Questions about Taguchi style Designed Experiments (DOE)

Have some question I have to answer for uni so any help greatly appreciated.
1. In a taguchi style designed experiment a confirmation run is necessary to,
a) identify secondary control factors
b) confound the control factors
c) disaggregate any interactions present
d) check the accuracy of the prediction

2. Taguchi style designed experiments are called
a) full factorials
b) fractional factorials
c) EVOP
d) response surfaces

thanks in advance to anyone who can help.
Dom

[Claes Gefvenberg]

Hello Dom,

I am not very good at Taguchi, but while we wait for someone who is, I wish you welcome to the Cove

/Claes

[Wes Bucey]

Originally Posted by Claes Gefvenberg

Hello Dom,

I am not very good at Taguchi, but while we wait for someone who is, I wish you welcome to the Cove

/Claes

s
Nutek, Inc. Quality Engineering Seminar, Software, and Consulting ( Since 1987) offers this description (found at http://www.rkroy.com/wp-sps.html):
Standardized Application Steps for Taguchi Experiments
PLAN – Experiment planning is the necessary first step in conducting experimental investigation. The problem solving team spends a day with experienced facilitator (consultant) to determine parameters and scopes of the study.
PRESCRIBE – Based on the discussions in the planning session, an experiment is designed. The prescription also includes individual test recipes and the data collection schemes.
PERFORM – Experiments are carried out as per the recipes and the results recorded.
PREDICT – Results of the experiments are analyzed and the condition for best solution is identified. Improvement expected from the new solution is also estimated.
PROVE - The predicted solution is tested with additional tests and prediction confirmed before changes are incorporated.

Nutek uses the term "full factorial" but I doubt they use it, since Taguchi's innovation had to do with fractional factorials - actually VERY fascinating to know and to use.

my advice:
Google the following search terms:
taguchi +experiments
taguchi +factorials

Comment:
Personally, I dislike this type of test or homework question from a professor because having the answer does not imply you have a clue what the answer means. If I were a student, I would want the professor to help me learn the difference between full and fractional factorials. I would want to know "why" I should go into proving or confirming my experimental results before committing resources to full production.

I know the answers on Taguchi, but I firmly believe in the philosophy
"Give a man a fish, you feed him for a day;
Teach a man to fish, you feed him for a lifetime."

(When you pay for schooling, you pay for knowledge, not the grade.)

Oh, yes. Welcome to the Cove, a place where we help you learn, but rarely just give simple answers you can get from an encyclopedia.

[Tim Folkerts]

Dominic,

Following up on what Wes said, simply getting an answer to a multiple choice question won't really help you understand. I think you'll find that most people here are happy to discuss topics, but unwilling to simply do your homework .

As a compromise, why don't you tell us what you think? Explain what you understand about the question and the different possible answers. Then we will know where you are coming from and perhaps someone can provide you with some real knowledge.

Oh, and welcome to the Cove. If you stick around and get involved, you will discover that there are a lot a people here with a lot of first-rate knowledge covering just about every facet of quality.


Tim F

For the second question I think the answer is fractional because one has no effect on the other.
But for question one i really dont have the faintest idea. I understand everyones comments but I have notes from all his lessons and half the questinos he did not even talk about.
So im finding it a challenge.
I know I should go and talk to him, but he is never around.
Please confirm if my answer to question 2 is correct or not.
thankyou

[Wes Bucey]

Originally Posted by Dominic

For the second question I think the answer is fractional because one has no effect on the other.
But for question one i really dont have the faintest idea. I understand everyones comments but I have notes from all his lessons and half the questinos he did not even talk about.
So im finding it a challenge.
I know I should go and talk to him, but he is never around.
Please confirm if my answer to question 2 is correct or not.
thankyou

If you are still guessing for Question #2, you need to read some more outside material about Taguchi and his "shortcuts" for design of experiments (DOE) - do the Google thing as I suggested. Especially make sure you get a feel for what is meant by a "Latin Square." It may take some time, but ultimately it will be more valuable to you than us giving you the answer.

Even with the minimum information contained in my first post, you should be able to get the answer to Question 1.

[alfawei]

I like this words:
When you pay for schooling, you pay for knowledge, not the grade

these questions were found in China's Sixsigma Book

Originally Posted by Wes Bucey

s
Nutek, Inc. Quality Engineering Seminar, Software, and Consulting ( Since 1987) offers this description (found at http://www.rkroy.com/wp-sps.html):
Standardized Application Steps for Taguchi Experiments
PLAN – Experiment planning is the necessary first step in conducting experimental investigation. The problem solving team spends a day with experienced facilitator (consultant) to determine parameters and scopes of the study.
PRESCRIBE – Based on the discussions in the planning session, an experiment is designed. The prescription also includes individual test recipes and the data collection schemes.
PERFORM – Experiments are carried out as per the recipes and the results recorded.
PREDICT – Results of the experiments are analyzed and the condition for best solution is identified. Improvement expected from the new solution is also estimated.
PROVE - The predicted solution is tested with additional tests and prediction confirmed before changes are incorporated.

Nutek uses the term "full factorial" but I doubt they use it, since Taguchi's innovation had to do with fractional factorials - actually VERY fascinating to know and to use.

my advice:
Google the following search terms:
taguchi +experiments
taguchi +factorials

Comment:
Personally, I dislike this type of test or homework question from a professor because having the answer does not imply you have a clue what the answer means. If I were a student, I would want the professor to help me learn the difference between full and fractional factorials. I would want to know "why" I should go into proving or confirming my experimental results before committing resources to full production.

I know the answers on Taguchi, but I firmly believe in the philosophy
"Give a man a fish, you feed him for a day;
Teach a man to fish, you feed him for a lifetime."

(When you pay for schooling, you pay for knowledge, not the grade.)

Oh, yes. Welcome to the Cove, a place where we help you learn, but rarely just give simple answers you can get from an encyclopedia.