Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
29 Cards in this Set
- Front
- Back
In a Design of Experiments (DOE), what is the effect of aliasing?
|
You cannot distinguish between effects of different runs. (Also called totally counfounded)
|
|
What process is followed when designing experiments?
|
The 12-step process. It gives a roadmap and questions to answer as we proceed through the stages of experiments
|
|
What is an orthogonal design?
|
A design (test) matrix which has both vertical and horizontal balance.
|
|
What are the five steps in Design of Experiments (DOE)?
|
Planning
Designing Conducting Analyzing Confirming |
|
What are the two main STRATEGIES (or uses) for Design of Experiments (DOE)?
|
Screening: when a large number of factors (six or more) to narrow focus to the critical few
Modeling: for fewer factors and a complete analysis |
|
What does the R-squared (R2) measure?
|
The proportion of variation in the data (Y) that is explained by the model (X).
|
|
When confirming, how close should we be to the predicted value?
|
Within +/-3 standard errors
|
|
In a regression output, what should our F Value be?
|
Greater than six
|
|
What is the definition of Design of Experiments (DOE)?
|
The purposeful changing of the inputs to a process in order to observe corresponding changes to the output.
|
|
In Design of Experiments, what does the K Value represent?
|
K is the number of input factors.
|
|
If your DOE model contains A, B, and C factors, each with a high and low value, with all possible combinations run, how would you descrie it?
|
A two-level, three-factor, full factorial experiment
|
|
In an interaction graph, factors that don't influence the output will be:
|
Parallel to each other
|
|
What does a Y-hat experiment predict?
|
the affect the factors will have on the mean (shift)
|
|
What does an S-hat experiment predict?
|
the affect the factors will have on the standard deviation (squeeze)
Look at this factor first! |
|
Which tool would you use in SPC XL to find the area of a section beneath a normal distribution (bell curve)
|
Continuous Distributions > Normal
(Inverse normal finds the percentile of a given result) |
|
In DOE, what does the Rule of Heirarchy state?
|
If an interaction is significant, the individual factors of that interaction cannot be omitted, even if they are not significant (p-value) individually.
|
|
In fractional factorials, what is the minimum resolution that should be considered effective
|
Five
|
|
How often will two-way interactions have significant influence in a DOE? Three-way?
|
Two-way: approximately 20%
Three-way: approximately 1% This is significant when considering which columns are aliased in a fractional factorial DOE. |
|
How do you determine if a DOE is HORIZONTALLY balanced?
|
It is horizontally balanced if the multiple of any two columns results in zero
|
|
In a DOE, if you have no p-values in the 's-hat' portion of your regression analysis, how can you determine which factors are significant to variation?
|
Use the Constant Coefficient (s-bar) and divide by two. Compare to all factors and interactions. If they are greater than 1/2 s-bar, they are significant.
|
|
How do we measure the degree of confounding in a model?
|
The tolerance scale.
A tolerance of one means not confounded; zero means totally confounded. (all values should be >/= .5, half should be >/= .7) |
|
How do you create a Y-hat prediction model?
|
Y-hat = Y-bar + (coeff)Z + ..., where z equals each significant factor.
|
|
How do we measure the degree of confounding in a model?
|
The tolerance scale.
A tolerance of one means not confounded; zero means totally confounded. (all values should be >/= .5, half should be >/= .7) |
|
How do you create a Y-hat prediction model?
|
Y-hat = Y-bar + (coeff)Z + ..., where z equals each significant factor.
|
|
How do you determine the Overall Equipment Effectiveness (OEE)?
|
Multiply:
% equipment availability X performance rate X quality rate X 100 |
|
What is the definition of flow?
|
The continuous progressive adding of value.
Starts at receipt of request, ends at delivery to customer |
|
In a Value Stream Map, how is PROCESS lead time defined?
|
The sum of all cycle time for the process.
|
|
What are the five steps in Value Stream Mapping?
|
1. Identify customer requirements
2. Peform an upstream walk 3. Record information in the data boxes 4. Add information flow data 5. Add timeline information. |
|
What are geometric designs and how are they used?
|
designs that follow the n=2^k pattern. They're used for modeling.
Non-geometric designs are used for screening. |