06control Chart 1
-
Upload
abishank09 -
Category
Documents
-
view
216 -
download
0
Transcript of 06control Chart 1
-
7/29/2019 06control Chart 1
1/37
Process Capability
-
7/29/2019 06control Chart 1
2/37
Process Capability
Process Capabilityis an important conceptin SPC. Process capability examines
-- the variability in process characteristics
-- whether the process is capable ofproducing products which conforms to
specifications
-
7/29/2019 06control Chart 1
3/37
Process Capability
Process capability studies distinguish betweenconformance to control limits and conformanceto specification limits (also calledtolerancelimits)
-- if the process is in control, then virtually all
points will remain within control limits
-- staying within control limits does not necessarily
mean that specification limits are satisfied-- specification limits are usually dictated by
customers
-
7/29/2019 06control Chart 1
4/37
Process Capability: Concepts
The following distributions show differentprocess scenarios. Note the relative positions
of the control limits and specification limits.
-
7/29/2019 06control Chart 1
5/37
Process Capability: Concepts
In control and productmeets specifications.
Control limits are within
specification limits
UCL: Upper Control Limits
LCL: Lower Control Limits
USL: Upper Specification Limits
LSL: Lower Specification Limits
-
7/29/2019 06control Chart 1
6/37
Process Capability: Concepts
In control but someproducts do not meet
specifications.
Specification limits arewithin control limits
-
7/29/2019 06control Chart 1
7/37
Data from process with low capability
Data from process with
mediumcapability
Data from process with highcapability
relative capability
-
7/29/2019 06control Chart 1
8/37
Process capability: capability index
The capability index is defined as: Cp = (allowablerange)/6 =T (Tolerance)/ 6= (USL - LSL)/6s
The distribution of process quality
is often assumed to be
approximated
with a normal distribution.
Px3=99.73%,
-
7/29/2019 06control Chart 1
9/37
Normal Distribution
x
f(x)
0
=0.5
=1
=2
0
f(x)
x1 2
The normal distribution N(,2) has several
distinct properties:
--The normal distribution is bell-shaped and is
symmetric
--The mean, , is located at the centre
-- is the standard deviation of the data
The smallerthe steeper the curve
For same changing the value of
is to move the curve without anychange in its shape
-
7/29/2019 06control Chart 1
10/37
3 Principle
XN(,2) P{-
-
7/29/2019 06control Chart 1
11/37
Process capability: capability index
The capability index (T/6) show how well a processis able to meet specifications. The higher the value
of the index, the more capable is the process:
-- Cp < 1 (process is unsatisfactory)
-- 1 < Cp < 1.6 ( process is of medium relativecapability)
-- Cp > 1.6 (process shows high relative capability):
better to analyze the actual specifications
(Tolerance) and process technics to save resources
in enhancing the process capability, such as the
increased accuracy of equipment.
-
7/29/2019 06control Chart 1
12/37
Process capability:process performance index
The capability index-- considers only the spread of the
characteristic in relation to specification limits
-- assumes two-sided specification limits
The product can be bad if the mean is not setappropriately. The process performanceindex takes account of the mean () and is
defined as:Cpk = min[ (USL - )/3, ( - LSL)/3]
-
7/29/2019 06control Chart 1
13/37
Process capability:process performance index
The process performance index can alsoaccommodate one sided specification limits
-- for upper specification limit:
Cpk = (USL - )/3
-- for lower specification limit:
Cpk = ( - LSL)/ 3
-
7/29/2019 06control Chart 1
14/37
Process capability: the message
The message from process capability studiesis:
-- first reduce the variation in the process
-- then shift the mean of the process towardsthe target
This procedure is illustrated in the diagram
below:
-
7/29/2019 06control Chart 1
15/37
-
7/29/2019 06control Chart 1
16/37
Basic Forms of Statistical Sampling for
Quality Control
Sampling to determine if the process is within
acceptable limits (Statistical Process
Control).
Sampling to accept or reject the immediate
lot ofproductat hand (Acceptance
Sampling).
-
7/29/2019 06control Chart 1
17/37
Process Control
Process Control is concerned withmonitoring quality while the production orservice is being conducted.
-
7/29/2019 06control Chart 1
18/37
UCL
LCL
Samples
over time
1 2 3 4 5 6
UCL
LCL
Samples
over time
1 2 3 4 5 6
UCL
LCL
Samples
over time
1 2 3 4 5 6
Normal Behavior
Possible problem, investigate
Possible problem, investigate
Statistical Process Control
-- Control Charts
-
7/29/2019 06control Chart 1
19/37
Control charts
Processes that are notin a state ofstatistical control-- show excessive variations
-- exhibit variations that change with time
A process in a state of statistical control is said to be
statistically stable. Control charts are used to detectwhether a process is statistically stable. Controlcharts differentiates between variations
-- that is normally expected of the process due chance
orcommon causes-- that change over time due to assignable orspecial
causes
-
7/29/2019 06control Chart 1
20/37
Control charts: common cause variations
Variations due to common causes-- have small effect on the process
-- are inherent in the process because of:
the nature of the system
the way the system is managed
the way the process is organized and operated
-- can only be removed by
making modifications to the process changing the process
-- are the responsibility of higher management
-
7/29/2019 06control Chart 1
21/37
Control charts: special cause variations
Variations due to special causes are-- localized in nature
-- exceptions to the system
-- considered abnormalities
-- often specific to a certain operator
certain machine
certain batch of material, etc.
Investigation and removal of variations due to specialcauses are key to process improvement
Note: Sometimes the delineation between common andspecial causes may not be very clear.
-
7/29/2019 06control Chart 1
22/37
Control charts: how they work
The principles behind the application of control charts
are very simple and are based on the combined use of-- run charts
-- hypothesis testing
The control limits most commonly used in
organizations are plus and minus three standarddeviations. We know from statistics that the chancethat a sample mean will exceed three standaddeviations, in either direction, due simply to chance
variation, is less than 0.3 percent (i.e., 3 times per1000 samples). Thus, the chance that a sample willfall above the UCL, or below the LCL because ofnatural random causes is so small that this occurrenceis strong evidence of assignable variation.
-
7/29/2019 06control Chart 1
23/37
Control charts: how they work
The procedure is to-- sample the process at regular intervals
-- plot the statistic(or some measure ofperformance), e.g.
mean range
variable
number of defects, etc.
-- check (graphically) if the process is understatistical control
-- if the process is not under statistical control, dosomething about it
-
7/29/2019 06control Chart 1
24/37
Control charts: types of charts
Different charts are used depending on the nature of the charteddata. Commonly used charts are:
-- forcontinuous (variables) data
Shewhart sample mean ( -chart)
Shewhart sample range (R-chart)
Shewhart sample (X-chart) Cumulative sum (CUSUM)
Exponentially Weighted Moving Average (EWMA) chart
Moving-average and range charts
-- fordiscrete (attributes and countable) data
sample proportion defective (p-chart) sample number of defectives (np-chart)
sample number of defects (c-chart)
sample number of defects per unit (u-chart)
x
-
7/29/2019 06control Chart 1
25/37
Control charts: assumptions
Control charts make assumptions about theplotted statistic, namely
-- it is independent, i.e. a value is not influenced
by its past value and will not affect futurevalues
-- it is normallydistributed, i.e. the data has a
normal probability density function
-
7/29/2019 06control Chart 1
26/37
Normal Probability Density Function
The assumptions of normality and independence enable
predictions to be made about the data.
-
7/29/2019 06control Chart 1
27/37
Control charts: interpretation
Control charts are normal distributions with anadded time dimension.
-
7/29/2019 06control Chart 1
28/37
Control charts: interpretation
Control charts are run charts with superimposednormal distributions.
-
7/29/2019 06control Chart 1
29/37
Control charts: run rules
Run rules are rules that are used to indicate out-of-statistical control situations. Typical run rules forShewhart X-charts with control and warning limits are:
-- a point lying beyond the control limits
-- 2 consecutive points lying beyond the warning limits,namely within the area of 2 ~3 or even beyond thecontrol limits
-- 7 or more consecutive points lying on one side of themean
-- 5 or 6 or more consecutive points going in the samedirection (indicates a trend)
-- Other run rules can be formulated using similar principles
-
7/29/2019 06control Chart 1
30/37
-
7/29/2019 06control Chart 1
31/37
UCL
CL
LCL
X
LCL
CL
UCL
X
attention investigate Prompt action
Trend
-
7/29/2019 06control Chart 1
32/37
Control charts: run rules
If only chance variation is present in theprocess, the points plotted on a control
chart will not typically exhibit any pattern.
If the points exhibit some systematicpattern, this is an indication that
assignable variation may be present and
corrective action should be taken.
-
7/29/2019 06control Chart 1
33/37
Attribute Measurements (p-Chart)
p =T o ta l N u m b e r o f D e fe c tiv e s
T o ta l N u m b e r o f O b s e rv a tio n s
n
s
)p-(1p
=p
p
p
3-p=LCL3+p=UCLs
s
Given:
Compute control limits:
-
7/29/2019 06control Chart 1
34/37
Example of Constructing a p-chart:
Sample n Defectives p1 100 4 0.042 100 2 0.023 100 5 0.054 100 3 0.035 100 6 0.066 100 4 0.047 100 3 0.038 100 7 0.079 100 1 0.01
10 100 2 0.02
11 100 3 0.0312 100 2 0.0213 100 2 0.0214 100 8 0.0815 100 3 0.03
1. Calculate thesample proportions, p
(these are what can
be plotted on thep-
chart) for eachsample.
-
7/29/2019 06control Chart 1
35/37
Example of Constructing a p-chart:
2. Calculate the average of the sample proportions.
0.036=1500
55=p
3. Calculate the standard deviation of the sample
proportion
.0188=100
.036)-.036(1=)p-(1p=pn
s
-
7/29/2019 06control Chart 1
36/37
Example of Constructing a p-chart:
4. Calculate the control limits.
3(.0188).036
UCL = 0.0924LCL = -0.0204(or 0)
p
p
3-p=LCL
3+p=UCL
s
s
-
7/29/2019 06control Chart 1
37/37
Example of Constructing a p-chart:5. Plot the individual sample proportions and the controllimits
UCL
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Sample number
p
UCL
CL