Does any body know who to use Taguchi method for design of experiment?

Yes, Lots of people know how to do Taguchi DOEs.  

Designing, conducting and interpreting a DOE or any designed experiment can be very complicated.  There are way more ways to do it wrong than there are to do it right.  

First critical thing is what is the response or what are the responses you are going to measure.?  If you are going to the time, effort, and expense of a DOE, you want to measure as many responses as you can to gain maximum value.  

Second critical thing is how good are those measurements.  If you have not done so, it is highly recommended that you determine exactly how good (or bad) your measurement system is.  Do a GR&R or other statistical measurement system analysis. There is no reason to even start a DOE if your measurement system is not capable of repeatedly and reproducibly telling you the "truth".

I assume you are planning an experiment so you can change your process in some way.  You need to think this through so you understand exactly what the goal is.  Are you trying to move the average result in one direction or are you trying to keep the average result where it is and reduce the variation around that average or are you trying to both move the average and reduce the variation?

You also need to consider how the responses behave and what your "design space" is.  If the response/s and the interactions between input factors will always be linear, then you may not need to include center points.  If you don't know, it is recommended to include some conditions that will check for linearity.

Once you deal with these issues, then you can tackle the DOE design.  If you have never done a DOE, I suggest you start with a simple full factorial design with some replicates.

Good luck