Hi! I am interested in implementing Taguchi's method to reduce the number of cases to simulate with CFD in a multiphase problem but not sure if it would be feasible. Has anyone got any experience on this? Cheers Xavier

Hi

I've done it plenty of times with star-cd. It worked every time.

allan

Hi, And how do you do to determine the reduced array (for instance L8), do you use any especific sofware or did you create your own excel spreadsheet? do you have any recommendations of sotware, book, etc to understand how to calculate it? I am using FLUENT and want to use it in order to reduce the number of cases to simulate for multiphase flows for hydrodynamics. Cheers Xavi

Hi

I tried the Taguchi method in fluent and it failed to work, as every simulation I tried with different bounday conditions gave the same answer. I guessed that their code was a bit overdamped, or the proces was boundary condition independant (which was known not to be the case). However do not be discouraged. To determinbe the array size you need to know how many factors you think will effect your process and how many levels you can set them at. I chose three levels rather than two as this will determine if there are non-linearities or not. You also need to choose a response variable i.e. what you want to optomize. You can check out a website I created some years ago for textbooks:

http://www.swiftcfd.co.uk/

I didn't need any software to do this I did it by hand.

perhaps you could send me some more details to my email address, not on this form

regards

allan