Lean Six Sigma.txt

Rate at which system generates money through sales

• 1. One piece flow at a rate determined by the customer
• 2. Smooth and un-interrupted flow

Eliminating possible errors; forgetfulness, misunderstanding, identification, willful errors, lack of standards, indadvertant errors.

• 1. A method of material control.
• 2. A kanban card will give permission for the next operation to produce more parts.

• Chart to track variables like pr., temp, thk, customer complaints, etc at varying intervals.
• Variations within piece, piece-to-piece and time-to-time can be spotted.
• USL-Aim-LSL..

• 1. Overproduction
• 2. Inventory
• 3. Rework/Rejects
• 4. Motion
• 5. Processing
• 6. Waiting
• 7. Transport

• 1. Visual control device - shows current state of production, indicates problems, etc
• Green - No problems
• Yellow - Situation needs assistance
• Red - Pdn. stopped

• 1. Seiri - Sort - Get rid of what's not needed. (How often you use it?)
• 2. Seiton - Set in order - Organize, put in order and make things easy to find and use.
• 3. Seiso - Shine - Clean, repair and fix.
• 4. Seiketsu - Standardize - Standardize the new orderliness (checklists, photos, videos, etc)
• 5. Shitsuke - Sustain - Discipline to follow the new stds.

• Tool to organize complex info. into logical categories
• 

• Concise summary of a project on a single sheet of paper
• 1. Uses concise wordings
• 2. Uses charts to illustrate points
• 3. Restricts to 3-4 bullets per section
• 4. Provides substance, not fluff.

• A short summary of the strategic reasons for the project.
• 1. Design of a new project/process
• 2. Redesign of existing projects/process

• Team is aware of the project goals and boundaries
• Team will remain focussed on goals
• Team will work in alignment with the organization's goals
• Team champion will support the team and their goals

• Describing the flow of people, information or material in almost any type of process.
• 

• 1. Cycle time reduction
• 2. Reduction in finished inventory
• 3. Reduction in WIP
• 4. Order stabilization by customers
• 5. Pricing stabilization

• 1. Dissatisfiers (Basic reqmts., "must be")
• 2. Satisfiers (variable reqmts., "more is better")
• 3. Delighters (latent reqmts.)

• Focuses on key metrics of customer satisfaction
• 

• Enables the organization to:
• 1. Make decisions on products and services
• 2. Identify product features and specs.
• 3. Focus on improvmenet plans
• 4. Develop baseline metrics
• 5. Identify customer satisfaction drivers

• Prioritize problems based on their frequency of occurance
• "Vital Few" and "Trivial Many"
• 20% 80%

• 1. Do Nothing: Process limits within spec. limits
• 2. Change Spec: If customer agrees
• 3. Center the process: If process spread matches spec. spread, but with a shift
• 4. Reduce Variability: DOE used to find variations
• 5. Accept the losses: Handle scrap and rework efficiently

• Bar graphs giving the distribution of data, used to find mean, median, std. deviation, etc.
• 

Measure of the smallest change in measurand that can be sensed by the msmt. device

Smallest msmt. change that can be represented by the display of the msmt. device

• 1. Measure of the ability of the msmt. process to repeatedly get the same results keeping all factors same.
• 2. Same gage and same operator

• 1. Repeatability and reproducibility
• 2. Describes how close successive msmt. results fall when repeated.
• 

• 1. Measure of ability of msmt. process to obtain same results under varying conditions.
• 2. Same gage, different operators - gage reproducibility

Systematic difference between 2 successive msmts. of the same thing

• 1. Degree of closeness of a msmt. to its true value
• 2. Lack of bias
• 

• 1. Too many digits of the measured value
• 2. Insensitivity due to rounding offs
• 3. Errors due to entering of large sequence of data by operators. So:
• Code: X_c = X + C, Decode: X¹=X_c¹- C and sigma = sigma_c
• X_c = X/f, Decode: X¹ = X_c f and sigma = f x sigma_c

• 1. Charts that utilize location data
• 2. To answer the question 'where'.

• Takt Time = Available pdn. time / Rate of customer demand
• It is the beat/rythm of the process.

A business mgmt. strategy that seeks to improve the quality of process outputs by identifying and removing the causes of errors and variablilty in the processes.

Define-Measure-Analyze-Improve-Control

Define-Measure-Analyze-Design-Verify

• If one has 6 std. deviations b/w the process mean and the nearest spec. limit, practically no item fails to meet specs. At the max. such a process can have 3.4 defects per million in the long run.
• 

• Root Cause Analysis:
• Car not starting - Battery Dead - Alternator not working - Belt Broken - Old - No Maintenance

Tells whether the process is capable of meeting the specs or not. Statistical methods are used to find the mean and std. deviation of the data histogram.

Analysis of Variance - tells whether the means of two or more groups of data are equal or not.

Measure of repeatability and reproducability of msmt.

Helps in understanding how the value of the dependent variable changes with variation of any one of the independent variables, while maintaining all others fixed.

• Indicates the strength and direction of a relationship b/w two random variables.
• 

Analysis of the potential failures within a system for classification by severity.

• Tables used to record and analyze the relationship b/w two or more 'categorical' variables.
• 

• Methods used to conduct hypothesis tests of means of 'non-normal' data.
• These methods are based on 'Rank Ordering' of the response.
• Softwares to conducts such tests are: Wilcoxon Sign Ranked test, etc

Deciding on the number of data samples to be tested to have the results of the statistical analysis within a given confidence interval.

Conducting experiments to decide on the relationship b/w a dependent variable and independent variables by having various levels of the independent variables (factors) and monitoring the response of the system.

• Initiates pdn. in response to 'anticipated demand'
• It is a "ready or not, here I come" system. Ex: McDonalds

• Initiates demand in response to 'present demand'.
• It is a "don't call me, I'll call you" system. Ex: Wendy's

The time b/w the 'last good piece' from the previous pdn. run and the 'first good piece' from the new pdn. run.

• 1. Breakdowns
• 2. Set-ups and adjustments
• 3. Idling and minor stoppages
• 4. Reduced speeds
• 5. Quality defects and reworks
• 6. Startup losses

• 1. Objects of pdn. - the products
• 2. Agents of pdn. - the people making the products; tools, etc
• 3. Methods - means by which actions are performed
• 4. Space - where action is performed
• 5. Time - how long action takes for completion

• In supermarket pull systems, customer schedule is sent to only one pdn. process in the door-to-door value stream. This process is called the pacemaker process.
• 

• When continuous flow is not possible due to very fast or very slow processes or other logistics issues, then a supermarket pull system is used to eliminate the need to schedule production for the processes that are not linked.
• Simply put, SM is the installation of a pull system where continuous flow is interrupted and the upstream process must still operate in a batch mode.
• 

Pitch time might be the time to produce the amount of material typically moved at one time, like one skid of material.

• Structured method emphasizing team synergy. "The team as a whole is better and smarter than the quality sum of the individuals.
• 

• a) Used to determine whether a population mean (u) is equal to a hypothesized value.
• b) Used with continuous data from a single random sample, normally distributed.

• CI is a range of likely values for a population parameter (say u) that is based on the sample data.
• CI quantifies the precision of the estimate.

• Power is the ability of a test to detect a difference when one exists (Probability of rejecting Ho when its false)
• PA would answer one of these questions:
• How many samples are needed?
• Is the sample size good enough?
• How large a difference can be detected by the test?
• Are the results trustworthy?

• Used to determine whether there is a statistical difference between two population means.
• Continuous data from 2 independent samples
• Seperate normality tests for each.

• Used to determine whether there is a statistical difference b/w two paired populations.
• One element in common b/w samples
• Normality test for diff. b/w samples.

• Used to see what proportion of the population has defects.
• This test determines whether adequate evidence exists that the proportion of events (defects, etc) is different from a hypothesized value.

NOTE: Success/Reject data (as opposed to continuous msmts.). do not capture detailed info about the sampled parts, sample sizes are commonly large.

A test of two proportions determines whether sufficient evidence exists that two proportions are different from one another.

helps to compare two independent msmt systems so that we can eliminate either one if there is a sound correlation.

• Helps in knowing how important is X in predicting Y.
• How would Y vary with X?

• Generalization of a 2 sample T test.
• However, more than 2 groups of samples can be analyzed.
• Comparing means and variances between sample groups.