This project concentrates on removing rivet height variation in process of rivet production at a production unit in Bangalore. Here, we apply six sigma methodology to improve the process.
3. Problem Statement: The rejection percentage
of rivets manufactured in an engineering
company (Turbo India), at Bangalore is
0.466%.
Goal/Objective: To achieve six sigma level in
the process of manufacturing rivets in order
to reduce the variation in critical height.
4. In Scope: Production, Inspection, Operator,
Performance, Machine efficiency, Raw Material
Quality.
Out of Scope: Logistics and Material
Handling.
5. Project Champion: Johanan Daniel
Project Team Leader: Muthuraman AR
Project Core Team Member: i) Divyank R Jain
ii) Apoorva
iii) Suyash Kashyap
6. High Level Process Map (SIPOC)
SUPPLIERS INPUT PROCESS OUTPUT CUSTOMERS
Akash Steels Raw Material
(Brass rods)
Manufacturing
the rivets
Rivets
according to
specification
ADM Motors
BESCOM Electricity
Mrini Tools
Supervisor CAMS
Rajiv
Lubricants
Working Fluid
(Oil)
9. Project Y (or Ys) in Y=f(x)
Variation in height of the rivet which is critical to
quality (CTQ) for the customer.
Critical Height Specification is 1.00±0.1 mm.
Y’s can not be controlled but they can only be
monitored. It is the X’s or the input factors that can be
controlled in order to achieve CTQ. So, we will try to
find out all possible X’s (factors) in the project.
Note: All dimensions are measured in millimeter (mm)
20. Theory: Measuring error could be one of the
reasons for the current DPMO and sigma
level. So, we need to test this by a proper
comparison technique.
H0: Mean difference between population is
zero.
H1: Mean difference between the population is
not zero.
24. N Mean StDev SE Mean
Manual Measure 10 1.0460 0.0151 0.0048
Machine Measure 10 1.0190 0.0137 0.0043
Difference = μ (Manual Measure) - μ (Machine Measure)
Estimate for difference: 0.02700
95% CI for difference: (0.01342, 0.04058)
T-Test of difference = 0 (vs ≠): T-Value = 4.19 P-Value = 0.001
DF = 17
Conclusion
Since, p value is less than 0.05 the null hypothesis is rejected.
Therefore, there is difference between the two populations.
Measuring error can be reduced with the use of testing machines.
25. Theory 2
Theory: Unclean collet is suspected to be the
cause of deviation in the critical height of the
rivets. So, we need to test this by appropriate
comparison technique.
H0: Difference between the mean of two
processes is zero.
H1: Difference between the mean of two
processes is not zero.
26. Theory 2: Data Collected
SL.N0 Before Cleaning After Cleaning
1 1.02 0.99
2 1.08 1.01
3 1.08 1.00
4 0.98 0.99
5 1.06 0.99
6 1.07 0.99
7 1.03 1.00
8 0.98 1.01
9 1.04 1.00
10 1.00 0.99
27.
28.
29. Two-sample T for before cleaning vs after cleaning
N Mean StDev SE Mean
Before cleaning 10 1.0340 0.0386 0.012
After cleaning 10 0.99700 0.00823 0.0026
Difference = μ (before cleaning) - μ (after cleaning)
Estimate for difference: 0.0370
95% CI for difference: (0.0087, 0.0653)
T-Test of difference = 0 (vs ≠): T-Value = 2.96 P-Value = 0.016
DF = 9
Conclusion
Since, p value is less than 0.05 the null hypothesis is rejected.
There is a difference between the two means. Therefore, collet
cleaning reduces the variability in the critical height of rivets.
31. Improvement Strategies for Proven Xs
Proven Xs (Causes): Strategies:
1) Measurement Error We try to improve measurement
standards by focusing on mechanized
Inspection of rivets.
If the measurement is done manually
highly calibrated and maintained
measurement instruments should be
used.
2) Unclean collet Since in our study we found that unclean
collet is a major reason for rivet height
deviations so, workers should ensure
cleanliness of collet regularly checked by
an appropriate time plan by the
supervisor.
3) Raw Material Defects R.M straightness is checked before
usage. Inspection chart is used for this
purpose.
32. Descriptions of Possible Solutions
(Pros and Cons)
Possible Solution: Strengths (Pros): Weaknesses (Cons):
Measurement should be done
by inspection machines.
Eliminates any kind of error in
measurement compared to
manual measures.
Skilled and technically trained
staffs required.
Machines are susceptible to
software errors.
Make a measurement
schedule in which operator
has to mention the readings.
Measurement can be ensured
at regular intervals and
effective control on them can
be established.
Time consuming.
Prepare a regular collet clean
schedule and it must be
administered by a senior
technical staff of organization.
Reduction in variation of rivet
height.
Time consuming.
Supervisor’s concentration
gets deviated.
33. Updated Flow Process Diagram
Start
Customer sends
order
Procure Raw
Materials
CAM design M/C Setting
Machining
Component
Cleaning
Inspection using
automated system
Stop
Dispatch
Check straightness
of RM Collet cleaning
34. Possible
factor
Potential
failure
Severity
Rating
Potential causes Occurrence
Rating
Control Detection
Rating
RPN
Raw Material Defective
component
6 • Bent rods
• Poor
specification &
quality
4 • Inspection & QC 2 48
Collet
(uncleaned)
Defective
component
3 • Collet not
cleaned by
worker
5 • Inspection & QC 4 60
Measurement
errors
Wrong
Measurement
8 • Wrong
measurement
tool
• Instrument not
calibrated
6 • Inspection & QC 4 192
Updated Failure Mode Effect Analysis (FMEA)
35. Raw Material Inspection Chart
• Raw material ovality is a vital reason for defects in the rivets. It can
be eliminated by careful examination of each Raw material, whether
it is straight or not.
• A Raw material inspection chart is used to monitor the straightness
of raw material.
• The chart is prepared date-wise in which quantity of raw material is
recorded by the assigned person. Then he is supposed to inspect all
the materials and record the no. of accepted quantity and the no. of
rejected quantity. Then the corresponding dates are to be signed by
the person and by the supervisor.
• This method ensures that only straightened raw materials are
accepted for manufacturing thereby reducing the no. of defectives.
The model of raw material chart is in the subsequent slide.
36. SL.NO Date Qty
(no. of rods)
Accepted
(QTY)
Rejected
(QTY)
Inspectors
Sign
Supervisor
Sign
Raw Material Inspection Chart
37. Collet Cleaning Schedule Chart
• Unclean collet is also an important reason behind the deviation
of critical height of the rivet. Due to constant rotation of brass
rods in the lathes, brass particles wear out and when dirt is
stuck in the collet during production of rivets, the Raw materials
aren't held properly between the jaws of the collet. So due to the
presence of dirt in collet the raw material cant be fed properly in
the lathe and deviation in critical height of rivets would occur.
• To avoid this the operators should clean the collet at regular
intervals (every alternate days) using kerosene and ensure that
there isn’t any presence of dirt in collet. This would reduce the
number of defectives considerably.
• Day-wise cleaning schedule chart can be maintained. In this
chart, operator’s name is mentioned with the respective
machines he is working on. The collet is cleaned by the operator
and puts a signature at the authorized place. The supervisor
reviews the same.
38. COLLET CLEANING SCHEDULE
M/C OPERATOR NAME
1 3 5 7 9 11 13 15
OS SS OS SS OS SS OS SS OS SS OS SS OS SS OS SS
GT-9
GT-10
GT-11
GT-12
GT-13
GT-14
M/C OPERATOR NAME
17 19 21 23 25 27 29 31
OS SS OS SS OS SS OS SS OS SS OS SS OS SS OS SS
GT-9
GT-10
GT-11
GT-12
GT-13
GT-14
40. Training Plans
No special operator training is required for the
control phase, they need to follow the Raw Material
Inspection Chart and Collet Cleaning Schedule to
achieve the desired sigma level.
Operators also need to make sure that Automated
Inspection System is used to measure rivet height.
Usage of Vernier Calipers should be avoided since it
leads to variation in the readings.
We sincerely believe that if mentioned plans are
followed then it will lead to high process capability
with respect to process of manufacturing rivets.
42. • From the updated process capability chart we
find the Cpk value to be equal to 1.44 which
can be used to calculate the improved sigma
level.
• Current Sigma level = Cpk ×3+ 1.5
Sigma Level = 1.44(3)+1.5=5.82
• From the updated histogram, we see the
variation is controlled within the specification
limits.
• The achieved mean is 1.00067 is much closer
to expected mean of 1.0000 compared to the
earlier mean of 1.02067
45. We learned many concepts of manufacturing processes in an
industry. Though we have theoretical knowledge about the subject,
witnessing the operations in front of our eyes in real-time was
once in a life time experience.
The project helped us better as an individual in dealing with
different people in varied number of situations. We learned to find
effective solutions to the problems that would help us in
accomplishing our future endeavors in life.
We would like to thank all the people involved directly or indirectly
in this project. Without you this project wouldn’t have been
possible.
Lessons Learned