Six Sigma at Cytec - Innovation and Six Sigma

1

Six Sigma at Cytec

Innovation and Six Sigma

2

MotivationSix Sigma Training: Total Trained vs. Spending - 2001-2005

720707622

9701000

0102030405060708090

100110120

2001 2002 2003 2004 2005

Nu

mb

er t

rain

ed

0

200

400

600

800

1000

1200

Budget (K$ USD)

Managers

Green Belts

Black Belts

Master BlackBelts

0102030405060708090

100110120

2001 2002 2003 2004 2005

0

200

400

600

800

1000

1200

0102030405060708090

100110120

2001 2002 2003 2004 2005

0

200

400

600

800

1000

1200

0102030405060708090

100110120

2001 2002 2003 2004 2005

0

200

400

600

800

1000

1200

0102030405060708090

100110120

2001 2002 2003 2004 2005

0

200

400

600

800

1000

1200

0102030405060708090

100110120

2001 2002 2003 2004 2005

0

200

400

600

800

1000

1200

3

Guiding Thoughts

1. To make more money, we should learn more about our products and processes.

2. To learn, we should employ the scientific method (a.k.a. induction/deduction).

3. To employ the scientific method, we should select a tool set and methodology based on critical thinking.

4

Guiding Thoughts

TOOLS LEARNING PROFIT

5

Guiding Thoughts




6

Critical Factors

All factors:

Y1 = f(x1, x2, x3…xn)

Y2 = f(x4, x5, x6…xm)

...

Critical factors:

Y* = f(x1*, x3*, x7*)

Y = f(x)

7

Science and Technology

• To paraphrase David Lilley, CEO…

– We are Cytec – our name indicates our focus on technology.

– We develop technically advanced solutions that differentiate our products from our competitors’ products.

– We sell technology in addition to products.

• The generation of knowledge is an important part of Cytec’s culture and success.

8

Learning and Financial Gain

Project goals

Financial gain

9


Project goals

Financial gain

Benefits offinancial gain

10


Project goals

Financial gain


11


Project goals

Financial gain Knowledge gain


12


Project goals

Financial gain Knowledge gain


Benefits ofknowledge gain

13

EISTC Level in CS2 recovered

0

5

10

15

20

25

30

35

1 20 39 58 77 96 115 134 153 172 191 210 229 248 267 286 305 324 343 362 381

Number of Batch

EIS

TC

leve

l % w

/w Implementation of recommendations

14

When is Six Sigma Necessary?

Mike Farrington – Whirlpool Corp.

15



16



17



18



19

The Challenge for Six Sigma

• Tie together all the pieces that support…


20




Guiding Thoughts

21

The Scientific Method

Statistics for Experimenters, Box, Hunter, and Hunter, Wiley, 1978.

22

DMAIC

Define

Measure

Analyze

Improve

Control

23

Science and Critical Thinking

“Science is not any particular method or set of techniques. It is a way of reasoning.

The standards are intellectual rather than procedural. The method of observation, formalization, and testing must vary with

the nature of the problem.”

Scientific Method and Social Problems, Edgar G. Miller, in Science 109, 290, 1949.

24

What are we trying to accomplish?

• We want to learn and innovate, so we can be more profitable.

• We want to take advantage of the power of Six Sigma:– Focus on variability reduction– Statistical experimentation– Consistent project management

• We want to foster creativity and innovation through critical thinking.

25

DMAIC

Define

Measure

Analyze

Improve

Control

26

Critical Factors

All factors:

Y1 = f(x1, x2, x3…xn)

Y2 = f(x4, x5, x6…xm)

...

Critical factors:

Y* = f(x1*, x3*, x7*)

Y = f(x) gainknowledge

27




Guiding Thoughts

28

Alternatives to DMAIC

• The thought map…

(Six Sigma Associates)

LEGEND:

ProblemQuestions/issues

THEORIESPredictions

Data/informationCollection

Results

ActionsComments

Links

29

Thought Map Model

P rob lem

T H E O R Y

D ata C o llec tionS tra teg ies (P ro 's/C on 's)

P red ic tion s

R esu lts/C on clu sion s

S o lu tion s

T H E O R Y

D ata C o llec tionS tra teg ies (P ro 's/C on 's)

P red ic tion s

A n aly zeD ata

E xecu tion

(Q u estion s/O b servation s)

. . . .

MO

DIF

IED

TH

EO

RIE

S

30

Advantages ofThought Maps

• Drives learning, captures ideas (good and bad)

• Flexible, robust

• Captures the journey as well the results

– Example: What didn’t work?– Example: What else might we have done?– Example: What unexpected things happened?

31

Advantages ofThought Maps

• Transparent and accessible

• Excellent group tool, promotes discussion

• Not hard to learn

• Hard to do “wrong”

32

Disadvantages ofThought Maps

• Messy

• Potentially unfocused

• Difficult to standardize

• Difficult to file

• Difficult to evaluate

• Not sequential (in time) in the usual sense

• Not time based

33

Core Processes

Learn

Develop people

Develop the Organization

Short Courses

Green Belt training

Black Belt training

Master Black Belt training

Project Leadership training

Mentoring

$$

Manager training

Project Selection

Candidate Selection

Project Tracking

Project Closure

Certification

Rewards and recognition

Managing by data and metrics

34

Thought Map Example

35

WE CAN MAKE CRUDE XAND PURIFY BY

RECRYSTALLIZATIONAND STEAM

DISTILLATION

Path 1

This route has been fullydeveloped technically (lab

scale). However, capitalinvestiment required at site

makes price of productunacceptably high.

This path is not feasiblewith current customer and

specifications. Seemonthly reports of (people)

from (dates) for moreinformation.

How can the process for production of X from A be developed to meet customerspecifications for a 50% solution in water by (date)?

APHA color < 10pH 7.0-9.0

chloride content <1200 ppmcost < (price)

Others: metals, bromide, alkylhalide, etc. - refer to customer documentation.

Route currently under investigation for processdevelopment:

A + B + C --> X + Y

Other routes have been investigatedextensively (see monthly reports 1974-2002)

Can product from thecurrent lab process be

purified to meet thespecifications? IMPURITIES CONTAINING

Cl- (P, Q) WILL BEPRESENT IN ALL

SAMPLES of A

Path 2

CoV work showed that theamount of Cl- impuritiesdepends heavily on the

solvent used for storage.Variation in the range of10000-20000 ppm was

observed.

P, Q ARISE AS ACONSEQUENCE OF

EXTENDED STORAGE OFA IN SOLVENT. PURE A

SHOULD BE USEDRATHER THAN A THATHAS BEEN STORED INSOLVENT - THIS WILL

REDUCE THE MEAN ANDVARIATION IN IMPURITY

LEVELS.

Does the starting materialhave impurities?

MS IS “REPORTEDLY”VALID AND AGREES WITH

CUSTOMER MS, BUTSHOULD BE CHECKED.

Path 3

Initial MSE foundmeasurement error was

~500 ppm. This isunacceptable for this

application.

CoV study indicated asystematic effect related to

the stirring time, and thetime between sample

preparation and analysis.

Is the Cl- measurementsystem valid?

CoV study to determinethe levels of chloride indifferent lots of starting

material A

CoV study to determinethe variability in chloridelevels in different lots of

pure A.

Variability is less than thatof the MS (<12 ppm, 6

sigma)

Embed MSE in Cov study

ALL SAMPLES SHOULD BESTIRRED FOR AT LEAST

(TIME) AT (SETTING)BEFORE ANALYSIS.

Second MSE to evaluate theeffect of changes in

measurement procedure

Measurement error reducedto ~12 ppm (6 sigma) -

sufficient for thisapplication. N13.

Path 4

Can the standard variablesbe used to increase purity to

sufficient levels withoutfurther purification steps?

Path 5

Can Cl- be removedchemically by addition of

(reagent)?

Path 6

Can (person)’s method forchloride removal be

employed?

Consider optimizing:

mode of additionstoichiometry

stirringreaction temperaure

reaction timetype of solvent

amount of solventfiltration procedurerotavap procedure

others?

Could this route beimplemented at another

plant?

See path 4 detail

Current best conditionsproduce chloride levels of1100 ppm with variation of

~50 ppm (6 sigma) andsuccessfully meet color and

pH specifications.

STANDARD VARIABLESCAN BE OPTIMIZED TO

REDUCE CHLORIDECONTENT OF PRODUCTBEFORE PURIFICATION.

THIS METHOD HAS BEENDEMONSTRATED FOR Br-,AND MAY WORK FOR Cl-AS WELL. SHOULD NOTAFFECT pH OR COLOR.

Study factor effects usingsequential DOEs

Study this treament optionusing whole plot/split plot

DOE designs

This treatment had noeffect on chloride or pH

(DOE2), but removed colorto give final product thatbeat color specification.

Include this treatment instandard procedure from

now on.

Stir all samples for (time) at(setting) before analysis.

TREATMENT WITH(reagent) DOES NOTAFFECT CHLORIDE

CONTENT OR pH, BUTDOES REMOVE COLORED

IMPURITY.

THIS METHOD HAS BEENDEMONSTRATED FOR

RELATED CASES, BUT ISUNKNOWN WITH X AND A.

Study this treament optionusing whole plot/split plot DOE

designs

See path 6 detail

This treatment had no effect onchloride under a wide variety of

conditions.

STILL ROOM FORIMPROVEMENT IN THIS

AREA. (SEE PATH 4 DETAIL)

Best previous procedure typically saw <10000ppm chloride, but often as high as 50000 ppm

before purification. Had seen as low as 3000 ppmonce. APHA color ranged from 75-150 units. pH

was approximately 4.5.

36





37






A + B + C --> X + Y





38






A + B + C --> X + Y



purified to meet thespecifications?






(reagent)?


employed?




39






A + B + C --> X + Y



purified to meet thespecifications?






(reagent)?


employed?




LEGEND:

ProblemQuestions/issues

THEORIESPredictions

Data/informationCollection

Results

ActionsComments

Links

40



DISTILLATION

Path 1






A + B + C --> X + Y





SAMPLES of A

Path 2




Path 3


Path 4



Path 5


(reagent)?

Path 6


employed?









41



DISTILLATION

Path 1












A + B + C --> X + Y





SAMPLES of A

Path 2




Path 3


Path 4



Path 5


(reagent)?

Path 6


employed?


plant?









42



DISTILLATION

Path 1












A + B + C --> X + Y





SAMPLES of A

Path 2




Path 3



material A


Path 4



Path 5


(reagent)?

Path 6


employed?


plant?









43



DISTILLATION

Path 1












A + B + C --> X + Y





SAMPLES of A

Path 2



observed.




Path 3



application.



material A


Path 4



Path 5


(reagent)?

Path 6


employed?


plant?









44



DISTILLATION

Path 1












A + B + C --> X + Y





SAMPLES of A

Path 2



observed.





LEVELS.




Path 3



application.






material A




Path 4



Path 5


(reagent)?

Path 6


employed?


plant?









45



DISTILLATION

Path 1












A + B + C --> X + Y





SAMPLES of A

Path 2



observed.





LEVELS.




Path 3



application.






material A


pure A.






Path 4



Path 5


(reagent)?

Path 6


employed?


plant?









46



DISTILLATION

Path 1












A + B + C --> X + Y





SAMPLES of A

Path 2



observed.





LEVELS.




Path 3



application.






material A


pure A.


sigma)








Path 4



Path 5


(reagent)?

Path 6


employed?


plant?










47



DISTILLATION

Path 1












A + B + C --> X + Y





SAMPLES of A

Path 2



observed.





LEVELS.




Path 3



application.






material A


pure A.


sigma)








Path 4



Path 5


(reagent)?

Path 6


employed?






others?


plant?










48



DISTILLATION

Path 1












A + B + C --> X + Y





SAMPLES of A

Path 2



observed.





LEVELS.




Path 3



application.






material A


pure A.


sigma)








Path 4



Path 5


(reagent)?

Path 6


employed?






others?


plant?






DOE designs





designs




49



DISTILLATION

Path 1












A + B + C --> X + Y





SAMPLES of A

Path 2



observed.





LEVELS.




Path 3



application.






material A


pure A.


sigma)








Path 4



Path 5


(reagent)?

Path 6


employed?






others?


plant?






DOE designs




now on.




IMPURITY.




designs




50



DISTILLATION

Path 1












A + B + C --> X + Y





SAMPLES of A

Path 2



observed.





LEVELS.




Path 3



application.






material A


pure A.


sigma)








Path 4



Path 5


(reagent)?

Path 6


employed?






others?


plant?






DOE designs




now on.




IMPURITY.




designs

See path 6 detail


conditions.




51



DISTILLATION

Path 1












A + B + C --> X + Y





SAMPLES of A

Path 2



observed.





LEVELS.




Path 3



application.






material A


pure A.


sigma)








Path 4



Path 5


(reagent)?

Path 6


employed?






others?


plant?

See path 4 detail



pH specifications.






DOE designs




now on.




IMPURITY.




designs

See path 6 detail


conditions.




52



DISTILLATION

Path 1












A + B + C --> X + Y





SAMPLES of A

Path 2



observed.





LEVELS.




Path 3



application.






material A


pure A.


sigma)








Path 4



Path 5


(reagent)?

Path 6


employed?






others?


plant?

See path 4 detail



pH specifications.






DOE designs




now on.




IMPURITY.




designs

See path 6 detail


conditions.

STILL ROOM FORIMPROVEMENT IN THIS

AREA. (SEE PATH 4 DETAIL)




53

54

55

56

57

Conclusions

• Thought maps are a viable alternative to DMAIC. There are probably others.

• The Cytec Six Sigma program may not look like Six Sigma programs at other companies, but it meets our needs, matches our culture, and achieves excellent results.

• It is necessary to both learn from others and innovate.

58

Guiding Thoughts




59

Guiding Thoughts


60

Thank You

andrew.downard@cytec.com

Six Sigma at Cytec - Innovation and Six Sigma

Education

sigma necessary

scientific method statistics

critical factors

particular method

method of observation

critical thinking science

sigma training

sigma associates

Six Sigma Revealed From International Six Sigma Institute

Six Sigma Courses - Professional development · Six Sigma.....

Six Sigma Project Six Sigma Project

Six Sigma Continuous Improvement Training Six Sigma...

Six Sigma Business Excellence Executive Overview · SIX...

Q-LEARNING Six Sigma DMAIC E-LEARNING · inhaltsverzeichnis...

Q-LEARNING Six Sigma DMAIC E-LEARNING · 2019-09-20 ·...

Six Sigma Yellow Belt Part I -...

Six Sigma and DFSS Training NSS1: Six Sigma Executive...

Improving Your Business Results Six Sigma Qualtec Six Sigma....

SIMULATION-BASED LEAN SIX-SIGMA AND DESIGN FOR...

Six Sigma and Lean Six Sigma - IIM Bangalore Sigma and Lean....