【正文】 tool used is called the MultiVari chart. Measurement Time or Condition Change Overtime or Condition Within Piece Schematic of a Multivari Chart Between Pieces 9 Why Is It Useful? When Is it used ?Multivari analysis can help you narrow a list of potential causes to a much smaller list suitable for detailed investigation, through designed experiments, for example. It helps you focus your attention as you funnel down to find potential variables that you may need to study. ?In the Analyze step to plement process analysis and stratification. It usually follows stratification analysis. 10 When Should It Be Used? ?In the Analyze step to plement process analysis and stratification. It usually follows stratification analysis. 11 Planning a Multivari Study: Overview A. Preliminary work B. Main work Choose product characteristics relevant to the problem (from Define or Measure steps) Develop list of potential factors (causes) of variation in the characteristic of interest 7. Analyze the results 1. Define scope or domain of the study 2. Define time profile of the sample selection 3. Make a detailed plan for collection of the product samples 4. Collect the samples 5. Measure the samples 6. Plot the sample data 12 Families (Components) of Variation Typical families of variation in a manufacturing environment can be clustered together: Unit families: ?Within Unit variation: different measurements or results at different points within the unit or product ?UnittoUnit: one unit to the next (within subgroup variation) ?Subgroup to subgroup ?Streamtostream: suppliertosupplier, linetoline, machinetomachine, tooltotool, operatortooperator, etc. 13 Families (Components) of Variation, cont. Time families ?Shortterm to Longterm Measurement family ? Gage Ramp。R 14 Exercise: Possible Applications of MultiVari Charts in Your Work List the possible sources of variation for a work situation where you are interested in reducing variation Situation: Possible Sources: 15 Exercise 2: MultiVari Display B a ck F r o n t0 . 0 50 . 0 40 . 0 30 . 0 20 . 0 1 Finish Thickness D a y 1 D a y 2AM PM AM PME x a m p l e o f M u l t i v a r i d i sp l a y2 3 1 Panel 1 2 3 1 2 3 1 2 3 1 2 3 Sample, 3 consecutive panels taken at one time. This multivari chart shows four families of variation. We plot measures of film thickness on front and back sides of panels produced on a paint line. 1. Withinpanel variation 2. Unittounit variation 3. Morningtoafternoon variation 4. Daytoday variation 16 Exercise 2: MultiVari Display What can you say about: ?Within panel variation? ?Panel to panel variation? ?AM/PM variation? ?Day to day variation? 17 Exercise 2: MultiVari Display, Answer 1. The difference in film thickness within units is roughly constant across all sampling periods (unit to unit, morning to afternoon and day to day) and neither face dominates. 2. The panel to panel variation is about the same across time of day and days. 3. On each day, the group of samples is relatively higher in the morning than in the afternoon. 4. The range of values on day 1 is about the same as on day 2. 18 Exercise 3: Interpreting a MultiVari Chart Continuous X’s This data was collected on a Folio cutter. It is desirable to reduce variation in cut length. Different speeds and braking methods were run. 2 4 . 9252 5 . 12 5 . 22 5 . 32 5 . 42 5 . 52 5 . 62 5 . 7brake set ting24 34 44 54 64 24 34 44 54 64 24 34 44 54 64 24 34 44 54m a x t h i c k n e s sm i n t h i c k n e s sManual Braking Speed = 300 fps Automatic Braking Speed = 300 fps Manual Braking Speed = 400 fps Automatic Braking Speed = 400 fps max length min length 19 Exercise 3: Interpreting a MultiVari Chart, Answer ?Automatic braking gives consistently lower variation ?When braking manually, 64 lbs. of pressure gives lower variation ?With automatic braking at the speed of 400 fps., the pressure doesn’t affect the variability ?The higher speed (400 fps.) results in reduced variation 20 Planning a MultiVari Study 1. Decide on the scope or domain of the study ?Identify the largest process family you will study (facilities? Production lines? Machines? Tools? Shifts?) ?If there are multiple streams ( lines), consider taking only the extreme performers to save effort. ?Determine whether you will sample within units 21 Planning a MultiVari Study, cont. 2. Decide on the time profile of the study ?Find out if units can be sampled consecutively ?Determine what constitutes a short interval ?Determine what constitutes a long interval ?Determine if you can use an interval between short and long term 3. Create the data collection plan (refer to the Data Collection plan format discussed in Week 1.) 22 Planning a MultiVari Study, cont. 4. Collect the samples. 5. Measure the samples 6. Plot the sample data. ?Use StatQuality Too