chap 17 statistical quality control
TRANSCRIPT
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 1/37
© 2008 Prentice-Hall, Inc.
Chapter 17
To accompanyQuantitative Analysis for Management , Tenth Edition ,
by Render, Stair, and Hanna Power Point slides created by Jeff Heyl
Statistical Quality Control
© 2009 Prentice-Hall, Inc.
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 2/37
© 2009 Prentice-Hall, Inc. 17 – 2
Learning Objectives
After completing this chapter, students will be able to:
Define the quality of a product or service
Develop four types of control charts: x, R, p, and c
Understand the basic theoretical underpinnings ofstatistical quality control, including the centrallimit theorem
Know whether a process is in control
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 3/37
© 2009 Prentice-Hall, Inc. 17 – 3
Chapter Outline
17.1 Introduction
17.2 Defining Quality and TQM
17.3 Statistical Process Control
17.4 Control Charts for Variables
17.5 Control Charts for Attributes
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 4/37
© 2009 Prentice-Hall, Inc. 17 – 4
Introduction
Quality is often the major issue in a purchasedecision
Poor quality can be expensive for both the
producing firm and the customer Quality management, or quality control (QC ),
is critical throughout the organization
Quality is important for manufacturing andservices
We will be dealing with the most importantstatistical methodology, statistical process control (SPC )
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 5/37
© 2009 Prentice-Hall, Inc. 17 – 5
Quality of a product or service is the degree towhich the product or service meets specifications
Increasingly, definitions of quality include anadded emphasis on meeting the customer‟s
needs Total quality management (TQM ) refers to a
quality emphasis that encompasses the entireorganization from supplier to customer
Meeting the customer‟s expectations requires anemphasis on TQM if the firm is to complete as aleader in world markets
Defining Quality and TQM
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 6/37
© 2009 Prentice-Hall, Inc. 17 – 6
Several definitions of quality “Quality is the degree to which a specific product
conforms to a design or specification.” (Gilmore, 1974)
“Quality is the totality of features and characteristics of
a product or service that bears on its ability to satisfystated or implied needs.” (Johnson and Winchell, 1989)
“Quality is fitness for use.” (Juran, 1974)
“Quality is defined by the customer; customers wantproducts and services that, throughout their lives, meetcustomers‟ needs and expectations at a cost thatrepresents value.” (Ford, 1991)
“Even though quality cannot be defined, you know whatit is.” (Pirsig, 1974)
Defining Quality and TQM
Table 17.1
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 7/37 © 2009 Prentice-Hall, Inc. 17 – 7
Statistical Process Control
Statistical process control involves establishingand monitoring standards, makingmeasurements, and taking corrective action as aproduct or service is being produced
Samples of process output are examined If they fall outside certain specific ranges, the
process is stopped and the assignable cause islocated and removed
A control chart is a graphical presentation of dataover time and shows upper and lower limits of theprocess we want to control
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 8/37 © 2009 Prentice-Hall, Inc. 17 – 8
Statistical Process Control
Patterns to look for in control charts
Normal behavior
Uppercontrol
limit
Target
Lowercontrollimit
One plot out above.Investigate for cause.
One plot out below.
Investigate for cause.
Figure 17.1
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 9/37 © 2009 Prentice-Hall, Inc. 17 – 9
Uppercontrollimit
Target
Lowercontrollimit
Uppercontrollimit
Target
Lowercontrollimit
Statistical Process Control
Patterns to look for in control charts
Two plots near lower control.Investigate for cause.
Run of 5 below central
line. Investigate for cause.
Two plots near upper controlInvestigate for cause.
Run of 5 above central line.
Investigate for cause.
Figure 17.1
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 10/37 © 2009 Prentice-Hall, Inc. 17 – 10
Statistical Process Control
Patterns to look for in control charts
Trends in either direction 5plots. Investigate for causeof progressive change.
Uppercontrol
limit
Target
Lowercontrollimit Erratic behavior.
Investigate.
Figure 17.1
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 11/37 © 2009 Prentice-Hall, Inc. 17 – 11
Statistical Process Control
Building control charts Control charts are built using averages of
small samples
The purpose of control charts is to distinguishbetween natural variations and variations due to assignable causes
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 12/37 © 2009 Prentice-Hall, Inc. 17 – 12
Statistical Process Control
Natural variations Natural variations affect almost every
production process and are to be expected,even when the process is n statistical control
They are random and uncontrollable When the distribution of this variation is
normal it will have two parameters
Mean, (the measure of central tendency of
the average) Standard deviation, (the amount by which
smaller values differ from the larger ones)
As long as the distribution remains withinspecified limits it is said to be “in control”
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 13/37 © 2009 Prentice-Hall, Inc. 17 – 13
Statistical Process Control
Assignable variations When a process is not in control, we must
detect and eliminate special (assignable )causes of variation
The variations are not random and can becontrolled
Control charts help pinpoint where a problemmay lie
The objective of a process control system is to provide a statistical signal when assignable causes of variation are present
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 14/37 © 2009 Prentice-Hall, Inc. 17 – 14
Control Charts for Variables
The x-chart (mean) and R-chart (range) arethe control charts used for processes thatare measured in continuous units
The x-chart tells us when changes haveoccurred in the central tendency of theprocess
The R-chart tells us when there has been a
change in the uniformity of the process Both charts must be used when monitoring
variables
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 15/37
© 2009 Prentice-Hall, Inc. 17 – 15
The Central Limit Theorem
The central limit theorem is the foundationfor x-charts
The central limit theorem says that the
distribution of sample means will follow anormal distribution as the sample sizegrows large
Even with small sample sizes the
distribution is nearly normal
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 16/37
© 2009 Prentice-Hall, Inc. 17 – 16
The Central Limit Theorem
The central limit theorem says
1. The mean of the distribution will equal thepopulation mean
2. The standard deviation of the samplingdistribution will equal the population standarddeviation divided by the square root of thesample size
n μx
x x
and
We often estimate and with the average of allsample means ( )
x
x
μ
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 17/37
© 2009 Prentice-Hall, Inc. 17 – 17
The Central Limit Theorem
x
Figure 17.2 shows three possible populationdistributions, each with their own mean () andstandard deviation ( )
If a series of random samples ( and
so on) each of size n is drawn from any of these,the resulting distribution of the „s will appear asin the bottom graph in the figure
Because this is a normal distribution1. 99.7% of the time the sample averages will fall between
±3 if the process has only random variations2. 95.5% of the time the sample averages will fall between±2 if the process has only random variations
If a point falls outside the±3 control limit, weare 99.7% sure the process has changed
,,,,4321
x x x x
i x
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 18/37
© 2009 Prentice-Hall, Inc. 17 – 18
|
–3 x
|
–1 x
|
+1 x
|
+2 x
|
–2 x
|
+3 x
|
x =
(mean) n
x
x
errorStandard
The Central Limit Theorem
Population and sampling distributions
Sampling Distribution of Sample Means (Always Normal)
Normal Beta Uniform
= (mean)
x = S.D.
= (mean)
x = S.D.
= (mean)
x = S.D.
99.7% of all xfall within±3 x
95.5% of all x fall within±2 x
Figure 17.2
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 19/37
© 2009 Prentice-Hall, Inc. 17 – 19
Setting the x-Chart Limits
If we know the standard deviation of the process,we can set the control limits using
x z xUCL )(limitcontrolUpper
x z xUCL )(limitcontrolLower
where
x
x
= mean of the sample means z = number of normal standard deviations
= standard deviation of the samplingdistribution of the sample means =
n
x
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 20/37
© 2009 Prentice-Hall, Inc. 17 – 20
Box Filling Example
ounces1711636
2316
x xz xUCL
ounces15116
36
2316
x xz x LCL
336216 z n x x ,,,
A large production lot of boxes of cornflakes issampled every hour
To set control limits that include 99.7% of thesample, 36 boxes are randomly selected andweighed
The standard deviation is estimated to be 2ounces and the average mean of all the samplestaken is 16 ounces
So and the control limits
are
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 21/37
© 2009 Prentice-Hall, Inc. 17 – 21
Box Filling Example
If the process standard deviation is not availableor difficult to compute (a common situation) theprevious equations are impractical
In practice the calculation of the control limits isbased on the average range rather than thestandard deviation
R A xUCL x 2
R A x LCL x 2
where
= average of the samples A2 = value found in Table 17.2
= mean of the sample means
R
x
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 22/37
© 2009 Prentice-Hall, Inc. 17 – 22
Factors for Computing Control Chart Limits
SAMPLE SIZE, n MEAN FACTOR, A2 UPPER RANGE, D4 LOWER RANGE, D3
2 1.880 3.268 0
3 1.023 2.574 0
4 0.729 2.282 0
5 0.577 2.115 0
6 0.483 2.004 07 0.419 1.924 0.076
8 0.373 1.864 0.136
9 0.337 1.816 0.184
10 0.308 1.777 0.223
12 0.266 1.716 0.284
14 0.235 1.671 0.329
16 0.212 1.636 0.364
18 0.194 1.608 0.392
20 0.180 1.586 0.414
25 0.153 1.541 0.459
Table 17.2
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 23/37
© 2009 Prentice-Hall, Inc. 17 – 23
Super Cola Example
Super Cola bottles are labeled “net weight 16ounces”
The overall process mean is 16.01 ounces andthe average range is 0.25 ounces
What are the upper and lower control limits forthis process?
R A xUCL x 2
16.01 + (0.577)(0.25) 16.01 + 0.144 16.154
R A x LCL x 2
16.01 – (0.577)(0.25) 16.01 – 0.144 15.866
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 24/37
© 2009 Prentice-Hall, Inc. 17 – 24
Setting Range Chart Limits
R
We have determined upper and lower controllimits for the process average
We are also interested in the dispersion orvariability of the process
Averages can remain the same even if variabilitychanges
A control chart for ranges is commonly used tomonitor process variability
Limits are set at±3 for the average range
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 25/37
© 2009 Prentice-Hall, Inc. 17 – 25
Setting Range Chart Limits
We can set the upper and lower controls using
where
UCL R = upper control chart limit for the range LCL R = lower control chart limit for the range D4 and D3 = values from Table 17.2
R DUCL R 4
R D LCL R 3
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 26/37
© 2009 Prentice-Hall, Inc. 17 – 26
A process has an average range of 53 pounds
If the sample size is 5, what are the upper andlower control limits?
From Table 17.2, D4
= 2.114 and D3
= 0
Range Example
R DUCL R 4
(2.114)(53 pounds)
112.042 pounds
R D LCL R 3
(0)(53 pounds)
0
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 27/37
© 2009 Prentice-Hall, Inc. 17 – 27
Five Steps to Follow in Using x and R-Charts
1. Collect 20 to 25 samples of n = 4 or n = 5 from astable process and compute the mean and rangeof each
2. Compute the overall means ( and ), set
appropriate control limits, usually at 99.7% leveland calculate the preliminary upper and lowercontrol limits. If process not currently stable, usethe desired mean, m, instead of to calculatelimits.
3. Graph the sample means and ranges on theirrespective control charts and determine whetherthey fall outside the acceptable limits
x
x
R
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 28/37
© 2009 Prentice-Hall, Inc. 17 – 28
Five Steps to Follow in Using x and R-Charts
4. Investigate points or patterns that indicate theprocess is out of control. Try to assign causes forthe variation and then resume the process.
5. Collect additional samples and, if necessary,
revalidate the control limits using the new data
chart x R-chart
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 29/37
© 2009 Prentice-Hall, Inc. 17 – 29
Control Charts for Attributes
We need a different type of chart tomeasure attributes
These attributes are often classified as
defective or nondefective There are two kinds of attribute control
charts1. Charts that measure the percent defective in
a sample are called p-charts 2. Charts that count the number of defects in a
sample are called c-charts
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 30/37
© 2009 Prentice-Hall, Inc. 17 – 30
p-Charts
Attributes that are good or bad typically followthe binomial distribution
If the sample size is large enough a normaldistribution can be used to calculate the controllimits
p p z pUCL
p p z p LCL
where
= mean proportion or fraction defective in the sample
z = number of standard deviations
= standard deviation of the sampling distribution whichis estimated by where n is the size of each sample
p
p
pˆ
n
p p p
)1(ˆ
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 31/37
© 2009 Prentice-Hall, Inc. 17 – 31
ARCO p-Chart Example
Performance of data-entry clerks at ARCO ( n = 100)
SAMPLENUMBER
NUMBER OFERRORS
FRACTIONDEFECTIVE
SAMPLENUMBER
NUMBER OFERRORS
FRACTIONDEFECTIVE
1 6 0.06 11 6 0.06
2 5 0.05 12 1 0.013 0 0.00 13 8 0.08
4 1 0.01 14 7 0.07
5 4 0.04 15 5 0.05
6 2 0.02 16 4 0.04
7 5 0.05 17 11 0.11
8 3 0.03 18 3 0.03
9 3 0.03 19 0 0.00
10 2 0.02 20 4 0.04
80
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 32/37
© 2009 Prentice-Hall, Inc. 17 – 32
ARCO p-Chart Example
We want to set the control limits at 99.7% of therandom variation present when the process is incontrol so z = 3
04020100
80
examinedrecordsofnumberTotal
errorsofnumberTotal
.))((
p
020100
0401040.
).)(.(ˆ p
1000203040 .).(.ˆ
p p z pUCL
00203040 ).(.ˆ p p z p LCL
Percentage can‟t be negative
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 33/37
© 2009 Prentice-Hall, Inc. 17 – 33
ARCO p-Chart Example
p-chart for data entry
UCL p = 0.10
LCL p = 0.00
040. p
0.12 –
0.11 –
0.10 – 0.09 –
0.08 –
0.07 –
0.06 –
0.05 –
0.04 –
0.03 – 0.02 –
0.01 –
0.00 –
F r a c t i o n
D e f e c t i v e
|
1
|
2
|
3
|
4
|
5
|
6
|
7
|
8
|
9
|
10
|
11
|
12
|
13
|
14
|
15
|
16
|
17
|
18
|
19
|
20
Sample NumberFigure 17.3
Out of Control
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 34/37
© 2009 Prentice-Hall, Inc. 17 – 34
ARCO p-Chart Example
Excel QM‟s p-chart program applied to the ARCOdata showing input data and formulas
Program 17.1A
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 35/37
© 2009 Prentice-Hall, Inc. 17 – 35
ARCO p-Chart Example
Output from Excel QM‟s p-chart analysis of theARCO data
Program 17.1B
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 36/37
© 2009 Prentice-Hall, Inc. 17 – 36
c-Charts
In the previous example we counted the number ofdefective records entered in the database
But records may contain more than one defect
We use c-charts to control the number of defects
per unit of output c-charts are based on the Poisson distribution
which has its variance equal to its mean
The mean is and the standard deviation is equalto
To compute the control limits we use
c
c
c c 3
8/3/2019 Chap 17 Statistical Quality Control
http://slidepdf.com/reader/full/chap-17-statistical-quality-control 37/37
Red Top Cab Company c-Chart Example
The company receives several complaints eachday about the behavior of its drivers
Over a nine-day period the owner received 3, 0, 8,9, 6, 7, 4, 9, 8 calls from irate passengers for atotal of 54 complaints
To compute the control limits
daypercomplaints69
54 c
Thus
3513452366363 .).( c cUCL c
0452366363 ).( c c LCL c