Download as pdf or txt
Download as pdf or txt
You are on page 1of 53

12/5/2012

Product and Process Design Improvement Study -


Application Roadmaps

Ranjit k. Roy
Nutek, Inc.
3 829 Quarton Road,Bloomfield Hills, MI 48302. USA.
Email: Support@Nutek-us.com www.Nutek-us.com

This document presents a roadmap for application of the design of


experiment technique (DOE/Taguchi) to scientifically investigate
product and process designs. In general, the object of such studies
are to improve process performance by optimizing designs or
finding solutions to issues with production based on experimental
data.

Application steps including content of the discussions for


experimental steps are described in this document.

Content:

- Project Identification
- System definition
- Experiment planning
- Etc.

Slide 2 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

1
12/5/2012

Thoughts for the day . .


“Give a man a fish and you feed him for a day. Teach a
man to fish and you feed him for a lifetime.”

-Chinese Proverb

Slide # 3

Driving Beliefs & Values for Project Studies

◦ Team over individual


◦ Consensus over opinion
◦ Project team effort directed toward
meeting company objectives

Slide # 4

2
12/5/2012

Project Initiation & Background Activities


1. Identify Projects Design Development Production

2. Appoint Project Appoint Team Leader


Leader

Solve Problems Optimize Designs


3. Establish Project ◦ Technical issues • Option selection
Objectives
◦ Warranty • Robustness
◦ Reject • Process Studies

4. Review Process
Review Process Flow and Define
Flow System for Study

Form Team
• First-hand knowledge
5. Form Team
• Customer/supplier
• Stake holders
Slide # 5

Identify Project

Select projects that:


- Have higher return on investment
- Are of smaller in size which can be carried out
within available time period
- Are of important to the stakeholders

Slide 6 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

3
12/5/2012

Define System

System may include all or part of the products or


processes. It must always include suspect areas (sub-
(sub-
processes or components) of the subject process under
study. Incorporate problem knowledge, when available, to
establish system boundaries.

Slide 7 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

System definition utilizes knowledge about problem


and probable causes.

Consider the process of baking cakes (shown next).

Once the system is defined (boundaries), proceed to


describe:

- Outputs (Objectives)

- How outputs are measured


- etc. (Planning discussions)

Slide 8 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

4
12/5/2012

System View of Process


(Cake Baking Process)

Gather Bake Cake in


Mix Ingredients
Ingredients Oven
(Make Batter)

System may be defined one of the three different ways shown


above (Dashed lined rectangles).

Input Output

System

Note: OUTPUT of previous sub-process is INPUT to the next. For example: Batter is output
of MIXING process, but input to the BAKING process.

Ref. Page N/A


Flowchart
“A flowchart (also spelled flow-chart and flow chart) is a schematic representation
of an algorithm or a process.” - Wikepedia

The process flowchart was first introduced by Frnak Gilbreth in 1921. He used it
show as a graphical and structured method for documenting process activities.

Figure 1. Clutch Plate Fabrication Process

Stamping / Deburring Rust


Hobbing Clutch plates Inhibitor
Clutch plate are tumbled in Parts are
made from a large submerged in
1/16 inch thick container to a chemical
rolled steel remove sharp bath
edges
Cleaned and dried parts
are boxed for shipping.

Slide 10 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

5
12/5/2012

Experiment Planning Meeting


- Readiness Checklist

Meeting Room needs: Date:


- Computer Projector Time
- One or Two Flipcharts Location:

Checklist of required tasks (TO DO)


Project Team Facilitator/Consultant
1. Project identified 1. Help team leader with project and
2. Project Leader/owner appointed team selection (Phone/email)
3. Team selected 2. Share meeting agenda and this
4. SYSTEM defined (if possible) presentation with the leader
5. Collect PROCESS FLOW and
CAUSE & EFFECT diagrams
6. Meeting scheduled and team
members invited

Slide 11 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

Experiment Planning Discussions with Project Team


Project Team = Project Leader + Group 1 + Group 2
Group 1: Management and stakeholders
Group 2: All individuals with direct knowledge of the process under study.

Agenda

Meeting 1: (Participation of Group 1 required)

- Project and Objective Description Definition

Meeting 2: (Participation of Group 2 required)


Brainstorming for Ideas leading to factor selection

- Interaction and Noise (Robustness)

- Project Logistics

Slide 12 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

6
12/5/2012

Ref. Page N/A

5 Steps Process Study Roadmap


5
Run Tests to
Confirm
Advisor & Team – on site

Advisor – off site


4 Solutions

Analyze Test
Project Team & Advisor 3 Results and
Prescribe
Carry Out Solutions
Planned Tests
and Collect
2 Results

Design and
1 Describe Test
Recipes
Hold
Experiment
Planning
Discussions

Slide 13 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

1 of 5 Steps in the Study

1 Hold Experiment Planning Discussions


This necessary session is facilitated by Nutek specialists. All project team
members need to participate in most sessions of the planning
discussions.

Experiment planning discussion is:


- most critical phase in the application process
- done on-site in which Nutek specialists work with project team
-requires that project team members to set up meetings (Part A and
Part B)
- where project goals are defined and study factors are identified
- follows a proven structure developed by Nutek.
(See task details in next few slides)

Slide 14 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

7
12/5/2012

Ref. Page 2-20

Example Experiment Planning Meeting Topics

Project Title: Pound Cake Baking Process Optimization Study


Objective: Determine the recipe of the “overall best” cake.
Are there more than one objectives?
How are the objectives evaluated, measured, and quantified?
What are the criteria of evaluation?
What are the relative weighting of these criteria?

Criteria Description Worst Best QC Relative


Reading Reading Weight (Wt)

C1: Taste 0 8 B 60
C2: Moistness 25 – 70 gms 40 gms. N 25
C3: Voids/Smoothness 6 0 S 15

Slide # 15

Ref. Page 2-21

Example Experiment –Evaluation Criteria

Criteria Description Worst Best QC Relative Sample Readings


Reading Reading Weight (Wt) Sample 1 Sample 2
C1: Taste 0 8 B 60 5 6
C2: Moistness 25 – 70 gms 40 gms. N 25 46 35
C3: Voids/Smoothness 6 0 S 15 4 5

OEC formulation require that (1) All criteria have the same direction of desirability, (2) Readings are
expressed in ratios, and that (3) Each criteria gets include its own weighting.

C1 C2 – C2-nom C3
OEC = x Wt1 + ( 1 - ) x Wt2 + (1 - )x Wt2
C1range C2b – C2-nom C3range

Calculation of OEC’s with Sample Readings:


OEC-1 = (5/8) x 60 + [1 – (46-40)/(70-40) ] x 25 - ( 1 – 4/6) x 15
= 37.5 + 20.0 + 5 = 62.5

OEC-2 = (6/8) x 60 + [1 – (40-35)/(70-40) ] x 25 - ( 1 – 5/6) x 15


= 45 + 20.83 + 2.5 = 68.33

Slide # 16

8
12/5/2012

Ref. Page 2-22

Factors Identification and Qualification (Planning)


Long List Qualified List Study List
1. Sugar 1. Sugar 3 Determine scopes of experiment
2. Butter 2. Butter (number of experiments possible
4 based on time and money)
3. Sifted Cake Flour 3. Flour
4. Egg 4. Egg 1
5. Baking Powder… 5. Baking Powder… Suppose 8 -10 experiments.
5
6. Granulated Sugar 6. Granulated Sugar Select L-8, which means 7 2-level
6 factors can be studied.
7. Vegetable Coloring 7. Vegetable Coloring
……. ………. 7
Select 7 out of 13 factors by team
12. Smell of Cake 12. Vanilla Extract . consensus.
13. Browning of cake 13. Brandy
… =============== .
22. Type of Oven .
23. Kitchen Temp. 22. Type of Oven
.
24. Vanilla Extract 23. Kitchen Temp.
25. Brandy .
26. Mixing Time Note: Scrutinize list to select. factors (input
variables)

Slide # 17

Ref. Page 2-34


Application Tasks Review (Basic Designs)

Plan Experiment
Agree on a Title
Define objectives Design Experiment
◦ Evaluation criteria &QC Select appropriate orthogonal
Analyze Results
◦ Relative weighting array Compute average and
◦ Table of Eval. Criteria Assign factors to the columns standard deviation
Brainstorm for factors Readjust array selection if Calculate grand average
◦ Long LIST necessary Calculate factor averages
◦ Qualified List (Ordered) Describe trial conditions Plot factor average effects
◦ Study List Establish number of samples Analyze results
tested in each trial condition
Establish Factor levels ִ Factor influence
Create DATA COLLECTION
◦ How many levels ִ Optimum condition
sheet
◦ 2-level strategy ִ Predicted performance
Prepare any special
◦ 3-level strategy improvement
instruction for test and data
Identify Interactions handling
◦ Two factor interaction Determine other
recommendations and
◦ Strategy
conduct CONFIRMATION
Consider Robust Design TESTS
◦ Noise factors
Assign TASKS for project
completion
◦ Who does what?

Slide # 18

9
12/5/2012

Project Initiation (Background Tasks)

1. Select Project with key characteristics:


• Significant financial loss
• Higher potential ROI
• Important to management
2. Appoint Team Leader
• Responsible for project output
• Interested in improvement
3. Form Project Team (leader + Group I + Group II)
• persons with first hand knowledge
• People involved in incorporating changes
• Individual responsible for test machines and equipments
• Group-I (objective and measure), Group-II (Factors and
settings)

(Schedule planning discussion)

Slide 19 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

Details of Activities in Step 1

Slide 20 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

10
12/5/2012

1a. Project Definition


Part A (First session, 2 – 4 hours with Group-I)

Project Title
• Contains process description (Spells out goals)
• Brief description of problem – PROBLEM STATEMENT
Objectives
• What we are after (one or more)
• How are they evaluated
• What is the units of measurement
• What direction is the results desired
• If multiple objectives, details of each and their relative weights
Data Reduction Scheme for Multiple Objectives
• Pareto chart for objectives
• Need for a single index (OEC)

Slide 21 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

1b. Brainstorming for Factors


1. Sources of ideas
• 6 M’s for Manufacturing (man, machine, method, materials,
measurement & mother nature)
• Process flow diagram, cause and effect charts
2. Long list of factors
• Discuss process flow
• Brainstorm for factor (new and known inputs)
3. Qualified list of factors
• Input, controllable, easy to adjust..
• Not output (may be uncontrollable)
• Releasable
• Identify uncontrollable
4. Paretoized qualified list
• Order by team consensus
5. Scopes of study
• Consider test size based on cost and time
• Include confirmation samples
• Test machine availability
6. Study list of factors
• Select top factors based on 2-level

Slide 22 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

11
12/5/2012

Example: High Heating Cost


(List of factors/causes)

Materials Machine
• Natural Gas • Furnace
• Filters
• Propane • Humidifies
• Oil/Hot water

Method Man
• Duct Cleanliness • Temperature Setting
• Fireplace
• Vents Open & Close • Late night Stay
• Space Heaters • Lack of Worm Clothes
• Insulation • Excessive door open/close

• Window Glass

Mother Nature
Measurement • Too many clod days
• Meter Reading Error • Windy nights
• Leaky Gas Tube • Storm and Rain
• Thermostat Control

Slide 23 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

Example: High Heating Cost

Slide 24 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

12
12/5/2012

Example: Factors Identification and Qualification

Long List Qualified List Study List


1. Sugar 1. Sugar 3 Determine scopes of experiment
2. Butter 2. Butter (number of experiments possible
4 based on time and money)
3. Sifted Cake Flour 3. Flour
4. Egg 4. Egg 1
5. Baking Powder… 5. Baking Powder… Suppose 8 -10 experiments.
5
6. Granulated Sugar 6. Granulated Sugar Select L-8, which means 7 2-level
6 factors can be studied.
7. Vegetable Coloring 7. Vegetable Coloring
……. ………. 7
Select 7 out of 13 factors by team
12. Smell of Cake 12. Vanilla Extract . consensus.
13. Browning of cake 13. Brandy
… =============== .
22. Type of Oven .
23. Kitchen Temp. 22. Type of Oven
.
24. Vanilla Extract 23. Kitchen Temp.
25. Brandy .
26. Mixing Time Note: Scrutinize list to select. factors (input
variables)

1c. Factor Level Definitions


1. Establish levels of Study List of Factors one at a
time by using available information on:
1. Factor influence
2. Current working level
3. Range of values possible to test and
release

Establish all factors at TWO levels unless


required to set it to three or four levels.

Slide 26 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

13
12/5/2012

1d. Interaction & Noise Factors


1. Interaction study
• Number of possible interactions
• Interaction between 2-level factors [ n (n-1)/2 ]
• Prior knowledge of interaction
• Sacrifices for interaction study
• Practical strategy “Dig wide ,not deep”

2. Noise Factor & robustness


• Identified noise factors
• Unknown noise
• Robust design strategy
• Dealing with noise

Slide 27 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

Ref. Page 9 - 32
Create Process Diagram (or P-Diagram)
Double-click here to start defining
DC parameters (shows first screen
below).
System Configuration or Process Diagram (P-Diagram)

Double-click on Signal to define Signal and


Noise factors and levels (next three screens)

Slide # 28

14
12/5/2012

1e. Test Scopes & Project Logistics

1. Test requirements (Sample and recipes)


2. Who does what
3. What to do about factors not included in the study
4. Summary Test requirements
• Number of objectives
• Number of test conditions (Trials)
• Number of samples/trial
• Total number of test samples
• Results collection and reporting

Slide 29 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

Experiment Planning Meeting


- Desired Output

Project Title, Project Objectives, Factors, Levels, Interactions,


Evaluation Criteria, Overall Evaluation Noise Factors, Size of
Criterion Experiments, etc.

Knowledge and Responsibilities


Team and Leader to decide Next Steps
1. Materials and test equipments 1. Consultant/Specialist to
2. Who will set up tests and DESIGN experiments and
when? prescribe test plan and DATA
3. Who will collect data? COLLECTION scheme
4. Who will forward
authenticated test data to
Nutek and when?
5. Etc.

Slide 30 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

15
12/5/2012

Planning Discussion Document -1

Slide 31 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

Planning Discussion Document -2

Slide 32 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

16
12/5/2012

Details of Activities in Step 2

Slide 33 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

2 of 5 Steps in the Study

2 Design and Describe Test Recipes


Work in this phase is generally done specialist (away from the
team). Based on the planning discussion, an experiment is
designed and the individual trial conditions are described. A
written report should be provided to project leader:

• Prescribe the recipe of each separate tests to be conducted

• Describe method to evaluate multiple objectives if applicable

• Defined method of data collections,

Slide 34 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

17
12/5/2012

2a. Data Collection Sheet


1. Prescribe collection sheet (Example shown below)
2. Specify result reporting requirements

Data Collection Sheet (L-8 design, 3 samples & 2 Criteria)


Trial Sample 1 Sample 2 Sample 3
# Criteria 1 Criteria 2 Criteria 1 Criteria 2 Criteria 1 Criteria 2

1 x x x x x x

2 x x x x x x
1 x x x x x x

2 x x x x x x

1 x x x x x x

2 x x x x x x
1 x x x x x x

2 x x x x x x

Slide 35 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

Details of Activities in
Step 3

Slide 36 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

18
12/5/2012

3 of 5 Steps in the Study

3 Carry Out Planned Tests and Collect Results


Tests are run by the project team. The collected results are sent to
Nutek for analysis. Tests are conducted by the team but managed by
Nutek specialist.

• Review all test conditions for possible conflict


• Randomize and sequence test
• Confirm test conditions before run (factors not included in the
study should be fixed for all tests)
• Collect results for each samples and record results in data collection
sheet.
• Note any variation in method of evaluation
• Forward results for analysis

Slide 37 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

Details of Activities in
Step 4

Slide 38 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

19
12/5/2012

4 of 5 Steps in the Study


4 Analyze Test Results and Prescribe Solutions
Test results vast amount of information. Information in FOUR basic
categories is obtained from: (1) Factor Influence (Factor average
effect or Main effect), (2) Relative influence of the factors to the
variation of results (ANOVA, which is short form for Analysis of
Variance), (3) Optimum condition, and (4) Expected performance at
the optimum condition.

Report of analyses will contain:


• Complete analysis (as described above)
• Observations and findings (formal report)
• Optimum design
• Expected performance with confidence interval
• Relative influence of factors
• Interactions possibilities between factors

Slide 39 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

Details of Activities in
Step 5

Slide 40 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

20
12/5/2012

5 of 5 Steps in the Study

5 Run Tests to Confirm Solutions


This final step is strongly recommended before adopting the
solutions in updating the design.

• Review solution (optimum condition)


• Run 5 or more samples at optimum condition
• Establish range of expected performance (confidence interval)
• Compare sample average performance with expected range
• Establish if solution is confirmed or not
• Incorporate recommended design changes in production process

Slide 41 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

Organizational DFSS Strategy (IDOV)

Identify Design Optimize Verify

Application Roadmap for Robust Design (DOE)

Identify Projects Appoint Leader Set Objectives Define System Form Team
- Higher ROI - Project - Achievable - Clear boundary - First-hand
- Can be done ownership - Implementable - Relates to knowledge
- Value to - Process - Team consensus issues - Customer &
stakeholders knowledge - Keeps focus supplier
- Work as team - Narrower - Stakeholders

Plan Experiment Design Expts Run Tests Analyze Results Confirm Recmd.
-Results & QC - Array - Order of tests - Main effects - Sample size
- Criteria of - Test description - Sample size - ANOVA - Validation
evaluations - Samples/trial - Results - Optimum & C.I. - Lessons learned
- Factors & Levels - Data collection recordings - Improvement - Future tasks
- Interaction - Multiple criteria - Loss/Savings
-- Noise factors
-- Logistics

Slide 42 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

21
12/5/2012

“Robust Design and Problem Solving


Application Support “
from
Nutek, Inc.
http://nutek-us.com/wp-Problem_Solving.html

Slide 43 Nutek, Inc. Process Study Application Steps http://Nutek--us.com


http://Nutek

Web site links:


All Seminar Topics: http://Nutek-us.com/wp-sem.html
Training & Workshop Design for Six Sigma (DFSS) http://nutek-us.com/wp-dfss.html
Assistance with Taguchi/DOE Seminar: http://nutek-us.com/wp-s4d.html
Onsite DOE seminar: http://nutek-us.com/wp-ons.html
application DOE Application Support: http://nutek-us.com/wp-support.html
Text Books Consulting & Problem Solving: http://nutek-us.com/wp-sps.html
Qualitek-4 Software: http://nutek-us.com/wp-q4w.html
Software (Free DEMO
download) Project Management Training: http://nutek-us.com/wp-pmp.html

Testimonials: http://nutek-us.com/wp-q4w-eval.html
Nutek Client List: http://nutek-us.com/wp-nut.html

FREE Game and References: http://nutek-us.com/wp-free.html

Nutek, Inc.
3 829 Quarton Road
Bloomfield Hills, MI 48302. Tel : 1-248-812-2071
Email: Support@Nutek-us.com www.Nutek-us.com

22
12/5/2012

Thoughts for the day . .


“Give a man a fish and you feed him for a day. Teach a
man to fish and you feed him for a lifetime.”

-Chinese Proverb

Slide # 45

23
12/5/2012

The Taguchi Design of Experiments (DOE) for


Robust Product and Process Designs

Application Case Study -


Examples

1
12/5/2012

Slide 2

2
12/5/2012

Slide 3

3
12/5/2012

Circuit Board Flow Soldering Process Study


(Process flow is studied to identify separate areas of studies)

Notation Factor Description Level 1 Level 2


A: Stencil Design Design 1 Design 2
B: Stencil Material Material 1 Material 2
C: Reflow ramp-up 0.5 ºC/sec 1.0 ºC/sec
D: Reflow peak temp Max. ºC Min. ºC

E: Board contamination Low exposure Long exposure

F: Stencil Cleanliness 5-Cycles 20 Cycles

G: PCB Support (SMT) Present Absent


Slide 4

4
12/5/2012

Collagen Coating of Bypass


Tubes Made of Fabric

Slide 5

5
12/5/2012

HYBRID SPRING DISENGAGEMENT FORCE STUDY

Slide 6

6
12/5/2012

Example Case Study# S1: (Production Problem Solving)

I. Experiment Planning
Project Title - Adhesive Bonding of Car Window Bracket

An assembly plant of certain luxury car vehicle experienced


frequent failure of one of the bonded plastic bracket for power
window mechanism.The cause of the failure was identified to be
inadequate strength of the adhesive used for the bonding.

Objective & Result - Increase Bonding Strength


Bonding tensile (pull) strengths were measured in three axial
directions. Minimum force requirements were available from
standards set earlier.
Quality Characteristics - Bigger is better (B)

Factors and Level Descriptions


Bracket design, Type of adhesive, Cleaning method, Priming time,
Curing temperature, etc.

II. Experiment Design & Results


Six different process parameters were quickly studied by
experiments designed using an L-8 array.

Slide # 7

7
12/5/2012

Example Case Study# S2: (Casting Process Optimization)

I. Experiment Planning
Project Title - Die-Casting Process Parameter Study (CsEx-01)

In a die casting process, metal (generally alloys of Aluminum, Zinc &


Magnesium) parts are formed by flowing molten metals (at 1200 – 1300
deg F) in the cavities of the dies made of steel.
Objective & Result Reduce Scraps
Quality Characteristics - Smaller is better (S)
# Criteria Descriptions Worst - Best Reading QC Rel. Weig
1 Crack and Tear (length) 10 mm 20 mm S 20
2 Heat Sinks (diameter) 15 mm 0 mm S 30
3 Lamination (area) 5 sq.cm 0 sq.cm S 25
4 Non-Fill (area of void) 2 sq.cm 0 sq.cm S 25

Commonly Observed Characteristics


There are many types of observed defects that result in scrapped parts.
The common defects observed are, Surface abnormalities (Cold flaw,
Cold lap, Chill swirls, Non-fill, etc.), Lamination (layers of metal on
inside or outside surface), Gas Porosity, Blister, Shrinkage Porosity,
Heat sinks, Crack & tears, Drags, Gate porosity, Driving ejector pins,
etc.

II. Experiment Design & Results


An L-12 array was used to design the
experiment to study 10 2-level factors.
Factors are assigned to the column in
random order. The results of each criteria
of evaluations were analyzed separately.

Slide # 8

8
12/5/2012

Example Case Study # S3: (Production Problem Solving)

I. Experiment Planning
Project Title - Clutch Plate Rust Inhibition Process Optimization Study
(CsEx-05)

The Clutch plate is one of the many precision components used in the
automotive transmission assembly. The part is about 12 inches in
diameter and is made from 1/8-inch thick mild steel.
Objective & Result - Reduce Rusts and Sticky
(a) Sticky Parts – During the assembly process, parts were found to be
stuck together with one or more parts.
(b) Rust Spots – Operators involved in the assembly reported
unusually higher rust spots on the clutch during certain period in the
year.

Factors and Level Descriptions


Rust inhibitor process parameters was the area of study.

Figure 1. Clutch Plate Fabrication Process

Stamping / Deburring Rust


Hobbing Inhibitor
Clutch plates
Clutch plate are tumbled Parts are
made from in a large submerged
1/16 inch container to in a
thick rolled remove sharp chemical
steel edges bath
Cleaned and dried parts
are boxed for shipping.

II. Experiment Design & Results


One 4-level factor and four 2-level factors in this experiment were
studied using a modified L-8 array. The 4-level factor was assigned to
column 1 modified using original columns 1, 2, and 3.

Slide # 9

9
12/5/2012

Example Case Study # S4: (Product Design Optimization)

I. Experiment Planning
Project Title - Piston Bearing Durability Life Optimization Study

Recent warranty study indicated high rate of field failure of


the Piston Bearing (type DX30) due to excessive wear and
vibration related malfunction.

Objective & Result - Increase Durability Life


Experimental bearing samples were tested and performance
evaluated by observing durability life (in hours, weighted at 60%)
and by measuring force generated due to unbalanced vibration (g
force, weighted at 40%). These two criteria of evaluation were
combined to produce an overall result used for the analysis.

Factors and Level Descriptions

II. Experiment Design & Results


Four 2-level factors and an interaction between two factors were
studied using an L-8 orthogonal array.

Slide # 10

10
12/5/2012

Example Case Study # S5: (Robust Process Design )

I. Experiment Planning
Project Title - Adhesive Bonded Joint Strength Study

Manufacturer of a popular brand of audio speakers was


experiencing higher than normal field failure with one of their
newer products. The source of the problem was identified as
bonded piece of magnetic component. Most failures were reported
from distribution centers in the tropical weather (Hot & humid).

Objective & Result - Increase the bonding strength under direct


tensile (pull) loading.
QC - Bigger is better (B)

Factors and Level Descriptions


FACTORS Level – I Level - II
A: Contact Plate Galvanized Brass Plated
B: Adhesive Thickness Current Specs. Thicker
C: Base Surface Machined Rough
D: Curing Temperature Ambient Air Convection

II. Experiment Design & Results


Four 2-level factors and three interactions
among factors were studied using an L-8 array.
Interactions AxB, BxC, abd AxC were studied
by reserving the columns prescribed by the
2-level Triangular table.

FIND MORE CASE STUDIES IN 16 Steps to Pruduct and process


Improvement …….

Slide # 11

11
12/5/2012

Example Case Study # S6: (Chemical Process Optimization)

I. Experiment Planning
Project Title - Study of Tenacity of Staple Viscose Fiber

Viscose is the man made fiber with excellent properties that can be engineered and
optimized for different textile and non-woven applications. The name was adopted
(1924), in preference to "artificial silk", by the U.S. Dept. of Commerce and various
commercial associations. Cellulose from wood pulp, washed, bleached, and pressed
into sheets, is dissolved by chemicals, then forced under pressure through minute
holes in a metal cap (spinneret), emerging as filaments that unite to form one
continuous strand solidified by passage through a suitable liquid or warm air.
Various stages of viscose fiber offer opportunities for process optimizations.
Objective & Result - Increase tensile strength of fiber
Quality Characteristics - Bigger is better (B)
The tensile strength of individual fiber is measured in terms of gram force/denier
(typically 2.5 - 2.7 gm weight/denier). A denier is a measure of weight of 9000 meter
long strand of fiber in grams (typically 1.2 - 1.5 gm/9000m).
Factors and Level Descriptions

II. Experiment Design & Results


Four different process parameters and interactions among them were
studied by experiments designed using an L-8 array.

Slide # 12

12
12/5/2012

Example Case Study # S7: (Healthcare Product Optimization)

I. Experiment Planning
Project Title - Study of In Vitro Strength of CORTOSS Bone Void Filler

CORTOSS is a high-strength, bone-bonding, self-setting


composite engineered to mimic the strength characteristics of
human cortical bone. Clinical studies are carried out during
development stage for multiple indications of CORTOSS,
including the augmentation of joints and vertebral fractures. DOE
was used to determine the formulation for CORTOSS that
produced the most desirable strength characteristic.

Objective & Result - Increase Fatigue Strength


Long term strength against tensile and impact loading are
measured.
Quality Characteristics - Bigger is better (B)

Reinforcing Factors Studied:


Synthetic combeite glass-ceramic particles,
Baria-boroalumino-silicate glass (BBAS)

radiopacity, Silica (SiO2), Silaine, etc.


II. Experiment Design & Results
The Process ingredients were studied by experiments designed using an L-8
array.
Cortical bone in the spine, arms and legs can become
damaged due to trauma or degenerative disease such as
osteoporosis. It then is often surgically repaired using wires,
hooks, metal plates and screws. CORTOSS offers an
alternative treatment to cortical bone which is structural and
weight bearing in nature. The new treatments with
CORTOSS provide sufficient strength and bonds directly to
bone to reinforce the weakened area and allow the bone to
Slide # 13 heal naturally.

13
12/5/2012

Example Case Study # S8: (Healthcare Product Optimization)

I. Experiment Planning
Project Title - Wound Care Product Design Study

A leading manufacturer of a broad range of wound care products


(bandages, needles and syringes, electrodes, specialized paper,
vascular therapy, urological care, incontinence care, and nursing
care products) carried out a study to determine specifications for
dressings for moist wound healing.

Objective & Result - Increase peel strength and functional life.


Peel strength close to a target value was desired. Longer lasting
bandage with minimum wear under normal use was also sought.
Quality Characteristics Nominal is the best & Bigger is better

Factors and Level Descriptions


A:Type of Fabric (Coarse, Flexible, Textured)
B:Adhesive Conc. (Low, Normal, Strong)
C:Curing Process (Extended, Current, Absent)
D:Packaging (Metal Wrap, Paper, Plastic)

II. Experiment Design & Results


Four different process parameters at three levels each were
studied using an L-9 orthogonal array..

Slide # 14

14
12/5/2012

Example Case Study # S9: ( Sporting Event Optimization)

I. Experiment Planning
Project Title - Bow & Arrow Tuning Study

There are a number of structural and geometrical factors in bow and


arrow that determines how well the arrow fly. A contestant for Olympic
archery competition planned to use DOE to lay out a set of experiments
to determine the best bow and arrow setting for best performance.

Objective & Result: Improve accuracy of hitting the bullseye. The


accuracy can be measured in terms of radial distance of the hit from the
center of the bullseye.
Quality Characteristics:
Radial distance measured in inches. Smaller is better

Factors and Level Descriptions:


A:Arrow Stiffness (Force required to pull the string)
B:Draw Length
C:Draw Weight (Force when not linearly proportional with draw length)
D:Point Weight (Weight of point of arrow, steel, 90 - 110 grams)
E:Plunger Button Tension (Compression force at guide - arrow rest)
F:Center shot (Horizontal location of the guide)
G:String Type (Plastic, Kevlar, etc.)
H:Knocking Location (Location of the arrow nock on string)
Interactions: AxE, ExF, AxB, & AxC
II. Experiment Design & Results
This experiment is designed using an L-16 array to study 8 factors and
4 interactions.

Slide # 15

15
12/5/2012

Example Case Study # S10: ( Race Car Optimization)

I. Experiment Planning
Project Title - Race Car Suspension Parameter Optimization

To achieve highest performance, major suspension parameters of


race cars like those for Daytona Superspeedway (2.5 mile oval; 31
degrees banking in 1-4 turns) are fine tuned for the track. Test
vehicle components can be evaluated by laying out simple
experiments to determine the most desirable combination.

Objective & Result: Determine the best combination of suspension


parameters for the race car.
Quality Characteristics: Time to complete the track. Smaller is better.

Factors and Level Descriptions: (Source: USA Today, February 15, 2002)
A:Right Front Tire Pressure (23 - 55 psi) Green = Superspeedway
B:Left Front Tire Pressure (15 - 30 psi)
C:Right Rear Tire Pressure (20 - 50 psi)
D:Left Rear Tire Pressure (15 - 30 psi)
E:Right Front Spring Rate (1,900 - 800 lbs/in)
F:Left Front Spring Rate (700 - 800 lbs/in)
G:Right Rear Spring Rate (225 - 350 lbs/in)
H:Left Rear Spring Rate (15 - 30 lbs/in)
I:Rear Spoiler Angle (0 - 55 degrees)
II. Experiment Design & Results
Up to 11 factors as shown above can be studied by designing an
experiment using an L-12 array.

Slide # 16

16
12/5/2012

AUTOMOBILE WHEEL PAD FLATNESS


MACHINING STUDY

Background: Lack of flatness in wheel pad was


suspected to be one of the reasons for uneven brake pad
wear. Improved machining of wheel pad was sought to
correct the brake performance.

Objectives: Determine machining parameters that


improves wheel pad flatness.

Method of Evaluation: The pad flatness was measured


by taking 42 readings on the pad surface for each wheel
sample. Range and standard deviation of such readings
were used as the sample performance data.

Summary Findings and Recommendations:

Based on the results of 63 (7 samples for each of 9 trial


conditions from L-9 experiment) experimental sample
wheels, here are some of the observations and
recommendations. The results (range and standard
deviation data) were analyzed both by standard method
(using average of results) as well as by using S/N ratios
of trial results. Conclusions from standard analysis have
been found to be fully supported by S/N analysis.

Slide 17

17
12/5/2012

Key observations:
1. The combination of parameters (Optimum Condition) which is likely to produce
minimum variation in pad flatness is
OPTIMUM CONDITION (2000 rpm, 4mm nose radius, .008/in feed, and 2 passes at
.010in)

OPTIMUM CONDITION (2000 rpm, 4mm nose radius, .008/in feed, and 2 passes at .010in)

2. The influential factors are (four factors tested are shown in order of their influence):
Feed Rate (63.4%)
Check Speed (12.7%)
Tool Nose Radius (7.3%)
Finish Passes (4.3%)
The percentages within the parentheses represent the influence of the factors to the
variation of the result.
3. Chuck Speed (1000, 1500, and 2000 rpm) affects results in a nonlinear manner. For
both Range and Std. Deviation, 1500 rpm is worse than either of 1000 or 2000 rpm. All
other factors appear to influence results in a linear manner (see plot of Main effects).
4. When the new process design is implemented, the reduced variation will result in less
number of wheel require rework (or rejected). Using the average (of all 9 trial condition)
performance as a reference, the cost savings expected is,
Expected Savings = 80 %
The above represents savings in cents for every dollar that would be spent when the
performance is at a level equal to the average of all the 9 trial conditions
Recommendation:
To confirm the prediction based on the experimental results, please setup the process
condition identified above (optimum condition) and fabricate 7 or more wheels in the
same manner as the original samples. Evaluate and record results as before (Range and
std. deviation).

Expected Std. Dev. = 0 - 3.0 (95% confidence


level)Expected Range = 4 - 12 (95% confidence level)

Slide 18

18
12/5/2012

Reference Calculations and Findings

Slide 19

19
12/5/2012

Type of Analysis
Standard and S/N analysis performed.
Both types of analyses confirmed the conclusion about the optimum
condition.
The expected values (Std. Dev. And Range) at the optimum condition are
specified based on conservative estimate from both analyses.

Slide 20

20
12/5/2012

Slide 21

21
12/5/2012

Slide 22

22
12/5/2012

Slide 23

23
12/5/2012

Solar Wafer Slicing Process Parameter Study

Slide 24

24
12/5/2012

Solar Wafer Final Cleaning Process Study

Slide 25

25
12/5/2012

Slide 26

26
12/5/2012

Slide 27

27
12/5/2012

Slide 28

28
12/5/2012

Slide 29

29
12/5/2012

Notes:

Slide 30

30

You might also like