Designing optimal component tolerances of a mechanical assembly by incorporating Taguchi's view of quality

In the past four decades, a substantial number of researchers have been trying to apply cost-tolerance functions to search for minimum-cost tolerance allocations in a mechanical assembly. However, most of the investigations have not taken quality cost into account, resulting in less than optimal tolerances. Therefore, the primary objective of this research was to examine the effect of tolerance allocation on the quality evaluation of a mechanical assembly and to introduce a method for the economical design of product tolerancing. It was shown that component quality did play a vital role when allocating assembly tolerance to its individual components.; The assembly total cost function, combining cost-tolerance and quality loss functions, was shown to be strictly convex. Therefore, any local minimum in a nonlinear assembly total cost function is guaranteed to be a global minimum over the feasible region. The determination of optimal component allowances, however, does not always require the use of nonlinear programming. An algorithm was developed to determine component optimal allowances that satisfy worst-case or statistical tolerance constraint.; The limitations of traditional quality viewpoint in selecting a tolerance constraint were identified, and it was shown that the above algorithm also incorporates Taguchi's quality viewpoint in selecting the most economical tolerance constraint.; The importance of variance reduction in lowering of component quality loss has been known for a long time. It was shown that more component variance reductions result in more savings as measured by quality loss. However, variance-reduction costs must be taken into account so that proper actions can be taken where net savings can be realized. An example provided a step-by-step sequence for parameter design implementation that shows parameter design is cost-effective in reducing assembly variance.