Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery.

1
Copyright (c) 2013 John Wiley & Sons, Inc.
 Learning Objectives

Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 2

Copyright (c) 2013 John Wiley & Sons, Inc.
Chapter 9 Introduction to Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 3
Copyright (c) 2012 John Wiley & Sons, Inc.
 8.1 The Cumulative Sum Control Chart

Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 4

Copyright (c) 2013 John Wiley & Sons, Inc.
Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 5
Copyright (c) 2013 John Wiley & Sons, Inc.
 The Cumulative Sum Control Chart

Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 6

Copyright (c) 2013 John Wiley & Sons, Inc.
• Of course, the cusum plot in Fig. 9.2 is not a control
chart because it lacks statistical control limits.
• There are two ways to represent cusums,
– the tabular (or algorithmic) cusum, and the
– V-mask form of the cusum.
• Of the two representations, the tabular cusum is
preferable.

Chapter 9 Introduction to Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 7

Copyright (c) 2012 John Wiley & Sons, Inc.
 The Tabular Cusum
• The tabular cusum works by accumulating derivations
from m0 that are above target with one statistic C+
and accumulating derivations from m0 that are below
target with another statistic C−.
• The statistics C+ and C− are called one-sided upper
and lower cusums, respectively.

One-sided upper and

lower CUSUMs

Chapter 9 8
 The Tabular Cusum

One-sided upper and

lower CUSUMs

Chapter 9 9
• The action taken following an out-of-control signal
on a cusum control scheme is identical to that with
any control chart; one should search for the
assignable cause, take any corrective action required,
and then reinitialize the cusum at zero.
• The cusum is particularly helpful in determining
when the assignable cause has occurred;
– just count backward from the out-of-control signal to the
time period when the cusum lifted above zero to find the
first period following the process shift. The counters N+
and N− are used in this capacity.

Chapter 9 Introduction to Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 10

Copyright (c) 2012 John Wiley & Sons, Inc.
Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 11
Copyright (c) 2013 John Wiley & Sons, Inc.
Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 12
Copyright (c) 2013 John Wiley & Sons, Inc.
Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 13
Copyright (c) 2013 John Wiley & Sons, Inc.
 Cusum Status Chart (Figure 9.3a)

• Each vertical bar represents the value of C+ i and C− i in period i.

• The observations xi for each period on the cusum status chart are plotted as
the solid dots.
Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 14
Copyright (c) 2013 John Wiley & Sons, Inc.
 MINITAB Version of Cusum Status Chart

Minitab calculates the lower Cusum this way

Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 15

Copyright (c) 2013 John Wiley & Sons, Inc.
Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 16
Copyright (c) 2013 John Wiley & Sons, Inc.
• Recommendations for Cusum Design
• The tabular cusum is designed by choosing values for the
reference value K and the decision interval H.
• It is usually recommended that these parameters be selected to
provide good average run length performance.
• There have been many analytical studies of cusum ARL
performance. Based on these studies, we may give some
general recommendations for selecting H and K.
• Define H = hs and K = ks, where s is the standard deviation of
the sample variable used in forming the cusum. Using h = 4 or
h = 5 and k = 1/2 will generally provide a cusum that has good
ARL properties against a shift of about 1s in the process mean.

Chapter 9 Introduction to Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 17

Copyright (c) 2012 John Wiley & Sons, Inc.
 Recommendations for Cusum Design

Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 18

Copyright (c) 2013 John Wiley & Sons, Inc.
Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 19
Copyright (c) 2013 John Wiley & Sons, Inc.
 The Standardized Cusum

Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 20

Copyright (c) 2013 John Wiley & Sons, Inc.
 Improving Cusum Performance for Large Shifts: The
Combined Shewhart-Cusum Scheme

• We have observed that the cusum control chart is very

effective in detecting small shifts.
• However, the cusum control chart is not as effective as the
Shewhart chart in detecting large shifts.
• An approach to improving the ability of the cusum control
chart to detect large process shifts is to use a combined
cusum-Shewhart procedure for on-line control.
• Adding the Shewhart control is a very simple modification of
the cumulative sum control procedure.
• The Shewhart control limits should be located approximately
3.5 standard deviations from the center line or target value
m0.
• An out-of-control signal on either (or both) charts constitute
an action signal.

Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 21

Copyright (c) 2013 John Wiley & Sons, Inc.
 Improving Cusum Performance for Large Shifts: The
Combined Shewhart-Cusum Scheme

Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 22

Copyright (c) 2013 John Wiley & Sons, Inc.
 The Fast Initial Response (FIR) Cusum

• This procedure was devised by Lucas and Crosier

(1982) to improve the sensitivity of a cusum at
process start-up.
• Increased sensitivity at process start-up would be
desirable if the corrective action did not reset the
mean to the target value.
• The fast initial response (FIR) or headstart
essentially just sets the starting values C0+ and C 0−
equal to some nonzero value, typically H/2. This is
called a 50% headstart.

Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 23

Copyright (c) 2013 John Wiley & Sons, Inc.
 The Fast Initial Response (FIR) Cusum

Target value = 100

K = 3, H = 12, headstart = H/2 = 6

Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 24

Copyright (c) 2013 John Wiley & Sons, Inc.
 H = 12 implies that the cusum signals at sample 3

Without the headstart, it would not signal until sample 6

Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 25

Copyright (c) 2013 John Wiley & Sons, Inc.
Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 26
Copyright (c) 2013 John Wiley & Sons, Inc.
 More on Cusums
• Cusums are often used to determine if a process has
shifted off a specified target because it is easy to
calculate the required adjustment
• One-sided cusums are often useful
• Cusums can also be used to monitor variability
• Cusums are available for other sample statistics
(ranges, standard deviations, counts, proportions)
• Rational subgroups and cusums

Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 27

Copyright (c) 2013 John Wiley & Sons, Inc.
Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 28
Copyright (c) 2013 John Wiley & Sons, Inc.
Chapter 9 Introduction to Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 29
Copyright (c) 2012 John Wiley & Sons, Inc.
Chapter 9 Introduction to Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 30
Copyright (c) 2012 John Wiley & Sons, Inc.
Chapter 9 Introduction to Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 31
Copyright (c) 2012 John Wiley & Sons, Inc.
Chapter 9 Introduction to Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 32
Copyright (c) 2012 John Wiley & Sons, Inc.
 9.2 The Exponentially Weighted Moving
Average Control Chart

The EWMA is

Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 39

Copyright (c) 2013 John Wiley & Sons, Inc.
Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 40
Copyright (c) 2013 John Wiley & Sons, Inc.
 Steady-state
control limits

Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 41

Copyright (c) 2013 John Wiley & Sons, Inc.
Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 42
Copyright (c) 2013 John Wiley & Sons, Inc.
Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 43
Copyright (c) 2013 John Wiley & Sons, Inc.
Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 44
Copyright (c) 2013 John Wiley & Sons, Inc.
 Design of the EWMA

Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 45

Copyright (c) 2013 John Wiley & Sons, Inc.
Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 46
Copyright (c) 2013 John Wiley & Sons, Inc.
Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 47
Copyright (c) 2013 John Wiley & Sons, Inc.
 Robustness of EWMA to
Non-normal Process Data

Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 48

Copyright (c) 2013 John Wiley & Sons, Inc.
Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 49
Copyright (c) 2013 John Wiley & Sons, Inc.
Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 50
Copyright (c) 2013 John Wiley & Sons, Inc.
Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 55
Copyright (c) 2013 John Wiley & Sons, Inc.
 Learning Objectives

Chapter 9 Statistical Quality Control, 7th Edition by Douglas C. Montgomery. 56

Copyright (c) 2013 John Wiley & Sons, Inc.