Work In Progress (WIP)

Work in progress (WIP) Partially finished goods waiting for

Work in queue (WIQ) completion and eventual sale of these
items.
Touch time
These items are either just being
Takt time fabricated or waiting for further
Cycle time processing.
Throughput

Work in Queue (WIQ) Touch Time

Material at work station waiting to be Time that material is actually being
processed. worked on.
Some of the Work in Progress may not This excludes moving, waiting time.
be in the queue.

Takt time is the time in which a product
needs to be produced in order to satisfy
customer demand.
Takt time =
Net Operating time per period
Customer demand per period
Takt Time
For example
Customer wants 10 units per week,
Plant operates 40 hour per week
Takt Time = 40/10 = 4 hrs.
Takt time is not related to how much time it
takes to make the unit.

Takt Time Cycle Time

For example Actual time to produce one piece.

Customer wants 10 units per week,

Plant operates 40 hour per week Process 2 Process 3
Two breaks of 10 minutes and 40 minutes
lunch break daily. (5 days week) 3.5 Hours 3.5 Hours 3.5 Hours
Net operating time = 40 hrs 5x(1 hr)= 35
hrs
Takt Time = 35/10 = 3.5 hrs.
Takt time = 3.5 hrs
Takt time is not related to how much time it Cycle time = 10.5 hrs
takes to make the unit.
 Cycle Time Cycle Time
Actual time to produce one piece.
Cycle time is value added time only.
Process 2 Process 3

1 Hours 3.5 Hours 3.5 Hours

Process 2 Process 3

1 Hours 3.5 Hours 1 Hours

Takt time = 3.5 hrs
Cycle time = 8 hrs

Lead Time
Total time a unit takes from the time the
work is requested to the time it is
delivered.
Includes processing and waiting time.
Lead Time = WIP/Throughput
Example:
WIP is 50 units
Throughput is 10 units per day
Lead Time = 50/10 = 5 days
Throughput is the average number of
units processed per time unit.
Throughput rate = 1/ Cycle time
Example:
Cycle time = 20 minutes
Throughput rate = 1/20 per minutes
Or 60/20 per hour = 3 per hour
Or 3x8 per day (for 8 hour shift) = 24 per
day
 Process Analysis Tools Flow Chart
Flowcharts Shows the processes as boxes connected
Process maps by arrows and decision boxes.
Value stream maps
Work instructions
Spaghetti diagrams
Circle diagrams
Gemba walk

Types of Flowcharts
Basic Flowcharts Flow chart and process map are used
Swim Lane interchangeably
Value Stream Maps
Process mapping is the process of
creating a diagram; the diagram itself is
called a flow chart.
 Value Stream Maps Value Stream Maps
Type of Flow Chart showing how value
flows through the organization
Flow of material
Flow of information

Value Stream Maps Value Stream Maps

SigmaXL demonstration
Symbols
Current Lean Future
State Principles State
 Work Instructions Spaghetti diagrams
Steps to do a job Trace the path of a product/activity
Brings consistency To reduce the transport and motion waste
Helps in orienting new hire
Generally in text but could be in the Waiting X-ray Reception Doctor
form of a flow chart, illustrations as well Exam
Should be easy to understand Waiting

Exam Doctor

Circle Diagrams Gemba Walk

Venn Diagram Gemba = work area or shop floor
Pie Chart gembutsu , which

Gemba walk is the action of going to see

Knowledge
the process, observe, understand, ask
10%
9%
1st Qtr questions and learn.
2nd Qtr
58% 3rd Qtr
23%
Experience Attitude 4th Qtr
 Qualitative vs Quantitative Continuous vs Discrete
Continuous Data
Data Measurements: Length, height, time
More information with less samples
More sensitive
Quantitative Data Provide more information
More expensive to collect
Discrete Data
Description Numbers Count: Number of students, Number of
heads

Measurement Scales Measurement Scales

Data Data

Ordinal Interval Ratio Nominal Ordinal Ratio

Example:
Color: Blue, Green, Red
 Measurement Scales Measurement Scales

Data Data

Nominal Ordinal Ratio Nominal Ordinal Ratio

Example: Example:
Pass/Fail Temperature: Celsius
Good, Bad, Worst

Measurement Scales Measurement Scales

Nominal Ordinal Interval Ratio

Ordered N Y Y Y
Data Difference N N Y Y
Absolute Zero N N N Y
Example Red, Blue Good, Bad, Temperature Length,
Nominal Ordinal Ratio Worst : Degree C Weight
Central Tendency Mode Mode, Mode, Mode,
Measurement Median Median, Median,
Example: Mean Mean

Height, mass, volume

 Qualitative vs Quantitative Sampling
Sample: Part
Population: Sampling of population
Complete Process
Data collection to
be studied

Nominal Interval Ratio

Parameter Statistic
Qualitative Data Quantitative Data Characteristic of Characteristic
a population of a sample
N number of members n
mean
standard deviation

Sampling Why Sampling?

Sample: Part
Population: Sampling of population
Complete Process
collection to
be studied

Statistic
Characteristic of Characteristic
a population of a sample
N number of members n

standard deviation
Sampling Sampling
Simple Random Sampling

Systematic Random Sampling

Probability
Sampling
Cluster Sampling

Probability
Sampling
Judgemental Sampling

Simple Random Sampling
Non
Systematic Random Sampling
Select elements at regular intervals through
that ordered list.
Example: Checking every 6th piece produced
by the machine.

Simple Random Sampling

Systematic Random Sampling

Probability
Sampling

Stratified Random Sampling

Cluster Sampling
Stratified Random Sampling
Used to ensure that sub-groups within a
population are represented proportionally
in the sample.
Example: If 10 people are drawn to
represent a country, 5 of them are male and
5 females to avoid the sex bias.
Simple Random Sampling
Sometimes it is more cost-effective to select
respondents in groups ('clusters'). Sampling
is often clustered by geography, or by time
periods.
Example: Survey all customers visiting
particular stores on particular days.
Simple Random Sampling

Systematic Random Sampling

Probability
Sampling

Stratified Random Sampling

Cluster Sampling

Convenience Sampling Judgmental Sampling

Accidental / Convenience Sampling Probability Accidental / Convenience Sampling

Sampling
Judgemental Sampling
Non

Quota Sampling Quota Sampling

 Quota Sampling Data Collection Plan
Why you need to collect data?
Goal and Objective
Operational Definition
How much? How? Where? When? Etc.
Type of data NOIR
Manual or Automatic
Past data vs Future
Accidental / Convenience Sampling
Is data reliable?

Quota Sampling

Data Coding
Measurement Operational How is it Type of Data Sample size Who? Data Comments
Definition measured? Recording
Form
Time to Time from Using a stop Continuous Every 10th Operator Assembly
assemble picking up watch piece Record
the first Ratio F-0156
piece to
placing the
assembled
item in tray
 Data Coding Data Coding
Multiplying or dividing
Example: 1.05, 1.03, 1.02, 1.10
Multiply 100 to each: 105, 103, 102, 110
Coded mean: 105
Un-coded mean: 105 / 100 = 1.05
Standard deviation need to divided by you
multiplied for coding.
For coded data s = 3.559
For original data s = 3.559/100 = 0.03559

Data Cleaning Missing Data Measurement System Analysis (MSA)

In statistics, imputation is the process of
replacing missing data with substituted
values.
Missing data can introduce bias.
Missing randomly
Reason for missing

Delete the row

Replace with the average value
 Measurement System Analysis (MSA) Measurement System Analysis (MSA)
Measurement System includes
operator, measuring instrument and
procedures.

Resolution Resolution
Resolution/ Discrimination
Smallest readable unit of the measuring
instrument.
10 to 1 Rule of Thumb:

discrimination (resolution) of the measuring

instrument should divide the tolerance of
the characteristic to be measured into ten
parts.
 Resolution Measurement System Analysis (MSA)
10 to 1 Rule of Thumb:
Which of these two would you use if the
part tolerance is 52.00 +/- 0.05 (51.95
52.05)
Tolerance Range = 0.10
Minimum Reading of Digital Vernier = 0.01
Digital Vernier divides the tolerance into 10
parts, hence acceptable.

Accuracy vs Precision Measurement System Analysis (MSA)

Accuracy
accepted reference value.
Bias
Linearity
Stability
Precision
other
Repeatability
Reproducibility
 Bias Bias
Bias is the difference between the Bias is the difference between the
observed average of measurements observed average of measurements
and the reference value. and the reference value.
Reference Measured
Value (psi) Value (psi)
Average
100 100 Reference
101 Value
100 102
100 102
100 101
100
Average 101 Bias
Bias = 101 psi -100 psi = 1 psi

Bias Linearity
Bias is the systematic error. Linearity measures the bias across
Bias is addressed by calibration. the operating range of a tool or
instrument.
Reference Average Bias BIAS
Value (psi) Measured 2.5

Value (psi)
2

0 0 0 1.5

50 50.5 0.5
1

100 101 1
0.5
150 151.5 1.5
0
200 202 2 0 50 10 0 15 0 20 0 25 0
 Stability Accuracy vs Precision
Stability measures the bias over time. Accuracy
Also known as drift.
accepted reference value.
Bias
Time Linearity
Stability

Time 2 Precision
Bias 2
other
Repeatability
Reproducibility
Time 1
Bias 1

Repeatability
Variation in measurements obtained
with one measuring instrument when
used several times by an appraiser.
Also called Equipment Variation (EV)
gauge to
produce consistent results.
Reproducibility
Variation in the average of the
measurements made by different
appraisers using the same gage
Also called Appraiser Variation (AV)
appraiser to
produce consistent results.
 Gage R&R (GRR) Gage R&R (GRR)
Combined estimate of repeatability and Combined estimate of repeatability and
reproducibility. reproducibility.

Precision to Tolerance Ratio Precision to Tolerance Ratio

How capable your measurement system P/T ratio is the most common
is? estimate of measurement system
Precision/Tolerance (P/T) is the ratio precision Measurement
between the estimated measurement Tolerance System
error (precision) and the tolerance of Variation
the characteristic being measured. P/T = 10%

P/T = 100%

P/T = 200%
LSL USL
 Precision to Tolerance Ratio Gage R&R (GRR)
PTR Gage R&R Studies
PTR = 5.15 ms Repeatability and reproducibility
USL-LSL (R&R) studies are a method for
determining the variation of a
measurement system.
Why not 6 ms ? There are three methods of
6 sigma includes 99.73% area conducting GRR Studies.
5.15 sigma includes 99% area Range
Average and Range
ANOVA

Gage R&R (GRR) Gage R&R (GRR) Range Method

The range method quantifies both Provides quick estimate of GRR.
repeatability and reproducibility This method does not split the GRR into
together. Repeatability and Reproducibility
The average and range method Two operators usually measure 5 parts
provides separate estimate of
repeatability and reproducibility. one time each.
The analysis of variance method
(ANOVA) estimates variation due to
repeatability and reproducibility and
parts, and also looks at the
interaction between those these.
 Gage R&R (GRR) Range Method GRR Average and Range Method
Two operators usually measure 5 parts Average and Range method provides the
one time each. estimate of both repeatability and
GRR = R-bar/d2 = 1.6/1.19 = 1.34 Part Appraiser A Appraiser B Range reproducibility.
#
Process standard deviation = 2.23 1 100 101 1 Interaction between appraiser and gage
(From previous study) 2 99 102 3 is not computed (ANOVA method does
3 97 99 2 that)
%GRR = 100 (GRR/process std dev) 4 103 103 0
%GRR = 100 x 1.34 / 2.23 = 0.60 5 101 103 2
R-bar = 1.6
Measurement variation consumes 60%
of the process variation. Measurement
system needs improvement.

GRR ANOVA Method Number of Distinct Categories (NDC)

ANOVA is Analysis of Variance Resolution:
Using a meter stick to measure the height of
the students: (100cm to 130cm heights)
Round to height to the nearest meter
All students will have the same height.
Round to the nearest decimeter
We are 3 different heights.
Round to the nearest centimeter
We are 30 different heights.
 Number of Distinct Categories (NDC)
NDC = 1.414 x parts
gage
Number of Distinct Categories (NDC)
Value of NDC
Less than 2, the measurement system is of
no value for controlling the process, since
one part cannot be distinguished from
another.
Equal to 2, the data can be divided into two
groups, say high and low.
Equal to 3, the data can be divided into 3
groups, say low, middle and high.
More than 5, denotes an acceptable
measurement system.
Number of Distinct Categories (NDC)
NDC is the number of buckets your data
values would naturally drop into.
What if The NDC is low?
Analyze more distinct parts that truly
represent the entire range of the process.
Increase the precision on your
measurement tool.
Measurement System Analysis (MSA)
Crossed gage R&R study
A study in which each operator
measures each part. (Non-destructive)
Nested gage R&R study
A study in which only one operator
measures each part, usually because
the test destroys the part. (Destructive)
 Measurement Systems
Organizations need measurement
systems to track the progress.
These are also called KPIs Key
Performance Indicators.
All functions in the business have
measurement systems in place.
Focus on those measurements which
ensure the overall business success, not
the one which are easy to measure.
Measurement Systems - Purchase
Average number of suppliers for key
components
Number of strategic alliances
Rejection rate
Measurement Systems - Sales
Number of customer complaints
Percent of items returned
Percent sales from returning customers
Sales growth
Market share
Customer satisfaction rating
Measurement Systems - Production
Value added per person
Internal repair/rejection rate
Work in Progress
Equipment Utilization
 Measurement Systems Metrology
Ensure that your measurements are free Metrology is the science of
from bias measurements. (Meteorology is science
of weather forecasting)
Repeatability and Reproducible of
measurements
Involves three Main activities:
Definition of internationally accepted units
of measurement
Realisation of these units of measurement
in practice
Application of chains of traceability linking
measurements made in practice to
reference standards

Seven Base Units Seven Base Units

Unit name Unit symbol Quantity name Dimension symbol Unit name Unit symbol

metre m length L metre The distance travelled by light in vacuum in 1/299792458 second.

kilogram kg mass M kilogram The mass of the international prototype kilogram.

The duration of 9192631770 periods of the radiation corresponding to
second s time T second the transition between the two hyperfine levels of the ground state of the
caesium-133 atom.
The constant current which, if maintained in two straight parallel
ampere A electric current I conductors of infinite length, of negligible circular cross-section, and
ampere
placed 1 m apart in vacuum, would produce between these conductors a
force equal to 2×10 newtons per metre of length.
thermodynamic
kelvin K
temperature kelvin 1/273.16 of the thermodynamic temperature of the triple point of water
mole mol amount of substance N The amount of substance of a system which contains as many elementary
mole
entities as there are atoms in 0.012 kilogram of carbon-12
candela cd luminous intensity J The luminous intensity, in a given direction, of a source that emits
candela monochromatic radiation of frequency 5.4×1014 hertz and that has a
radiant intensity in that direction of 1/683 watt per steradian.
 Derived Units Calibration
The derived units in the SI are formed by Purpose:
powers, products or quotients of the To ensure consistency
base units. To determine the accuracy readings
For example velocity is a quantity that is Can measurements be trusted
derived from the base quantities of time
and length, so in SI the derived unit is
metres per second (symbol m/s).

Traceability Traceability

whereby the result can be related to a SI Units

reference through a documented
unbroken chain of calibrations, each National Labs
contributing to the measurement
Third Party Labs

Metrology Department

Working Instruments
 Basic Statistical Terms Notations
Population Sample
Sample: Part Parameters Statistics
Population: Sampling of population Mean
Complete Process Standard Deviation s
collection to Variance 2 s2
be studied Proportion of population P p
having an attribute
Inference Proportion of population not Q q
Parameter Statistic having an attribute (=1-P) (=1-p)

Characteristic of Characteristic Correlation coefficient r

a population of a sample Number of elements N n
N number of members n
mean
standard deviation s

Central Limit Theorem Central Limit Theorem

 Central Limit Theorem Central Limit Theorem

Central Limit Theorem Standard Error of Means

Standard deviation of the sampling
Distribution of
distribution of the sample mean
Individual Observations
Distribution of
Sample Means
 Mean
Descriptive
Statistics Also known as Average Central
Tendency

Central
Affected by extreme values Mean Mode Median
Variability
Tendency
Example: 10, 11, 14, 9, 6
Mean = (10+11+14+9+6)/5 = 50/5 = 10
Quartile

Mean

Standard
Mode
Deviation

Interquartile
Median
Range

Percentile

Mode Median
Most occurring item Central
Tendency Middle value when put in ascending or Central
Tendency

descending order.
Example: 10, 11, 14, 9, 6, 10
Example: 10, 11, 14, 9, 6
Mean Mode Median Mean Mode Median

Mode = 10 Quartile In ascending order - 6,9,10,11,14 Quartile

Median = 10

Example: 10, 11, 14, 9, 6, 11

In order - 6,9,10,11, 11,14
Median = 10.5
 Percentile Percentile/Quartile Steps
Median divides the data in two equal Central
Tendency Arrange in ascending or descending Central
Tendency

parts when arranged in ascending or order

descending order Mean Mode Percentile
Calculate location(i) = P.(n)/100 Mean Mode Percentile

Percentile divides data in 99 parts Quartile P=percentile, n=numbers in data set Quartile

Quartile divides data in 4 parts If i is whole number Percentile is

average of (i)th and (i+1)th location
Example: 6,9,10,11, 11,14
If i Percentile
Q1=9, Q2=10.5, Q3=11 is located at (i+1)th whole-num.
Example: 6,9,10,11, 11,14
Q1=9, Q2=10.5, Q3=11

Descriptive Statistics
Difference between lowest and the Variability

highest value.
Variability Interquartile Standard
Example: 6,9,10,11, 11,14
Range
Range Deviation

Range = 14-6 = 8
Interquartile Standard
Range
Range Deviation
 Interquartile Range Standard Deviation
Range of middle 50% data Variance = average of squared deviation Variability

IQR = Q3-Q1 about the arithmetic mean.

Interquartile Standard Standard
Square root of variance is standard
Range Range
Range Deviation Deviation
Example: 6,9,10,11, 11,14
deviation
Q1=9, Q2=10.5, Q3=11
IQR = 11-9 = 2
Box-and-Whisker Plot

Standard Deviation
x Box-and-whisker plots
100 0 0
S2 = Scatter diagrams
101 1 1 n-1
99 -1 1 Histograms
102 2 4 S 2 = 10/5 = 2 Normal probability plots
98 -2 4
S = 2 = 1.414 Frequency distributions
100 0 0
Cumulative frequency distributions.
 Box and Whisker Plots Box and Whisker Plots
Demonstration: SigmaXL > Customer
Data > Average number of
orders/month

70 70
60 60
50 Median 50 Median

25th 25th
40 40
75th 75th
30 Mean
30 Mean

20 Outliers 20 Outliers

10 Avg No. of 10 Avg No. of

orders per orders per
0 mo 0 mo

Scatter Diagram
One of seven basic quality tools Graphical representation of the
To see relationship between two distribution of numerical data
variables
Relationship should make practical frequency for each bin is plotted.
sense
Temperature(X) vs Ice cream sale (Y)
Some times relationship between two
variables is because of a third variable.
(ice cream sale vs heat stroke cases)
Correlation/Regression is covered in the
Analyze Phase
 Histograms
Demonstration Using SigmaXL and MS
Excel 30

25
20

20
15

Frequency
Frequency
15

10 Frequency
10

5
5

0 0

7.1
12.58
18.06
23.54
29.02
34.5
39.98
45.46
50.94
56.42
More
11.7

16.2

20.8

25.4

30.0

34.5

39.1

43.7

48.2

52.8

57.4
7.1
Avg No. of orders per mo

Valid Statistical Conclusions

Hypothesis: Null Hypothesis and Alternate
A claim that we want to test Hypothesis are pair and cover all
possibilities.
Null Hypothesis - H0 Only one of these has to stand and not
Default position / Currently accepted both.
position / Assumed / Status Quo
Alternate Hypothesis Ha
Claim to be tested. Also known as Research
Hypothesis or the other option.
 Types of Errors Probability
True State of Nature
Classic Model
H0 Ha
Is true Is true Number of outcomes in which the event occurs
Total Number of possible outcomes of an experiment
Support H0 /
Reject Ha Correct Type II Error
Conclusion
Conclusion Support Ha /
Reject H0 Type I Error Correct
Conclusion

Probability
Relative Frequency of Occurrence

Number of times an event occurred

Total number of opportunities for an event to occur
 Probability Probability
Union: Probability that events A or B Mutually Exclusive Events: When two
occur: P(A B) events cannot occur at the same time

Intersection: Probability that events A Independent Events: The occurrence of

and Event A does not change the probability
of Event B

Complementary Events: The probability

that Event A will NOT occur is denoted
by P(A').

Probability
Rule of Addition
The probability that Event A or Event B occurs
=
Probability that Event A occurs
+
Probability that Event B occurs
-
Probability that both Events A and B occur

P(A B) = P(A) + P(B) -

 Probability Probability
Independent Events Dependent Events

Permutation/ Combination
Factorial of a non-negative integer n,
denoted by n!, is the product of all
positive integers less than or equal to n
 Continuous Probability Distributions Continuous Probability Distributions
Normal probability distribution In case of a Continuous Distribution:

Student's t distribution The probability of a specific value is zero?

Chi-square distribution You should be looking for the probability of
F distribution a range in case of Continuous Distribution.

Normal Probability Distribution

Symmetrically distributed

Long Tails / Bell Shaped

Mean/ Mode and Median are same

 Normal Probability Distribution
About 68% of the area under the curve
falls within 1 standard deviation of the
mean.
About 95% of the area under the curve
falls within 2 standard deviations of the
mean.
About 99.7% of the area under the
curve falls within 3 standard deviations
of the mean.
Normal Probability Distribution
The total area under the normal curve =
1.
The probability of any particular value is
0.
The probability that X is greater than or
less than a value = area under the
normal curve in that direction

Normal Probability Distribution

The value of the random variable Y is:

sqrt -(x -

where X is a normal random variable,

e is approximately 2.71828.
Z Table Z Table in Excel
NORMSDIST(z) returns p

NORMSINV(p) returns z.

Continuous Probability Distributions Continuous vs Discrete Variable

If a variable can take on any value
between two specified values, it is called
a continuous variable; otherwise, it is
called a discrete variable.

Binomial Probability Distribution
Bernoulli Distribution
Hypergeometric Probability Distribution
Geometric Distribution
Negative Geometric Distribution
Poisson Probability Distribution
Binomial Probability Distribution
Binomial Probability Distribution
A binomial experiment has the
following properties:
The experiment consists of n repeated
trials.
Each trial can result in just two possible
outcomes. We call one of these outcomes a
success and the other, a failure.
The probability of success, denoted by p, is
the same on every trial.
The trials are independent; that is, the
outcome on one trial does not affect the
outcome on other trials.
Binomial Probability Distribution
x: The number of successes that result from the
binomial experiment.
n: The number of trials in the binomial
experiment.
p: The probability of success on an individual
trial.
q: The probability of failure on an individual
trial. (This is equal to 1 - p.)
n!: The factorial of n (also known as n factorial).
P(x) : Binomial probability - the probability that
an n-trial binomial experiment results in exactly
x successes, when the probability of success on
an individual trial is p.
nCx: The number of combinations of n things,
taken x at a time.
 Binomial Probability Distribution Binomial Probability Distribution
The binomial probability refers to the The mean of the distribution ( x) is
probability that a binomial experiment n.p
results in exactly x successes.
Suppose a binomial experiment consists
x) is
2
of n trials and results in x successes. If
the probability of success on an n.p.(1-p)
individual trial is p, then the binomial
probability is:
The standard deviation ( x) is
P(x) = nCx . px . (1 - p)n - x
or sqrt[ n . p . ( 1 - p ) ]
P(x) = { n! / [ x! (n - x)! ] } . px . (1 - p)n - x

Five Conditions - Binomial Bernoulli Distribution

1. There is a fixed number, n , of Distribution of successes on a single
identical trials. trial.
2. For each trial, there are only two What is the probability of getting head in
tossing of a coin once?
possible outcomes (success/failure).
3. The probability of success, p, remains
the same for each trial.
4. The trials are independent of each
other.
5. x = the number of successes observed
for the n trials.
 Hypergeometric Distribution Hypergeometric Distribution
There is a fixed number, n , of identical N: size of population
trials. A: number of successes in population
For each trial, there are only two possible x: The number of successes that result from the
outcomes (success/failure). experiment.
The probability of success, p, remains the n: The number of trials without replacement.
same for each trial. p: The probability of success on an individual
The trials are independent of each other. trial.
Finite and known population without q: The probability of failure on an individual
replacement. trial. (This is equal to 1 - p.)
Number of successes in population are P(x) : The probability that an n-trial experiment
known results in exactly x successes
x = the number of successes observed for nCx: The number of combinations of n things,
the n trials. taken x at a time.

Hypergeometric Distribution Geometric Distribution

Out of 10 people (6M, 4F), 3 people are P(x) = ACx . N-ACn-x / NCn
selected without replacement. What is
the probability that two of them are
females?
P(2) = 4C2 . 10-4C3-2 / 10C3
= 4C2 . 6C1 / 10C3 = 6x6/120 = 0.3

In Excel use: HYPGEOM.DIST function

When sample size is less than 5%
population then can use Binomial.
 Negative Binomial Distribution
Generalization of the Geometric A Poisson experiment has the following
distribution properties:
Number of trials needed to get the first The experiment results in outcomes that can
number of successes. be classified as successes or failures.
What is the probability that if the coin is
tossed repeatedly the first third time head occurs in a specified region is known.
appears on 5th trial? Outcomes are random. Occurrence of one
outcome does not influence the chance of
In Binomial distribution trials are fixed, another outcome of interest.
in Negative Binomial number of The outcomes of interest are rare relative to
successes are fixed. the possible outcomes.

Poisson Distribution
e: A constant equal to approximately - ) ( x) / x! Poisson Formula. Suppose we conduct a
2.71828. (Actually, e is the base of the Poisson experiment, in which the average
natural logarithm system) number of successes within a given region is

occur in a specified region.

x: The actual number of successes that P(x - ) ( x) / x!
occur in a specified region.
P(x Poisson probability that
exactly x successes occur in a Poisson where x is the actual number of successes
experiment, when the mean number of that result from the experiment, and e is
approximately equal to 2.71828.
 Poisson Distribution
Process Capability Indices
Ratio of the spread between the process
specifications to the spread of the
process values, (6 process standard
deviations) .
Poisson Distribution
On a booking counter on the average 3.6
people come every 10 minute on
weekends. What is the probability of
getting 7 people in 10 minutes?
P(x - ) ( x) / x! = (e-3.6) (3.67) / 7!
=0.02732 x 7836.41 / 5040 = 0.0424
LSL Lower Specification Limit
USL - Upper Specification Limit
LCL Lower Control Limit
UCL - Upper Control Limit
 Process Capability Indices
Process Capability Indices
Why to do Process Capability study?
Understand the behaviors of new/repaired/
adjusted equipment
Review of tolerances
Allocation of equipment
Process Capability Indices
Cp = (USL LSL)/(6* within)
CpL = (Process Mean LSL)/(3* within)
CpU = (USL Process Mean)/(3* within)
Cpk = Min (CpU, CpL )
Conditions to be met:
Sample to represent the population
Normal distribution of data
The process must be in statistical control
Sample size must be sufficient
 Process Capability Indices Process Performance Indices
Conditions to be met:
Sample to represent the population
Normal distribution of data
Cp 1.00 1.33 1.66 2.00 The process must be in statistical control
Rejects 0.27 % 64 ppm 0.6 ppm 2 ppb Sample size must be sufficient

Process Performance Indices

Pp = (USL LSL)/(6* overall) Cpk
deviation, while Ppk
standard deviation.
PpL = (Process Mean LSL)/(3* overall)
Cpk is for short term and Ppk is for long
term.
PpU = (USL Process Mean)/(3* overall)

Ppk = Min (PpU, PpL )

 Taguchi Capability Index - Cpm
Process Capability Attribute Data
When your data is pass/fail instead of
values, you are looking for Attribute
Data Process Capability.
Common way to do this is to use p, np, c
and u charts. We will cover these charts
later in the Control section.
Process Capability Studies
Select the process
Data Collection Plan
Measurement System Analysis
Gather data
Confirm normality of data
Confirm that the process is in control
Estimate the process capability
Continually improve process
What if the data is not normal?
Check for Normality Anderson Darling
Test
Transformation Basics
Box-Cox Power Transformation
 25
20

Check for Normality Transformation Basics

Frequency
15
10
5
0

46
804
1561
2319
3077
3834
4592
5349
6107
Cycle Time (Minutes)

1.83

NSCORE
0.83

-0.17

-1.17

-2.17

0.00

2000.00

4000.00

6000.00

8000.00
-4000.00

-2000.00
Cycle Time (Minutes)

Anderson-Darling Normality Test 5.357

P-Value (A-D Test) 0.0000

Transformation Basics Transformation Basics

 SigmaXL Demonstration

Process Performance Metrics Percent Defectives

Percent Defectives Percent of parts having one or more
PPM defects
DPMO 2 percent 2 pieces per 100 pieces
DPU
Rolled Through Yield
 Parts per Million (PPM)
Defective parts per million.
2 percent 2 pieces per 100 pieces
0.02 x 1,000,000 = 20,000 PPM
Defect Opportunity
Circumstances in which CTQ can fail to
meet.
Number of defect opportunities relate to
complexity of unit.
Complex units Greater opportunities of
defect than simple units
Examples:
A units has 5 parts, and in each part there are 3
opportunities of defects Total defect
opportunities are 5 x 3 = 15
A nonconforming unit is a defective
unit
Defect is nonconformance on one of
many possible quality characteristics of
a unit that causes customer
dissatisfaction.
Defects Per Opportunity (DPO)
Number of defects divided by number of
defect opportunities
Examples:
In previous case (15 defect opportunities), if 10
units have 2 defects.
DPO = 2 / (15 x 10) = 0.0133333
 Defect Per Million Opportunities (DPMO)
DPO multiplies by one million Number of Defects / Number of Units
Examples:
In previous case (15 defect opportunities), if 10 In 3,000 welds defects observed were:
units have 2 defects.
10 Cracks
15 Porosity
DPO = 2 / (15 x 10) = 0.0133333 5 Undercut
DPMO = 0.013333333 x 1,000,000 = 13,333

DPU = (10+15+5)/3,000 = 30/3,000

13,333 DPMO is 3.7 Sigma
Six Sigma performance is 3.4 DPMO
= 1/100 = 0.01

Rolled Through Yield Short Term vs Long Term Capability

Units entering a process = P Cp Cpk Short term
Defective Units = D
Yield = (P-D)/P Pp Ppk Long term
Y1 = 0.99, Y2 =0.95, Y3=0.98
RTY = Y1 . Y2 . Y3 = 0.99x0.95x0.98 In both of these measurements the
=0.92169 difference lies in the calculation of
sigma.