Optimal Tolerance Allocation for Tolerance Stack-Ups
Abstract The allocation of individual tolerances that form critical stack-ups is an important task in mechanical design. It is desirable, but difficult in practice, to allocate tolerances to obtain all required stack-ups at minimum cost. A minimum-cost allocation method is proposed here that works for both a single tolerance stack-up and for multiple tolerance stack-ups that share one or more individual tolerances. Tolerances can be optimally allocated for both worst case and a variety of 6σ statistical cases. The method is applicable to one-dimensional stack-ups and to multi-dimensional stack-ups with known sensitivity functions. It is a numerical Lagrange multiplier method that is more general than the Lagrange multiplier methods that have often been proposed. The basic method will almost always provide the lowest cost result when the manufacturing process to produce each toleranced dimension has been firmly established in advance. An exact method for efficiently extending the basic method to determine the lowest cost process for producing each dimension is also introduced.