CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

37
CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control

description

World Class Quality (Shewhart..1960) “On-Target with Minimum Variance.” Operating “On-Target” requires a different way of thinking about our processes. Operating with “Minimum Variance” is achieved only when a process displays a reasonable degree of statistical control. “Good Parts” vs “Bad Parts” 3

Transcript of CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

Page 1: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

CBE 491 / CBE 433 3 Dec 2012

1

Statistical Process Control

Page 2: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

World Class Quality (Shewhart..1960)

• “On-Target with Minimum Variance.”• Operating “On-Target” requires a different way of thinking about

our processes.• Operating with “Minimum Variance” is achieved only when a

process displays a reasonable degree of statistical control.

“Good Parts” vs “Bad Parts”

2

Page 3: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

World Class Quality (Shewhart..1960)

• “On-Target with Minimum Variance.”• Operating “On-Target” requires a different way of thinking about

our processes.• Operating with “Minimum Variance” is achieved only when a

process displays a reasonable degree of statistical control.

“Good Parts” vs “Bad Parts”

3

Page 4: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

4

World Class Quality (Shewhart..with an in-control process)

MAKE PACK SHIP

Page 5: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

5

World Class Quality (Shewhart..w/in-control process)

MAKE PACK SHIP

Challenges:

Examples from industry:• Polymer gum in fiberpac• Corn sweetener

“Crystal Clear”

Page 6: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

Shewhart: “While every process displays variation, some processes display controlled variation while others display uncontrolled variation.”

• Controlled Variation: Stable, consistent pattern of variation over time. “Chance” causes.

6

Page 7: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

7

Page 8: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

Shewhart: “While every process displays variation, some processes display controlled variation while others display uncontrolled variation.”

• Controlled Variation: Stable, consistent pattern of variation over time. “Chance” causes.

• Uncontrolled Variation: Pattern of variation that changes over time. “Assignable” causes.

8

Page 9: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

9

Page 10: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

Shewhart: “While every process displays variation, some processes display controlled variation while others display uncontrolled variation.”

• Controlled Variation: Stable, consistent pattern of variation over time. “Chance” causes.

• Uncontrolled Variation: Pattern of variation that changes over time. “Assignable” causes.

Shewhart: “A phenomenon will be said to be controlled when, through the use of past experience, we can predict, at least within limits, how the phenomenon may be expected to vary in the future.”

10

Page 11: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

Statistical Control

• Predictability• Use statistics to highlight uncontrolled variations• Eliminate uncontrolled variations

Next: Control Charts as a way to see uncontrolled variations.

11

Page 12: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

Statistical Measures

• Location: What can we use?

• Dispersion: What can we use?

12

Page 13: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

Statistical Measures

• Location: What can we use?

• Dispersion: What can we use?

Look at example (Camshaft Bearings)

13

Page 14: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

= 1.375146 = 1.374978 = 1.374820X R = 0.00080 = 0.000750 = 0.00120

S = 0.0001697 = 0.0001697 = 0.0003576

14

Page 15: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

15

Page 16: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

Controlled Variation

16

Subgroups used (4-5)

Reason: use of mean tends to normalize information (distribution of the subgroup average)

Shewhart: pick subgroups to organize the data into a rational manner.

Page 17: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

Uncontrolled Variation

17

Gather 20 – 30 subgroups before calculating the control limits

Page 18: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

18

Page 19: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

19

Page 20: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

20

Page 21: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

21

Page 22: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

22

Page 23: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

23

Page 24: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

24

PS Exercise: Control Charts

Page 25: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

25

SPC Quick Summary

Page 26: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

26

Page 27: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

27

Control Chart ExamplesData from 99 samples of the silicon content drawn from a blast furnace. Data organized into 33 subgroups of size 3.

Page 28: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

28

Control Chart Examples

Page 29: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

29

Control Chart Examples

Page 30: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

30

Control Chart ExamplesData from 99 samples of the silicon content drawn from a blast furnace. Upon further inspection it was discovered that each sample was the first tap from each shift…thus 3 silicon values for each day.

The sample subgrouping was supposed to represent each day, however it was recorded in columns not rows, thus 144, 150 and 180 was the 1st day’s samples. The data was reorganized and rerun.

Page 31: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

31

Control Chart Examples

Page 32: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

32

Control Chart Examples

Page 33: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

33

Control Chart ExamplesSubgroups of 5 samples from each shift.

Page 34: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

34

Control Chart ExamplesSubgroups of 5 samples from each shift.

Page 35: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

35

Control Chart ExamplesExtruded vinyl sheet to be used for car dashboard. Thickness controlled by automatic system.

Page 36: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

36

Questions?

Page 37: CBE 491 / CBE 433 3 Dec 2012 1 Statistical Process Control.

37

Questions?