Statistical Process Control Using Control Charts

1
 Control Charts

• A control chart is a graphical tool for monitoring the activity of an ongoing

process.
• Control charts are sometimes referred to as Shewhart control charts,
because Walter A. Shewhart first proposed their general theory.
• The values of the quality characteristic are plotted along the vertical axis,
and the horizontal axis represents the samples, or subgroups (in order of
time), from which the quality characteristic is found.
• Samples of a certain size (say, four or five observations) are selected, and
the quality characteristic (say, average length) is calculated based on the
number of observations in the sample.
• These characteristics are then plotted in the order in which the samples
were taken.

2
Typical control chart

3
• Examples of quality characteristics include average length, average waiting
time, average tensile strength, average resistance, and average service time.
These characteristics are variables, and numerical values can be obtained for
each.

• The term attribute applies to such quality characteristics as the proportion of

nonconforming items, the number of nonconformities in a unit, and the
number of demerits per unit.

• Three lines are indicated on the control chart.

• The centerline, which typically represents the average value of the
characteristic being plotted, is an indication of where the process is centered.
Two limits, the upper control limit and the lower control limit, are used to
make decisions regarding the process.

4
• If the points plot within the control limits and do not exhibit any
identifiable pattern, the process is said to be in statistical control.

• If a point plots outside the control limits or if an identifiable nonrandom

pattern exists (such as 12 out of 14 successive points plotting above the
centerline), the process is said to be out of statistical control.

5
• Several benefits can be realized by using control charts. Such charts indicate
the following:

• 1. When to take corrective action. A control chart indicates when something

may be wrong so that corrective action can be taken.
• 2. Type of remedial action necessary. The pattern of the plot on a control chart
diagnoses possible causes and hence indicates possible remedial actions.
• 3. When to leave a process alone. Variation is part of any process. A control
chart shows when an exhibited variability is normal and inherent such that no
corrective action is necessary.
• 4. Process capability. If the control chart shows a process to be in statistical
control, we can estimate the capability of the process and hence its ability to
meet customer requirements. This helps product and process design
• 5. Possible means of quality improvement. The control chart provides a
baseline for instituting and measuring quality improvement. Control charts
also provide useful information regarding actions to take for quality
improvement.

6
 CAUSES OF VARIATION

• Variability is a part of any process, no matter how sophisticated ,so

management and employees must understand it.

• Several factors over which we have some control, such as methods,

equipment, people, materials, and policies, influence variability.

• Environmental factors also contribute to variability. The causes of variation

can be subdivided into two groups: common causes and special causes.

• Control of a process is achieved through the elimination of special causes.

Improvement of a process is accomplished through the reduction of
common causes.

7
 Special Causes

• Variability caused by special or assignable causes is something that is not

inherent in the process.
• That is, it is not part of the process as designed and does not affect all items.
• Special causes can be the use of wrong tool, an improper raw material, or an
incorrect procedure.
• If an observation falls outside the control limits or a nonrandom pattern is
exhibited, special causes are assumed to exist, and the process is said to be
out of control.
• One objective of a control chart is to detect the presence of special causes as
soon as possible to allow appropriate corrective action.
• Once the special causes are eliminated through remedial actions, the process
is again brought to a state of statistical control.
• Deming believed that 15% of all problems are due to special causes. Actions
on the part of both management and employees will reduce the occurrence of
such causes.

8
 Common Causes
• Variability due to common or chance causes is something inherent to a process.
• It exists as long as the process is not changed and is referred to as the natural
variation in a process.
• It is an inherent part of the process design and affects all items.
• This variation is the effect of many small causes and cannot be totally
eliminated.
• When this variation is random, we have what is known as a stable system of
common causes.
• A process operating under a stable system of common causes is said to be in
statistical control.
• Examples include inherent variation in incoming raw material from a qualified
vendor, the vibration of machines, and fluctuations in working conditions.
• Management alone is responsible for common causes. Deming believed that
about 85% of all problems are due to common causes and hence can be solved
only by action on the part of management.

9
 STATISTICAL BASIS FOR CONTROL CHARTS

• A control chart has a centerline and lower and upper control limits.
• The centerline is usually found in accordance with the data in the samples.
• It is an indication of the mean of a process and is usually found by taking
the average of the values in the sample.
• Normal distributions play an important role in the use of control charts
(Duncan 1986).The values of the statistic plotted on a control chart(e.g.,
average diameter) are assumed to have an approximately normal
distribution
• The control limits are two lines, one above and one below the centerline,
that aid in the decision-making process. These limits are chosen so that the
probability of the sample points falling between them is almost 1 (usually
about 99.7% for 3σ limits) if the process is in statistical control.

10
• If a system is operating under a stable system of common causes, it is
assumed to be in statistical control.
• Typical control limits are placed at three standard deviations away from the
mean of the statistic being plotted.
• Normal distribution theory states that a sample statistic will fall within the
limits 99.74% of the time if the process is in control.
• If a point falls outside the control limits, there is a reason to believe that a
special cause exists in the system.
• We must then try to identify the special cause and take corrective action to
bring the process back to control.
• If a process is under statistical control, control chart information can estimate
such process parameters as the mean, standard deviation, and proportion of
nonconforming (also known as fallout). These estimates can then be used to
determine the capability of the process.

11
• Control charts for variables are constructed to show measures of central
tendency as well as dispersion.

• Variable control charts display such information as sample mean, sample

range, sample standard deviation, cumulative sum, individual values, and
moving average.

• Attributes, on the other hand, indicate the presence or absence of a

condition.

• Typical attribute charts deal with the fraction of nonconforming items, the
number of nonconforming items, the total number of nonconformities, the
number of nonconformities per unit, or the number of demerits per unit.

12
 Selection of Control Limits

• Let θ represent a quality characteristic of interest and represent an estimate

of represent an estimate of θ.
• For example, if θ is the mean diameter of parts produced by a process,
would be the sample mean diameter of a set of parts chosen from the
process.
• Let E …† represent the mean, or expected value, and let σ … be the
standard deviation of the estimator
• The centerline and control limits for this arrangement are given by:

13
• Where k represents the number of standard deviations of the sample statistic
that the control limits are placed from the center line.

• Typically, the value of k is chosen to be 3 (hence the name 3σ limits).

• If the sample statistic is assumed to have an approximately normal

distribution, a value of k = 3 implies that there is a probability of only 0.0026
of a sample statistic falling outside the control limits if the process is in
control.

14
• Ques. A semiautomatic turret lathe machines the thickness of a part that is
subsequently used in an assembly. The process mean is known to be 30 mm
with a standard deviation of 1.5mm. Construct a control chart for the
average thickness using 3σ limits if samples of size 5 are randomly selected
from the process. Table below shows the average thickness of 15 samples
selected from the process. Plot these on a control chart and make
inferences.

15
• Ans The centerline is CL ˆ 30 mm
• The standard deviation of the sample mean X̅ is given by

• Assuming a normal distribution of the sample mean thickness, the value of k in

eq. is selected as 3.
• The control limits are calculated as follows:

16
• The centerline and control limits are shown in figure below. This also
shows that all of the sample means are within the control limits. Also, the
pattern of the plot does not exhibit any nonrandom behavior. Thus, we
conclude that the process is in control.

17
 Errors in Making Inferences from Control Charts

• Making inferences from a control chart is analogous to testing a

hypothesis. Suppose that we are interested in testing the null hypothesis
that the average diameter θ of a part from a particular process is 25 mm.
• This situation is represented by the null hypothesis H0:θ=25; the alternative
hypothesis is Ha: θ≠ 25.
• The rejection region of the null hypothesis is thus two tailed.
• The control limits are the critical points that separate the rejection and
acceptance regions. If a sample value (sample average diameter, in this
case) falls above the upper control limit or below the lower control limit,
we reject the null hypothesis.
• In such a case, we conclude that the process mean differs from 25 mm and
the process is therefore out of control.

18
 Type I Errors

• Type I errors result from inferring that a process is out of control when it is
actually in control.
• The probability of a type I error is denoted by α.
• Suppose that a process is in control. If a point on the control chart falls
outside the control limits, we assume that the process is out of control.
• However, since the control limits are a finite distance (usually, three
standard deviations)from the mean, there is a small chance of a sample
statistic falling outside the control limits.
• In such instances, inferring that the process is out of control is a wrong
conclusion.
• Probability of making a type I error in control charts is the sum of the two
tail areas outside the control limits.

19
20
 Type II Errors

• Type II errors result from inferring that a process is in control when it is

really out of control.
• If no observations fall outside the control limits and a random pattern is
exhibited, we conclude that the process is in control.
• Suppose, however, that a process is actually out of control. Perhaps the
process mean has changed (say, an operator has inadvertently changed a
depth of cut or the quality of raw materials has decreased). Or, the process
could go out of control because the process variability has changed (due to
the presence of a new operator).
• Under such circumstances, a sample statistic could fall within the control
limits, yet the process would be out of control—this is a type II error.

21
Figure depicts a process going out of control due to a change in the process mean
from A to B.
For this situation, the correct conclusion is that the process is out of control.
However, there is a strong possibility of the sample statistic falling within the
control limits (as indicated by the shaded area), in which case we would conclude
that the process is in control and thus make a type II error. 22
• Ques1(a). A control chart is to be constructed for the average breaking
strength of nylon fibers. Samples of size 5 are randomly chosen from the
process. The process mean and standard deviation are estimated to be 120
kg and 8 kg, respectively.
• (a) If the control limits are placed three standard deviations from the
process mean, what is the probability of a type I error?

23
• Ans 1. From the problem statement, μ̂ˆ =120 and σ̂ ˆ= 8. The centerline for
the control chart is at 120 kg. The control limits are

• For these Z-values in the standard normal table in Appendix A-3, each tail
area is found to be 0.0013.

• The probability of a type I error, as shown by the shaded tail areas in Figure
in next slide, is therefore 0.0026.

24
Similarly, the Z-value at the lower control limit is 3.00

25
26
Q (b) If the process mean shifts to 125 kg, what is the probability of
concluding that the process is in control and hence making a type II error on
the first sample plotted after the shift?

Ans b. The process mean shifts to 125 kg. Assuming that the process standard
deviation is the same as before, the distribution of the sample means is as
shown.

27
• The probability of concluding that the process is in control is equivalent to
finding the area between the control limits under the distribution.
• We find the standardized normal value at the upper control limit as

• From Appendix A-3, the tail area below the lower control limit is
approximately 0.0000. The area between the control limits is 1-(0.0548+
0.0000)=0.9452. Hence, the probability of concluding that the process is
in control and making a type II error is 0.9452, or 94.52%.

28