Aulas praticas (exercicios 9)

Soluções

1.
a) Calculate the 3 control limits for X-bar and R charts based on the first 12 samples reflecting
the process before any problems were denounced.

Sample Screw 1 Screw 2 Screw 3 Mean (x-bar) Range (R)

1 0.276 0.238 0.261 0.258 0.038
2 0.249 0.263 0.234 0.249 0.030
3 0.264 0.255 0.258 0.259 0.009
4 0.255 0.279 0.269 0.268 0.024
5 0.262 0.273 0.234 0.256 0.039
6 0.268 0.267 0.270 0.268 0.003
7 0.266 0.244 0.270 0.260 0.026
8 0.232 0.261 0.278 0.257 0.046
9 0.242 0.277 0.253 0.257 0.035
10 0.246 0.253 0.236 0.245 0.017
11 0.279 0.230 0.235 0.248 0.049
12 0.238 0.243 0.237 0.239 0.005
Average: 0.255 0.027

x-double-bar = 0.255
R-bar = 0.027

From page 310 (CJA):

A2 = 1.023
D3 = 0
D4 = 2.574

So, calculate the control limits (as explained in book on pages 760ff.):

X-bar-chart
UCLX = x-double-bar + A2 x R-bar = 0.255 + 1.023 (0.027) = 0.283
LCLX = x-double-bar - A2 x R-bar = 0.255 – 1.023 (0.027) =0.228

R-chart
UCLR = D4 x R-bar = 2.574 x 0.027 = 0.069
LCLR = D3 x R-bar = 0 x 0.027 = 0.000

SolucoesAP9.DOC 1
Aulas praticas (exercicios 9)
Soluções

b) Plot X-bar and R charts labeling the data points, upper and lower control limits, and center
lines on both charts. Plot the means/ranges of all 18 samples, but use the control limits and
center lines calculated for the first 12 samples.

Sample Screw 1 Screw 2 Screw 3 Mean (x-bar) Range (R)

13 0.220 0.261 0.250 0.243 0.042
14 0.293 0.300 0.284 0.292 0.016
15 0.286 0.293 0.211 0.263 0.082
16 0.227 0.216 0.210 0.218 0.017
17 0.290 0.258 0.239 0.262 0.051
18 0.283 0.227 0.223 0.244 0.060

X-B ar C hart
0.300
0.290
0.280
0.270
0.260
x -ba r

0.250
0.240
0.230
0.220
0.210
0.200
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
S a m ple Num be r

Mean (x-bar) C enter UC L LC L

R C hart

0.090
0.080
0.070
0.060
0.050
r-ba r

0.040
0.030
0.020
0.010
0.000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
S a m ple Num be r

R ange (R) Center UC L LC L

SolucoesAP9.DOC 2
Aulas praticas (exercicios 9)
Soluções

c) Do you feel that the screw production process is in control? Is there something suspicious?
The graphs confirm that the customers complaints are justified – it really seems as if the
variation in the screw diameters has increased significantly. The R-chart suggests significant
differences even within single samples! Clearly some corrective measures have to be taken to
bring this process back under control!

2.
a) Calculate the 3 control limits for the supplier’s manufacturing process based on the first 15
weeks (i.e. weeks 1-15, when the quality of the alloy did not seem to be an issue).

Week 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Flaws per 200 8 9 11 8 10 11 12 10 9 12 14 13 16 13 18
Proportion (P) 0.040 0.045 0.055 0.040 0.050 0.055 0.060 0.050 0.045 0.060 0.070 0.065 0.080 0.065 0.090

n = 200
p-bar = average of all proportions = .058
σ= (p(1 − p) / n) = .058(1 − .058) / 200 = .01653
UCLp-bar = p-bar + 3 = .058 + 3(.01653) = 0.10758
LCLp-bar = p-bar – 3 = .058 - 3(.01653) =0.0084

b) Create the SPC chart including the weekly data, control limits, and the center line. Plot the
defective fractions of all 20 weeks, but use the control limits and the center line of the first 15
weeks!! Interpret the chart – what does it suggest?

Week 16 17 18 19 20
Flaws per 200 19 17 18 22 21
Proportion (P) 0.095 0.085 0.090 0.110 0.105

P-Chart
0.120
Proportion Defectives

0.100

0.080

0.060

0.040

0.020

0.000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Week
Proportion (P) Center UCL LCL

SolucoesAP9.DOC 3
Aulas praticas (exercicios 9)
Soluções

Clearly there is an upward tendency in the fraction of defective packages of alloy. Even though
the problem was not noticed before week 15, it seems as if this process started already in week 9.
The magnitude of the fluctuations is not as alarming as is the steady upward tendency. Even
though only the very last weeks reveal defective fractions outside the control limits, the graph
gives the impression that the fractions will keep on increasing in future. It seems as if more and
more sulphur was used in blending the alloy. Whatever the reasons for this increase might be, it
leads to severe quality problems of Screwed’s final products and cannot be accepted. Definitely
this problem should be looked into more carefully …

3.
From page 310 (CJA), n=4, A2=0.729; D3=0; D4=2.282.

We can calculate UCLr, LCLr, UCLx, and LCLx using the following expressions:
UCLr = D4 * R
LCLr = D3 * R
UCLx = x + A2 * R
LCLx = x – A2 * R

Sample
Number 1 2 3 4 A2= 0.729
1 3.51 3.63 3.39 3.35 D3= 0
2 3.5 3.56 3.42 3.64 D4= 2.282
3 3.68 3.49 3.53 3.62 LCLx= 3.469
4 3.45 3.33 3.47 3.55 UCLx= 3.731
5 3.7 3.58 3.64 3.68 LCLr= 0
Overall UCLr= 0.41076

x-bar
Sample Sample Center
Number LCLx Mean Line UCLx
1 3.469 3.4700 3.6 3.731
2 3.469 3.5300 3.6 3.731
3 3.469 3.5800 3.6 3.731
4 3.469 3.4500 3.6 3.731
5 3.469 3.6500 3.6 3.731

R-bar
Sample
Number LCLr Sample R R-bar UCLr
1 0 0.2800 0.18 0.41076
2 0 0.2200 0.18 0.41076
3 0 0.1900 0.18 0.41076
4 0 0.2200 0.18 0.41076
5 0 0.1200 0.18 0.41076

SolucoesAP9.DOC 4
Aulas praticas (exercicios 9)
Soluções

X-bar

3.800
3.750
3.700
3.650 LCLx
3.600
Sample Mean
3.550
Center Line
3.500
3.450 UCLx
3.400
3.350
3.300
1 2 3 4 5

R-chart

0.45
0.4
0.35
0.3 LCLr
0.25 Sample R
0.2 R-bar
0.15 UCLr
0.1
0.05
0
1 2 3 4 5

b) The process is not in control because the sample mean for sample 4 (in the X-bar graph) falls
outside the control limits.

4.

a)
Incorrect
Sample Orders Propotion
1 10 0.050 p= 0.055
2 12 0.060 n= 200
3 15 0.075 sigma= 0.01612
4 2 0.010 3sigma= 0.04836
5 19 0.095

SolucoesAP9.DOC 5
Aulas praticas (exercicios 9)
Soluções

6 8 0.040
7 24 0.120
8 7 0.035
9 10 0.050
10 7 0.035 LCL= 0.00664
11 15 0.075 UCL= 0.1034
12 3 0.015
Total 132 0.660
Average 11 0.055

Proportion Center
Sample LCL (p) Line UCL
1 0 0.050 0.055 0.1034
2 0 0.060 0.055 0.1034
3 0 0.075 0.055 0.1034
4 0 0.010 0.055 0.1034
5 0 0.095 0.055 0.1034
6 0 0.040 0.055 0.1034
7 0 0.120 0.055 0.1034
8 0 0.035 0.055 0.1034
9 0 0.050 0.055 0.1034
10 0 0.035 0.055 0.1034
11 0 0.075 0.055 0.1034
12 0 0.015 0.055 0.1034

b)
The data in part a (proportion of sample) shows that the sample in week 7 exceeds the upper
control limit. Therefore, the process is seen to be out of control. The reasons for the poor
performance that week should be analyzed.

SPC Chart

0.14
0.12
LCL
0.1
0.08 Propotion (p)
0.06 Center Line
0.04
UCL
0.02
0
1 2 3 4 5 6 7 8 9 10 11 12

SolucoesAP9.DOC 6