WITH MODIFICATIONS
1.7822
1.4021
4
I-MR CHART
PROCESS NORMALITY PLOT
The Normality plot shows a scatter plot of the measurements and the line of best fit. More
points are on and closer to the line of best fit comparatively. The P-Value is 0.703 which
proves our distribution is normal.
PROCESS CAPABILITY PLOT
The Process Capability test shows that the Cpk is 0.51 which is better than our previous Cpk
value. However the process is still not capable .
Xbar-R CHART
•
The X bar chart shows
 The points are the average of each
subgroup
 The red control limits which shows the
process in under control as none of the
points are outside the UCL and LCL
 The green line is the overall average
which is the mean X bar which is
78.537
•
The R bar chart shows
 The points as difference in the largest
and the smallest measurement within
each sub group
 The green line is the grand average of
each points which is the mean R bar
which is 0.753
 The red lines are the upper and lower
control limits and here none of the
points are outside the UCL and LCL
CURRENT SIGMA LEVEL
• Sigma Level = Cpk x 3
• Current Cpk = 0.51
• Current Sigma Level = 0.51 x 3 = 1.53
CONCLUSION
• All the Data points in the Xbar and R chart are
within the UCL and LCL hence the process is in
control.
• The Cpk of the process has increased from 7.82 to 0.51.
• Further Analysis is required to increase the
Cpk and Sigma level.