ECE 480. Six Sigma Overview & Introduction to Design for Six Sigma

Transcription

1 ECE 480 Six Sigma Overview & Introduction to Design for Six Sigma

2 Six Sigma Companies 3M Delphi Allied Signal Armstrong Astrazeneca Avnet Bayer Black & Decker Boston Financial Services Celanese Conoco Covenant Health Decoma Dow Corning Eastman Kodak Emerson Ford General Motors GlaxoSmithKline Harley Davidson ADT Security American Standard Armstrong World Industries Atlantic Health Bank of America BC Hydro Boeing Calloway Golf City Bank Corning Crompton Degussa Deutsche Bank DuPont Eaton Corp. Energizer Fortis Health Georgia Pacific Goodrich Goodyear HP Air Products Americhem Asahi Kasai Avery Denison BASF Becton Dickenson Bombardier Caterpiller Chlorox Cott Beverages Dannon Dell Dow Chemical Eastman Chemical Eli Lily Florida Light & Power General Electric Gillette Hitachi

3 More Six Sigma Companies Honeywell ITT Industries Johns Manville Corp. Kohler Corp. Lord Mckesson National Semiconductor Noranda Omnova Solutions Pitney Bowes Raytheon Royal Bank of Canada SAS Inst. Sherwin Williams Sony Sun Chemical Timken Transfreight Visteon IBM ITW Johnson & Johnson LG Chemical Lubrizol Moog NBC News Northrop Grumman Owens Corning PPG Industries Rogers Corp. Saint-Gobain Scott Seeds Siemens Sprint Sunbeam Toyota TRW WR Grace Intel John Deere Kellogg Lockheed Martin Maytag Motorola NCR Corp. Noveon Phillips Praxair Rhom & Haas Samsung Seagate Technologies Silicon Graphics Square D Swagelok Trane US Filter Xerox

4 Lecture Six Sigma Tour Lecture Title WOWing Customers with Six Sigma Products - DFSS Understanding the Customer s Viewpoint - VOC Homework Assignment Quality Function Deployment (QFD) Customer Driven Development Function Definition & Analysis Powerful Problem Analysis Technique Homework Assignment Six Sigma Optimization - MAIC G. G. A. A. Motter, Motter, &

5 WOWing Customers with Six Sigma Products... via Design for Six Sigma

6 DFSS Discussion Objectives Define Quality, Defect, and Sigma Level Describe generic DFSS Process flow Highlight the DFSS Process with an example Explain What s different about DFSS from traditional Engineering Design approach?

7 Six Sigma is... A Systematic Data Driven Approach for: Continuous Improvement MAIC Problem Solving DFSS

8 Quality and Sigma Level Defined Quality : degree of excellence of a Product, Process, Software, IT System, or Service.... from the Customer s Viewpoint Every process has Variation. If the outcome is too far from target value (beyond a spec limit), a Defect occurs Standard deviation is a measure of statistical variation (spread) about the mean Sigma Level of a process is an indication of how often defects are likely to occur = Spec Width / 2 (Std. Deviation)

9 Matching Product Requirements and Process Capability Manufacturing Sigma Level % Out of Spec Defects Voice of Process Lower Spec Limit Mean Upper Spec Limit Voice of Customer Source: Implementing Six Sigma, F. Breyfogle III

10 Optimize or Design? Define the Project Std. Engin. Design Product Exist? Yes D D No M F S No Breakthru Needed? No A I S C Yes Top Bus. Oppty? No Source: Six Sigma for Growth, E. Abramowich

11 Why DFSS?: Revolutionize Product Development Reactive Design Quality DFSS Predictive Design Quality From Evolving design requirements Extensive design rework Product performance assessed by build and test Performance & manufacturability problems fixed after product in use; fire fights Quality tested in To Disciplined CCR flowdown Controlled design parameters Product performance modeled and simulated Designed for robust performance & manufacturability Quality designed in

12 Requirements Flow Down from Customer and Design Capabilities Flow Up Customer System Requirements Flow Down System Design Subsystem Design Assembly Design Part Design Capability Flow Up Design for Robustness, Reliability & Manufacturability at Specified Sigma Level Source: Design for Six Sigma, C. Creveling

13 Define Identify CCRs Design Optimize for 6 Validate Define the Project Business Case, Opportunity Statement, Goal, Scope and Boundaries Capture & Analyze Voice of Customer Identify Critical Customer Requirements (CCRs) & Establish System Specifications via QFD 1 Identify Conceptual Design Determine System Functionality Map CCRs to System Functions via QFD2 Develop Detailed Design Map Functions to Design Parameters via QFD3 Design for Robust Performance Minimize Sensitivity to Design & Operating Variations Design for Manufacturability Minimize Sensitivity to Mfg Variations Predict Quality Predict Iterate to Meet Quality Target OK Test & Validate Assess Performance, Reliability, Mfg,... OK Deliver to Customer Typical DFSS Process (DIDOV) Not OK Not OK Optimize Design Statistical Design Understand & Control Variation Maximize Probability of Meeting Performance, Reliability & Manufacturability Goals Y = f (x) Source: Design for Six Sigma, K. Yang

14 Voice of the Customer and Quality Function Deployment Identify Target Direction Customer Needs Whats Voice of the Customer Surveys Focus groups Conjoint Analysis Customer Importance High Power Low Oper. Cost Meet Range Req'ts Long Life High Payload Easy To Maintain Easy To Troubleshoot Hows Thrust Tot Maint Cost Mission Fuel Burn Cycle Life Limits Weight Time To Remove Disk Burst Speed More is better Less is better Targeted amount Relationships Strong - 9 Medium - 3 Weak - 1 System Level Importance Lb $/Flt Hr Gal/Flt Flt Cycles Lb Minutes RPM Intense Focus on what the Customer wants

15 System Design Design System Locomotive Platform Engine Generator Inverter Traction Truck Subsystems Control Console Coupler Sander Requirements Flow Down Assemblies Parts Customers buy System Performance and Reliability Design Decisions are made at Subsystem, Assembly and Parts Systems engineering allows Flow Down of Customer Requirements to lower design levels Rational Design Decisions to achieve system-level goals

16 Design for Robust Performance Optimize Quantify relationships between CCRs & Design Parameters First principles models Numerical models (finite elements, lumped parameter, ) Designed experiments (DOE) QFD DOE Main Effects Plot Response Surface Fuel Economy Octane Level Air Temp X 2 Y = 50 Optimum Steep gradient, high sensitivity to variabilities in X s Regression to obtain Transfer Function: Y = f(x 1, X 2, X 3 ) a 0 + a 1 X 1 + a 2 X 2 + a 3 X 3 + a 4 X 1 X 2 + a 5 X 1 X 3 + a 6 X 2 X X 1 Main Effects 2-Way Interactions Capture knowledge in intransfer Function libraries & design templates

17 Statistical vs Deterministic Design: Switching Power Supply Example V in = Vac V o = 5 Vdc, +/- 5% Input Filter Isolated Switching Converter System Requirements: V in : V V o : 5 V, +/- 5% 6 quality Low cost Feedback V o Baseline design Isolated switching converter/ feedback section OPTO R 2 Low cost, combine power MOSFET and control circuit in a 3-pin package PWM IC R1 CTRL V ref I b R 1

18 Deterministic Design Analysis: Transfer function V o = V ref + R 2 V ref + I b R 1 ( ) Choose values for design parameters: Design Parameter Value LM 431I ref voltage, V ref (volts) R 1 (ohms) R 2 (ohms) Bias current, I b (amps) 5.0E-06 Substituting: Output voltage = 5.04 volts Baseline design meets 5V, 5V, +/- +/-5% performance requirement But, But, quality level is isnot notyet yetdetermined

19 Statistical Design Analysis: Transfer function Design parameters are statistical in nature. Engineer selects mean values and a measure of variability (e.g., standard deviation, based on component tolerance). V o = V ref + V ref R 2 ( R 1 + I b ) Design Parameter Mean Std Dev Tolerances Lower Upper LM 431I V ref (volts) R 1 (ohms) % 1% R 2 (ohms) % 1% Bias current, I b (amps) 5.0E E E E-06 Do a statistical analysis (e.g., Monte Carlo), using the Transfer Function and the statistical parameter values Results: V o mean V o std dev Defects/million 5.04 volts volts 188 (5.06 ) Probability Volts Baseline design meets 5V, +/- 5% performance But quality level is only 5

20 Statistical Design: Approaching 6 Design optimization analysis: Use transfer function to understand the shape of the response surface and the output voltage s sensitivity to each design parameter Reduce defect rate by shifting mean values or reducing variances of the design parameters Design Parameter Mean Std Dev Sensitivity LM 431I V ref (volts) R 1 (ohms) R 2 (ohms) Bias current, I b (amps) 5.0E E Design Mod 1: Center the distribution by increasing R 1 to ohms Results: V o mean V o std dev Defects/million 5.00 volts volts 20 (5.61 ) Probability Base Centered Volts

21 Statistical Design: Reaching 6 Design Mod 2: Mod 1 plus reduce variance by using 0.1% resistors Design Mod 3: Mod 2 plus LM 431AI MOSFET to reduce V ref variance Centered 0.1% Resistors Base 0.1% Resistors MOSFET Upgrade Probability Volts Probability Volts Summary Mean Std Dev DPMO Z ST Cost Baseline Design % Mod 1: Centered via R % Mod 2: 0.1% Resistors % Mod 3: LM 431AI ~ % Statistical design enables prediction of performance, quality and cost during the design process

22 What s Different About DFSS? Disciplined, comprehensive process applicable to all Designs Line of Sight from Customer Needs to all System Design levels Statistical design to understand... and reduce Variation New tools: QFD, Function Analysis, TRIZ, DOE, DFM, statistical tolerance, Robust Design, multi-variable optimizations Quality prediction throughout development Dedicated Team can develop a Breakthrough Design in months But, does not replace need for sound Engineering Judgment

23 Questions & Discussion

24 Appendix

25 Mapping of Common Tools to DFSS Stages I Voice of Customer Market Research & Brand Analysis QFD Kano Model Bench Marking Quality History: Surveys, Ratings, etc. Quality History: Warranty, etc. Quality Loss Function D O Concept Generation & Selection Numeric/ Heuristic Optimization Designed Experiment System & Functional Diagrams Parameter Design P- Diagram Toleranc e Design FMEA Analytical Reliability & Robustne ss (AR&R) Axiomatic Design Statistical Tolerance Robust Engineeri ng Design and R&R Checklist Dimension Variation Analysis Process Capability Gage R&R Control Plan FDVS: Target Setting & Verification DFSS Scorecard V Design Verification Plan & Report Robustness/Reliability Demonstration