Robust Tolerance Synthesis With the Design of Experiments Approach

Abstract

Design of tolerances impacts quality, cost, and cycle time of a product. Most literature on deterministic tolerance design has focused on developing exact and heuristic algorithms to minimize manufacturing cost. Some research has been published on probabilistic tolerance synthesis and optimization. This paper presents the design of experiments (DOE) approach for concurrent selection of component tolerances and the corresponding manufacturing processes. The objective is to minimize the variation of tolerance stackups. Numerical examples illustrate the methodology. The Monte Carlo simulation approach is used to obtain component tolerances and tolerance stackups. Process shift, the worst case and root sum square tolerance stackup constraints, and setup reduction constraints have been incorporated into the proposed methodology. Benefits of the proposed DOE approach over exact algorithms are discussed. [S1087-1357(00)00202-1]