The Essentials of 2-Level Design

of Experiments
Part I: The Essentials of Full Factorial

Developed by Don Edwards, John Grego and James Lynch
Center for Reliability and Quality Sciences
Department of Statistics
University of South Carolina

Part I: The Essentials of Full
Factorial Designs
 Some Motivation and Background
 Two Important Advantages
of Factorial Experiments
 The Essentials of 2-Cubed
 Full Factorial Designs

1 Some Motivation Arno Penzias Chief Scientist and VP for Research. Bell Labs & Nobel Laureate-Physics “Teaching Statistics to Engineers. I.” Science Editorial. June 2 1989  Statistical Tools Are Needed In Industry – Competitive Position Demands It – Optimizing Complex Technological Manufacturing Processes Requires It .

I.1 Some Motivation  Leaders In Quality – Use Statistics At All Process Stages For Quality and Optimization Purposes – Provide The Necessary Statistical Training To Do This .

” .1 Some Motivation QS9000  QS9000 required that “The supplier shall demonstrate knowledge in Design of Experiments (DOE) and use it as appropriate. I.

g. I.  Sara Lee Hosiery Division has used simple designs in a number of settings. e. . Several have resulted in considerable annual savings..1 Some Motivation Examples of DOE Applications  NCR has used factorial designs in a variety of situations. to analyze computer performance and to compare different soldering methods.

1 Some Motivation Examples of DOE Applications  Ohio Brass has conducted several fractional factorial designs which have had significant impact. Another enabled them to reduce the dimensions of two key components which resulted in annual savings of $50K. . I. One study resulted in an annual savings of $25K by modifying an existing process and avoided a capital investment of a 1/4 to 1/2 million dollars in new equipment.

.1 Some Motivation Examples of DOE Applications  Michelin has used designs to determine maintenance programs for some of their machinery. I.

1 Background Why Should You Use DOE?  Intelligent Decisions Should Be Based On "Informed Observation And Directed Experimentation" (George Box) – It is consistent with the Scientific Method which is fundamental to the quality management philosophy (The Deming- Shewhart PDSA Cycle) – DOE is a formalism that forces you to organize your thoughts (so you don't do things haphazardly) . I.

1 Background Why Should You Use DOE?  DOE Concentrates Your Efforts – Screening designs aid in identifying the vital/critical factors that may affect the (process) response of interest – More refined design techniques determine the factor levels that optimize the response . I.

. I. This knowledge helps to choose factor settings that are cost effective but don’t compromise quality (constrained optimization).1 Background Why Should You Use DOE?  DOE Concentrates Your Efforts – DOE helps you to understand how factors affect the process.

1 Background Quality Management Philosophy  Some Tenets Related to These Components – All processes have variation – Different types of variation  e. a common cause variation system) ..e..g. I. i. common cause system verses special causes being present – Management needs predictable/stable processes to make decisions (process needs to be in control.

1 Background Quality Management Philosophy  Implications for DOE – The smaller the effects you are trying to detect relative to the background variation. the more replication you need or a different design (blocking) – Data from an out-of-control process is suspect . I.

1 Background Contrasting SPC and DOE  Statistical Process Control (SPC) – SPC is used to determine if a process is in control – An Out-of-Control process that is brought into control is not process improvement (Juran) . I.

DOE Allows You to Converse With It William Hunter .1 Background Contrasting SPC and DOE  Design of Experiments (DOE) – A methodology useful for determining  what factors may affect a response  what factor settings are feasible  SPCLets You Listen to the Process. I.

I.1 Background Experimentation  Experiment – A series of trials or tests which produce quantifiable outcomes.  Quantifiable Outcome – Some Outcome Measurement of Interest – “Response Variable” (y) .

1 Background Examples of Responses  Yield  Viscosity  Computer Performance  Breaking Strength of Fiber  Smoothness of Polyurethane Sheets  Bowing of a Molding  Chain Length in Polymer  Number of Flaws . I.

Most Most Severe Severe Bowing Bowing –– Bottom Bottom -.Flat. I. Flat. No No Bowing Bowing .Bowing of a Molding  Three Three Moldings Moldings –– Top Top -.1 Background Responses.

Bowing of a Molding True versus Substitute Quality Characteristics  The The Displacement Displacement D D –– Substitute Substitute Quality Quality Characteristic Characteristic for for Bowing Bowing –– Measurable Measurable . I.1 Background Responses.

g. Flow rate of a raw material – B. Presence/Absence of a Catalyst – D. Process temperature – C. 1.1 Background Factors  Experimental(Variable) Conditions That May Affect the Response. or 3) . – A. Raw Material Supplier (e.2. I.

1 Background Factors  Factors May Be – Continuous (A and B Above) – Discrete (C and D Above) . I.

I...1 Background First Motivation To Experiment  To“Improve” The Response...  Optimize average response  Minimize variability in response  Minimize susceptibility to uncontrollable “noise” factors .

I.1 Background Best Motivation  To“Understand” The Response! (George Box)  Levels of Understanding – Which? – How? – Why? .

Yellowfin Tuna Growth  Traditional Traditional Theoretical Theoretical Growth Growth Models Indian Ocean Yellowfin Tuna Models Allow Allow For For Age versus Fork Length Only Only One One Point Point of of 120 Inflection Inflection 110 100 (Two Growth (Two Growth 90 Stages) Stages) Fork 80 Length 70 (cm) 60 50 40 30 0 500 1000 Age (days) .1 Background Levels of Understanding An Example . I.

I.1 Background Levels of Understanding : “How” Stage An Example .Yellowfin Tuna Growth  Lowess Lowess Fit Fit Suggests Suggests Indian Ocean Yellowfin Tuna –– Two Two Points Points of of Age versus Fork Length Inflection Inflection 120 110 –– Rethink Rethink Theory Theory 100 90 Fork 80 Length 70 (cm) 60 50 40 30 0 500 1000 Age (days) .

I.Yellowfin Tuna Growth  More More Pronounced Pronounced In In Atlantic Ocean Yellowfin Tuna The The Atlantic Atlantic Age versus Fork Length Ocean Ocean 110 Yellowfin Yellowfin Tuna Tuna 100 90 Fork Length 80 (cm) 70 60 50 400 500 600 700 800 900 Age (days) .1 Background Levels of Understanding: “How” Stage An Example .

Yellowfin Tuna Growth Atlantic Ocean Yellowfin Tuna Age versus Fork Length 110 Gonadal 100 Stage 90 Fork Length 80 (cm) 70 Somatic 60 Stages 50 400 500 600 700 800 900 Age (days) . Levels of Understanding: “Why” Stage An Example .