SESSION 12: YOUR SIX SIGMA PROJECT: THE CONTROL PHASE

1
SESSION 12 YOUR SIX SIGMA PROJECTTHE
CONTROL PHASE

• INTRODUCTION TO
• IMPROVING BUSINESS PERFORMANCE SIX SIGMA, LEVEL
  1
• APRIL 16 - 18, 2007

2
SESSION OBJECTIVE

• Outline steps in the Control Phase

3
RECAP

• We have
• Defined what project
• Measured current state
• Analyzed - identified vital few Xs that cause
  the defects
• Improved - and established the relationship for
  CTQ or Y and the vital few Xs
• Now we are ready t control those Xs to ensure a
  sustained predictable Y the only purpose for
  the Control Phase

4
CONTROL PHASE OVERVIEW
5
CONTROL PHASE QUESTIONS

• Whats your plan to enforce policy/procedures?
• What chart will you use to show how the process
  is working and ensure control?
• Who is responsible for maintaining the solutions
  and control plans?
• What are the consequences when the process is out
  of control?
• Whats your communication plan or training plan
  to sustain control?
• Who will document and implement the monitoring
  plan?
• What will be the standardized process?
• How will you share knowledge so you can
  institutionalize the improvements?
• Whats your review plan?

6
Control Phase Considerations

• The Control Phase ensures new process conditions
  are documented and monitored via process control
  methods.
• After an interval for stabilization you should go
  back and reassess the process capability
• Two critical methods are generally used in the
  control phase
• Statistical Process Control (SPC)
• Mistake proofing
• Their purpose is to complete the cycle of finding
  the controls for the solution and maintaining

7
Statistical Process Control SPC

• Developed by Walter Shewart in 1924
• SPC is a statistically based graphing technique
  that compares current process data with a set of
  stable control limits established from normal
  process variation
• When data points go outside the limits the
  process is considered to be out-of-control
  (meaning out of STATISTICAL control)

8
SPC Why do we need it?

• We have already identified the vital few - we
  understand that Y X
• However if we only collect data on Y the output
  we are being reactive we would be practicing
  statistical process monitoring
• SPC enables to be proactive by collecting data on
  the inputs our Xs

9
SPC Benefits

• SPC groups data - enables us to statistically
  test the data
• Provides insight into behavior of
  processes/products and/or process characteristics
• Facilitates understanding of underlying cause
  system
• Provides a graphical representation of products
  and process performance
• Detects assignable causes that affect the central
  tendency and/or variability of the sources
• Serves as a decision making tool based on
  probabilities, acts a practical tool for
  detecting changes in process performance, and
  points out when actions are needed

10
Step 1 Pick Variable to Control

• Pick a critical one to control look at the
  results of your experiments
• What types of values do you have for your Xs?
  Which one had the greatest impact on your Ys?
• Remember there are two types of data
• Attribute data discrete (either or) data
  yes/no clean/dirty high/low 0-1
• Variable data data that changes and is measured
  on a continuous scale temperature pressure
• Data type is used in selecting control chart to
  be used

11
Step 2 Select the Type of Control Chart to Use

• The rules for calculating control limits vary
  according to type of data you have select you
  chart accordingly
• Data type
• Sample size
• For variable data use
• Xi X-bar and R and X bar and sigma charts
• For attribute data use C,U, NP or P charts
• Then to decide on actual chart consider sampling
  to be done

12
Step 3 Determine Rational Subgroup Size and
Sampling Interval/Frequency

• Rational subgroups are basis of control charts
  you cannot do random sampling, you must sample
  rationally by sampling groups of items that come
  through the process under the same conditions.
• Rational subgrouping is a method for determining
  size and frequency of sampling cso each sample
  represents a point in time or space
• It is way to collect data that best represents a
  process state that is unknown.

13
DATA TYPES AND CONTROL CHARTS
14
Step 4 Determine Measurement Methods and Criteria

• Measuring methods will depend on the types of
  data you will collect attribute data will need
  a check sheet variable data will require
  measuring tools scales etc

15
Step 5 Calculate Parameters of the Control Chart

• Basics
• After collecting at least 20 data points using
  the methods developed in Steps 3 and 4, you can
  create a control chart and then plot data on the
  type of chart you selected in step 2.
• The purpose of the control chart is to track
  changes over time
• Plot data collected chronologically left to right
• Calculate process average and draw horizontal
  center-line to represent that average
• Calculate UCL and LCL draw them above and below
  the line
• You need 20 data points to calculate the trial
  control limits 3 standard deviations from the
  mean (the 3 is historically based) At that
  level if a process is in control then nearly all
  data points will fall within the control limits
  and if new data points continue to be in range
  then the process is considered under control -
  if all points are inside then the process was
  always in control and the control limits are
  suitable for controlling the process
• If they fall out then the process is considered
  unstable, out of control and unpredictable

16
Step 6 Develop a Control PlanControl Plan
Components
17
Step 6 Develop a Control PlanControl Plan
Checklist
18
Step 7 Train People and Use Charts

• You must train your personnel to use the Control
  Charts you have developed so they are able to
  implement and analyze the information being
  monitored/collected
• Effective use will require periodic revisions to
  reset control limits and center-lines revision
  can be time-bound or on sample/batch sizes
• Control chart limits are very critical to SPC
• REMEMBER IN CONTROL DOES NOT ALWAYS MEAN
  ACCEPTABLE IT MEANS A PROCESS IS STABLE AND
  PREDICTBLE NOT NECESSARILY ACCEPTABLE TO YOU OR
  CUSTOMERS

19
PRECONTROL A VISUAL FORM OF SPC

RED ZONE
Superior limit
YELLOW ZONE
GREEN ZONE
Center limit
Individual Value
55 52 48 46 43
YELLOW ZONE
Inferior limit
RED ZONE
10
15
5
20
Observation Number
20
Using the Pre-control Chart

• Simplest form of SPC
• Works like traffic light
• If you are in green zone continue operating
  normally
• If the process is in yellow check if you need
  adjustment
• If two units in same yellow adjust process
• If in different yellow adjust in opposite way
• If in red stop the process
• When 5 consecutive units fall in green zone
  process is considered qualified

21
Step 8 Phase Gate Analysis

• As with other phases Conduct your Phase Gate
  Analysis

22
Mistake Proofing

• Typically the last tool used
• Purpose is to determine methods that will ensure
  a process will not allow defects
• Should be used for those processes where steps
  can be bypassed or performed out of order
• It seeks to ensure the steps cannot be leftout
• REMEMBER A MISTAKE IS AN ACTION THAT RESULTS IN
  DEFECTS
• Examples of mistake proofing electric kettles
  with auto-shut-offs medicines in tamper-proof
  packaging computers ask you are you sure?

23
Implementing Mistake Proofing

• Look at repetitive processes
• the 10 most common errors
• Processing omissions
• Processing Errors
• Error in set-up
• Assemble omissions
• Inclusion of wrong parts
• Wrong work-piece
• Operating errors
• Adjustment, measurement dimensional errors
• Equipment maintenance
• Error in preparation of setup or tool adjustments

24
SUMMARY

• In the Control Phase, the team works to maintain
  changes in Xs to sustain the improvements in the
  Ys.
• You need a 5 part Control Plan consisting of
  training for the operators, document plan,
  monitoring plan, response plan and
  institutionalization plan
• INSTITUTIONALIZATION is the culmination of the
  DMAIC process and should set the stage for
  sustaining the improvements you have made