Six Sigma

1Six Sigma for Chemical Engineers

George Liebermann

Senior Engineering Fellow / Black Belt

Xerox Research Centre of Canada

AIChE WebinarJuly 8, 2009

[email protected]

2Outline

1. Six Sigma Methodologies Overview

Six Sigma and Lean Six Sigma

Design for Six Sigma

2. Six Sigma / DMAIC

3. Implementing Six Sigma in Chemical Industry

4. Six Sigma Tool Box for Chemical Engineers

5. Summary

3Six Sigma for Chemical Engineers

Six Sigma methodologies are normally implemented by companies via training by consultancies and a company-wide organization is

usually put in place for implementation and continuous practice

There is a large body of literature: books as well as internet sites that provide information at various levels

The goal of this short course is to provide an introduction to

Six Sigma:

- mainly for practicing chemical engineer and

- a guide to use the most relevant tools for chemical

engineering applications

4 The Greek symbol Sigma which means standard deviation; a measure of variation

As Sigma (, s) gets smaller the error in predicting the result gets smaller

= Standard Deviation

A statistical measure of a processs ability to meet customer requirements (CTQs)

Process Six Sigma equates to 3.4 defects per million opportunities or a Cpk =2

= Measure of Process

Capability

Standard Deviation and Process Capability

Yield DPMO Sigma

69.2% 308,000 2

93.3% 66,800 3

99.4% 6,210 4

99.977% 230 5

99.99966% 3.4 6

Yield DPMO Sigma

Cpk = ~1.33

Cpk = 2

Six Sigma Methodologies Overview

LSLxor

xUSLMinC pk

33

-6 -3 -2 -1 x +1 +2 +3 +6

2

1

1

)

n

ii

n

xx2

1(

i 2

1

1

)

n

ii

n

xx2

1(

i

Sigma Percent

+/-1 sigma 68.27

+/-2 sigma 95.45

+/-3 sigma 99.73

5Six Sigma Methodology

Evolved from TQM methods, procedures to reduce defects, statistics

Structured methodology for product and process improvement / optimization and problem solving

Key features:

Customer requirements: specifications

Process thinking: understand the process

Data driven: reduction of variations, statistics, statistical process control

Structured approach: defined phases (roadmap), checkpoints

Tool box: tools to select from

Generally: - top down (management championing),

- projects lead by trained Six Sigma Green Belts and Black Belts

- team approach

Standard Six Sigma methodology is based on a 5 phase roadmap:

DMAICDefineDefine MeasureMeasureMeasureMeasure AnalyzeAnalyzeAnalyzeAnalyze ImproveImproveImproveImprove ControlControlControlControl


6Lean

Types of Wastes

Waste of Production

Waste of Time

Waste in Transportation

Waste of Processing

Waste of Inventory

Waste of People Movement/Time

Waste of Product (defective)

Lean: Methodology with tools/techniques to

reduce/eliminate the waste in order to reduce cost


7Lean Six Sigma (LSS):

Integration of Two Improvement Approaches

Goal: Less Waste, Higher Speed

Focus: Identify and Eliminate Waste and Delays

Method: Team Event/Inspection

Goal: Improve Performance on Customer Requirements

Focus: - Eliminate Variation- Optimize output

Method: - Use of DMAIC & Tools - Use of trained individuals (Black Belts, Green Belts)

Six SigmaCulture + Quality

LeanLower Cost

After

Before

BeforeAfter


time Parameter (output)

8Design for Six Sigma (DfSS)

Success of Six Sigma in improvement of existing

processes/products lead to a search and development

of a similar methodology for new processes/products

DFSS (or DfLSS) Alphabet soup

While Six Sigma/DMAIC is well established, there are several versions of DfLSS methodologies, for example:

DMEDI Define/Measure/Explore/Develop/Implement

DMADV - Define/Measure/Analyze/Design/Verify

IDOV Identify/Design/Optimize/Validate

ICOV Identify/Characterize/Optimize/Validate


DfSS uses many tools used in Six Sigma/DMAIC

DfSS not discussed in this webinar

9Six SigmaDMAIC Improvement Methodology

Define

Measure

Analyze

Improve

Control

Define the problem or opportunity vs customer requirements

Search for the key factors (critical Xs) with the biggest impact and determine the root causes

Measure the process/product performance

Develop and demonstrate solutions to improve the critical Xs

Implement and monitor the solution

Six Sigma / DMAIC

DefineDefine MeasureMeasureMeasureMeasure AnalyzeAnalyzeAnalyzeAnalyze ImproveImproveImproveImprove ControlControlControlControl

10

Six SigmaDMAIC Improvement Methodology

Guidelines for each phase:- Objective- Methodology- Tool box- Review readiness check

Measurable Objectives agreed upon in the Define phase to be monitored throughout the project

Six Sigma / DMAIC

DefineDefine MeasureMeasureMeasureMeasure AnalyzeAnalyzeAnalyzeAnalyze ImproveImproveImproveImprove ControlControlControlControl

Tollgate (Phase Gate) Reviews after each phase

Structured, well suitable for chemical engineers

11

Six Sigma in Chemical Industry

Six Sigma/DMAIC:

problem solving / improvement in existing processes and/or products

Chemical process/product (manufacturing)

Process demonstration, including pilot plant

Pilot plant/semi-works interim production

Technical Support Activities

- Analytical Methods and Quality Control

- Maintenance

- Utilities

- Others

Other Processes (work process improvements)

- Engineering

- Purchasing

- Inventory Management and Sales/Marketing

- Others

Implementing Six Sigma in Chemical Industry

12

Six Sigma in Chemical Industry

Typical Six Sigma opportunities for chemical engineers

(Chemical) Process/Product Problem Solving Yield Cycle time Product or Intermediate Quality Equipment or Instrumentation/Control Performance

(Chemical) Process/Product Optimization or Improvement (same as above)

Off-line Process/Product related Analytical/QC problem solving or improvements Maintenance, Utilities and other support processes Raw Materials

Implementing Six Sigma in Chemical Industry

13

Six Sigma

DMAIC Define Phase Tasks

1. Define why, what, goal, how, when, who: Charter

2. Review the process Walk the process

Supplier-Input-Process-Output-Customer (SIPOC) Map

Update PFDs, P&IDs High-level Map, Value Stream Map

3. Review specifications/targets Voice of Customer (VOC) and/or Voice of Business

Critical Customer Requirements (CCR) key deliveries/results needed

4. Identify and secure resources (material and personnel)

5. Sketch out Project Plan/Schedule (eg in Microsoft Project)

6. Assess expected financial benefit

Define MeasureMeasureMeasureMeasure AnalyzeAnalyzeAnalyzeAnalyze ImproveImproveImproveImprove ControlControlControlControl

Six Sigma Tool Box for Chemical Engineers

Six Sigma terms are in blue and red

Red items further details

14

Guidelines for Charter Content

Project Title Date:

Project Sponsor: Deployment Manager:

Stakeholders: Black Belt: Green Belt:

Business Case/Impact

Opportunity/Problem Statement

Goal Statement

Project Scope

Project Plan

Resources

Why should this be doneWhat is the benefit Does this project align with the business needsWhat is the projected value of the project ($)

What is the concern and impactWhat is wrong or not working

Opportunity lost or potential benefit

Measurable Goal definitionWhat are the deliverables

How will the project be doneWhat are the boundaries of the project

What is not within scope

When will it be doneMajor milestones (tollgates)

Proposed date of completion

Who is in the team - roles & responsibilitiesTime allocation

Expected expense and capital spending

Normally the Black Belt/Green Belt, with help from team, develops the Charter.


WWGHWW

15

SIPOC

Supplier-Input-Process-Output-Customer Map

Process Name: DBPProcess Name: DBP

Process : Process : Building CBuilding C

Suppliers Inputs Process Output(s) Customers

Occidental Chem.

Vista Chemical Eastman Chemical Penta Manuf.

Goldschmidt Chem.

Calgon

Local Disttributor

Plant facilities

Phthalic anhydride

Butanol

pTSA

Activated carbon

Sodium Hydroxide

Steam

Cooling Water

Vacuum

Raw Materials Charging

- Molten PA addition

Reaction

- Heat-up, reaction,

reactor discharge

Washing steps

-Neutralization

-Water washes

Steam Stripping

- Continuous distillation

Act. Carbon Treatment

and Filtration

-Treatment

-Filtration and pump to

storage

Dibutyl Phthalate

Wet butanol

Aqueous waste

Internal

QC Lab Distribution Sales

External

Bulk customers Distributors

Very useful for overall review of process/product. Supports:

- High-level Process Map (Process Flow Diagram)- Value Stream Map (P&ID or enhanced PFD)

- Identification of Key Outputs (Ys) and Inputs (Xs) towards Y = f(X)(Cause end Effect info)


Problem: Dibutyl Phthalate color is sometimes out of specification

16

Value Stream Map (VSM)

Goal:

Update of all process information required for the project

Update existing PFD/P&IDs

Break down process steps, materials quantities and flow

Record all process steps/times (value-add and non value-add)

Add all measurement points, samples for QC, etc

Record all known info/values: process parameters, metrics

Team check: walk the process

VSM: An updated/enhanced version of

Process Flow Diagram (PFD)

or Process & Instrumentation Diagram (P&ID)


17

Six Sigma

DMAIC Measure Phase Tasks

1. Review Key Process & Product Parameters

(Key Input, Process and Output Variables)

2. Establish what needs to be measured/analyzed and HOW

(Operational Definitions )

3. Validate Measurement Methods Measurement System Analysis (MSA)

4. Data Collection Plan (historical and new)

5. Collect Data Data Organizing Tools

6. Baseline (current) Process Capability Analysis

Define MeasureMeasureMeasure AnalyzeAnalyzeAnalyzeAnalyze ImproveImproveImproveImprove ControlControlControlControlMeasure


Red items further detailsSix Sigma Tool Box for Chemical Engineers

18

Select what to measure

Key Input, Process and Output Variables

X1: Raw Materials Quality (x11, x12) Y1: DBP ColorX2: Process Conditions (x21, x22) Y2: DBP PurityX3: Equipment conditions (x31, x32) Y3: Process Cycle Time (cost factor)X4: Individual Process Steps Time (x41, x42)

Yi= f(X1, X2..)

Process Name: DBPProcess Name: DBP

Process : Process : Building CBuilding C

Suppliers Inputs Process Output(s) Customers

Occidental Chem.

Vista Chemical Eastman Chemical Penta Manuf.

Goldschmidt Chem.

Calgon

Local Disttributor

Plant facilities

Phthalic anhydride

Butanol

pTSA

Activated carbon

Sodium Hydroxide

Steam

Cooling Water Vacuum

Raw Materials Charging

- Molten PA addition

Reaction

- Heat-up, reaction,

reactor discharge

Washing steps

-Neutralization

-Water washes

Steam Stripping

- Continuous distillation

Act. Carbon Treatment

and Filtration

-Treatment

-Filtration and pump to

storage

Dibutyl Phthalate

Wet butanol

Aqueous waste

Internal

QC Lab Distribution Sales

External

Bulk customers Distributors

Input Metrics Process Metrics Output Metrics

Quality

Speed

Cost

Reaction time Number of washes Stripping steam rate AC treatment and filtration time

UMC (rework yes/no)

Phthalic Anhydride

Melt Color Acid numberButanol

Color % Assay (GC)pTSA

% assay

Weight of raws Reaction temperature Vacuum Acid Number (end of reaction)

NaOH amount Acid Number (after washes)

Stripping Temperature, Vacuum

% Assay (GC) Specific Gravity Color (APHA) Acid Number (mg KOH/g) Flash point (oC)

Overall process time

Process Time Rework Time, if any

Tools for Y and X selection:

SIPOC,

Value Stream Map

Cause and Effect Diagram

In a chemical engineering project the

key output (Y) is usually well defined,

however several secondary Ys may

need to be addressed


Problem:

Dibutyl Phthalate color

is sometimes out of spec

19

Ishikawa Cause & Effect (Fishbone) DiagramTool to Determine Key Input, and Process Variables

MaterialsProcessEquipment/Facilities

Procedures/Policies Measurement:

-

Effect

(Y)

,

Potential cause

People

Potential cause

Potential cause

Potential cause

Potential causes (Xs)

Cause Categories

Potential cause

Potential cause

Potential cause Potential cause

Raw MaterialsSynthesis ConditionsPost-synthesis

Equipment Color Measurement:

DBP

Color

,

Operator

Lab bias

APHA/Hazen

CIELAB

N2 purging procedure

Reactor bottom valve

Reactor seal

Sources

Reducing

Agents

Decolorizing

agents

Neutral. Temp.pTSA color

PA color

Inerting procedure

Inert Gas

Reaction temperature

Reaction time

Reactor

Discharging

Temperature

RM addition

PA impurities

Procedures/Execution

Stripping column

RM addition

Butanol Imp.

Team contribution:

brainstorming, available info,

assumptions


Problem:

Dibutyl Phthalate color

is sometimes out of spec

20

Validate Measurement Methods - MSA

Are the measurements methods good enough?

A Measurement System Analysis (MSA) assesses how much variation is associated with the measurement system and compares it with the process variation (and specifications)

Main Tool: Gage R&R assesses the percentage measurement variation (repeatability and reproducibility) represents of the variation of the process.

Gage R&R results:

Under 10%: Acceptable measurement system.

10% to 30%: May be acceptable.

Over 30%: Unacceptable measurement system

Repeatability same operatorReproducibility different operators

Software packages like Minitab and JMP have standard MSA (Gage R&R) programs

Less expensive versions available from Six Sigma Consultancies (Excel based)


21

Data Organizing Tools

Control Charts

Lots

Yie

ld,

%

71645750433629221581

100

90

80

70

60

_X=81.42

UCL=98.70

LCL=64.14

Historical yields

Pareto Charts

Yield target =90%

X-bar = 81.4%

Process in control, but Mean (X-bar)

for yield significantly under target process is not capable

0

10

20

30

40

50

60

Butanol

impur.

Butanol

color

Time over

8hr

pTSA color PA color PA Impur.

Co

un

t

In Excel:

Bar (Column) Chart using Sort

In Minitab

Ranking based on frequency of occurrenceUpper and Lower Control Limits

+/- 3 sigma

Helps select vital few from trivial many 80/20 rule

For Analyze phase


22

Baseline Process Capability Analysis

Process Capability index - Cp

Cp =

Process Capability:

Voice of the Customer (spec. limits) vs Status of Business (control limits)

Accepted Variation (Specification)

Process Variation

6

LSL -USLCp

Clarifies current status and will allow analysis/comparison of future improvements

LSL USL

Process Variation - 99.65% of values

Target

Cannot be used unless there are both lower (LSL) and upper specification (USL) limits

If process mean not centered on target results can be misleading

Cpk is the most common index for process capability

For cases where both LSL and USL is specified, the smaller Cpk number is reported

Typical manufacturing goal for Cpk is 1.33 - Sigma Qual. Index (SQL) of 4

LSLxor

xUSLMinC pk

33

X

_

A SQL is equal to a Cpk of 2.0

6


23

Six Sigma

DMAIC Analyze Phase Tasks

1. Data and Process Analysis Tool box

2. Lean Opportunities - Non-Value Added Analysis

3. Identify most likely causes of the problem - Prioritize Critical Input and Process Variables (Xs) for Y = f(X)

4. Root Cause Proof/Experimentation

5. Summary/documentation of cause/effect findings

Define MeasureMeasureMeasure AnalyzeAnalyzeAnalyzeAnalyze ImproveImproveImproveImprove ControlControlControlControlMeasure Analyze



24

Data and Process Analysis

Root Cause Analysis Hypotheses for root causes of problem

Data: Statistical and other Six Sigma tools

Descriptive statistics and Data distribution - Excel

Control charts Excel applications

Pareto Charts - Excel (column charts sorted)

Regression Analysis Excel y=f(x) with R2

Hypothesis Testing (t-test and ANOVA) - Excel


Process: Six Sigma tools

Brainstorming

Process Maps PFD, P&ID

Cause and Effect (Fishbone) Diagram

Failure Mode and Effects Analysis (FMEA)

Non-Value Added Analysis

25

Hypothesis Testingstatistical treatment

: 210 H

1. 1 sample t-test: Compares sample mean to historical mean or target

2. 2 sample t-test: Compares means of two samples

1. One way ANOVA: Compares means of more than two samples

2. Two way ANOVA: Compares two samples with multiple factors

Ho - null hypothesis: the mean of a sample/population is equal to.

Ha - alternative hypothesis: the mean of a sample/population is not equal to.

e.g. for two equal means the

difference is zero

: 21 aH

nothing changed


26

lean tool

Non-Value Added Analysis

Identifies Process Steps or Process Time which does not add value to the Process/Product

Categories:

Customer Value Added (CVA) meets a customer need

Business Value Added (BVA) required to be in business

Non-Value Added (NVA) not required for customer/business needs - waste

Customer Value Added (CVA)

customer sees the value, e.g. synthesis, purification

should be considered for optimization

Business Value Added (BVA)

required to complete the process/product, e.g. QC, maintenance, change-over, transfers

should be optimized and reduced

Non-Value Added (NVA)

not required, waste in the process, e.g. rework, storage/inventory, holding/delays, extra process time

should be eliminated


27

Experimentation for Proof of Cause-Effect

Experimentation to prove hypothesis:

Confirm information from analysis of existing data

preferably in lab or pilot plant, sometimes at manufacturing scale

Experimentation

one factor at a time (OFAT)

design of experiments - DOE (designed experimentation)

Design of Experiments The variation of controllable factors(independent variables, Inputs, Xs) at different levels in order to assess their effect on the response (dependent variable, Output, Ys)


28

Six Sigma

6. DMAIC Improve Phase Tasks

1. Generate/Develop Potential Solution(s)

2. Establish Evaluation Criteria Quality, Cost, Delivery, Safety

3. Select Best Solution(s) Pair-wise Comparison, Pugh Matrix

4. Risk Assessment Failure Modes and Effects Analysis (FMEA)

5. Pilot/Demonstrate Solution

Define MeasureMeasureMeasure AnalyzeAnalyzeAnalyzeAnalyze ImproveImproveImproveImprove ControlControlControlControlMeasure Improve



29

Pair-wise Comparison/Ranking:

Pairwise Comparison Example

Distillation Yield Improvement

Solution Ranking/Preference

Replace all valves 1 1 1 1 4

Install blanks 1 1 1 3

Change Control Program 0

Install check valves 1 1 2

Replace Column 1 1

Compare Solutions in

Pairs

Solution Selection Tools

Pugh Matrix:

1. List Requirements/Evaluation

Criteria

2. Assign Weight to each Evaluation

Criteria

3. List Solutions to be assessed

(< 8-10)

4. Current process is the Baseline

5. Fill in Pugh Matrix

6. Rate Solutions vs baseline:

+, 0, -, or ++, +, 0, -, --

7. Compute

8. Compare Totals (Pareto)

9. Select solution

Evaluation Criteria (VOC/VOB)

Quality, Cost, Delivery

Total

Weight

(Importance

Ranking)

(1-5)

Solution 4

Pugh Matrix - Which solution is the best

Baseline

(Current

process)

Solution 1 Solution 2 Solution 3

Solutions


30

Failure Modes and Effects Analysis (FMEA)

Proposed

Solution/Change

Potential

Failure Mode

Potential Failure

Effects

Potential

Causes

Current

Controls

R

P

N

Actions

Recommended

Resp.

Description of

change as per

proposed

solution

Negative

Impact,

Failure, or

Error

Impact on

Output

Variables

(identified risk

items )

What could

cause the

problem

Existing

controls and

procedures

To avoid/reduce

occurrence

and/or improve

detection

S

E

V

E

R

I

T

Y

O

C

C

U

R

A

N

C

E

D

E

T

E

C

T

I

O

N

Process / Product FMEA in Improve

Rating Scale:

1 not severe, not likely to occur, likely to detect, to10 very severe, very likely to occur, not likely to detect

Severity - Importance of the effect on customer requirements.

Occurrence - Frequency (or probability that it will occur).

Detection - The ability of current controls to detect, before or after occurrence

In Improve:

List of Proposed Changes

Recommended Actions if

problem occurs

Responsibility


31

Pilot/Demonstrate Solution

Performed on a small scale (equipment or time)

Confirms with data the expected improvements

Reduces risk of failure, exposes issues

Will make the full scale implementation more effective

Compare results with baseline Process/Product

Tools:

Descriptive statistics, histograms

2 sample t-test compare mean of Improve results vs baseline

ANOVA (ANalysis Of Variance) compare mean of Improve results vs 2 or more baseline set of results

Process capability (Cpk) vs baseline


32

Six Sigma

7. DMAIC Control Phase Tasks

1. Modify/Develop Standard Operating Procedures (SOP)

2. Develop a Control Plan

3. Implement Process Changes and Controls

4. Perform and Monitor the Improved Process

5. Transition Project to Process Owner

6. Recalculate Expected Financial Benefits

Define MeasureMeasureMeasure AnalyzeAnalyzeAnalyzeAnalyze ImproveImproveImproveImprove ControlControlControlControlMeasure Control

Beyond Control Phase: sustain the improved results

Six Sigma terms are in blue


33

Control Plan

Measurements: What to measure, Measurement Methods

Sampling Plan

Control Plan Tools:

comparison to data prior to improvement

- control charts

- process capability (Cpk, Sigma Quality Level)

- (2 sample t-test)

- risk management

- mistake-proofing: redundancies

Control Plan Documents


34

Six Sigma and Chemical Engineers

Chemical industry and chemical engineering are data driven, very well suited for applying Six Sigma

Apply the tools creatively, where and when deemed to help, and in conjunction with sound technical judgment

Normally Six Sigma is a corporate/organization wide initiative- management driven, training

Do other options exist ?Projects may be conducted according to Six Sigma methodology without having a formal Black Belt or Green Belt project or full training

Not that complex use of phase methodology and associated tool box

36

Select Six Sigma Websites

iSix Sigma - http://www.isixsigma.com

Soc. of Manuf. Engineers - http://www.sme.org/sixsigma

Six Sigma.us - http://www.6sigma.us

Knowware International - http://www.qimacros.com/index.html

Aveta - http://www.sixsigmaonline.org/index.html

Pyzdek Institute - http://www.pyzdek.com

Many others

37

Select Six Sigma Books

Six Sigma Demystified A Self-Teaching Guide by Paul Keller, 2005 ISBN: 007144544-7

Lean Six Sigma Demystified A Self-Teaching Guide by Jay Arthur, 2006 ISBN: 007148650X

The Six Sigma Handbook: The Complete Guide for Greenbelts, Blackbelts, and Managers at All Levels, Revised and Expanded Editionby Thomas Pyzdek, 2003 ISBN: 0071410155

The Lean Six Sigma Pocket Toolbook: A Quick Reference Guide to 100 Tools for Improving Quality and Speedby Michael L. George, John Maxey, David T. Rowlands, Mark Price2004 - ISBN: 0071441190

Six Sigma for Dummiesby Craig Gygi, Neil DeCarlo, Bruce Williams2005 - ISBN: 0764567985

38

Descriptive Statistics (mean, standard deviation, variance, median, confidence intervals)

Histograms

Run Charts

Control Charts

Scatter Plots

Bar Charts

Gage R&R (MSA)

Cp & Cpk

Pareto Chart

Box plots

Design of Experiments (DOE)

Regression Analysis

Analysis of Variance (ANOVA)

Main Effects Plots

Interaction Plots

Hypothesis Testing (e.g. t-test)

Response Surface Methodology

Six Sigma Statistical Tools

39

Basic Six Sigma ToolsHost of Tools That Can Be Used When Relevant Tool Box

Statistical Tools

SIPOC

Process Mapping

Value Analysis

Brainstorming

Quality Function Development (QFD HOQ)

Fishbone Diagram

Operational Definitions

Measuring System Analysis (MSA)

Affinity Diagrams

Interrelationship Diagraphs

Force Field Analysis

Tree Diagrams

Cause & Effect Matrix

Failure Mode and Effects Analysis

Piloting

Standard Operating Procedure (SOP)

Control Plan