Sei sulla pagina 1di 40

Statistical basis of the control chart

1. Basic principles 2. Choice of control limits 3. Sample size and sampling frequency 4. Rational subgroups 5. Analysis of patterns on control charts 6. Discussion of sensitizing rules for control chart 7. Phase I and Phase II of control chart application 1. Basic principles The control chart contains a center line (CL) that represents the average value of the quality characteristic. Two other horizontal lines, called the upper control limit (UCL) and the lower control limit (LCL) are also shown on the chart. If all of the sample points fall between the UCL and LCL, the process is assumed to be in control and no action is necessary. If a point falls outside of the UCL and LCL, the process is assumed to be out of control and investigation and corrective action are required.

It is customary to connect the sample points on the control chart with straight line.
If all the points plot inside the control limits and if they behave in a systematic or nonrandom manner, then this could be an indication that the process is out of control. General model of a control chart

If x is a sample statistic that measures some quality characteristic of interest, then the center line (CL), the upper control limit (UCL) and the lower control limit (LCL) become:

UCL x L x CL x LCL x L x
Where L is the distance of the control limits from the center line expressed in standard deviation units. Classification of control chart Control chart may be classified into two general types: (a) Variable control chart (b) Attribute control chart

If the quality characteristic can be measured and expressed as a number on some continuous scale of measurement, then a variable control chart is used.

Many quality characteristics are not measured on a continuous scale, In these cases, we may judge each unit of product as either conforming or non-conforming on the basis of whether or not it possesses certain attributes, then an attribute control chart is used.
2. Choice of control limits

By moving the control limits farther from the center line, we decrease the risk of a type I error, that is, the risk of a point falling beyond the control limits, indicating an out of control condition when no assignable cause is present.
Widening the control limits also increase the risk of a type II error, that is, the risk of a point falling between the control limits when the process is really out of control. By moving the control limits closer to the center line, the opposite effect is obtained, the risk of type I error is increased while the risk of type II error is decreased. Regardless of the distribution of the quality characteristic, it is standard practice in the United States to use the three sigma control limits.

Warning limits on control charts Some analysts suggest using two sets of limits on control charts. The outer limits usually at three sigma are known as action limits. The inner limits usually at two are known as warning limits. If one or more points fall between the warning limits and the action limits or very close to the warning limits, we should be suspicious that the process may not be operating properly. In this case it is necessary to increase the sampling frequency and/or the sample size so that more information about the process can be obtained quickly.

3. Sample size and sampling frequency


In general, larger samples will make it easier to detect small shifts in the process. If the process shift is relatively large, then we use smaller sample size. Current industry practice tends to favor smaller, more frequent samples, particularly in high-volume manufacturing processes, or where a great many types of assignable causes can occur.

4. Rational subgroups The rational subgroup means that subgroups or samples should be selected so that if assignable causes are present, the chance for differences between subgroups will be maximized, while the chance for differences due to these assignable causes within a subgroup will be minimized. Two general approaches to constructing rational subgroups are used. In the first approach, each sample consists of units that are produced at the same time (or as closely together as possible). In the second approach, each sample consists of units of product that are representative of all units that have been produced since the last sample was taken. If a process consists of several machines that pool their output into a common stream. A logical approach to rational subgrouping here is to apply control chart techniques to the output for each individual machine.

5. Analysis of patterns on control charts A control chart may indicate an out of control condition when: (a) One or more points fall beyond the control limits, (b) Plotted points exhibit some nonrandom pattern of behavior, (c) Plotted points show an unusual run up or run down, (d) Plotted points show a cyclic behavior, a trend and a sudden jump. The Western Electric Handbook (1956) suggests a set of decision rules for detecting nonrandom patterns on control charts. The process is out of control if either: (a) One point plots outside the three-sigma control limits, (b) Two out of three consecutive points plot beyond the two-sigma warning limits, (c) Four out of five consecutive points plot at a distance of one-sigma or beyond from the center line, (d) Eight consecutive points plot on one side of the center line.

The process in the figure shows out of control since four of five consecutive points (last four points) fall in zone B or beyond. 6. Discussion of sensitizing rules for control charts Several criteria may be applied simultaneously to a control chart to determine whether the process is out of control. These criteria are known as sensitizing rule:

(a) One or more points outside of the control limits, (b) Two of three consecutive points outside the two-sigma warning limits but still inside the control limits, (c) Four of five consecutive points beyond the one-sigma limits, (d) A run of eight consecutive points on one side of the center line, (e) Six points in a row steadily increasing or decreasing, (f) Fifteen points in a row in zone C (both above and below the center line), (g) Fourteen points in a row alternating up and down, (h) Eight points in a row on both sides of the center line with none in zone C, (i) An unusual or nonrandom pattern in the data, (j) One or more points near a warning or control limit. 7. Phase I and Phase II of control application

In Phase I, a set of process data is gathered and analyzed. Trial control limits are determined based on the data and a reliable control limits are established to monitor future production.
In Phase II, we use the control chart to monitor the process for each successive sample.

Control charts for variable


When dealing with a quality characteristic that is a variable, it is usually necessary to monitor both the mean value of the quality characteristic and its variability. Control of the process mean is usually done with control chart for mean or the

x chart.

Process variability can be monitored with either a control chart for the standard deviation, called the s chart, or a control chart for the range, called an R chart. The R chart is more widely used than s chart. Control charts for x and R Charts based on standard values When it is possible to specify standard values for the process mean and standard deviation, we may use these standards to establish the control chart without analysis of past data.

Suppose that the standards given are and . Then the control limits of the
UCL 3 CL LCL 3

chart are

The quantity table.

3 A is a constant that depends on n, which is obtained from the given n

So the control limits can be written as

UCL A CL LCL A
The control limits of the R chart can be written as

UCL D2 CL d 2 LCL D1

Charts based on standard values not given


Suppose that a quality characteristic is normally distributed with mean and standard deviation .

In practice, we do not know and . Therefore they must be estimated from samples or subgroups.
Suppose that m samples are available, each containing n observations on the quality characteristics. Typically, m should be at least 20 to 25 samples and n should be 4, 5 or 6. Let x1 , x2 ,, xm be the average of each sample. Then the best estimator of is

x1 x2 xm m
R1 R2 Rm m

Let R1 , R2 ,, Rm be the ranges of the m samples. The average range is

Then the control limits of the

x chart are
UCL x A2 R CL x LCL x A2 R

The quantity A2 is a constant that depends on n, which is obtained from the given table.

The control limits of the R chart can be written as

UCL D4 R CL R LCL D3 R
Process standard deviation can be estimated as

R d2

Example: Suppose in an industry we want to establish control chart of the process using mean and range charts. We have 25 samples each of 5 size. The data are given as follows:

When setting up x and R control chart, it is best to begin with the R chart. Because the control limits on the x chart depend on the process variability, unless process variability is in control, these limits will not have much meaning.
The center line for the R chart is

For samples of n = 5, we find that D3 0 and D4 2.114 . Therefore, the limits for the R chart are:

Since the R chart indicates that process variability is in control, we may now construct the mean chart.
The center line is

The upper and lower control limits are

Since both charts exhibit control, we would conclude that the process is in control.

The process standard deviation can be estimated as

Revision of control limits and center lines The effective use of any control chart will require periodic revision of the control limits and center lines. Some practitioners establish regular periods for review and revision of control chart limits, such as every week, every month or every 25, 50 or 100 samples.

When the R chart is out of control, we eliminate the out of control points and recompute a revised value of R. This value is then used to determine new limits and center line on the R chart and new limits on the x chart.

Example: Establish statistical control?

and R control charts for the following data. Is the process in

20

86 84

15

82 80

10

78 76 74
0

72
0 2 4 6 8 10 12 14 16 18 20

10

12

14

16

18

20

function y=controlchart(data); s=size(data); x=1:s(1); % value of the x axis meandata=mean(data'); % mean of the observations rangedata=range(data'); % range of the observations gmean=mean(meandata); % grand mean grange=mean(rangedata); % grand range
% control limits of the mean chart uclx=gmean+.577*grange; clx=gmean; lclx=gmean-.577*grange; % control limits of the range chart uclr=2.114*grange; clr=grange; lclr=0*grange;

% mamimum and minimum value of the mean and range data maxx=max(meandata); maxr=max(rangedata); minx=min(meandata); minr=min(rangedata);

% range chart figure(1); hold on; axis([0 s(1) minr-5 maxr+5]); line([0 s(1)],[clr clr], 'color','m'); line([0 s(1)],[lclr lclr], 'color','k'); line([0 s(1)],[uclr uclr], 'color','k'); plot(x,rangedata,'-*b'); % mean chart figure(2); hold on; axis([0 s(1) minx-5 maxx+5]); line([0 s(1)],[clx clx], 'color','m'); line([0 s(1)],[lclx lclx], 'color','k'); line([0 s(1)],[uclx uclx], 'color','k'); plot(x,meandata,'-*b');

%axis([xmin xmax ymin ymax])

y=[uclr clr lclr;uclx clx lclx];

Guidelines for the design of

and R chart

If the x chart is being used primarily to detect moderate to large process shifts say, on the order of 2 or larger, then relatively small samples of size n = 4, 5 or 6 are reasonably effective. On the other hand, if we are trying to detect small shifts, than larger sample sizes of possibly n = 15 to n = 25 are needed.

The R chart is relatively insensitive to shifts in the process standard deviation for small samples. For large n say, n > 10 or 12, it is probably best to use a control chart for s 2 or s instead of the R chart.
From economic consideration, if the cost associated with producing defective items is high, smaller, more frequent samples are better that larger, less frequent ones. The use of 3 sigma control limits on the x and R control charts is a widespread practice. If type I errors (an out of signal is generated when the process is really in control) are very expensive to investigate, then it may be best to use wider control limits, perhaps at 3.5 sigma. However, if the process is such that out of signals are quickly and easily investigated with a minimum of lost time and cost, then narrower control limits, perhaps at 2.5 sigma are appropriate.

Changing sample size on the x and R chart There are two situations in which the sample size n is not constant. Each sample may consists of a different number of observations. A permanent change in sample size because of cost or because the process has exhibited good stability and fewer resources are being allocated for process monitoring. In the second situation it is easy to recompute the new control limits directly from the old ones without collecting additional samples based on the new sample size.

For the

chart the new control limits are

Where the center line

x is unchanged and the factor A2 is selected for the new sample size.

For the R chart the new control limits are

Where D3 and D4 are selected for the new sample size.

Example: Consider the previous example where the limits were based on the sample size of five. Since the process exhibits good control, the process engineering personnel want to reduce the sample size to three and to calculate the new control limits.

For the

chart the new control limits are

For the R chart the new control limits are

Interpretation of

x and R chart x chart, we must determine whether or not the R chart

In the interpreting patterns on the is in control.

If the both x and R charts exhibit a nonrandom pattern, the best strategy is to eliminate the R chart assignable causes first. In many cases, this will automatically eliminate the nonrandom pattern on the x chart. Never attempt to interpret the condition.

x chart

when the R chart indicates an out of control

Example:

Control limits of the mean and range chart:

Range chart:
14 12 10 8 6 4 2 0 -2 -4

10

15

20

Mean chart:

42 40 38 36 34 32 30 28 0 5 10 15 20
X: 12 Y: 38.6 X: 15 Y: 37.1

Control limits of the mean and range chart for 22 samples:

Range chart:

12 10 8 6 4 2 0 -2 -4

10

12

14

16

18

20

22

Mean chart:
40

38

36

34

32

30

28

10

12

14

16

18

20

22

42 40 38 36 34 32 30 28 26

42 40 38 36 34 32 30 28 26

42 40 38 36 34 32 30 28 26

10

15

20

10

15

20

10

15

20

Potrebbero piacerti anche