Tolerance allocation is a design task with a strong potential impact on manufacturing choices. In practice, however, it is often carried out with simple heuristics rather than with an optimization approach like those available in research literature. One reason could be the difficulty in predicting the economic benefits resulting from optimization. To allow for such considerations, the paper proposes a procedure to estimate the cost reduction that optimization allows compared to three traditional allocation methods (equal tolerances, precision factor, proportional to nominal). The chosen optimization method is based on the closed-form solution of a problem of cost minimization with a stackup constraint, using the extended reciprocal power cost-tolerance function. Compared to other methods, it provides analytical expressions of both the allocated tolerances and the associated costs. When applied to specific cases, these help recognize the conditions in which optimization allows a significant reduction in manufacturing costs. The results show that this occurs when the features of the same dimension chain have very different properties regarding a set of design variables with particular influence on the amount of machining required.
Read full abstract