TAGUCHI’S CONTRIBUTIONS TO EXPERIMENTAL DESIGN & QUALITY ENGINEERING

Taguchi’s Contribution  In the early 1980s, Prof. Genechi Taguchi introduced his approach to using experimental design for 1)Designing products or processes so that they are robust to environmental conditions. 2)Designing/developing products so that they are robust to component variation. 3)Minimizing variation around a target value.

Taguchi’s Contribution  By robust, we mean that the product or process performs consistently on target and is relatively insensitive to factors that are difficult to control.

Taguchi Philosophy  3 stages in a product’s (or process’s) development: 1)System design: uses scientific and engineering principles to determine the basic configuration. 2)Parameter design: specific values for the system parameters are determined. 3)Tolerance design: determine the best tolerances for the parameters.

Taguchi Philosophy  Recommends: statistical experimental design methods have to be used for quality improvement, particularly during parameter and tolerance design phases.  Key component: reduce the variability around the target (nominal) value.  Imposes a quadratic loss function of the form: L(y) = k(y – T) 2

Taguchi Philosophy  Involves 3 central ideas: 1)Products and processes should be designed so that they are robust to external sources of variability. 2)Experimental design methods are an engineering tool to help accomplish this objective. 3)Operation on target value is more important than conformance to specifications.

Taguchi Approach to Parameter Design  There are controllable factors and uncontrollable noise factors that affect the system.  Aim: to find the levels of the controllable factors that are least influenced by the noise factors and yield the maximum (or minimum) response.

Taguchi Approach to Parameter Design  Select one experimental design for the controllable factors and another design for the noise factors.  Thus defines orthogonal arrays.  Inner array: array containing the controllable factors  Outer array: array containing the noise factors  Test each run from the inner array across the runs from the outer array called a crossed array design.

Taguchi Approach to Parameter Design  Then analyze the mean response for each run in the inner array.  Also analyze the variation by the use of appropriate signal-to-noise (SN) ratio derived from the loss function.

Taguchi Approach to Parameter Design  Widely used SN ratios are: 1)Nominal the best: used to reduce variability used to reduce variability around a target value around a target value 2)Larger the better: used to max. response used to max. response 3)Smaller the better: used to min. Response. used to min. Response.

Critiques  Ignores interactions  It leads to a very large experiment while interactions cannot be estimated.  In general: crossed array approach is often unnecessary. Better strategy: use a combined design that incorporates both controllable and noise factors.  Also there are critiques about the use of SN ratios (see Montgomery)