Chapter 24 Flashcards
Reason for DOE
Reduce the time to design/develop new products and processes
Improve performance of existing processes
Improve reliability and performance of products
Achieve product and process robustness
Perform evaluation of materials, design alternatives, setting component and system tolerances
Methods to understand the relationship between inputs and outputs
Modelling
- understand the process (inputs vs outputs)
Optimisation
- get the output that we want
Control
- what changes in input to move to the new target value
Nuisance inputs
Variables that can’t be influenced, but affects the output
Controlled inputs
Variation and replication
Use our control over them to vary the inputs and repeat the experiment in a systematic way
Called factors
Changed to specific ‘levels’
Uncontrolled but Observed Inputs (u)
Use Blocking
Groups experiments into blocks, each block having a fixed value
Uncontrolled and Unobserved Inputs (v)
Use Randomisation
When to use Fractional Factorial Design?
Used when time and resources are limited
Advantages: simplicity and economy
Narrowing the Xs down to direct focus on the vital few
Design consists of a fraction of the runs of a full factorial design
Typically used for screening experiments in order to identify a few important factors and the lower order interactions
Runs are chosen so that if some factors are confounded with others, the fractional factorial becomes a higher resolution design for remaining factors
Design on Experiments
A scientific process of planning an experiment that will yield statistically useful results
An experiment
the deliberate variation of one or more process variables while observing the effect on one or more response variables
An observation
We observe both the process variables and response variables
Goal of DoE
to get the most information from the least amount of data
Advantage of DoE
allows the experimenter to study the effect of many factors that may influence the product or process simulatanously
Types of DoE
1) Fractional Factorial DoE
2) Full Factorial DoE
3) Response Surface Modelling DoE
Fractional Factorial DoE
Screening DoEs
To find the key factors that affect a response
“Screen” many factors at one time in order to determine which are worthy of deeper investigation
Full Factorial DoE
Test factors across all possible combinations in order to determine which factors are statistically significant
Characterisation studies
Take longer to complete and cost more money
