Report

DMAIC project template Roberto lopez LSSMBB [email protected] 6 sigma steps (DMAIC) • • • • • Define Measure Analyze Improve Control • • • • • Define your business problem Measure your process (Y) performance Find the root causes (X’s) of the problem Improve, implement new solution Deliver Y performance over time Define Business Problem 1 Pb Statement Description of the problem (what? where? since when? based on which data? Based on which facts?) Voice Of Customer (Customer perception) Internal perception Competitors benchmark 2 Goal Statement Develop Goal statement/opportunity or objective in clear, concise, measurable terms Demonstrate alignment of the project with the company/business strategy Define Business Case 3 Business Case Why to do the project? What if project not done? Financial Business Case What Comment Cost K€ Project Labor Project Materials Project Contractors Equipment investment Yearly operation cost TOTAL Costs Incremental Revenues Labor Savings Material Savings TOTAL Benefits (1 year) ROI 0 0 0 Define Project charter 4 Project scope Project: out of scope Project: in scope Green Belt/Black Belt role: Team members/role: 5 Team Manager: Key stakeholders from Business team: Customers: Others: Black Belt: Master Black Belt: VP Quality / Quality Manager: 6 Planning D Dates: Gantt M A I C Process map & CTQ’s Define 7 (SIPOC) process as felt by the customer Supplier = My Company Input Step 1 Step 2 Step 3 Output My Customer measurement CTQ* Definition Specification limit Performance targets *Critical To Quality Measure M-step 1: Selected CTQs CTQ details Selected Project CTQ(s): Discrete/continuous CTQ? CTQ Operational Definition(s): What is it (measurement starting point and end point) ? How to measure it? M-step 2: Define Performance standard What is a defect? If applicable to your CTQ, LSL* contractual value or limit defined by customer: If applicable to your CTQ USL* contractual value or limit defined by customer: *LSL Lower Specification Limit *USL Upper Specification Limit Measure M-step 3: Measurement System Analysis Gauge R&R Measurement system analysis Data Collection plan for Repeatability & Reproductibility study: Data sample size = Date of data collection: Where data have been collected? How data have been collected? How many operators? How many repeats per operator? Gage R&R study – Continuous data: -Tolerance (USL-LSL): -Short method (yes/no): -Gage R&R as %contribution: Gage R&R as %tolerance: Nb of distinct categories: Gage R&R study – discrete data: If we have a Known sample (a reference), then how was built the known sample? Conclusion/decision taken (evidence that the measurement of our CTQ is reliable): Measure M-step 4: Baseline current (As Is) process capability CTQ data collection plan CTQ data collection plan •Data sample size = •Date of data collection: •Where data have been collected? •How data have been collected? •Will you take the opportunity to collect obvious X’s when collecting your CTQ measurements? •If yes, which X’s: •Display an extract of the data collected: Measure M-step 4: Baseline current (As Is) process capability Photography of current performance Continuous CTQ •Sample size= •Short term data or Long term data? •Histogram with Normality test result and descriptive statistics: •Type of the distribution (normal, other): •Trend represented by its (m, median, Q1, Q3, other): •Variability represented by its (s, P95, Q1/Q3, other): PROCESS CAPABILITY MEASURE: NORMAL Distribution Observed DPMO= Potential DPMO= ZLT = ZST (ZLT+1.5)= Conclusion •Problem with trend, variability or both? Non NORMAL Distribution Observed DPMO= Equivalent ZLT= Equivalent ZST (ZLT+1.5)= Measure M-step 4: Baseline current (As Is) process capability Photography of current performance Discrete CTQ •Sample size: •Short term data or Long term data? •Process capability •Yield (%)= •%defects= •DPMO= •Opportunities per unit= •DPU= •Equivalent ZLT= •Equivalent ZST (ZLT+1.5)= •95% confidence interval on DPMO: Analyze A-step 5: As Is process Graph Analysis Graph analysis Continuous CTQ •Normal probability plot: Conclusion (can we detect a true process and an unstable process?): All type of CTQ •Stability over time/Run chart: Conclusion: •Pb of technology (short term common causes)? •Pb of control (long term shift due to special causes)? •Both pbs? Analyze A-step 6: Define Performance Objectives of To Be process Set expected performance Improvement Goal Target (from benchmark) for: Trend: Variability: LT DPMO or LT %defects= Yield (%)= ZLT= ZST=(ZLT+1.5)= Gap (Improvement – Actual) DPMO gap or Defect reduction factor: Z gap= 6 Sigma tools necessary to meet performance objectives: • Ground fruit: Logic and Intuition • Low hanging fruit: Basic tools (Process Map, CE/CNX fishbone diagram, SOP, FMEA, Pareto, Histogram, Box plot) • Bulk of Fruit Process Characterization and Optimization • Sweet Fruit: Transform the project into a DFSS project Analyze A-step 7: Identify X’s (root causes) Process analysis Process (AS IS) detail analysis: •Detailed process map Conclusion •CTQ Intermediate distribution plot (continuous CTQ only) Conclusion Standard Operation Procedures analysis: •Exist: •Are applied: •Incitation to apply them exist: •Training has been given to operators: Policy analysis: •Exist •Are applied Analyze A-step 7: Identify X’s (root causes) Potential root causes List of potential significant X’s Label C/N/X on your fishbone, then rate every X impact vs implementation (table below) Fishbone diagram of my Y Measurements Materials Men Y/CTQ Output Impact Environment Methods Machines High Low Implementation Easy 1 2 Hard 3 4 Analyze A-step 7: Identify X’s (root causes) Segmentation analysis Results of segmentation (continuous CTQ only) : (Box plots, Scatter plots, Distributions per X) Conclusion: Analyze A-step 7: Identify X’s (root causes) Pareto analysis Pareto chart of the CTQ categories: Conclusion: Pareto chart of the defect categories: Conclusion: FMEA: Conclusion: Analyze A-step 7: Identify X’s (root causes) Hypothesis testing Test Statistical Tests Performed X Factors p value one sample or two sample T-test Annova one way Homogeneity of variance Chi square test Simple regression Non parametric data test ANOVA 2 ways GLM DOE screening Conclusion/Prioritized list of significant X’s: Financial benefits confirmation (Update your Business case): %contribution conclusion Improve I-step 8: Screen potential causes. List improvement actions/projects Improvement actions List of Vital few Xs (screening DOE): Improvement action plan (Who What When): Improve I-step 9: Discover Variable relationships (if necessary) Improved performance Transfer function: Proposed Solution: Optimal Settings Confirmation runs capability: I13 I-step 10: Establish Operating Tolerances (if necessary) X tolerance settings: Control C-step 11: Define and Validate Measurement System on X’s (if necessary) Gage R&R on X’s Control on one X (yes/no): If yes, X measurement system analysis: •Data Collection plan for Repeatability & Reproductibility study: Data sample size = Date of data collection: Where data have been collected? How data have been collected? How many operators? How many repeats per operator? •Gage R&R study – Continuous data: -Tolerance (USL-LSL): -Short method (yes/no): -Gage R&R as %contribution: categories: Gage R&R as %tolerance: Nb of distinct •Gage R&R study – discrete data: If we have a Known sample (a reference), then how was built the known sample? •Conclusion/decision taken (evidence that the measurement of our X is reliable): Control C-step 12: Determine new Process Capability Confirmation of improved performance Process or Product new capability: Continuous CTQ: NORMAL Distribution Observed DPMO= Potential DPMO= ZLT = ZST (ZLT+1.5)= Non NORMAL Distribution Observed DPMO= Equivalent ZLT= Equivalent ZST (ZLT+1.5)= Discrete CTQ: •Sample size:= •Short term data or Long term data? •Process capability •Yield (%)= •%defects= •DPMO= •Opportunities per unit= •DPU= •Equivalent ZLT= •Equivalent ZST (ZLT+1.5)= •95% confidence interval on DPMO Statistical Confirmation of Improvement (by statistical test): Control C-step 13: Implement Process Control Process controls Control Charts (SPC) in place/review mechanisms: Mistake Proofing action taken: FMEA control actions in place: