Measuring KCs using Scorecards Practical Application Ian James Chief Engineer Electronics & Software Technology Aero Engine Controls February 2009 A Rolls-Royce and Goodrich Corporation joint venture Measuring KCs Using Scorecards Contents of Lecture Introduction Customer Expectations What is a Scorecard ? How to Compile A DFSS Scorecard A Little More About Goodrich Engine Controls EEC Scorecard Example Summary 1 Customer Expectations The Customer Expects … Supplier will design and manufacture a product that …. Meets all requirements defined by design specifications Demonstrates Six Sigma Levels of quality The following list summarizes the bounds of customer expectation for 6σ levels of quality Safety Expected Operation (performance) Life / Reliability Maintainability (Cost of Ownership) Cost / Weight Producible / Cost Effective (Rework) Customer Expectations Key Characteristics [Customer View] Key Characteristics (KCs) may be provided by Customer Critical To Quality (CTQs) characteristics Critical Characteristics Major Characteristics Characteristics can be identified via … Quality Function Deployment (QFD) Failure Mode and Effects Analysis (FMEA) Historical Quality Data 2 Customer Expectations Key Characteristics [Supplier View] Further Key Characteristics may be defined ‘internally’ Critical To Cost (CTC) characteristics Key Control Characteristics (KCC) Additional QFD and FME(C)A may be conducted jointly by both parties to agree a definitive list of KCs prior to project implementation implementation. Customer Expectations Capability Metrics Customer expects all Key Characteristics 6σ by E.I.S. (Cp=2.0, Cpk=1.5, DPMO=3.4, Zst = 6.0) Supplier predicts capability level for each KC at E.I.S. Prediction based upon similar equipment and processes Combination of Data Analysis, Simulation and Modelling If predictions do not indicate 6σ levels by E.I.S. then a jointly developed plan will be created to drive the levels to 6σ. 3 Customer Expectations Monitoring Capability for Key Characteristics The DFSS Scorecard is a tool that can be used to assess product & process performance to establish if customer requirements are met at all stages of the design process. The Scorecard can be used to Locate areas for improvement Recognise key missing issues & problem areas Improve & optimise the design and manufacturing processes The Scorecard is a live document that should be updated and reviewed at each phase of the project. Customer Expectations Monitoring Capability for Key Characteristics A DFSS Scorecard may be enforced by a customer to facilitate the monitoring of predicted capability The scorecard is a live document and will be updated as a minimum for presentation at: Proposal Submittal Preliminary Function Review (PFR) Preliminary Design Review (PDR) Detailed Design Review (DDR) Pre-Production Manufacturing and Quality Review (PM&QR) Production Readiness Review (PRR) 4 What Is A Scorecard ? The Scorecard lists all the Key Characteristics of a design For each Key Characteristic … Identify the ‘Specification Limits’ or ‘Targets’ Identify the Process Behaviour i.e. Mean, Variation, Defects As a result….. knowledge can be gained about: Comparison to determine whether CUSTOMER EXPECTATIONS Robustness of product (at all levels) are being met Manufacturing process performance What Is A Scorecard ? The DFSS Scorecard Consists of Two Parts Product & Sub-assembly / Component level level. DFSS Scorecard The DFSS Scorecard consists of two parts: Product Scorecard Subassembly/ Component Scorecard Assesses the overall robustness of the design and its ability to meet customer expectations – estimate the defects experienced by customer Assesses the quality of all components and sub-assemblies (& subsequently the processes) used in the design – estimates internal/supplier defect rate (Customer Perspective) (Internal Perspective) 5 What Is A Scorecard ? How To Compile a DFSS Scorecard Determine the Key Characteristics 1 • Information generated at all stages of this process are recorded on the DFSS Scorecard • The same process is used to compile both parts of the scorecard Perform Measurements Systems Analysis on the Key Characteristics For each Key Characteristic, identify the target and specification limits 2 3 Collect ‘performance d t ’ for data’ f each h Key K Characteristic Analyse data to establish if the performance of each Key Characteristic is acceptable 4 5 How To Compile A DFSS Scorecard Step 1: Determine the Key Characteristics For the Product Scorecard Tips - Ways to identify Key Characteristics - Customer Requirements documents Obtain Customer Requirements Document - Brainstorm by project team Use these requirements to define KCs - Use list of Key Characteristics on similar products Assign metrics and units For the Component Scorecard Note : The list of Key Characteristics will grow as the project progresses & better understanding is gained about product performance Identify all the subassemblies, functions and Components that may impact the Product KCs Accordingly, derive the Key Characteristics for each of these subassemblies and components. Assign metrics and units. 6 How To Compile A DFSS Scorecard Example : Identifying Key Characteristics The Design g Team developing p g a flight g control actuator for use on the A380 used the following method to identify the key characteristics for their product. Since the team had no historic data to draw from, they decided to hold a brainstorming session. The Brainstorming Session was structured to focus on the key characteristics of the overall product and those related to the top level subassemblies How To Compile A DFSS Scorecard Example : Identifying Key Characteristics (Cont’d) Key Output Characteristics Brainstormed Key Input Characteristics Brainstormed (Customer Requirements Document used as a starting point.) At each level, the team • Brainstormed all the key output characteristics • Brainstormed all the key input characteristics • Developed an input-output matrix t i and d ranked k d th the correlation between the inputs and the outputs using 1, 3, 9 or none • Used the matrix to prioritise the key input parameters Numbers used to ensure all key input characteristics have been identified and prioritise them 7 How To Compile A DFSS Scorecard Step 2: Perform measurement system analysis BEFORE gathering performance data for each KC Ensure that the performance of the measurement system that will be used to collect the data is acceptable Define All assumptions Step 3: Determine Target & Specification Limits May be straightforward …. From F customer t requirements i t Previous similar products May require some additional analysis / estimation Tolerance apportionment at the System Level Justification by reasoned argument How To Compile A DFSS Scorecard Example : Define Metric and Specification Limits for Key Characteristic For the A380 Flight Control Actuator, the mechanical balance of the rotor was identified as a Key Characteristic during brainstorm activities. Key Characteristic: Mechanical Balance of Rotor Gage name: Date of study : Reported by : Tolerance: Misc: Gauge R&R Study of Balancing Machine 1. Amount of Unbalance in the NonDrive End of the Rotor (Plane 1) Ralph August 20001 Rakhee Kar 4 gmm Xbar Chart Sample Mean 4 2. Amount of Unbalance in the Drive End of the Rotor ((Plane 2)) 1 2 3 4 5 6 By Part 7 8 9 4 10 3 3 3.0SL=2.902 X=2.710 -3.0SL=2.517 2 2 1 1 2 3 4 5 6 7 8 9 10 1 2 3 Part 1 2 3 4 5 4 5 6 7 8 9 10 Part R Chart Ch 0.8 6 C Components off Variation V i i 7 8 9 10 %Total Var 100 %Study Var 0.7 3.0SL=0.4840 0.4 0.3 0.2 Percent Sample Range %Toler METRIC 0.6 0.5 50 R=0.1880 0.1 -3.0SL=0.00E+00 0.0 0 1 2 3 4 5 6 7 8 9 10 Gage R&R Part-to-Part Part The Amount of Unbalance should be less than 4 gmm in each plane MEASUREMENT SYSTEM: Balancing Machine MEASUREMENT SYSTEM ANALYSIS confirms that measurement system is acceptable SPECIFICATION LIMIT 8 How To Compile A DFSS Scorecard Step 4: Collect Data for each Key Characteristic Data can be obtained from f om man many sources. so ces Prototype Test Data, Simulation, Analysis, SPC Customer Complaints, Audit Results, Warranty issues Historic information for similar products, components and processes Step 5: Assess the performance of each Key Characteristic Check the data for Normality Determine Mean and Standard Deviation Compile the Scorecard using pre-designed spreadsheet How To Compile A DFSS Scorecard Example : For Each Key Characteristic Collate Data In a Spreadsheet How variable is the critical parameter ? The solution assumes that data is normally distributed Sigma Metric How centered is it ? Expected Defect Rate DFSS SCORE CARD Customer Requirement Level Key Characteristic List of Critical to Quality Requirements for the product, subassemblies & components Unit Data Type Expected Performance for the critical parameter Target Performance USL LSL Mean Information that must be entered using current performance data for the critical parameter Standard Deviation Z Score (ST) Cp Cpk Expected Long Term Performance (PPM) These are the ‘Performance Indicators’ and show how well the critical requirements are being met. (Values automatically generated by the spreadsheet.) 9 How To Compile A DFSS Scorecard Example : Using Performance Indicators To Drive Design Improvements Below is an extract of the DFSS Scorecard for Product X Generator. DFSS SCORE CARD Customer Requirement Level Target Our Performance Key Characteristic Unit Data Type USL Product Vibration at 15150 RPM (Load) mm Variable - 0.434 Product Vibration at 171871 RPM (Off Load) mm Variable - 0.434 LSL SIX SIGMA Project raised for Product X Vibration Mean Standard Deviation Z Score (ST) Cp Cpk Expected Long Term Performance (PPM) - 0.24 0.05 5.38 1.45 1.29 53.0 - 0.328 0.07 3.014 1.03 0.505 648978.1 Scorecard shows the vibration performance of the generator is unacceptable at 171871 RPM (Off Load) How To Compile A DFSS Scorecard Example : Using Performance Indicators To Drive Design Improvements For Those of You who Miss the Maths….. Vibration At 15,150 RPM Vibration At 171,871 RPM LSL = 0, USL = 0.434 LSL = 0, USL = 0.434 Mean, X = 0.328, Standard Deviation, σ = 0.07 Mean, X = 0.24, Standard Deviation, σ = 0.05 Cp = USL - LSL = 0.434 – 0 6σ 6 x 0.07 Cp = USL - LSL = 0.434 – 0 6σ 6 x 0.05 Cp = 0.434 = 1.033 0.42 Cp = 0.434 = 1.446 0.35 Cpk = Min X – LSL , USL – X 3σ 3σ Cpk = Min X – LSL , USL – X 3σ 3σ Cpk = Min 0.328 - 0 , 0.434 – .328 3 x .07 3 x .07 Cpk = Min 0.24 - 0 , 0.434 – .24 3 x .05 3 x .05 Cpk = Min 1.5619 , 0.5047 Cpk = Min 1.6 , 1.2933 Cpk = 0.505 Cpk = 1.293 10 How To Compile A DFSS Scorecard Example : Using Data To Refine The List Of Key Characteristics The data collected was used to investigate whether the amount of rotor run out had a significant influence on the Vibration of the Generator. Regression plot for Vibration Level versus Runout Y = 3.59E-02 + 0.736884X R-Sq = 36.2 % Regression Analysis shows that there is no statistically significant correlation between Vibration and Rotor Run out. Therefore, Rotor Run out out’ was ‘Rotor eliminated from the list of Key Characteristics. 0.6 0.5 Vibration n Level The concentricity of the rotor/stator was identified as a key input characteristic for the Product X Generator Generator. It was thought have an influence on the Vibration of the generator – a key output characteristic. As a result it was decided to collect data for the amount of rotor run out on all production units. 0.4 03 0.3 0.2 0.1 Regression 0.0 95% CI 95% PI -0.1 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 Rotor Runout A Little More About EECs Engine Control System Rolls-Royce Trent 700 Engine 11 A Little More About EECs Full Authority Digital Electronic Control (FADEC) A Little More About EECs VSVA Main Engine Pump VSVC FMU EEC, PCU and OPU 12 A Little More About EECs Computer Systems - Personal Computer [P.C] A Little More About EECs Computer Systems - Personal Computer [P.C] Flexible…..User Friendly…..Multi-Application…..Expandable Operating (System) Software MODEM Application Software Monitor Scanner Microprocessor and Support Circuitry Printer Mouse Keyboard Operating Software Device Drivers Standard Interfaces (e.g. USB, PCMCIA) Manages the interface between the computer hardware and its peripherals and is also the visual user interface. e.g. Windows, UNIX. Device Drivers are used to configure new peripherals into the system. Sound Card Application software CD Burner Harnesses the resources of computer system to fulfil practical ends. e.g. Word Processing, Accounting, Games, ….. Limited only by the power of the Computer system, e.g. Speed, Memory (and the imagination of the software house !) 13 A Little More About EECs Computer Systems - Electronic Engine Controller A Little More About EECs Computer Systems - Electronic Engine Controller Embedded…..Application Specific…..Limited Modification Capability Key Characteristic #1 Speed Probe KC Thermistor KC Resolver KC Switch KC Application Software Microprocessor and Support Circuitry Operating Software Speed Measurement Accuracy 0.01% of point over range X to Y KC Solenoid KC Torque Motor KC Lamp KC Spare Key Characteristic #2 Solenoid Drive Accuracy y +/- 0.5V over a range X to Y Key Characteristic #3 Torque Motor Drive Accuracy +/- 1mA over a range X to Y 14 A Little More About EECs Production Acceptance Testing (P.A.T.) Every Product will be subjected to a series of Tests For an E.E.C. the number of tests may exceed 5000 A computer-based piece of test equipment records all results Pass / Fail for Production Pass-Off Purposes A Range of products are likely contain Generic Circuit Blocks e.g. Speed Measurement, Solenoid Drivers, Torque Motor Analysis of results from existing product Test can be useful Indicate area of design marginality Provide performance data for new design KCs Determine capability of test equipment EEC Test Results are ideal for measuring system capability EEC Scorecard Example Example - EEC Scorecard Analysis Process Define KCs of new design Evolve hardware requirements from KCs (e.g use of FMEA, QFD) Select existing functional circuit blocks which meet requirements From PAT schedule determine test step for each function selected Review applicability of data for direct comparison with relevant KCs Apply and record relevant assumptions to enable comparison Extract PAT data for each relevant test step Maximise data: Use ALL tests and ALL test equipment Carry out Capability Analysis for each Key Characteristic Record results in Scorecard Feedback in the case where analysis shows KCs cannot be satisfied: Inappropriate Analysis or Ineffective Design ? 15 EEC Scorecard Example Example - EEC Scorecard Analysis Process Key characteristic are, in fact, hardware requirements Data available for Ambient, Hot and Cold PAT test Testing carried out on a selection of (identical) test rigs Typically 10 tests per test point For each test point, max, min and average recorded in test results Max, Min and Average to fall between limits for a pass EEC Scorecard Example Example - EEC Scorecard Analysis Process Torque Motor Drive Interface Existing Design – Product X Torque motor drive output is tested at +40mA and -40mA Accuracy limits +/- 1mA Current is measured using internal current monitor Key Characteristic Torque Motor Driver Output accuracy - +/- 1mA Capability Overview Combined accuracy of driver and monitor Sample of 332 Data Points – @ +40mA: Mean = 0.346, StDev(LT) = 0.129 and Cpk = 2.84 – @ -40mA: Mean = 0.385, StDev(LT) = 0.151 and Cpk = 2.3 16 EEC Scorecard Example Normal Probability Plot Normal Probability Plot @ +40mA .999 .99 Probability .95 .80 .50 .20 .05 .01 .001 5.98 5.99 6.00 6.01 6v Excite Average: 6.00006 StDev: 0.0066425 N: 332 Anderson-Darling Normality Test A-Squared: 0.458 P-Value: 0.263 EEC Scorecard Example Capability Analysis Process Capability Analysis for Error for +4 @ +40mA LSL Process Data USL Target LSL Mean Sample N USL ST LT 1.00000 * -1.00000 0.34608 332 StDev (ST) 0.076725 StDev (LT) 0.129089 Potential (ST) Capability Cp 4.34 CPU 2.84 CPL 5.85 Cpk 2.84 * Cpm Overall (LT) Capability -1.0 -0.5 0.0 Observed Performance 0.5 Expected ST Performance 1.0 Expected LT Performance Pp 2.58 PPM < LSL 0.00 PPM < LSL 0.00 PPM < LSL 0.00 PPU 1.69 PPM > USL 0.00 PPM > USL 0.00 PPM > USL 0.20 PPL 3.48 PPM Total 0.00 PPM Total 0.00 PPM Total 0.20 Ppk 1.69 17 EEC Scorecard Example Capability Analysis Process Capability Analysis for Error for -4 @ -40mA LSL Process Data USL USL ST LT 1.00000 Target LSL * -1.00000 Mean 0.38525 Sample N 332 StDev (ST) 0.089271 StDev (LT) 0.151432 Potential (ST) Capability p Cp 3.73 CPU 2.30 CPL 5.17 Cpk 2.30 Cpm * Overall (LT) Capability -1.0 -0.5 0.0 Observed Performance 0.5 Expected ST Performance 1.0 Expected LT Performance Pp 2.20 PPM < LSL 0.00 PPM < LSL 0.00 PPM < LSL 0.00 PPU 1.35 PPM > USL 0.00 PPM > USL 0.00 PPM > USL 24.58 PPL 3.05 PPM Total 0.00 PPM Total 0.00 PPM Total 24.58 Ppk 1.35 Summary Summary Customers Expect ….. 6σ quality levels as well as functional & performance compliance A means of tracking this data …. and a part in the process Scorecards provide a monitor for KCs throughout design process KCs can be defined by the customer, supplier or both Typical 6σ metrics used to gauge performance (Cp, Cpk, Zst, DPMO) Scorecard Makes all design assumptions and targets explicit Collation process provides the means to refine KCs Locate areas for improvement in design / manufacturing process Provides insight into prior products as well as new design 18 DFSS Definitions Characteristic Those dimensional dimensional, visual, visual functional and material features or properties which describe and constitute the design of the item and can be measured, observed or identified to determine conformance to the design requirements. Key Characteristic (KC) The Select few, measurable features of a specific part, drawing, specification or process whose variation can significantly impact customer satisfaction, manufacturability, durability or performance. These characteristics include, but are not limited to the following: CTQs, Critical Characteristics, Major Characteristics, CTCs and KCCs DFSS Definitions (Con’t) Critical To Quality (CTQ) Characteristic The select few few, measurable key characteristics of a specific part/drawing/specification where reduced variation will have a positive impact on the customer. Critical Characteristic Those characteristics of an item which, if nonconforming, may result in hazardous or unsafe conditions for personnel using maintaining or depending of the unit-of-product; or which may prevent or seriously affect the satisfactory operation or function of the unit-of-product. The applicability of this definition should be evaluated at a point 50% beyond the specification limits. 19 DFSS Definitions (Con’t) Major Characteristic Those characteristics of an item item, other than critical critical, which which, if nonconforming, may result in operational or functional failure of the item or which materially reduce the usability, physical or functional interchangeability or durability of the unit-of-product for its intended purpose. The applicability of this definition should be evaluated at a point 50% beyond the specification limit(s). Critical to Cost (CTC) Characteristic Those characteristics which drive manufacturing losses upward. They are usually characteristics that have repetitive MRB activity, excessive rework loops, low process yields, high scrap, etc. DFSS Definitions (Con’t) Key Control Characteristic (KCC) A specific characteristic for a specific part or process that will be monitored by SPC or another appropriate monitoring method, in order to assure conforming CTQs, KQCs (Key Quality Characteristics) or CTC. It may be a process parameter or an output characteristic. Quality Function Deployment (QFD) A communication tool to translate customer expectations / requirements into specific design and manufacturing requirements. 20