1、1,APS Training Module QF 6 Sigma,Autoliv 14-April-2006 QF-TM-004-C2.0,2,3,Why this training,During this training you will learn: What 6 Sigma is A measurement of quality A problem solving method Examples of tools in 6 Sigma,4,What is 6 Sigma,Sigma is the Greek letter used to represent variationA sig
2、ma level is a measurement of quality A 6 Sigma level process creates close to no failures A sigma level is a measurement of how well the process meets customer specifications6 Sigma is a problem solving method Customer centered Systematic Data driven,5,A Measurement of Quality,6,2s 308,5373s 66,8074
3、s 6,2105s 2336s 3.4,A Measurement of Quality,A 6 Sigma process has only about 3.4 ppm!,Sigma Level,PPM,7,A Measurement of Quality,Out of 365 Rounds per Year,2s 6 missed putts per round,3 1 missed putt per round,4 1 missed putt every 9 rounds,5 1 missed putt in 1.5 years,6 1 missed putt in 45 years,8
4、,A Problem Solving Method,9,APS Toolbox,The APS Toolbox is expanded with the 6 Sigma Tools to solve more complex problems with increased efficiency,10,Customer Centered,Who is the customer?What is critical to the quality (CTQ) for the customer?How are the customers expectations not being met?,Action
5、s based on data are the key to Customer Satisfaction,11,Systematic,6 Sigma follows a roadmap To link tools more powerfully together To guide the team through the problem,12,The DMAIC Method,Define project or problem,Measure current situation Process behavior charts Capability analysis Pareto charts,
6、Create Cause & Effect Matrix,Select corrective actions,Document Results,Develop Project Plan Select team Business case SMART objective Project scope Key metric Operational definition,Conduct Measurement System Analysis (MSA),Create VMEA,Implement Corrective actions,Implement control method Control p
7、lan Standards Training matrix SWI Audit SPC,Create Process Map or Fishbone Diagram,Root cause analysis process VMEA Actions Testing potential root causes OFAT Tests DOE Tests,Verify corrective actions Capability analysis Process behavior charts OFAT Tests DOE Confirmatory,Document Lessons Learned,De
8、termine customers CTQs,Optimize Process DOE Tests Response Optimizer Response surface models (RSM),Confirm root cause Reproduce failure OFAT Tests DOE Tests,13,Systematic,6 Sigma is about reducing variation Understand the system Y=f(x) Determine to dominate source of variation Control the dominate s
9、ource,14,The DMAIC Method,Process Map Collects all known input and output variables Cause and Effect Matrix Prioritizes the variables most likely to have a major impact VMEA Studies how selected variables can cause the process to failStatistical tools are then used to quantify the relationship betwe
10、en these selected input and output variables,15,Data Driven,Key Concepts“Variable always trumps categorical” Variable data is more powerful than categorical because it contains more information “Data is innocent until proven guilty” Variation in a process is considered to be random until evidence sh
11、ows otherwise “A graph is worth a thousand data” The correct graph or statistical tool can objectively make conclusions much better than raw or tabulated data,16,Data Driven,Attempting to control variation without understanding and quantifying it, simply adds one more source to the problem,17,Data D
12、riven,Based on the data, what is this process doing? Getting better Getting worse Staying the sameWhat will the % scrap be in three months?,18,Data Driven,Based on the chart, what is this process doing? Getting better Getting worse Staying the sameWhat will the % scrap be in three months?,19,Data Dr
13、iven,Based on the chart, what is this process doing? Getting better Getting worse Staying the sameWhat will the % scrap be in three months?,20,Measurement System Analysis Gage R&R,Objective Evaluates whether a measuring system for variable data is adequate Reports Repeatability Amount of variation d
14、ue to gage and method Reproducibility Amount of variation due to changing operators % Study Variation Percent of observed variation that is due to the measurement system % Tolerance Compares the measurement system to the tolerance,Study Var %Study Var %Tolerance Source StdDev (SD) (5.15 * SD) (%SV)
15、(SV/Toler) Total Gage R&R 0.0089286 0.045982 19.37 9.20Repeatability 0.0082886 0.042686 17.98 8.54Reproducibility 0.0033197 0.017096 7.20 3.42Appraiser 0.0033197 0.017096 7.20 3.42 Part-To-Part 0.0452258 0.232913 98.11 46.58 Total Variation 0.0460987 0.237408 100.00 47.48Number of Distinct Categorie
16、s = 7,21,Measurement System Analysis Kappa Test,Objective Evaluates whether a measuring system for categorical data is adequate Reports Kappa Score .7 and .9 is excellent Percent Agreement and Confidence Intervals,Fleiss Kappa StatisticsResponse Kappa SE Kappa Z P(vs 0) Good 0.769656 0.0632456 12.16
17、93 0.0000 Short 0.769656 0.0632456 12.1693 0.0000,Fleiss Kappa StatisticsResponse Kappa SE Kappa Z P(vs 0) Good 0.253472 0.0632456 4.00775 0.0000 Short 0.253472 0.0632456 4.00775 0.0000,Before,After,22,Data Driven Example 1,Youre the plant manager You are reviewing an areas scrap rate What should yo
18、u do?,Month Scrap Rate,Give area a plaque for an all-time low in scrap!,Wished you had the plaque back.,Thats better!,Five months of steadily increasing scrap!,23,Data Driven Example 1,Process behavior charts Identify special cause variation from common cause variationCommon Cause The inherent rando
19、m variation caused by many inputs Special Cause The non-random changes in a process usually caused by one input,All changes in scrap rate were due to common cause variation!,24,Data Driven Example 2,Youre the production manager A machine supplier offers you a new machine to reduce cycle-time The new
20、 machine is $50,000 A reduction of 2 seconds in cycle-time could save $100,000 over a year The following data is given by the machine supplier,Machine Cycle Times Old New 8.7 9.4 19.9 13.9 6.3 5.9 21.2 16.8 12.3 11.813.68 11.56,Averages,25,Data Driven Example 2,T Test Distinguishes if differences be
21、tween samples are significant or simply by random chance Interval Plot Displays the confidence intervals around each sample mean and visually represents the t test evaluationOther Statistical Tests ANOVA Regression Chi-Squared Proportion Test of Equal Variance,T Test Results: P-Value = 0.284 Conclus
22、ion: The difference between machines is by random chance.,26,Data Driven Example 3,A team has implemented improvements that they believe will reduce the variability by at least 20% Management wants less than 5% chance of missing the potential improvement Management also wants less than 5% chance of
23、implementing the changes if the improvement is not real,27,Data Driven Example 3,Sample size evaluation Determines the samples needed to see changes based on: Risk The probability of making a wrong decision Difference The magnitude of the change to be tested compare to the normal variation,Samples f
24、or 20% reduction at 95% confidence,220 samples before and after changes!,28,Data Driven Example 4,The web sensitivity for a seatbelt model has too much variation The team needs to know where to set and how to control these component dimensions: Inertial mass weight Bearing height Spring rate Lever l
25、ength,29,Data Driven Example 4,Which factors have an affect?What is the optimal condition?Remember Reducing component tolerances costs money,Data are fictional values,30,Data Driven Example 4,DOE (Design of Experiments) A systematic set of tests that efficiently determine the effect of each input on
26、 the outputsInteraction How one input changes the effect of another input on the output (not seen in single factor tests)Results These results can be used to determine the settings of the inputs to optimize the output,Data are fictional values,31,Control,Control is the last step of DMAIC that ensure
27、s the improvements are continued Tools in Control SPC Poke Yoke Lessons LearnedThe idea is to continually improve,32,Design Engineering & Development,Support Services,Processing & Manufacturing,Shipping & Receiving,Product Use & Customer Experience,Where Does 6 Sigma Apply?,33,Process Improvement,34
28、,Summary,6 Sigma Insures all work centers on customer requirements Uses systematic methodology to solve problems Requires actions based on data and analysis Allows for efficient solutions to complex problems 6 Sigma is an additional tool in the APS Toolbox,35,Why this training,Can you answer these questions: How many PPM correspond to a 6 Sigma process? What does DMAIC stand for? What are some of the tools used to be data driven?,36,References,“What is Six Sigma” -Promontory Management Group (c) 2001 Promontory Management Group, Inc.,
