Professional Documents
Culture Documents
2
Paths to Cost Reduction
Ticket A Ticket B
$500 round trip $650 round trip
6 hours 3 hours
3 layovers 1 layover
Arriving in NYC Arriving in
Groton/New London
What does value mean to you?
Lean Six Sigma 9
Lean Philosophy
Value
More to value than just cost
“Defined by the ultimate customer” – Womack
Voice of the Customer (VOC)
Expressed in terms of
A specific product
A function or capability
Questions
What does the customer want to buy?
What would they pay extra for?
UNDER-UTILIZED
SKILLS
Certification Package
Not Required
Minimize Rework
10 Minutes
10 Minutes
10 Minutes
30+ Minutes for order of 10
Continuous Flow
Process Process Process
A B C
12 Minutes for
order of 10
Lean Six Sigma 14
Lean Approach From: NAVSEA VSA Training
Dept 1
Batch Continuous Flow
Dept 2 DONE 4 3
IN OUT
OUT
IN IN
Dept 3
1 2
Dept 4
OUT IN IN
OUT
DONE
Disassemble
Wait Transport
Transport Wait Set-up
Machine Machine Re-Install
Remove From
Ship
Start Time
Finish
Broken Repaired
Component Component
Traditional Focus
• Improve Value-Added
work steps
• Better tools, machines,
Time
Small Amount of
instructions
LARGE amount Time Eliminated • Result: Small time
savings
of time saved
Time savings have a direct
Lean Focus
impact on
• Make all of the Value Stream visible
• Cost • Capacity
• Reduce or eliminate Non-Value-
Added portions of the process • Schedule • Flexibility
• Result: Large time savings • Resources • Etc.
Lean Six Sigma 18
Reduce variation
Y=f(X)
Making decisions based on data
Practical Problem
Pass rate for Technicial Exams was declining
Statistical Problem
Y=f(X)
Y – Scores
X
Exam section
Place of training
How often skills are used (experience)
Elapsed time since training
Statistical Solution
3 sections of the exam are the highest trouble spots
Experience is the most significant factor in passing
Practical Solution
Lean Six Sigma
Focus training on 3 areas for inexperienced technicians 24
What is “Six Sigma”?
99.99966% of values are within specifications
m
LSL USL
3s Process
A6
Is “Six Sigma” Overkill?
99% (≈3.8s) 99.99966% (6s)
20,000 lost articles of 7 lost articles of mail per
mail per hour hour
15 minutes of unsafe 1 minute of unsafe
drinking water each day drinking water per 7
5,000 incorrect surgical months
operations per week
1.7 incorrect surgical
2 short or long landings
operations per week
at most major airports
each day 1 short or long landing
11 hours of no electricity every 5 years
per month
1 hour of no electricity
Lean Six Sigma
every 34 years 26
Pipe Welding Quality
Reduce concave & convex defects in pipe butt welds
Y=f(x)
Lean Six Sigma 28
31% 30%
Government Purchased
Material Material
Shipyard
Labor
39%
Lean Six
Sigma
4δ Process
45 minutes drive for a cup of coffee
5δ Process
5 minutes drive for for a cup of coffee
6δ Process
4 steps for a cup of coffee
Is 99% still not OK?
99% Process is OK
➢ Black Belt Line of Sight can cut across geographical regions or when several
processes intersect.
Lean Six Sigma Belt Levels
YELLOW
Knows the basic tools of LSS and acts as data analysts in cooperation with GB/BB
projects. Implements quick wins and LSS improvements in short span of time
GREEN
Primary implementer of the LSS methodology. He earns this title by taking classes
in Six Sigma, demonstrating a competence on statistical tests and implementing
projects using LSS Tools
BLACK
Has sufficient skills to allow him to act as an instructor, mentor and expert to
green belts. A BB is also competent in additional LSS tool-specific software
programs and statistics.
MASTER BLACK
Generally has management responsibility for the LSS organization. This could
include setting up training, measuring its effectiveness, coordinating efforts with
the rest of the organization and managing LSS people.
Tool Application Matrix
I.
Y f (x)
Define -> y
DMA AIC
Measure -> y
(capability
Analyze -> y, x
ACTIVATED
Improve -> x BY MEASUREMENT
(vital few)
Control ->x
Measurement Power
Data in Numbers
Statistics
Information
What is a Measure?
A measure is a quantified value or characteristic:
Here are few examples
Cycle Time ( seconds, minutes, hours)
Days (number of days)
Size (length, width, height)
Dollars (revenues, sales, profits, costs, loss overtime)
Attribute counts (types of loans, gender, products, days of
the week)
Defect counts (number of errors, complaints, late items)
Good vs Poor Measures
▪ Good Measures ▪ Poor Measures
▪ a) Sufficient – available to be ▪ a) Poor or non existent operational
measured regularly definitions.
▪ b) Relevant – help to ▪ b) Difficult measures
understand/isolate the problems
▪ c) Representative – of the process ▪ c) Poor sampling
across shifts and people.
▪ d) Contextual – collected with ▪ D) Lack of understanding of the
other relevant information that definitions
might explain process variability ▪ E) Inaccurate, insufficient or non-
calibrated measurement devices
BASELINE DATA
What is a Baseline Data?
Baseline data is a measurement of the current state of a process. It is
an assessment of the capability of the process to meet customer
requirements.
Examples:
▪ It takes an average of eight (8) hours to complete the Preventive
Maintenance of a machine.
▪ Customer claims average to sold $3,000 per month.
▪ Lead Time to deliver sample mold to clients exceeds due date by
an average of three (3) days
▪ Rejection of parts produced is at 10% vs 5% standard
Collecting Baseline Measurement
Select
measure(s)
Develop a data
collection plan
s
The Standard Deviation of sample Data
σ
The Standard Deviation of population Data
S2
The Variance of sample Data
Common Statistical Notations
σ2
The variance of population data
R
The range of data
Ř
The average range of data
k
Multi purpose notation i.e. # of subgroups, # of classes
Common Statistical Notations
|y|
The absolute value of some term
>, <
Greater than, less than
≥, ≤
Greater than or equal to, less than or equal to
X
An individual value, an observation
Common Statistical Notations
x1
A particular (1st) individual value
xi
For each, all, individual values
x̅
The mean, average of sample data
μ
The mean of population data
Common Statistical Notations
p
A proportion of sample data
P
A proportion of population data
n
Sample size
N
Population sample size
Measurement Scales
Nominal
Ordinal
Interval
Ratio
Continuous and Discrete Variables
Length of a part
Number of Defects
Measures of Central Tendency and Dispersion
Precision Accuracy
Reproducibilit
Repeatability Stability Bias Linearity
y
Precision vs Accuracy
Precision and Accuracy
Introduction to Measurement System Analysis
Critical
Dimensions
MSA
Pass or Fail
Attribute Go or No Go
Customer
Service
Response
Gage R&R Studies
Total GR&R, which can be broken into Repeatability and Reproducibility
Part-to-Part, which is the variability in measurements across different parts.
Ideally, very little of the variability should be due to repeatability and
reproducibility; instead, differences between parts (Part-to-Part) should account
for most of the variability. This would be shown by:
A small %Contribution for the sources GR&R
A large %Contribution for the source Part-to-Part
The most important information is found in the %contribution, %Study Var,
%Tolerance, and %Process columns. These columns show the sources of variation.
Typically, the Total gage R&R should equal less than 30% of the study variation, and
less than 10% would be ideal.
Repeatability
P.S. Think also of a name for your coffee shop and explain the reason behind the
name of your choice.
PARETO ANALYSIS
What is a Pareto Chart?
A bar chart where the bars are arranged in descending order of
magnitude. The bars nay represent defect categories, locations,
departments, and so on. The magnitude (length) of the bars may
represent frequencies, percentages, costs or times.
A problem-solving tool that involves ranking all potential problem areas
or sources of variation according to their contribution to cost or total
variation. Typically, 80% of the effects come from 20% of the possible
causes, so efforts are best spent in these “vital few” causes,
temporarily ignoring the “trivial many” causes.
The name Pareto chart is derived from Italian economist Vilfredo Pareto
(1848-1923), who theorized that is certain economies the majority of
the wealth was held by a disproportionately small segment of the
population. Quality engineers have observed that defects usually follow
a similar Pareto distribution
Pareto Chart Analysis Workshop 1
y y
WITH GROUPS
You are interested in how well a new formulation of your
company's camera batteries is meeting customers' needs. Market
research shows that customers become annoyed if they have to
wait longer than 5.25 seconds between flashes.
You collect samples of batteries (both old and new formulations)
that have been in use for varying amounts of time. You then
measure the voltage remaining in each battery immediately after
a flash (VoltsAfter), as well as the length of time required for the
battery to be able to flash again (flash recovery time, FlashRecov).
Create a scatterplot, grouped by formulation, to examine the
results. Include a reference line at the critical flash recovery time
of 5.25 seconds.
Scatter Plot Analysis Workshop 3
We have verbal and math SAT scores and first-year college grade-point
averages for 200 students and we wish to investigate the relatedness of these
variables. We use correlation with the default choice for displaying p-values.
Open the worksheet GRADES.MTW.
REGRESSION
What is Regression Analysis?
Consequently, when there is an evident correlation coefficient calculated, we
can use the x variable to predict the possible value of the y variable. This is with
the use of the regression analysis. The term regression can be tracked with the
works of Sir Francis Galton. He noted that the sons of very tall fathers tended to
be tall, but not quite as tall as their fathers. Also, sons of very short fathers
tended to be short, but not quite as short as their fathers. He called this
tendency, regression.
Regression Workshop
As part of a test of solar energy, you measure the total heat flux
from homes. You wish to examine whether total heat flux
(HeatFlux) can be predicted by the position of the focal points in
the east, south, and north directions. You found, using best
subsets regression, that the best two-predictor model included the
variables North and South and the best three-predictor added the
variable East. You evaluate the three-predictor model using
multiple regression.
Open the worksheet EXH_REGR.MTW.
Analyze Phase Challenges
Spending the time needed to explore root causes
Getting caught up in analysis paralysis
Jumping to solution without confirming root cause hypothesis
Collecting the right data to confirm or validate hypothesis
IMPROVE
Purpose of Improve Phase
The purpose of the improve phase is to learn from pilots of the selected
solution(s) and execute full scale implementation.
Key Steps in Improve
1. Develop potential solutions (Creating Flow, Poka Yoke, Visual
Management, 5S, etc.)
2. Evaluate, select and optimize best solution. (Criteria Based Matrix,
Economic Valuations, etc.)
3. Develop and implement pilot solution.
4. Confirm attainment of project goals.
5. Develop and execute full scale implementation plan.
6. Complete the Improve Phase Toll Gate Review.
Pillars of Lean
Define
Value
Map the
Pursuit of Value
Perfection Stream
Lean
Principles
Establish Create
Pull Flow
8W APPROACH
8W Approach
Defects
Extra
Over
Processi
Prod’n
ng
Waiting
Motion
Inventory Non-
utilized
Talent
Transpo
ITEMS
8W Approach Entry errors that cause the wrong actions like shipping too many or too few to the
wrong address
The process flow should not slow because of how the process is
designed. The customer should be able to move through the
process and not get stuck.
What happens when something blocks the flow?
What happens when there is a smooth flow?
How do we create flow?
SHINE SAFETY
SORT 6S SUSTAIN
6S System
• SORT – Involves removal and/or disposal of items that are not needed at
a workstation.
• SET IN ORDER – Having a place for everything and everything in its
place.
• SHINE – Involves cleaning the workstation and equipment. Clean
workstation and equipment allows for easy inspection.
• SAFETY – A neat and clean environment will lessen the chances of an
accident occurring.
• STANDARDIZE – Involves rules to maintain sorting, setting in order,
scrubbing and safety standards.
• SUSTAIN – Involves following the 6S System and may require training
from time to time as well as audit.
SORT
SET IN ORDER/STRAIGHTEN
SHINE/SWEEP
STANDARDIZE
REGRESSION &
OPTIMIZATION
How Can the Regression Equation Help
The orchard manager has been keeping track of the weight of peaches on a day
by day basis. He wants to know if the duration of peaches on the tree affects
their weight and what might be the optimal days on tree before picking them.
5.75
Weight (Ounces)
5.50
5.25
5.00
4.75
4.50
What can you say about the relationship of the peach weight
and its duration on tree?
Validate Predictive Model Significance
How to Optimize the Response (Y)
CONTROL
Purpose of Control Phase
The purpose of the Control Phase is to complete the project work and hand off
improved process to process owners, with procedures for maintaining gains.
Key Steps in Control
1. Launch Implementation
2. Develop Standard Procedures to Lock Up Gains
3. Develop a Process Control Plan
4. Develop a Statistical Process Control Monitoring Plan
5. Finalize Project
6. Validate Performance and Financial Results
7. Complete the Improve Phase Toll Gate Review
What is a Process Control Plan
Ensures that controls and measurements are in place to maintain long term
success
Defines what the key process measures are
Helps clearly identify where process measures will take place by whom and
how often
Helps to ensure that gains will be maintained
Consist of:
Map
Monitoring Plan
Control Plan
Control Plan
CONTROL CHARTS
What is a Control Chart?
• A statistical control chart is a line graph of the measurements of a product or
process over time that has statistically based control limits placed on it.
• The points that are plotted on a control chart may be the actual
measurements of a part characteristic or summary statistics from samples
(subgroups) of parts taken as they are produced over time.
• A control chart has control limits based upon process variation and a
centerline representing the average of all the measurements used to
construct the control chart.
• Summary statistics often plotted include the subgroup average, subgroup
range, subgroup standard deviation, percent defective, average number of
defects per unit, and so on.
• Key characteristics are examples of process output that can be monitored by
statistical control charts.
Control Chart Elements
Tests for Special Causes of Variation
Tests for Special Causes of Variation
Which Chart Should be Used?
Subgroup
Size
DATA: Camshaft
X Bar and S Chart
Use X bar and S charts to draw a control chart for subgroup means
(X bar chart) and a control chart for subgroup standard deviations
(S chart) on one page. Interpreting both charts together allows you
to track both process center and process variation and detect the
presence of special causes. The S chart must be in control in order
to interpret the chart because the control limits are calculated
considering both process spread and center.
If the S chart is out of control, then the control limits on the X bar
chart will be inaccurate and may falsely indicate an out-of-control
condition. In this case, the lack of control will be due to unstable
variation rather than actual changes in the process center.
When the S chart is in control, you can be sure that an out-of-
control X bar chart is due to changes in the process center.
X Bar and S Chart Sample
DATA: Canning
P Chart
Use P chart to study the proportion of defectives in each sample
and determine whether or not the process is in control.
An in-control process exhibits only random variation in the
proportion of defectives per sample.
An out-of-control process exhibits nonrandom variation in the
proportion of defectives per sample, which may be due to the
presence of special causes.
A defective is a product or service in which a nonconformity (or
flaw) renders the product or service unusable. Examples of
defectives include broken light bulbs, late deliveries, unanswered
calls, and blown fuses.
Use P charts when your sample sizes vary.
P Chart Sample
DATA: Telephone
U Chart
Use U charts to assess the number of defects per unit of measurement and
determine whether or not the process is in control.
An in control process exhibits only random variation in the number of defects
per unit due to the presence of special causes.
An out-of-control process exhibits unusual variation in the number of defects
per unit due to the presence of special causes.
Defects are nonconformities (or flaws) in part of a product or service that do
not render the product or service unusable. Examples of defects include air
bubbles in a sheet of glass, weaknesses in insulated wire, or blemishes in
cloth.
Use U chart when your sample sizes vary.
U Chart Sample
DATA: Transcription