Dynamic Response Analysis – “How to analyze 'non-linear ideal function”

[Ranjit Roy]

Hello sasun,

I’m glad you had the exposure to presentation by a leading expert on
the concept of Dr. Taguchi’s DYNAMIC CHARACTERISTIC. As perhaps you
already realize, it is relatively a difficult concept, but has high
return on investment. Naturally, you will need to spend some good deal
of time preparing and researching the background.

Considering the scope of this forum, I will help with some literature
that is part of my seminar handout. Please download and study document
on ROBUST DESIGN (Item# 1.1 Robust Design:DOE-II_Mod0-9_Free_Nutek in
http://nutek-us.com/wp-free.html ). You may also find some literature
covering the theory in GOOGLE SEARCH.

Further, you should download our free Qualitek-4 DEMO (www.nutek-
us.com/wp-q4w.html ) and review/analyze results of the example
experiment indicated below.
- Open file DC-AS400.Q4W experiment file
- Review experiment formulation carefully
- Analyze results by selecting DYNAMIC CHARACTERISTIC
- Scrutinize how results are treated and S/N calculated
- Etc.

In regards to formulating your own experiments, I assume you are after
ENGINE NOISE as response VS. ENGINE SPEED as input. To study non-
linearity, you must consider 3 or more levels of input and have
several noise (to the system) levels. This should make, a larger
number of results in each trial condition. (Say 3 signal x 4 noise
levels x 2 repaets = 24 sample results in each trial condition). Of
course, the total results would depend on the array size used for your
experiment design.
Good luck with your research project.

- Ranjit Roy, Nutek Inc. 11-0127

[lrsm...@peoplepc.com]

It is difficult to comment on this without more information. I have in the past had success with:

. If it makes sense, break the non-linear function into segments and look for factors that affect each segment.

. Define the ideal function and use the differences between actual performance and ideal as a response. Seek to minimize the differences.

In addition to Ranjit's comments, hope this helps!

Larry R. Smith
Sent from my BlackBerry® wireless handheld

-----Original Message-----
From: Ranjit Roy <rkr...@gmail.com>
Sender: tda...@googlegroups.com
Date: Thu, 27 Jan 2011 11:44:24
To: Taguchi DOE Application Discussion Group (TDA-DG)<TDA...@googlegroups.com>
Reply-To: tda...@googlegroups.com
Cc: <c.s...@tvsmotor.co.in>
Subject: Dynamic Response Analysis – “How to analyze 'non
-linear ideal function”