Chapter 3

Other Statistical Process-Monitoring and Control

Techniques
3.1. Cumulative Sum (CUSUM ) x-bar control charts
 The CUSUM (cumulative sum) chart directly
incorporates all the information in the sequence of
sample values by plotting the cumulative sums of the
deviations of the sample values from a target value.
 Suppose that samples of size n 1 are collected, and is
the average of the ith sample.

1
 Then if is the target for the process mean, the cumulative
sum control chart is formed by plotting the quantity,

Cj = against the sample number j. Cj is called the

cumulative sum(CUMSUM).
 Because they combine information from several samples,
cumulative sum charts are more effective than Shewhart
charts for detecting small process shifts.
 Furthermore, they are particularly effective with samples
of size n = 1.
2
 This makes the cumulative sum control chart a good
candidate for use in the chemical and process
industries where rational subgroups are frequently of
size one.

 If the process remains in control at the target value 𝜇0,

the cumulative sum defined in E(Cj) = 0 and Var(Ci) =

j
 If the process is in control the graph should be neither
increasing nor decreasing.
3
 However, if the mean shifts upward to some value; > ,
say, then an upward or positive drift will develop in the
cumulative sum Cj.

 Conversely, if the mean shifts downward to some < ,

then a downward or negative drift in Cj will develop.

 In short, a positive or negative slope implies the

process is out of control.

4
 Example: Consider the following data for 30 samples
each of size n=1
Sample number 1 2 3 4 5 6 7 8 9 10
𝑋𝑖 9.45 7.99 9.29 11.66 12.16 10.18 8.04 11.46 9.20 10.34

Sample number 11 12 13 14 15 16 17 18 19 20
𝑋𝑖 9.03 11.47 10.51 9.40 10.08 9.37 10.62 10.31 8.52 10.84

Sample number 21 22 23 24 25 26 27 28 29 30
𝑋𝑖 10.90 9.33 12.29 11.50 10.60 11.08 10.38 11.62 11.31 10.52

5
 Let the target value = 10, then the CUSUM
becomes:
Cj = = ( - 10) + = ( - 10) + Cj-1,

where the starting value for the CUSUM, C o, is

taken to be zero.

6
7
8
 There are two ways to represent CUSUMs, the tabular
(or algorithmic) CUSUM, and the V-mask form of the
CUSUM.
 Note that for the first 20 observations where 𝜇 = 10,
the CUSUM tends to drift slowly, in this case
maintaining values near zero. However, in the last 10
observations, where the mean has shifted to 𝜇 = 11, a
strong upward trend develops

9
3.2. The Exponential Weighted Moving Average
Control Chart(EWMA)
 The exponentially weighted moving average
(EWMA) control chart is also a good alternative to the
Shewhart control chart when we are interested in
detecting small shifts.
 The performance of the EWMA control chart is
approximately equivalent to that of the cumulative
sum control chart, and in some ways it is easier to set
up and operate. 10
 As with the CUSUM the EWMA is typically used with
individual observations, and also give the results for
rational subgroups of size n > 1.
 The EWMA control chart was introduced by Roberts
(1959).
 The exponentially weighted moving average is defined
as:
Zi = Xi + (1 - )Zi-1,
where 0 < 1 is a constant and the starting value of
(required with the first sample at i =1) is the process
11
 Sometimes the average of preliminary data is used as the
starting value of the EWMA, so that Zo=.

 To show that the EWMA Zi is a weighted average of all

previous sample means, we may substitute for Zi-1 on the

right-hand side of the above equation to obtain:
 Zi = Xi + (1 - )[Xi-1 + (1 - )Zi-2]

= Xi +(1 - )Xi-1 + (1 - )2Zi-2 continuing to substituting

for Zi-j , j=2, 3, 4, …, t, we obtain:

Zi = Xt-j + (1 - )iZ0.
12
 The weights, decrease geometrically with the age of the
sample mean. Furthermore, the weights sum to unity.
Choice of weighting factor:
 The parameter 𝜆 determines the rate at which 'older' data
enter into the calculation of the EWMA statisti
 A value of = 1 implies that only the most recent
measurement influences the EWMA (degrades to
Shewhart chart). Thus, a large value of = 1 gives more
weight to recent data and less weight to older data; a
small value of gives more weight to older data. 13
 Since the EWMA can be viewed as a weighted average
of all past and current observations, it is very insensitive
to the normality assumption.
 If the observations Xi are independent random variables

with variance then the variance of Zi is: .

 Therefore, the EWMA control chart would be
constructed by plotting Zi versus the sample number i (or
time). The center line and Control limits for the EWMA
control chart are as follows.
14
Exercise: Proof Variance of Zi equals to
.

UCL =
CL =
LCL = ,
where the factor L in the above equations is the width of
the control limits.

15
 Note that the term in the equations above approaches
unity as i gets larger. This means that after the EWMA
control chart has been running for several time
periods, the control limits will approach steady-state
values given by:
UCL = and
LCL =

16
Example: Recall the example discussed above for
CUSUM with the target value of the mean is =10 and the
standard deviation is =1. Apply EWMA control chart with
= 0.10 and L=2.7.
Solution: Then the first value of EWMA is:
Z1 = X1 + (1 - )Z0 = 0.1(9.45) + (1 – 0.1)(10)
= 9.945;
Z2 = X2 + (1 - )Z1 = 0.1(7.99) + (1 – 0.1)(9.945)
= 9.7495; etc.
17
Sample i 1 2 3 4 5 6 7 8 9 10
𝑋𝑖 9.45 7.99 9.29 11.66 12.16 10.18 8.04 11.46 9.20 10.34
Zi 9.95 9.75 9.7 9.9 10.13 10.13 9.92 10.08 9.99 10.02

Sample i 11 12 13 14 15 16 17 18 19 20
𝑋𝑖 9.03 11.47 10.51 9.40 10.08 9.37 10.62 10.31 8.52 10.84
Zi 9.92 10.08 10.12 10.05 10.05 9.98 10.05 10.07 9.92 10.01

18
Sample i 21 22 23 24 25 26 27 28 29 30
𝑋𝑖 10.90 9.33 12.29 11.50 10.60 11.08 10.38 11.62 11.31 10.52
Zi 10.1 10.02 10.25 10.37 10.4 10.47 10.46 10.57 10.65 10.63

UCL = and
LCL = ,
i.e., UCL = = 10.62 and

LCL = = 9.38.

19
 The jagged line is the EWMA statistic over time and
the control chart tells us that the process is in control
because all EWMA lie between the control limits.
However, there seems to be a trend upwards for the
last 5 or 6 periods. 20
3.3. Uni-variate Process Monitoring and
Control
Individual control charts
 It is the same as x-bar chart with n=1.
 The moving range is defined as

 For the control chart for individual measurements, the

lines plotted are:

21
where is the average of all the individuals and
is the average of all the moving ranges of two
observations. Keep in mind that either or both averages
may be replaced by a standard or target, if available.

(Note that 1.128 is the value of d2 for n = 2).

22
Example: The following example illustrates the control chart for
individual observations. A new process was studied in order to
monitor flow rate. The first 10 batches resulted in the following
table. Batch Number Flow rate x Moving Range R

1 49.6

2 47.6 2.0

3 49.9 2.3

4 51.3 1.4

5 47.8 3.5

6 51.2 3.4

7 52.6 1.4

8 52.4 0.2

9 53.6 1.2

10 52.1 1.5

= 50.81 = 1.8778

23
This yields the parameters below.

24
3.4. Multivariate Process Monitoring and Control
 There are many situations in which the simultaneous
monitoring or control of two or more related quality
characteristics is necessary.
 For example, suppose that a bearing has both an inner
diameter (x1) and an outer diameter (x2) that together
determine the usefulness of the part. Because both quality
characteristics are measurements, they could be monitored
by applying variable control chart and monitoring these two
quality characteristics independently can be very misleading.
25
 Rather when we examine the two variables simultaneously,
the unusual behavior of the point is fairly obvious.
 The most familiar multivariate process-monitoring and
control procedure is the Hotelling T2 control chart for
monitoring the mean vector of the process. It is a direct
analog of the univariate Shewhart X- bar chart.
 These multivariate control charts work well when the
number of process variables is not too large, say, 10 or
fewer.
26
27
Example:

28
29
30
31
Chi-square control chart
 The test statistic plotted on the chi-square
control chart for each sample is:

=n(x-(x–

The upper limit on the control chart is:

UCL =

32
Individual Observations (sample size n=1)
Let and S be the sample mean vector and covariance
matrix. (m=number of samples)
Hotelling T2 statistic
T2=(x-(x– and the control limits for this statistic are:

33
 When the number of preliminary samples m is
large--say, m>100--many practitioners use an
approximate control limit, either

34