DOE_Introduction-english.ppt

Evolution of Quality Assurance Methods Definition of an Experiment An experiment is a test or series of tests in which purposeful changes are made to the inputs to observe and identify reasons for change in the output response “Every experiment is a designed experiment, some are just better than others” - G.E.P. Box Objectives of an Experimental Design Obtain the maximum amount of information using a minimum amount of resources Determine which factors (inputs) shift the average response, which shift the variability and which have no effect Find factor setting that optimize the response and minimize the cost Build empirical models relating the response of interest to the input factors Knowledge With Leverage Six Sigma Principles Define problem (Understand what is critical to the customer) Establish appropriate measurement systems (MSA) Identify possible factors (X’s) which have an effect on the output (Y) Execute Designed Experiments Screening Phase Characterization Phase Optimization Phase Implement process control using SPC Error-proofing Shainin Principles of ProblemSolving Strategy Select projects from a Rolling Top 5 Pareto Understand the Physics and Geometry of the process Move from Effect to Cause, not from Cause to Effect Find and Leverage Contrasts Apply Elimination Techniques DOE Terminology Inputs Controlled Experimental Factors May be set at various levels Held Constant Factors Noise Factors Outputs Response(s) Designed Experiment DOE Levels of a Factor Input-Output Relationships Relationships exist between the control factors and the response Factors or groups of factors may have different effects upon the response Main Effects Interactions Typical Spike Main Effect Relationship Main Effect A factor has a Main Effect if the response changes as you move through the range of the input No Interactions Occur when the effect of one variable (or factor) depends upon the value of another Typical Interaction Relationship Typical Interaction: T