12
STATISTICS AND OUTLIERS Aaron Saks Process Advancement Leader Boise, Inc 10/26/2011

STATISTICS AND OUTLIERS Aaron Saks Process Advancement Leader Boise, Inc 10/26/2011

Embed Size (px)

Citation preview

Page 1: STATISTICS AND OUTLIERS Aaron Saks Process Advancement Leader Boise, Inc 10/26/2011

STATISTICS AND OUTLIERSAaron Saks

Process Advancement Leader

Boise, Inc

10/26/2011

Page 2: STATISTICS AND OUTLIERS Aaron Saks Process Advancement Leader Boise, Inc 10/26/2011

Bio

• Graduated in 2007 – ChemE and PSE• Started career with Boise Inc, Wallula WA as a process engineer

• October 2008 went to work for Envoy Development

• April 2010 returned to Boise as Process Advancement Leader, Wallula WA

• November 1st 2011 move to Boise, ID to become Project Manager for Boise Packaging

Page 3: STATISTICS AND OUTLIERS Aaron Saks Process Advancement Leader Boise, Inc 10/26/2011

Statistics and Data

• Why is it important?• Everything we really know, we know because of data.

• As a new engineer, we know how to apply math and logic to solve problems.• We don’t really know how anything works.

• By focusing on the data we can learn, solve problems, and teach others. • There is a lot of data out there.• Statistics = The language of data.

Page 4: STATISTICS AND OUTLIERS Aaron Saks Process Advancement Leader Boise, Inc 10/26/2011

Thinking Statistically

• We can’t just think in terms of the “average”. • Need to think in terms of the distribution of data, and the probability of events occurring.

• What are good statistical tools for a new engineer?• Six Sigma methods• Understanding the common probability distributions and

their mean and variance• Understanding histograms and pareto charts• Excel skills

Page 5: STATISTICS AND OUTLIERS Aaron Saks Process Advancement Leader Boise, Inc 10/26/2011

Tear Strength

0.0%

5.0%

10.0%

15.0%

20.0%

25.0%

30.0%

26

.7

27

.8

28

.9

30

.0

31

.2

32

.3

33

.4

34

.5

35

.6

36

.7

37

.8

38

.9

40

.0

41

.1

42

.2

43

.3

%

Freq%

Normdist%

4.1%

0.34%

Current Capability: 30.3

Current Spec:32

4.1% represents ~ 1640 tons below current spec

~40 tons (0.1%) rejected for below-spec MD tear

Current Target: 42

Page 6: STATISTICS AND OUTLIERS Aaron Saks Process Advancement Leader Boise, Inc 10/26/2011

Ppkm: Run to Target and Reduce Variation

• Historically, papermakers would run “in the warning” all day, as long as the tests are within specification limits.

• This resulted in running off target, with different means run to run (Poor Quality).

• Needed a way to encourage ($$$) running to targets, and reducing variation within the specs.

• Created a variation on the classical Process Capability metric “Cpk”.

• Result a financial incentive for improving Quality.

Page 7: STATISTICS AND OUTLIERS Aaron Saks Process Advancement Leader Boise, Inc 10/26/2011

Ppkm - Running to Target and Reducing Variation

• Ppkm is an example of a process performance metric.

• Ppkm captures both deviation from target and variation within specification limits.

• If Ppkm >= 1, then the process is Capable: running to target and variation is well within specification limits

• To improve Quality we award operators for the number for Key Product Properties that have a Ppkm value above 1.0.

0%

5%

10%

15%

20%

25%

30%

Low

er S

pec

Lim

it

Upp

er S

pec

Lim

it

Tar

get

Ave

rage

Variation

Deviation from Target

Out of Spec Paper

0%

5%

10%

15%

20%

25%

30%

Low

er S

pec

Lim

it

Upp

er S

pec

Lim

it

Tar

get

Ave

rage

Variation

Deviation from Target

Out of Spec Paper

0%

5%

10%

15%

20%

25%

30%

0%

5%

10%

15%

20%

25%

30%

Variation

Ave

rage

Upp

er S

pec

Lim

it

Tar

get

Low

er S

pec

Lim

it

0%

5%

10%

15%

20%

25%

30%

0%

5%

10%

15%

20%

25%

30%

0%

5%

10%

15%

20%

25%

30%

Variation

Ave

rage

Upp

er S

pec

Lim

it

Tar

get

Low

er S

pec

Lim

it

Page 8: STATISTICS AND OUTLIERS Aaron Saks Process Advancement Leader Boise, Inc 10/26/2011

2.3

2.35

2.4

2.45

2.5

2.55

2.6

Caliper

TGT

LRL

URL

Ppkm = 1.02

• Here is an example of a process that has a Ppkm>1, meaning it is fully capable of meeting customer expectations.

• Process is running to target – i.e. the mean is equal to the target

• Variation is well within the Specification (Red) limits.

8

Page 9: STATISTICS AND OUTLIERS Aaron Saks Process Advancement Leader Boise, Inc 10/26/2011

2.3

2.35

2.4

2.45

2.5

2.55

2.6

Caliper

TGT

LRL

URL

2.3

2.35

2.4

2.45

2.5

2.55

2.6

Caliper

TGT

LRL

URL

Ppkm = .65

9

Running to Target – Excessive variation

Acceptable variation – but off Target

Page 10: STATISTICS AND OUTLIERS Aaron Saks Process Advancement Leader Boise, Inc 10/26/2011

Outliers

• We spend most of our time working on the outliers:• Process problems• Lowering cost• Increasing production

• Outliers exist in people too• Top 20• Middle 70• Bottom 10

• The future leaders of tomorrow will be from the top 20.

Page 11: STATISTICS AND OUTLIERS Aaron Saks Process Advancement Leader Boise, Inc 10/26/2011

Thoughts for a New Engineer

• Understand what is expected • Focus on independent learning• Always speak from facts and data• Effective communication creates results • Focus on business results and accomplishments

• You don’t have to be at the top to be a leader

Page 12: STATISTICS AND OUTLIERS Aaron Saks Process Advancement Leader Boise, Inc 10/26/2011

Questions:

• What has been your biggest challenge?• If you could go back to school what additional classes would you take?

• What's a typical day like?• What kind of products does Boise make?• How long do projects last?• What ethical issues have you encountered?