You are on page 1of 63

QUALITY CONTROL

ENGINEERING
(TI 184517)

WEEK #4
DEWANTI ANGGRAHINI, EFFI LATIFFIANTI
MANUFACTURING SYSTEMS LABORATORY
DEPARTMENT OF INDUSTRIAL ENGINEERING
©2019
OBJECTIVES
Our objectives for this section are to learn how to use control charts to monitor continuous data,
and the assumptions behind the charts, their application, and their interpretation.
INTRODUCTION
• There is no two natural items in any category are the same.
• Variation may be quite large or very small.
• If variation very small, it may appear that items are identical, but precision instruments will show
differences.
• Control charts are powerful aids to understanding the performance of a process over time.
INTRODUCTION (2)
• Control Charts help us to determine whether a process is in a state of statistical control or out-of-
control.
• To monitor output, we use a control chart
• we check things like the mean, range, standard deviation
• To monitor a process, we typically use two control charts
• mean (or some other central tendency measure)
• variation (typically using range or standard deviation)
INTRODUCTION (3)
• Variable is a single quality characteristic that can be measured on a numerical scale. Examples:
length, size, weight, height, time, velocity.
• When working with variables, we should monitor both the mean value of the characteristic and
the variability associated with the characteristic.
• Since statistical control for continuous data depends on both the mean and the variability,
variables control charts are constructed to monitor each. The most commonly used chart to
monitor the mean is called chart. There are two commonly used charts used to monitor the
variability: the R chart and the s chart.

xത
CONTROL CHART COMPONENTS
• Centerline
• shows where the process average is centered or the central tendency of the data
• Upper control limit (UCL) and Lower control limit (LCL)
• describes the process spread

(stephen.mitchell)
EXAMPLE

(a) Mean and standard deviation at nominal levels


(b) Process mean µ1 > µ2
(c) Process standard deviation 1 > 2
HOW TO DEVELOP A CONTROL CHART?
Define the problem

Select a quality characteristic to be measured

Choose a subgroup to be sampled

Collect the data

Determine trial centreline

Determine trial control limits

Determining an alternative value for the standard


deviation
CONTROL CHARTS FOR x
ത AND R

Notation for variables control charts


• n - size of the sample (sometimes called a subgroup) chosen at a point in time
• m - number of samples selected
• x i = average of the observations in the i-th sample (where i = 1, 2, ..., m)

• x = grand average or “average of the averages (this value is used as the center
line of the control chart)
CONTROL CHARTS FOR x
ത AND R (2)

Notation and values


• Ri = range of the values in the i-th sample
Ri = xmax – xmin

• R = average range for all m samples


•  is the true process mean
•  is the true process standard deviation
CONTROL CHARTS FOR x
ത AND R (3)

Statistical Basis of the Charts


• Assume the quality characteristic of interest is normally distributed with mean ,
and standard deviation, .
• If x1, x2, …, xn is a sample of size n, then he average of this sample is

x1  x 2    x n
x
n

• x is normally distributed with mean, , and standard deviation,


x  / n
CONTROL CHARTS FOR x
ത AND R (4)
• The probability is 1 -  that any sample mean will fall between


  Z / 2 x    Z / 2
n
and

  Z / 2 x    Z / 2
n
• The above can be used as upper and lower control limits on a control chart
for sample means, if the process parameters are known.
CONTROL CHARTS FOR x
ത AND R (5)

Control Limits for the x chart

x1  x 2    x m
x UCL  x  A2 R
m
Center Line  x
R
R1  R2    Rm LCL  x  A2 R
m

• A2 is found in Appendix VI for various values of n.


CONTROL CHARTS FOR x
ത AND R (6)

Control Limits for the R chart

UCL  D4 R
Center Line  R
LCL  D3 R

• D3 and D4 are found in Appendix VI for various values of n.


CONTROL CHARTS FOR x
ത AND R (7)
Estimating the Process Standard Deviation
• The random variable W = R/ is called relative range.
• The parameters of distribution of W are a function of the sample size n.
• The mean of W is d2.
• The process standard deviation can be estimated using a function of the sample
average range.

R

 
d2

• This is an unbiased estimator of 


CONTROL CHARTS FOR x
ത AND R (8)
R
• If we use x as an estimator of µ and as an estimator of ,
d2
then the parameter of the x chart are

If we define
CONTROL CHARTS FOR x
ത AND R (9)

• Now consider the R chart. The center line will be R .

• To determine the control limits, we need an estimate of R . Assuming that the


quality characteristic is normally distributed, can be found from the distribution
of the relative range W = R/ . The standard deviation of W, say d3 , is a known
function of n. Thus

R=W

• the standard deviation of R is


 R = d3 
CONTROL CHARTS FOR x
ത AND R (10)
• Since  is unknown, we may estimate R by

• Consequently, the parameters of the R chart with usual three sigma control limits are

• If we let
CONTROL CHARTS FOR x
ത AND R (11)

Estimating Process Capability


• The x-bar and R charts give information about the capability of the process relative to its
specification limits.
• Assumes a stable process.
• We can estimate the fraction of nonconforming items for any process where specification
limits are involved.
• Assume the process is normally distributed, and x is normally distributed, the fraction
nonconforming can be found by solving:
P(x < LSL) + P(x > USL)
CONTROL CHARTS FOR x
ത AND R (12)
Process-Capability Ratios (Cp)
• Used to express process capability.
• For processes with both upper and lower control limits,
Use an estimate of  if it is unknown.

USL  LSL
Cp 
6
• If Cp > 1, then a low number of nonconforming items will be produced.
• If Cp = 1, (assume norm. dist) then we are producing about 0.27% nonconforming.
• If Cp < 1, then a large number of nonconforming items are being produced.
CONTROL CHARTS FOR x
ത AND R (13)

Process-Capability Ratios (Cp)


• The percentage of the specification band that the process uses up is denoted by

 1 
P̂   100%
C 
 p

**The Cp statistic assumes that the process mean is centered at the midpoint of the
specification band – it measures potential capability.
CASE STUDY 1

A hard bake process is used in conjunction with photolithography in semiconductor


manufacturing. We wish to establish statistical control of the flow width of the resist in this
process using xത and R charts. Twenty five samples, each of size five wafers, have been taken
when we think the process is in control. The interval of time between samples is one hour. The
flow width measurement data (in microns) from the samples are shown in table 5.1
CASE
STUDY
1
SOLUTION - CASE STUDY 1
OUTPUT MINITAB - CASE STUDY 1
ESTIMATING PROCESS CAPABILITY - CASE STUDY 1

Where the value of d2 for samples of size five is found in Appendix VI. The specification limits
on flow width are 1.50 ± 0.50 microns. The control chart data may be used to describe the
capability of the process to produce wafers relative to this specifications. Assuming that flow
width is a normally distributed random variable, with mean 1.5056 and standard deviation
0.1398, we may estimate the fraction of non-conforming wafers produced as
PROCESS CAPABILITY AND PROCESS CAPABILITY RATIO - CASE STUDY 1
PROCESS FALLOUT AND THE PROCESS CAPABILITY RATIO
PHASE II – CASE STUDY 1
PHASE II – CASE STUDY 1
Revision of control limits and center lines
PHASE II – CASE STUDY 1

Output minitab
CONTROL LIMITS, SPECIFICATION LIMITS, AND NATURAL TOLERANCE
LIMITS

• Control limits are functions of the


natural variability of the process
• Natural tolerance limits represent
the natural variability of the
process (usually set at 3-sigma
from the mean)
• Specification limits are
determined by
developers/designers.
• There is no mathematical
relationship between control
limits and specification limits.
RATIONAL SUBGROUPS
• X bar chart monitors the between sample variability
• R chart monitors the within sample variability.
• The estimate of  reflects only within sample variability, so it is not correct to estimate 
based on the usual quadratic estimator.
THE AVERAGE RUN LENGTH FOR THE xത CHART
• For any Shewhart control chart, we have noted previously that the ARL can be expressed as

*ARL adalah jumlah rata rata titik atau sampel yang harus digambarkan sebelum suatu titik/sampel
menyatakan keadaan tidak terkendali
AVERAGE TIME TO SIGNAL
• The average time to signal is the number of time periods that occur until a signal is generated on
the control chart. If samples are taken at equally spaced intervals of time h, then the average
time to signal or ATS is

• It may also be useful to express the ARL in terms of the expected number of individual units
sampled , I , rather than the number of samples taken to detect a shift. If the sample size is n, the
relationship between I and ARL is

*ATS adalah rata-rata panjang periode untuk mendapatkan 1 kondisi out of control
CONTROL CHART FOR xത AND S

• Although x ത and R charts are widely used, it is occasionally describe to estimate the process
standard deviation directly instead of indirectly through the use of the range R.
• This lead to control charts for xത and s, where s is the sample standard deviation. Generally,
and s charts are preferable to their more familiar counterparts, x
ത and R charts when either :
1. The sample size n is moderately large, n > 10 or 12.
2. The sample size n is variable.xത
CONSTRUCTION OF x
ത AND S CHARTS
• If the 2 is the unknown variance of a probability distribution, then an unbiased estimator of 2 is
the sample variance.

• If no standard is given for , then it must be estimated by analyzing past data. Suppose that m
preliminary samples are available, each of size n, and let si be the standard deviation of the i-th
sample. The average of the m standard deviation
CONSTRUCTION OF xത AND S CHARTS (2)

ҧ 4 is an unbiased estimator of  . Therefore, the parameters of the s chart would


• The statistic 𝑠/c
be

• We usually define the constants


CONSTRUCTION OF xത AND S CHARTS (3)
• Consequently, we may write the parameters of the s chart as

• Note that B4 = B6/c4 and B3 = B5/c4


CONSTRUCTION OF xത AND S CHARTS (4)
CASE
STUDY
2
SOLUTION - CASE STUDY 2
The grand average and the average standard deviation are

Consequently, the parameters for the xത chart are

and for the s chart


CONTROL CHARTS - CASE STUDY 2
ESTIMATION OF  - CASE STUDY 2

We can estimate the process standard deviation using the fact that s/c4
is an unbiased estimate of . Therefore, since c4 = 0.9400 for sample
size 5, out estimate of the process standard deviation is

𝑠ҧ
𝜎ො =
𝑐4
0.0094
= = 0.01
0.9400
THE ഥx AND S CONTROL CHARTS WITH VARIABLE SAMPLE SIZE
CASE
STUDY
3
SOLUTION - CASE STUDY 3
Consider the data in Table 5.4 which is a modification of the piston ring data used in Case Study
2. Note that the sample sizes vary from n = 3 to n = 5. Use the procedure decribed before to set
up the x bar and s control chart.
CONTROL CHARTS - CASE STUDY 3

Therefore the control limits for xത chart are

And for s chart


OUTPUT MINITAB - CASE STUDY 3
THE SHEWHART CONTROL CHART FOR INDIVIDUAL MEASUREMENT

There are many situations in which the sample size used for process monitoring is n = 1, that is, the
sample consists of an individual unit. Some examples of these situations are as follows:
1. Automated inspection and measurement technology is used, and every unit manufactured is
analysed so there is no basis for rational subgrouping.
2. Data comes available relatively slowly, and it is inconvenient to allow sample sizes of n > 1 to
accumulate before analysis. The long interval between observations will cause problems with
rational subgrouping. This occurs frequently in both manufacturing and non-manufacturing
situations.
3. Repeat measurements on the process differ only because of laboratory or analysis erros, as in
many chemical processes.
4. Multiple measurements are taken on the same unit of product, such as measuring oxide
thickness at several different locations on a wafer in semiconductor manufacturing.
5. etc
THE SHEWHART CONTROL CHART FOR INDIVIDUAL MEASUREMENT (2)
CASE
STUDY
4
SOLUTION - CASE STUDY 4

Viskositas cat primer pesawat merupakan karakteristik mutu penting. Produk ini diproduksi
dalam batch, dan karena setiap batch membutuhkan waktu beberapa jam untuk memproduksi,
tingkat produksi terlalu lambat untuk memungkinkan subkelompok rasional ukuran lebih besar dari
satu. Viskositas dari 20 batch sebelumnya disajikan pada tabel 5-6.

Untuk mengatur peta kendali untuk pengamatan individu, diketahui bahwa rata-rata sampel
dari 20 pembacaan viskositas x bar = 34,088 dan bahwa rata-rata moving range dari dua
pengamatan adalah MR = 0,573. untuk mengatur berbagai moving range chart, kita gunakan D3 dan
D4 = 0 = 3,267 untuk n = 2. Oleh karena itu, berbagai moving range chart memiliki garis tengah MR =
0,5726, LCL 0 =, dan UCL = D4 MR = (3,267) 0,5726 = 1,871. peta kendali (dari minitab) ditunjukkan
pada Gambar 5-19b. Perhatikan bahwa titik untuk sampel 4 berada di luar kendali.
SOLUTION - CASE STUDY 4
Untuk peta kendali untuk pengukuran individu, parameter yaitu:

jika moving range


sebesar n = 2
pengamatan
digunakan, maka
d2 = 1,128. untuk
data dalam tabel
5-6 kita memiliki
CONTROL CHART- CASE STUDY 4
PHASE II OPERATION
AND INTERPRETATION
OF THE CHART –
CASE STUDY 4
PHASE II OPERATION AND
INTERPRETATION OF THE
CHART –
CASE STUDY 4
PHASE II OPERATION AND
INTERPRETATION OF THE
CHART –
CASE STUDY 4
AVERAGE RUN LENGTH
- Crowder (1987b) menunjukkan bahwa ARL0 individu gabungan dan chart moving-range dengan
batas 3-sigma konvensional yang pada umumnya jauh lebih kecil daripada ARL0 (= 370) sebesar
standard Shewhart grafik pengendali

- Kemampuan individu untuk mendeteksi pergeseran kecil sangat buruk, Bukannya


mempersempit batas 3-sigma, pendekatan yang tepat untuk mendeteksi perubahan kecil adalah
jumlah-kumulatif atau secara grafik kendali eksponensial weighted moving-average (Bab 8)
CASE
STUDY
5
THANK YOU

You might also like