Some Hints for the Correct Use of the Taguchi Method in Product Design

The paper discusses the problem of the correct identification of the Objective Function and the associated SNR function that designers must choose when employing the Taguchi method in product design, considering this step as the basic element to quantify the uncertainty of the device performance prediction. During product design, when many design aspects must still be understood by the design team, it is important to identify the most suitable “loss function” that can be associated with the characteristic function. The second step considers the variability of the characteristic function. The Taguchi method considers many Signal to Noise Ratio functions whereas in the paper the use of a unique function is suggested for all kinds of loss function. The discussion is argued in the context of so-called parameter design, with the perspective of identifying the best ranges of variation of the parameters that designers have identified as influential on the characteristic function, and also to adjust those ranges in order to obtain twofold results: reduce Bias between the mean value of the characteristic function response and the target value; obtain less variability of the characteristic function. The discussion of a case of study will point out the approach and the use of a unique Noise Reduction function.