Free Lean Six Sigma Black Belt Improve Phase: DOE Questions and Answers — Questions and Answers
Question 1: A Black Belt is designing an experiment to optimize a chemical process with five factors. Due to budget and time constraints, running a full factorial (2^5 = 32 runs) is not feasible. The team needs to identify the most significant factors and two-factor interactions quickly. Which of the following is the most appropriate initial approach?
- A full factorial design to ensure all possible interactions are studied.
- Response Surface Methodology (RSM) to find the optimal process settings.
- A fractional factorial design to screen for the vital few factors and interactions with fewer runs. (Correct answer)
- A one-factor-at-a-time (OFAT) experiment to simplify the analysis.
Correct answer: A fractional factorial design to screen for the vital few factors and interactions with fewer runs.
A fractional factorial design is the most suitable approach when there are many factors to investigate and resources are limited. It allows for the efficient screening of main effects and some interactions by running a carefully selected subset of the runs from a full factorial design, which helps in identifying the most influential factors for further study.
Question 2: In a Design of Experiments (DOE), when the effect of one factor on the output response is dependent on the level of another factor, this phenomenon is known as:
- A main effect
- Confounding
- An interaction effect (Correct answer)
- Replication
Correct answer: An interaction effect
An interaction effect occurs when the influence of one factor on a response variable is affected by the setting or level of another factor. This is a crucial concept in DOE, as analyzing only main effects can lead to incorrect conclusions if significant interactions exist. Interaction plots are often used to visualize this, where non-parallel lines suggest the presence of an interaction.
Question 3: A project team adds several center points to their 2-level factorial design. What is the primary reason for including these runs?
- To increase the number of factors that can be studied in the experiment.
- To ensure the experiment is run in a random order.
- To estimate the experimental error with fewer replicated runs.
- To check for curvature in the relationship between factors and the response. (Correct answer)
Correct answer: To check for curvature in the relationship between factors and the response.
Center points are experimental runs where all factor levels are set halfway between the high and low settings. A primary reason for including them in a 2-level factorial design is to test for curvature. If the average response of the center points is significantly different from the average response of the factorial points, it indicates that a non-linear relationship exists, and a simple linear model is not adequate.
Question 4: Which of the following describes a situation of 'confounding' in a fractional factorial experiment?
- The effect of a main factor cannot be distinguished from the effect of a two-factor interaction. (Correct answer)
- The experiment produces a non-linear response surface.
- The process output is influenced by uncontrollable 'noise' variables.
- The measurement system variation is too high to trust the results.
Correct answer: The effect of a main factor cannot be distinguished from the effect of a two-factor interaction.
Confounding, also known as aliasing, is a key characteristic of fractional factorial designs. It occurs when the effects of two or more factors or interactions cannot be separated from each other. For example, a main effect might be confounded with a two-factor or higher-order interaction, meaning the analysis cannot distinguish which one is causing the observed change in the response.
Question 5: A Black Belt has completed a screening experiment and identified the two most critical factors influencing the process output. The next goal is to find the precise settings for these factors that will maximize the output. Which DOE technique is most appropriate for this optimization step?
- A new fractional factorial design with higher resolution.
- A Taguchi design to minimize variation.
- Response Surface Methodology (RSM). (Correct answer)
- A full factorial design with only two levels.
Correct answer: Response Surface Methodology (RSM).
Response Surface Methodology (RSM) is a collection of statistical and mathematical techniques used for modeling and analyzing problems where a response of interest is influenced by several variables. Its primary purpose is to optimize this response. After a screening experiment identifies the key factors, RSM is used to explore the relationship between the factors and the response in more detail, often to find the factor settings that result in a maximum or minimum response.
Question 6: When analyzing the results of a two-level full factorial experiment, what does a large 'main effect' for a specific factor indicate?
- The factor has a strong interaction with another factor.
- The relationship between the factor and the response is non-linear.
- Changing the factor from its low level to its high level causes a significant change in the average response. (Correct answer)
- The factor's effect is confounded with a block effect.
Correct answer: Changing the factor from its low level to its high level causes a significant change in the average response.
The main effect of a factor is the average change in the response variable produced by a change in the level of that factor. A large main effect signifies that altering the factor from its low setting to its high setting has a substantial impact on the process output, making it a statistically and practically significant variable.
Free Lean Six Sigma Black Belt Improve Phase: DOE Questions and Answers
A Black Belt is designing an experiment to optimize a chemical process with five factors.
Due to budget and time constraints, running a full factorial (2^5 = 32 runs) is not feasible.
The team needs to identify the most significant factors and two-factor interactions quickly.
Which of the following is the most appropriate initial approach?