in my taguchi design, a level of one factor is zero. This cause to obtain similar state in two runs. Suppose we have the following factors and levels:
levels...............1 ......... 2 .......... 3
factor A ......... 10 ....... 40 ........ 60
factor B ......... 0 ......... 5 ......... 15
in this example, changing the levels of factor A at the first level of factor B is unimportant. means A1B1 and A2B1 are the same runs (experiments).
Have anyone faced with this problem? If yes, please introduce me a reference.
Thank you so much
Dear sir,
Even in same run some possible response you will get..do it get the observation and conclusion ...
Dear Mina Sedighi,
I suggest you to see links and attached files in topics.
- 5.3.3.10. Three-level, mixed-level and fractional factorial designs
www.itl.nist.gov/div898/handbook/pri/section3/pri33a.htm
- Chapter 2 Introduction to Taguchi Method
www.ecs.umass.edu/mie/labs/mda/fea/sankar/chap2.html
- INTRODUCTION TO TAGUCHI METHOD - EE-IITB
https://www.ee.iitb.ac.in/~apte/CV_PRA_TAGUCHI_INTRO.htm
- Taguchi Orthogonal Array Designs
www.weibull.com/hotwire/issue131/hottopics131.htm
- Sustainable Manufacturing - Résultats Google Recherche de Livres
https://books.google.dz/books?isbn=1118622952
- Machining of Metal Matrix Composites
https://books.google.dz/books?isbn=0857299387
- Experimental Quality: A strategic approach to achieve and improve ...
https://books.google.dz/books?isbn=1461552931 -
Best regards
Dear Dr. Tamilarasan,
You mean I keep similar runs and add my responses identical for both?
Do you have any reference for this situation?
Thank you so much
Dear Sr. Lafifi,
Thank you for adding these reference files. I checked them but unfortunately, I could not find anything regarding my own situation. I will appreciate if you introduce a detailed topic.
Regards,
Mina Sedighi
Mina Sedighi: One approach is to *not* use zero as a level for Factor B, but rather some small, non-zero value. You did not tell us what your system is, but in chemical systems we never used zero as a level of a Factor because it could (and usually does) fundamentally change the system - the nature of the chemical reaction. This could also be the case for a non-chemical reaction system, a mixture or blending system for example.
Dear Dr. fischer,
Thank you for your reply. In fact, the level of mentioned factor is not zero. I have modified a metal organic framework using functional groups. So, the value of zero in this factor means MOF without functional groups. omission of this level cause to eliminate a number of desired states in MOF modification.
Regards
So Mina, if I understand you correctly, it appears the levels of Factor B are categorical, that the level of "zero" only means that the MOF without fictional groups is to be used in that run of the Designed Experiment? If that is the case, then use another (non-zero) designation of the level of Factor B.
I am not very familiar with Taguchi designs (the model associated to a given design is not explicit, so, statistical analysis can be not simple. It's why I prefer factorial or response surface designs). But, if I understand your problem, it's like if your 2 factors are, for example, the molecule weight of a product (4000 and 10000) and it's quantity (0% or 5%). Of course, 0% of MW 4000 or 0% of MW 10000 is the same experiment. In such a case, a solution is to combine the 2 factors in one with 3 levels: no product, 5%-4000, and 5%-10000.
Hope this can help you,
Pierre Chagnon
Dear Dr. Chagnon ,
Exactly. I think it is a good solution for a case with few factors. I have 5 factors for taguchi analysis. Eventually, I had to ignore the zero level, as Dr. Fischer suggested.
Thank you very much.
- Badges
- Science topic