Enhancing and Restoring Safety & Quality Cultures - Dave Litwiller - May 2024...
Ā
Six sigma
1. Six Sigma
Statistical Process Control
Having completed this module you will be able to:āØ
ā¦ Deļ¬ne what Statistical Process Control is.
ā¦ Describe Control Charts.
ā¦ Describe Control Charts for Attributes, and Control Charts for
Variables.
ā¦ Explain Control Chart patterns.
The key points from this module are:āØ
Statistical process control (SPC)āØ
-Monitors production process to prevent poor quality.āØ
āØ
Acceptance SamplingāØ
- Inspects random sample of product to determine if a lot is
acceptable.āØ
āØ
SampleāØ
- Subset of items produced to use for inspection.āØ
āØ
Control ChartsāØ
- Process is within statistical control limits.āØ
āØ
Production data ALWAYS has some variability.āØ
Chance and Assignable causes of Variation.āØ
āØ
Examples of Quality Characteristics:āØ
Painted surface, Thickness, Hardness, Resistance to fading or
chipping, Viscosity, Sweetness, Electrical resistance, Frequency....āØ
2. āØ
We can control only those characteristics that can be counted,
evaluated or measured.āØ
Engineering characteristics may show problems with accuracy or
with precision.
Variability
Random:Ā
- Common causes
- Inherent in a process
- Can be eliminated only through improvements in the system
Non-Random
- Special causes
- Due to identiļ¬able factors
- Can be modiļ¬ed through operator or management action
Assignable causes are controlled by SPC
- Take periodic samples from process
- Plot sample points on a control chart
- Determine if process is within limits
- Prevent quality problems
Control Charts
A key tool in SPC
Graph establishing process control limits
Charts for variables
- Mean (X-bar)
- Range (R), EWMA, CUSUM
Charts for Attributes
- p, np and c
Shewhart Control Chart
A time-ordered plot of sample statistics
When chart is within control limits
3. - Only random or common causes present
- We leave the process alone
Plot of each point is the test of hypothesis:
Ho: Process is "in control" vs.
H1: Process is out of control and requires investigation
A Process is "in control" if...
- No sample points are outside limit
- Most points near process average
- About equal number of points above and below centerline
- Points appear randomly distributed
- A process "in control" is supposed to be under the inļ¬uence of random
causes only
Potential reasons for Variation:
Operator: Training, Supervision, Technique
Method: Procedures, Set-up, Temperature, Cutting speeds
Material: Moisture content, Blending, Contamination
Machine: Set-up, Condition, Inherent precision
Management: Poor process management, Poor systems
Charts may signal incorrectly!!
Charts may repeatedly apply hypothesis testing
Type I error results with charts: Concluding that a process is not in control
when it actually is.
Type II error results with charts: Concluding that a process is in control
when it actually is not.
Two Types of Process Data
Variables: Things we measure
4. Length, Weight, Time, Temperature, Diameter, Tensile strength
Ā Ā
Attributes: Things we count
Number or percent of defective items in a lot; Number of defects per item;
Types of defects;
Value assigned to defects: Minor = 1, Major = 5, Critical = 10
SPC Applied to Services
Nature of defect is diļ¬erent in services.
Service defect is a failure to meet customer requirements.
Monitor times, and customer satisfaction.
Examples:
Hospitals: Timeliness, Responsiveness, Accuracy
Grocery Stores: Check-out time, Stocking, Cleanliness
Airlines: Luggage handling, Waiting times, Courtesy
Three Sigma Control Limits
The use of 3-sigma limits generally gives good results in practice.
If the distribution of the quality characteristic is reasonably well
approximated by the normal distribution, then the use of 3-sigma limits is
applicable.
These limits are often referred to as action limits.
Types of Shewhart Control Charts
Control charts for Variable Data
X-bar and R charts: For sample averages and ranges
X-bar and s charts: For sample means and standard deviations
5. Md and R charts: For sample medians and ranges
X-bar charts: For individual measures; uses moving ranges
Control charts for Attributes Data
p charts: Proportion of units nonconforming
np charts: Number of units nonconforming
c charts: Number of non-conformities
u charts: Number of non-conformities per unit
Understanding Control Charts
Having completed this module you will be able to:āØ
ā¦ Describe why every process displays some variability.
ā¦ Explain how control charts can identify abnormal variability.
ā¦ Describe why control charts may give false alarms.
ā¦ Explain why on-line control leads to early detection and
correction.
6. The key points from this module are:āØ
p Charts - Calculate percent defectives in sampleāØ
āØ
Use of p ChartsāØ
āØ
When two observations can be placed into two categories:āØ
- Good or badāØ
- Pass or failāØ
- Operate or don't operateāØ
āØ
When the data consists of multiple samples of several observations
each.āØ
āØ
Control chart for attributes are used to monitor the proportion of
Ā defectives in a process.āØ
āØ
c Charts - Count number of defects in itemāØ
āØ
Use of c ChartsāØ
āØ
Use only when the number of occurrences per unit of measure can
be Ā counted (non-occurrences cannot be counted)āØ
e.g. Scratches, chips, dents, cracks or faults per unit of distance.
Control Charts SummaryāØ
āØ
X-bar and R ChartsāØ
- Variables dataāØ
- Application of normal distributionāØ
7. āØ
p ChartsāØ
- Attributes data (defects per n observations)āØ
- Application of binomial distributionāØ
āØ
c ChartsāØ
- Attributes data (defects per inspection)āØ
- Application of Poisson distributionāØ
Control Chart Design IssuesāØ
- Basis for samplingāØ
- Sample or Subgroup size (n)āØ
- Frequency of samplingāØ
- Location of control limitsāØ
āØ
Charts for AttributesāØ
Fraction non-conforming (p chart)āØ
- Fixed sample sizeāØ
- Variable sample sizeāØ
āØ
np Chart for number non-conformingāØ
āØ
Charts for defectsāØ
- c ChartāØ
- u Chart
Statistical Process Control (SPC) - Implementation Requirements
- Process capability audit
- Top management commitment
- SPC Project champion - the process owner
- Initial workable project
- Employee education and training
- Accurate measurement system (Gage R&R)
8. Process improvement moves in steps:
Measurement: External and Internal
Analysis: Analyze Variation
Control: Adjust Process
Improvement: Reduce Variation
Innovation: Redesign Product/Process
Process capability: The inherent variability of process output relative Ā to
the variation allowed by the design or customer speciļ¬cation.
Process Capability Analysis
Diļ¬ers fundamentally from Control Charting
- Focuses on improvement, not control
- Variables, not attributes, data involved
- Capability studies address range of individual outputs
- Control charting addresses range of sample measures
Assumes Normal Distribution
- Remember the Empirical rule?
- Inherent capability (6 sigma) is compared to speciļ¬cation
Requires process ļ¬rst to be in statistical control
Why measure Process Capability?
Process variability can greatly impact customer satisfaction
Three common terms for variability:
1. Tolerances: Speciļ¬cation for range of acceptable values established Ā by
engineering design or customer requirements
2. Process variability: Natural or inherent variability in a process
3. Control limits: Statistical limits that reļ¬ect the inherent variation Ā of
sample statistics
Capability Analysis
9. Capability analysis determines whether the inherent variability of the
Ā process output falls within the acceptable range of the variability Ā allowed
by the design speciļ¬cation for the process output.
The range of possible solution:
1. Redesign the process so that it can achieve the desired output
2. Use an alternate process that can achieve the desired output
3. Retain the current process but attempt to eliminate unacceptable
Ā output using 100 percent inspection
4. Examine the speciļ¬cation to see whether they are necessary or could
Ā be relaxed without adversely aļ¬ecting customer satisfaction
10. Design of Experiments:
Having completed this module you will be able to:āØ
ā¦ Describe how to use Design of Experiments to understand and
optimize process settings.
ā¦ Explain and apply the statistical techniques used in Six Sigma.
ā¦ Deļ¬ne what an experiment is.
ā¦ Describe the various types of variables.
ā¦ Outline the steps in planning Design of Experiments.
ā¦ List the advantages of Design of Experiments.
ā¦ List and explain the three major approaches to using Design of
Experiment.
11. The key points from this module are:āØ
What is an Experiment?āØ
āØ
Systematically vary variables of interest e.g. Giving diļ¬erent drugs to
subjects In Ā each trial measure the responseĀ āØ
āØ
Critical concepts:āØ
ā¢ Variable must be manipulated by experimenterāØ
ā¢ Random assignment of experimental units to conditionsāØ
ā¢ Avoid confounding of variablesāØ
āØ
Approaches to ExperimentationāØ
1. Build-test-ļ¬x problemsāØ
2. One-factor-at-a-time (the classical approach)āØ
3. Designed experiments (DOE)
Kinds of Variables
Independent variables (factors) - What you manipulate?
Dependent variables (responses) - What you measure?
Control variables - What you hold constant?
Random (noise) variables - What you allow to vary randomly?
Confounding variable - Correlated with independent variable
Steps in Planning DOE
1. Deļ¬ne objective
2. Select the Response (Y)
3. Select the Factors (Xs)
4. Choose the factor levels
5. Select the Experimental Design
6. Run Experiment and Collect the Data
7. Analyze the data
8. Conclusions
12. 9. Perform a conļ¬rmation run
Illustrating the DOE Process
1. Determine the goals
2. Deļ¬ne the measures of success
3. Verify feasibility (rough estimate)
4. Design the experiment (precise estimate)
5. Run the experiment
6. Collect and analyze the data
7. Determine and verify the response
8. Act on the results
Advantages of Doing DOE
Statistical foundation of DOE yields a lot of information at relatively low
cost. Ā Provides main, secondary and interaction eļ¬ects of factors being
tested.Ā
Basic DOE Ā can be conducted and evaluated without signiļ¬cant statistical
knowledge or Ā expertise.Ā
DOE gives much more information than obtained from one-at-a-time
Ā experimentation.Ā
DOE - A pro-active tool for directing improvements.
Major Approaches to using
Factorial DesignĀ
Taguchi MethodĀ
Response Surface Design
Summary of DOE
DOE is a systematic, rigorous approach to engineering problem-solving.
Applies Ā principles and techniques at the data collection stage. Assures
the generation of Ā valid, defensible and supportable engineering
13. conclusions. All of this is carried out Ā under the constraint of minimal
expenditure of engineering runs, time and money.
The key points from this module are:āØ
14. Taguchi Methods (or Quality Engineering or Robust Design)āØ
āØ
Focus is on reducing variability of response to maximize robustness,
generally achieved through Orthogonal Array ExperimentsāØ
āØ
Historical PerspectiveāØ
George E. P. Box (born 1919) was a student of R. A. Fisher. He made
several advances to Fisherās work in DOE theory and statistics. The
founding chair of the University of Wisconsinās Department of
Statistics, Box was appointed the R. A. Fisher Professor of Statistics
at UW in 1971.āØ
āØ
āØ
Typical QE ApplicationsāØ
āØ
ā¢ In manufacturing - improve performance of a manufacturing
process.āØ
ā¢ In process development - improve yields, reduce variability and
cost.āØ
ā¢ In design - evaluation and comparison of basic conļ¬gurations,
materials and Ā Ā āØ
Ā Ā parameters.āØ
ā¢ The method is called Taguchi Methods.āØ
ā¢ The key tool is DOE (Design of Experiment).
C. R. Rao
You may know Rao from his Cramer-Rao Inequality. Rao is recognized
worldwide as a pioneer of modern multivariate theory and as one of the
worldās top statisticians, with distinctions as a mathematician, researcher,
scientist and teacher. Taught Taguchi. Author of 14 books and over 300
papers.
Genichi Taguchi
15. An engineer who developed an approach (now called Taguchi Methods)
involving statistically planned experiments to reduce variation in quality.
Learned DOE from Professor Rao. In 1960ās he applied his learning in
Japan. In 1980ās he introduced his ideas to US and AT&T.
What are Taguchiās Contributions?
ā¢ Quality Engineering Philosophy - Targets and Loss functions
ā¢ Methodology - System, Parameter, Tolerance design steps
ā¢ Experiment Design - Use of Orthogonal arrays
ā¢ Analysis - Use Signal-to-Noise ratios (S/N ratios)
Taguchiās key contribution is Robust Design
Robust Design - A Design that results in products or services that can
function over a broad range of usage and environmental conditions.
Taguchiās Product Design Approach has 3 steps:
1) System Design Choose the sub-systems, mechanisms, form of the
prototype - develop the basic design. This is similar to conventional
engineering design.Ā
2) Parameter Design Optimize the system design so that it improves
quality (robustness) and reduces cost.Ā
3) Tolerance Design Study the tradeoļ¬s that must be made and determine
what tolerances and grades of materials are acceptable.
Parameter Design (the Robust Design step)
Optimize the settings of the design parameters to minimize its sensitivity
to noise - ROBUSTNESS.
By highlighting ārobustnessā as a key quality requirement, Taguchi really
opened a whole area that previously had been talked about only by a few
very applied people.
His methodology is heavily dependent on design of experiments like
Fisherās and Boxās methods, but the diļ¬erence he made was that for
16. response he looked at not only the mean but also the variance of
performance.
Robust Design - How it is done
Identify Product/Process Design Parameters that:
ā¢ Have signiļ¬cant/little inļ¬uence on performance
ā¢ Minimize performance variation due to noise factors
ā¢ Minimize the processing cost Methodology: Design of Experiments
(DOE)Ā
Examples: Chocolate mix, Ina Tile Co., Sony TV
Taguchiās Experimental Factors Parameter design step identiļ¬es and
optimizes the Design Factors
Control Factors - Design factors that are to be set at optimal levels to
improve quality and reduce sensitivity to noise.
- Size of parts, type of material, Value of resistors etc.Ā
- Noise Factors - Factors that represent the noise that is expected in
production or in actual use of the product
- Dimensional variation
- Operating Temperature Adjustment Factor - Aļ¬ects the mean but not the
variance of a response
- Deposition time in silicon wafer fabrication Signal Factors - Set by the
user communicate desires of the user - Position of the gas pedal
Several diļ¬erent types of Experimental plans (ādesignā) are available to the
design engineer - Factorial, Fractional, Central Cuboid etc. Taguchi used
āOrthogonalā Designs
Focus: Handle many factors
Output: List of important Factors, Best Settings, Good design
Comments on Taguchi Arrays
17. Taguchi designs are large screening designs Assumes most interactions
are small and those that arenāt are known ahead of time Taguchi claims
that it is possible to eliminate interactions either by correctly specifying the
response and design factors or by using a sliding setting approach to
those factor levels.Ā
Doesnāt guarantee that we get āhighest resolutionā design. Instead of
designing the experiment to investigate potential interactions, Taguchi
prefers to use three-level factors to estimate curvature.
Designers should embrace Taguchiās philosophy of quality engineering. It
makes very good sense. Note, however, that a key weakness of Taguchi
method is its assumption of a āmain factor onlyā (or āadditiveā model) ā¦
Taguchi ignores interactions Therefore, rather than use inner - outer arrays,
we may use more eļ¬cient and exact methods that are no more diļ¬cult to
learn and apply to carry Taguchiās robust design philosophy into practiceā¦
You may use any of the various experimental and optimization techniques
available in the literature such as multiple regression/RSM to develop
robust designs.
Sensitivity Analysis vs. āRobust Designā
Practice common in conventional engineering design: sensitivity analysis
(SA). SA ļ¬nds likely changes expected in the designās performance due to
uncontrollable factors. For designs too sensitive, one uses the worst-case
scenario - to plan for the unexpected. However, worst-case projections
often unnecessary: a ārobust designā can greatly reduce oļ¬-target
performance (Taguchi, 1986)
Benchmarking
Having completed this module you will be able to:āØ
ā¦ Deļ¬ne benchmarking and list its beneļ¬ts.
ā¦ Describe the key elements of the Benchmarking Code of
Conduct.
ā¦ List the steps of widely used benchmarking procedures.
ā¦ Describe management's benchmarking challenge.
18. The key points from this module are:āØ
What is Benchmarking?āØ
āØ
A highly structured strategy for acquiring, assessing and applying
customer, competitor and enterprise intelligence for the purpose of
product, system or enterprise innovation and design.āØ
āØ
Benchmarking is ā¦āØ
A continuous process.āØ
A process of investigation that provides valuable information
stimulating improvement.āØ
A process of learning from others; a pragmatic search for ideas - that
are successfully working elsewhere.āØ
A time-consuming, labor-intensive process requiring discipline - the
task is not easy!āØ
A viable tool that provides useful information for improving virtually
any business process.āØ
The How-to-do in BenchmarkingāØ
āØ
1. Identify Process to BenchmarkāØ
2. Select Organization to BenchmarkāØ
3. Prepare for the VisitāØ
4. Visit the OrganizationāØ
5. Debrief and Develop an Action PlanāØ
6. Retain and CommunicateāØ
Sources of Benchmarking Information include:āØ
āØ
Library DatabasesāØ
Industry PublicationsāØ
Customer FeedbackāØ
Industry Experts
Ten Generic Benchmark Categories
19. 1. Customer service performance
2. Product/service performance
3. Core business process performance
4. Support process and services performance
5. Employee performance
6. Supplier performance
7. Technology performance
8. New product/service development and innovation performance
9. Cost performance
10. Financial performance
Best Practices Benchmarking can be described as the process of seeking
out and studying the best internal and external practices that produce
superior performance.
Donāt reinvent what others have learned to do better!
Borrow shamelessly!
Adopt, adapt, advance!
Imitate creatively!
Adapt innovatively!
Process benchmarking focuses on discrete work processes and operating
systems, such as the customer complaint process, the order-and-
fulļ¬llment process, or the strategic planning process.
Strategic benchmarking examines how companies compete and is seldom
industry-focused. It roves across industries seeking to identify the winning
strategies that have enabled high-performing companies to be successful
in their marketplaces.
Widely Acknowledged Benchmarking Beneļ¬ts
- Improves organizational quality - products and services
- Leads to lower cost positions
- Creates buy-in for change
- Exposes people to new ideas
- Broadens the organizationās operating perspective
- Creates a culture open to new ideas
20. - Serves as a catalyst for learning
- Tests the rigor of internal operating targets
- Creates an external business view
- Raises the organizationās level of maximum potential performance
The Benchmarking Code of Conduct Summary
In Benchmarking remember to:
- Keep it legal
- Be willing to give what you get
- Respect conļ¬dentiality
- Keep information internal
- Use benchmarking contact
- Donāt refer without permission
- Be prepared at initial contact
Lessons from Active Benchmarkers
1. Do not strive to benchmark everything as best-in-country or best-in-
world levels. No company can be the best in every function - focus on
processes and practices of strategic importance.
2. Seek best-in-class benchmarks for core processes and functions of the
highest strategic importance, the Pareto Principle wins again. Other
benchmarks can come from levels 2 through 5. World and country
leadership benchmarks require precious time, resources and eļ¬ort to
develop.
3. Seek internal, regional or industry benchmarks for secondary and
support processes. For some processes and business activities that are
not critical to the organizationās strategic advantage internal, regional or
competitive benchmarks may be most appropriate. Such benchmarks
produce incremental improvements that are substantial - even if not
radical or breakthrough in terms of the size of the expected improvement
beneļ¬ts.
Performance Measure Examples
Employee Performance Measures
Supplier Performance Measures
21. Cost Performance Measures
Financial Performance Measures
Benchmarking Critical Success Factors
Senior management support
Benchmarking training for the project team
Useful information technology systems
Cultural practices that encourage learning
Resource dedication - especially in the form of time, funding and useful
equipment
The Motorola Five-Step Benchmarking Process
1. Decide what to benchmark
2. Find companies to benchmark
3. Gather data
4. Analyze data and integrate results into action plans
5. Recalibrate and recycle the process
Managementās Benchmarking Challenge
- Commit required resources to key projects
- Provide focused training/facilitation to project participants
- Proactively manage the direction and momentum of benchmarking within
the organization
- Create visibility of the benchmarking process
- Recognize benchmarking team eļ¬orts
Six Sigma and Supply Chain
Having completed this module you will be able to:āØ
ā¦ Deļ¬ne what supply chain management is.
ā¦ Describe how Six Sigma can be used in supply chain
management.
ā¦ List examples of supply chains that have Six Sigma processes.
22. ā¦ Describe management's challenge of integrating Six Sigma
processes into supply chain systems.
The key points from this module are:āØ
A Supply Chain is a network of organizations.āØ
āØ
Goal: Produce value for the ultimate consumerāØ
āØ
Primary purpose:āØ
- Satisfy customer needsāØ
- Manage ļ¬ows of goods and informationāØ
- Generate proļ¬ts for itselfāØ
- Decide locations, which products to produce at what stage, how to
produce and how to distributeāØ
āØ
āØ
Integrated, coordinated network of value delivering business
processes that procure raw materials, transform them into ļ¬nal
products, services and delivers the product to the customers
involving:āØ
- ProcurementāØ
- Manufacturing/AssemblyāØ
- Inbound logisticsāØ
- WarehousingāØ
- DistributionāØ
- Outbound logistics
Goal is to design and manage a supply chain network that delivers high-
quality products to the right customers at the right time at minimum cost.
Good supply chain management aims at reducing inventory and freeing up
working capital without aļ¬ecting service levels.
23. A great deal of inventory piles up along the supply chain due to:
Poor quality of supplies - production may stop if there are no materials and
parts in the buļ¬er stock.
Poor management of logistics and no monitoring of lead times.
Machines break down resulting from poor quality culture, causing
production interruption - again buļ¬ers are needed.
Quality of ļ¬nished goods may not be high enough to support JIT mode of
operation in your customerās operations.
Examples of Supply Chains operating in India - All have buļ¬ers (ājust in
caseā)
Automotive - Telco, ALL, Mahindra, Maruti
Aerospace - ADA, HAL
Chemicals - Asian Paints, Apollo tyres, Reliance
Apparel - Madura Coats, Reliance
Food - Cadbury, Parle, Amul Products, HLL
Consumer durables - HLL, P & G
Forest Products - Papermills
Construction - L & T
Pharmaceutical - Ranbaxy, Glaxo
Electrochemical - Kirloskar, L & T
Tooling, HMT, Widia, Mico
PC and Computer - IBM, WIPRO, HCL, Intel
Each entity in a supply chain is both a supplier and a customer.
Thus, important to have a customer focused corporate vision while
implementing TQM and SCM practices both upstream and downstream.
Customer-driven vision can produce a number of competitive advantages
for the supply chain by:
- Helping improve productivity
- Reducing inventory and cycle time
- Boosting customer satisfaction, market share and proļ¬ts
24. Supply Chain Management (SCM) and Total Quality Management (TQM)
are two key tools used by manufacturers and service providers to make
their business competitive.
Quality: Value addition during production and delivery.
Sustaining quality throughout the chain has signiļ¬cant implications for
reducing costs and raising customer satisfaction, hence sales.
Many end-products producers now recognize the potential beneļ¬ts of
partnering with their suppliers in managing quality in their supply chains.
A key example is Chrysler:
In 1935 Chryslerās suppliers showed a quality level of 300 to 400 defective
parts per million (ppm).
To achieve < 100 ppm for its suppliers, Chrysler forced its 2nd and 3rd tier
suppliers to implement the same quality standards as its 1st tier suppliers.
Same procedures through the entire supply chain improved quality
performance of all suppliers.
Chrysler also used quality as a deciding factor to reduce the number of its
suppliers.
TQM philosophy to guide all plans, operations and activities:
- Top managementās direct involvement
- Strong internal/external customer orientation
- Systematic problem solving using statistics
- Everyone participates - suppliers, organization, customers
7 tools to solve problems at the ļ¬oor level: QS 9000 for culture
SPC to monitor variability in all that can be measured: Sampling to provide
evidence
DMAIC projects to focus improvement eļ¬orts in meeting requirements,
delivery times and cost/proļ¬t targets