Download as pdf or txt
Download as pdf or txt
You are on page 1of 6

Kurt Wiessmeyer

Manufacturing Engineering

Prof. Peter Rourke

2/26/19

Quality Engineering Report 1

In the following scenario, two sets of data are given for the diameter measurements of shafts turned on
two different lathes. The specifications call for a diameter of 0.258” +/- 0.005”.

Shown below is a detailed quality analysis of these two datasets.

N-5 subgroup
a b c d e f
0.262 0.260 0.262 0.261 0.265 0.257 1
0.263 0.256 0.260 0.259 0.260 0.260 2
0.262 0.256 0.257 0.258 0.263 0.255 3
0.261 0.266 0.262 0.261 0.260 0.262 4
0.259 0.258 0.257 0.259 0.256 0.260 5
0.257 0.263 0.260 0.260 0.259 0.258 6
0.257 0.256 0.261 0.259 0.259 0.261 7
0.258 0.257 0.260 0.260 0.261 0.259 8
0.261 0.260 0.262 0.260 0.265 0.256 9
0.256 0.259 0.262 0.258 0.254 0.260 10

N-7 subgroup
a b c d e f
0.256 0.254 0.253 0.255 0.255 0.255 1
0.257 0.256 0.257 0.253 0.254 0.255 2
0.257 0.257 0.256 0.258 0.256 0.257 3
0.258 0.260 0.258 0.256 0.257 0.258 4
0.260 0.262 0.257 0.257 0.258 0.259 5
0.260 0.258 0.258 0.256 0.258 0.258 6
0.264 0.258 0.259 0.262 0.260 0.261 7
0.259 0.263 0.257 0.261 0.259 0.260 8
0.262 0.260 0.263 0.263 0.261 0.262 9
0.263 0.265 0.262 0.267 0.263 0.264 10
Tables 1 & 2: Data tables given; measurements are in inches.
1. Qualitative Analysis: Create histograms for each process

Figures 1 & 2: Histograms of the N-5 and N-7 data sets

a. Without using statistics which process appears more capable?


Process N-5 appears more capable at a quick glance. This is because the data is more
bell-curve shaped, which means it is more precise. While the data is not centered about
the specifications and is not 100% accurate, it is more consistent than N-7 and it could be
adjusted, therefore it is more capable.
2. Quantitative Analysis
a. Grouped data analysis: Which process appears more capable from the Cp calculation?
Do either/both processes appear stable? Do either/both processes appear capable?
N-5 N-7
max = 0.2660 max = 0.2670
min = 0.2540 min = 0.2530
range = 0.0120 range = 0.0140
kurtosis = -0.0291 kurtosis = -0.3466
skewness = 0.1632 skewness = 0.3770
Xbar = 0.2596 Xbar = 0.2588
sx = 0.0025 sx = 0.0031
Cp = 0.6597 Cp = 0.5333
Cpk= 0.4508 Cpk= 0.4498

Tables 3 & 4: Tables containing grouped data analysis for N-5 and N-7

According to the above calculations, process N-5 does appear to be more capable than process N-7,
however not by a great deal. Both process capabilities are much less than the recommended 1.33,
therefore they do not appear capable.

Also, based on the range and average charts below, the N-7 process appears more stable than N-5
because the ranges are all closer to the average range value. The average range of N-7 is also lower than
that of N-5.

b. Create SPC charts: Xbar, R, and X charts. Do both processes appear stable: are there
any points that appear to be assignable causes?
See figures 3 through 6 below:
Average Chart (N-5)
0.264

0.263

0.262
Average Diameter (on.)

0.261

0.26 Average

0.259 UCLx
LCLx
0.258
Xbar
0.257

0.256

0.255
1 2 3 4 5 6 7 8 9 10
Subgroup

Average Chart (N-7)


0.266

0.264

0.262
Average Diameter (on.)

0.26
Average
0.258
UCLx
0.256 LCLx
Xbar
0.254

0.252

0.25
1 2 3 4 5 6 7 8 9 10
Subgroup

Figures 3 & 4: Xbar charts for N-5 and N-7


Range Chart (N-5)
0.014

0.012

0.01
Avg. Range (in.)

0.008
Range
0.006 Rbar
UCLr
0.004

0.002

0
1 2 3 4 5 6 7 8 9 10
Subgroup

Range Chart (N-7)


0.014

0.012

0.01
Avg. Range (in.)

0.008
Range
0.006 Rbar
UCLr
0.004

0.002

0
1 2 3 4 5 6 7 8 9 10
Subgroup

Figures 5 & 6: Rbar Charts for N-5 and N-7

As shown above, there is at least one assignable cause point in these datasets. Points 1,
2, and 10 from the N-7 set appear either less than or greater than the control limits,
therefore are assignable causes. One may believe that this is due to tool wear over time,
due to the fact that the parts become larger in diameter over time. This is predictable
behavior for a tool and can be resolved with predictive maintenance.
c. Calculate the Cp from the SPC charts (6(Rbar/d2)).
(note: do not remove assignable cause values, if any.)
d2 = 2.534 for subgroup size 6

N-5: σ0 = (.0065/2.534) → σ0 = .002565


Cp = (UCL – LCL)/6σ0 = (.263-.253)/(6*.002565)→ Cp = 0.650

N-7: σ0 = (.0042/2.534) → σ0 = .001657


Cp = (UCL – LCL)/6σ0 = (.263-.253)/(6*.001657)→ Cp = 1.006

d. Are the processes similarly capable? Which is "better"? Are either/both capable at all?
Both processes are similarly incapable, although process N-7 is a bit “better” than N-5.
When calculated from the SPC charts, the Cp of process N-5 remains the same as when
calculated via grouped data analysis. However, the Cp of process N-7 is greater this way
compared to the previous calculation. Either way, neither process passes the
recommended 1.33 Cp value, and neither appear to be capable at all.

You might also like