Chapter Contents
Chapter Contents
Previous
Previous
Next
Next
UCHART Statement

Example 41.3: Creating u Charts for Varying Numbers of Units

See SHWUEX3 in the SAS/QC Sample Library

In the fabric manufacturing process described in "Creating u Charts from Defect Count Data", each roll of fabric is 30 meters long, and an inspection unit is defined as one square meter. Thus, there are 30 inspection units in each subgroup sample. Suppose now that the length of each piece of fabric varies. The following statements create a SAS data set (FABRICS2) that contains the number of fabric defects and size (in square meters) of 25 pieces of fabric:

   data fabrics2;
      input roll defects sqmeters @@;
   datalines;
    1  7 30.0   2 11 27.6   3 15 30.4   4  6 34.8   5 11 26.0
    6 15 28.6   7  5 28.0   8 10 30.2   9  8 28.2  10  3 31.4
   11  3 30.3  12 14 27.8  13  3 27.0  14  9 30.0  15  7 32.1
   16  6 34.8  17  7 26.5  18  5 30.0  19 14 31.3  20 13 31.6
   21 11 29.4  22  6 28.6  23  6 27.5  24  9 32.6  25 11 31.7
   ;

A listing of FABRICS2 is shown in Output 41.3.1.

Output 41.3.1: The Data Set FABRICS2
 
Number of Fabric Defects

roll defects sqmeters
1 7 30.0
2 11 27.6
3 15 30.4
4 6 34.8
5 11 26.0
6 15 28.6
7 5 28.0
8 10 30.2
9 8 28.2
10 3 31.4
11 3 30.3
12 14 27.8
13 3 27.0
14 9 30.0
15 7 32.1
16 6 34.8
17 7 26.5
18 5 30.0
19 14 31.3
20 13 31.6
21 11 29.4
22 6 28.6
23 6 27.5
24 9 32.6
25 11 31.7

The variable ROLL contains the roll number, the variable DEFECTS contains the number of defects in each piece of fabric, and the variable SQMETERS contains the size of each piece.

The following statements request a u chart for the number of defects per square meter:

   title 'u Chart for Fabric Defects per Square Meter';
   symbol v=dot c=vig;
   proc shewhart data=fabrics2;
      uchart defects*roll / subgroupn = sqmeters
                            outlimits = flimits
                            cframe    = steel
                            cinfill   = ligr
                            coutfill  = yellow
                            cconnect  = vig;
   run;

The u chart is shown in Output 41.3.2, and the data set FLIMITS is listed in Output 41.3.3.

Output 41.3.2: A u Chart with Varying Number of Units per Subgroup
uex3b.gif (5950 bytes)

Note that the control limits vary with the number of units per subgroup (subgroup sample size). The legend in the lower left corner indicates the minimum and maximum subgroup sample sizes.

Output 41.3.3: The Control Limits Data Set FLIMITS
 
Control Limits for Fabric Defects

_VAR_ _SUBGRP_ _TYPE_ _LIMITN_ _ALPHA_ _SIGMAS_ _LCLU_ _U_ _UCLU_
defects roll ESTIMATE V V 3 V 0.28805 V

Output 41.3.3 shows that the variables _LIMITN_, _ALPHA_, _LCLU_, and _UCLU_ have the special missing value V, indicating that these variables vary with the sample size.

The following statements request a u chart with a fixed sample size of 30.0 for the control limits. In other words, the control limits are computed as if each piece of fabric were 30 meters long.

   proc shewhart data=fabrics2;
      uchart defects*roll / subgroupn = sqmeters
                            outlimits = flimits2
                            limitn    = 30
                            alln
                            nmarkers
                            cinfill   = ligr
                            cconnect  = cxfefefe
                            cframe    = steel;
   run;

The ALLN option specifies that points are to be displayed for all subgroups, regardless of their sample size. By default, when you specify the LIMITN= option, only points for subgroups whose sample size matches the LIMITN= value are displayed. The NMARKERS option requests special symbols that identify points for which the subgroup sample size differs from the nominal sample size of 30. The chart is shown in Output 41.3.4.

Output 41.3.4: Control Limits Based on Fixed Subgroup Sample Size
uex3d.gif (6419 bytes)

In Output 41.3.4, no points lie outside the control limits, indicating that the process is in control. However, you should be careful when interpreting charts that use a nominal sample size, since the fixed control limits based on this value are only an approximation. Output 41.3.5 lists the data set FLIMITS2, which contains the fixed control limits displayed in Output 41.3.4.

Output 41.3.5: The Fixed Control Limits Data Set FLIMITS2
 
Fixed Control Limits for Fabric Defects

_VAR_ _SUBGRP_ _TYPE_ _LIMITN_ _ALPHA_ _SIGMAS_ _LCLU_ _U_ _UCLU_
defects roll ESTIMATE 30 .002621618 3 0 0.28805 0.58201

Chapter Contents
Chapter Contents
Previous
Previous
Next
Next
Top
Top

Copyright © 1999 by SAS Institute Inc., Cary, NC, USA. All rights reserved.