Design of Experiments and Total Quality Management Flashcards
3 primary experimental objectives in DOE
Screening
Robustness testing
Optimisation
Commonly used DOEs
Full factorial
Orthogonal (Taguchi) array
Why do we use DOEs
Intelligent decisions should be based on informed observation and directed experimentation
Consistent with the scientific method
Organises and concentrates efforts
Helps you understand how factors affect a response
Improvements can only be measured by statistical data
Input conditions for DOEs
Number of factors
Factor ranges (levels)
Response
Experiment objective
Why is Taguchi used over full factorial in industry
F.F. designs are not practical for processes with many factors
Taguchi designs minimise the number of experiments needed for a given set of factors
Groups of quality gurus
Early 50s - Americans taking messages of quality to Japan
Late 50s - Japanese who developed new concepts in response to the Americans
70s - 80s - Western gurus who followed Japanese industrial success
Best known Quality Guru and what did they do
W Edwards Deming
American who was ignored in the states (due to the post-war boom in industry) so took his knowledge to Japan (Sony, Nissan, Mitsubishi and Toyota) and greatly assisted their improvements in quality presenting his ideas in a meaningful and understandable way down to the production worker level
Deming’s chain reaction/philosophy
Improve quality
Costs decrease
Productivity improves
Capture the market with better quality and lower prices
Deming’s quote
‘If I had to reduce my message to just a few words , I’d say it all had to do with reducing variation’.
Variation is reduced by reliability, predictability and consistency of products and services will all be improved
Main areas for variation
Variation of the raw material
Variation of the people
Variation of procedures
Variation of suppliers
Variation in the working environment
2 classifications of variation, according to Deming
Special or assignable causes:
Arise due to changes in the raw material, operator and procedure used
Common or unassignable causes:
All causes due to the process itself, its design and installation
Deming Cycle
Plan
Do
Check
Act
what are the 3 primary experimental objectives
Screening ( which factors are most influential)
Robustness testing ( how to adjust factors for robustness)
Optimisation
How many experiments are needed for a full factorial DOE?
level ^ number of factors
e.g 2 level DOE for 3 factors requires 2^3=8 experiments
when is action required in a process control chart?
Any point falls outside of control limit
A run of 7 points all above or below central line
A run of 7 points going up/down
