Beruflich Dokumente
Kultur Dokumente
17 1
Learning Objectives
After completing this chapter, students will be able to:
Define the quality of a product or service Determine the benefits derived from SQC Review statistical background Refresh topics- mean, median std.deviation, ave.
17 2
Chapter Outline
17.1 17.2 17.3 17.4 17.5 Introduction Defining Quality and Total Quality Mgt Statistical Process Control Control Charts for Variables Control Charts for Attributes
17 3
Introduction
Quality is often the major issue in a purchase decision Poor quality can be expensive for both the producing firm and the customer Quality management, or quality control (QC is QC), critical throughout the organization Quality is important for manufacturing and services We will be dealing with the most important statistical methodology, statistical process control (SPC SPC)
2009 Prentice-Hall, Inc. 17 4
17 7
17 8
Mean
It is the average of a series of results. The mean of a finite number of measurements, xl, x2, x3, x4,........, xn, is commonly represented as x. It may be calculated by taking the average of individual results as shown below :
17 9
Median
The median of an even number of results is nothing but the average of the two middle values provided the results are listed in order ; whereas for an odd number of results the median is the middle value itself. However, the mean and the median are exactly identical in the case of a truly symmetrical distribution. In short, median is an useful measure specifically when dealing with very small samples.
17 10
17 12
Standard Deviation
It is the distance from the mean to the point of inflexion of the normal distribution curve. In comparison to the average deviation the standard deviation is usually considered to be much more useful and meaningful statistically. For a finite number of values it is normally symbolised as S, and may be expressed as follows :
17 13
17 14
Range
The difference between the largest and smallest result in a series of measurements
17 15
Examples:
The normality of a solution of sodium hydroxide as determined by an analyst by FOUR different titrations are found to be 0.5038 ; 0.5049 ; 0.5042 ; and 0.5039. Calculate the mean, median, average deviation, relative average deviation, standard deviation and coefficient of variation.
17 16
Upper control limit One plot out above. Investigate for cause.
Target
Figure 17.1
Target
Lower control limit Trends in either direction 5 plots. Investigate for cause of progressive change.
17 19
17 20
17 21
the control charts used for processes that are measured in continuous units The x-chart is a data analysis technique for determining if a measurement process has gone out of statistical control. The R-chart tells us when there has been a change in the uniformity of the process Both charts must be used when monitoring variables
2009 Prentice-Hall, Inc. 17 23
3. Graph the sample means and ranges on their respective control charts and determine whether they fall outside the acceptable limits
17 24
x chart
R-chart
2009 Prentice-Hall, Inc. 17 25
17 26
sample means ( x )
2009 Prentice-Hall, Inc. 17 27
distributions, each with their own mean (Q) and standard deviation (W x) If a series of random samples ( x 1 , x 2 , x 3 , x 4 , and so on) each of size n is drawn from any of these, the resulting distribution of the x s will appear as i in the bottom graph in the figure Because this is a normal distribution
1. 99.7% of the time the sample averages will fall between s3W if the process has only random variations 2. 95.5% of the time the sample averages will fall between s2W if the process has only random variations
17 28
99.7% of all x fall within s3Wx 95.5% of all x fall within s2Wx
| 3Wx
| 2Wx
| 1Wx
| Qx = Q (mean)
| +1Wx
Wx n
| +2Wx
| +3Wx
Figure 17.2
Standard error ! W x !
17 29
If chosen 99.7% of the time the sample averages will fall between 3W if the process has only random variations
If chosen 95.5% of the time the sample averages will fall between 2W if the process has only random variations
17 31
sampled every hour To set control limits that include 99.7% of the sample, 36 boxes are randomly selected and weighed The standard deviation is estimated to be 2 ounces and the average mean of all the samples taken is 16 ounces So x ! 16,W x ! 2, n ! 36, z ! 3 and the control limits are 2 UCLx ! x zW x ! 16 3 ! 16 1 ! 17 ounces 36 2 LCLx ! x zW x ! 16 3 ! 16 1 ! 15 ounces 36
2009 Prentice-Hall, Inc. 17 32
or difficult to compute (a common situation) the previous equations are impractical In practice the calculation of the control limits is based on the average range rather than the standard deviation UCLx ! x A2 R LCLx ! x A2 R where R = average of the samples A2 = value found in Table 17.2 x = mean of the sample means
2009 Prentice-Hall, Inc. 17 33
Table 17.2
2009 Prentice-Hall, Inc. 17 34
17 35
limits for the process average We are also interested in the dispersion or variability of the process Averages can remain the same even if variability changes A control chart for ranges is commonly used to monitor process variability Limits are set at s3W for the average range R
17 36
UCLR ! D4 R LCLR ! D3 R where UCLR = upper control chart limit for the range LCLR = lower control chart limit for the range D4 and D3 = values from Table 17.2
17 37
Range Example
A process has an average range of 53 pounds If the sample size is 5, what are the upper and lower control limits? From Table 17.2, D4 = 2.114 and D3 = 0 UCLR ! D4 R ! (2.114)(53 pounds) ! 112.042 pounds LCLR ! D3 R ! (0)(53 pounds) !0
2009 Prentice-Hall, Inc. 17 38
Table 17.2
2009 Prentice-Hall, Inc. 17 39
Problems:
No. Insp. A 1 2 3 4 10.5 10.4 10.1 10.4 5 sample size of Volume of vials every 30 minutes B 10.3 10.3 10.3 10.1 C 10.6 10.1 10.5 10.3 D 10.7 10.3 10.7 10.7 E 10.2 10.5 10.6 10.2 mean R
a. Calculate the mean, average mean, range, average range b. Calculate the UCL and LCL of both mean and range 2009 Prentice-Hall, Inc. using 99.7% in normal distribution
17 40
p-Charts
Attributes that are good or bad typically follow the binomial distribution If the sample size is large enough a normal distribution can be used to calculate the control limits UCL p ! p zW p
where
LCL p ! p zW p
p = mean proportion or fraction defective in the sample z = number of standard deviations W p = standard deviation of the sampling distribution which is estimated by W p where n is the size of each sample Wp ! p(1 p ) n
2009 Prentice-Hall, Inc. 17 42
17 43
UCL p ! p zW p ! 0.04 3(0.02 ) ! 0.10 LCL p ! p zW p ! 0.04 3(0.02 ) ! 0 Percentage cant be negative
2009 Prentice-Hall, Inc. 17 44
Out of Control
UCLp = 0.10
Fraction Defective
p ! 0.04
LCLp = 0.00
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Sample Number
2009 Prentice-Hall, Inc. 17 45
Figure 17.3
Program 17.1A
2009 Prentice-Hall, Inc. 17 46
Program 17.1B
2009 Prentice-Hall, Inc. 17 47
c-Charts
In the previous example we counted the number of
defective records entered in the database But records may contain more than one defect We use c-charts to control the number of defects per unit of output c-charts are based on the Poisson distribution which has its variance equal to its mean The mean is c and the standard deviation is equal to c To compute the control limits we use
cs3 c
2009 Prentice-Hall, Inc. 17 48