Statistical

Process
Control

Prepared by: …..

December, 2008
Introductio
n
 Statistical process control
(SPC) is a collection of tools
that when used together can
result in process stability and
variance reduction.
Introductio
n
The seven major tools are
1) Histogram or Stem and Leaf plot
2) Check Sheet (Defect
concentration diagram)
3) Pareto Chart
4) Cause and Effect Diagram
5) Why-why Diagram
6) Scatter Diagram
7) Control Chart
 Introductio
n
Check sheet Pareto chart Histogram

Number

Number
Nonconformities Tally Total
Blur code 4
////
Scratches 2
//
…. 1
/ Type Measurement

Cause & effect diagram Why-why diagram

Possible
Solution
Why cause
Possible
cause Possible
Why Solution
Observation cause

Possible Why Possible

Solution
cause cause

Scatter Diagram Control chart

UCL
Factor 1

Value

CL

Factor 2 LCL
Time
Samples
To measure the process, we take samples
and analyze the sample statistics following
these steps
Each of these
(a) Samples of the represents one
sample of five
product, say five bottles of Nutrition
bottles of Nutrition
taken off the filling
Frequency
# #
machine line, vary # # #
from each other in
weight # # # #
# # # # # # #

# # # # # # # # # #

Weight
Samples
To measure the process, we take samples
and analyze the sample statistics following
these steps
The solid line
(b) After enough represents the
distribution
samples are taken
from a stable
process, they form
a pattern called a Frequency
distribution

Weight
Samples
To measure the process, we take samples
and analyze the sample statistics following
these steps
(c) There are many types of distributions, including the
normal (bell-shaped) distribution, but distributions do
differ in terms of central tendency (mean), standard
deviation or variance, and shape
Frequency

Central tendency Variation Shape

Weight Weight Weight

Samples
To measure the process, we take samples
and analyze the sample statistics following
these steps

(d) If only natural

causes of variation
are present, the Frequency
Prediction
output of a process
forms a distribution
that is stable over
time and is Tim
e
predictable Weight
Samples
To measure the process, we take samples
and analyze the sample statistics following
these steps
?
?? ??
? ?
(e) If assignable ?
?
? ?
?
?
causes are ??? ??
?
present, the Frequency
Prediction
process output
is not stable
over time and is
not predictable e
Tim
Weight
Control
chart
 Control chart is a primary
tool used for SPC
 Constructed from historical

data, the purpose of control

charts is to help distinguish
between natural variations
and variations due to
assignable causes
Introductio
n
Process improvement using the control chart

Process

Control
chart
Detect
Verify a follow
assignable
up
cause

Implement Identify root

corrective cause of
action problem
Introductio
n
A typical control chart has control limits set
at values such that if the process is in
control, nearly all points will lie within the
upper control limit (UCL) and the lower
control limit (LCL)
Upper Control
Limit

Central line

Lower Control
Limit
 Introductio
n
Popularity of control charts:
∞ A proven technique for improving productivity
∞ Effective in defect prevention
∞ Prevent unnecessary process adjustment
∞ Provide diagnostic information
∞ Provide information about process capability
General model of a
control chart
Let W be a sample statistic that measures some quality
characteristic of interest:
UCL = µw + kw
CL = µw
LCL = µw - kw
Where
k = distance of the control limit from the center line
µw = mean of the sample statistic
w = standard deviation of the statistic
Control limits can be set at 3 standard deviations from
the mean in both directions
Types of
Data
Variables Attributes
 Characteristics  Defect-related
that can take any characteristics
real value  Classify products
 May be in whole or as either good or
in fractional bad or count
numbers defects
 Continuous
 Categorical or
discrete random
random variables
variables
Steps In Creating
Control Charts
1. Take samples from the population and compute the
appropriate sample statistic
2. Use the sample statistic to calculate control limits
and draw the control chart
3. Plot sample results on the control chart and
determine the state of the process (in or out of
control)
4. Investigate possible assignable causes and take
any indicated actions
5. Continue sampling from the process and reset the
control limits when necessary
Control Charts for
Variables
 For variables that have continuous
dimensions
 Weight, speed, length, strength, etc.
 X-charts are to control the central tendency
of the process
 R-charts are to control the dispersion of the
process
 These two charts must be used together
Control Charts for
Attributes
 For variables that are categorical
 Good/bad, yes/no,
acceptable/unacceptable
 Measurement is typically counting
defectives
 Charts may measure
 Percent defective (p-chart)
 Average number of defects (u-chart)
Patterns in Control
Charts
@ Stable process: one which performs with a
high degree of consistency at an essentially
constant level for an extended period of
time
– “In-control”
@ A process that is not stable is referred to as
being in an “out-of-control” state
@ The performance of process can be showed
in control charts as patterns in next slides
 atterns in Control Charts
Patterns
How do you know a process is “out of control”?
The Western Electric rules include
- Points outside of three-sigma control limits
(3σ)
- 8 consecutive points on one side of center line
- 2 of 3 consecutive points outside the two-
sigma limits (2)
- 4 of 5 points outside the one-sigma limits (
1)
- 7 consecutive points trending up or down
Patterns in Control
Charts
Upper control limit

+ 2

+ 1
Target

- 1

- 2
Lower control limit

Normal behavior.
Process is “in control.”
Patterns in Control
Charts
Upper control limit

+ 2

+ 1
Target

- 1

- 2
Lower control limit

One plot out above (or below).

Investigate for cause. Process
is “out of control.”
Patterns in Control
Charts
Upper control limit

+ 2

+ 1
Target

- 1

- 2
Lower control limit

Run of 8 above (or below)

central line. Investigate for
cause. Process is “out of
control.”
Patterns in Control
Charts
Upper control limit

+ 2

+ 1
Target

- 1

- 2
Lower control limit

Two plots very near lower (or

upper) control or outside two-
sigma. Investigate for cause.
Process is “out of control.”
Patterns in Control
Charts
Upper control limit

+ 2

+ 1
Target

- 1

- 2
Lower control limit

Four plots outside one-sigma.

Investigate for cause. Process
is “out of control.”
Patterns in Control
Charts
Upper control limit

+ 2

+ 1
Target

- 1

- 2
Lower control limit

Trends in either direction, 7

plots. Investigate for cause of
progressive change.
Patterns in Control
Charts
Upper control limit

Target

Lower control limit

Erratic behavior. Investigate.

Variable Control
chart
۞ X-bar CHART

۞ R control CHART
X-bar and R Control
Charts
Example 1:
In the manufacturing processes of
Nutrition products, the hardness of
tablets are checked at the compression.
Samples are taken at every 30 minutes,
5 tablets are measured hardness for
each time. The hardness test results are
presented on the table at next slides.
 X-bar and R Control
Charts
Statistical calculation
n = size of subgroup (in this example: 5
samples/time)
m= number of subgroups (in this example:
n
20 times)
1
The mean of each subgroupX i  n  xi
i 1

The range of each subgroupr  xmax  xmin

m
1
The grand mean X   X i
m i 1
The average range 1 m
R   ri
m i 1
 X-bar and R Control
Charts
Statistical
calculation
X-bar chart:
UCL  X  A2 R R chart:
UCL D4 R

CL  X CL R
LCL D3 R
LCL  X  A2 R
A2, D4, D3 are constants tabulated for various sample sizes as
below tableSample A2 D4 D3
size
In this 2 1.880 3.267
exampl
e 3 1.023 2.575
Do not
4 0.729 2.282 apply
5 0.577 2.115
6 0.483 2.004
7 0.419 1.924 0.076
X-bar and R Control
Charts
X-bar and R Control
Charts
Results:
For X-bar chart: Points more than 3
sigma from center line are found at
6, 8, 11, 19
For R control chart: One point more
than 3 sigma from center line is
foundProcess
at 9 “out of
control”
Attribute Control
chart
۞ P CHART

۞ nP CHART

۞ U CHART
P Chart and nP Chart
Example 2:
In the filling line of Nutrition products, all
the finished goods are checked before
put into cases. Inspection is conducted
for every case, we assume that have 100
bottles in each case. The number of
defects returned for rework after filling in
a day are presented on the table at next
slides.
 P Chart and nP Chart
Statistical calculation
n = size of subgroup (in this example: 100 bottles)
np= number of defectives in each subgroup
The fraction of defectives
p np / n
The average fraction of defectives
p  np /  n

P UCL  p  3

p 1 p  nP UCL  pn  3 n p (1  p )
n
char CL  p char CL  pn
t: t: LCL  p  3 n p (1  p )
LCL  p  3

p 1 p 
n
P Chart and nP Chart

Process “out
of control”
U
Chart
Example 3:
Inspection for incoming bottle at
warehouse, QA take 5 bottles in each
case, we assume that have 20 cases for
this lot of bottles. The number of defects
(such as blur, scratch…) after inspection
are presented on the table at next slides.
 U
Chart
Statistical calculation
n = number of units in each subgroup

u = average number of defects per unit

The center line, upper and lower control limits for U
chart are:
u
UCL u  3
n
CL u
u
LCL u  3
n
U
Chart

Process “out
of control”
Application in our
processes
 In fact, all control limits are set up in
manufacturing documents (PS, PI, MI, RMS,
PCS) in forms of maximum value, minimum
value and target value
 Data collection timely and accurately
 Statistic calculation and plot data on control
charts
 Process performance evaluated by using control
chart
 Breakthrough thinking and efforts to identify and
reduce or eliminate common causes of variation
Question &
Discussion