One of the aim of statistical tolerance analysis is to evaluate a predicted quality level in the design stage. A method consists in computing the defect probability D P expressed in parts per million (ppm). It represents the probability that a functional requirement will not be satisfied in mass production. This paper focuses on the statistical tolerance analysis of over-constrained mechanism with gaps. In this case, the values of the functional characteristics depend on the gap situations, and are not explicitly formulated as a function of part deviations. To compute D P , two different methodologies will be presented and confronted. The first one is based on an optimization algorithm and Monte Carlo simulations. The second methodology uses system reliability methods. The whole approach is illustrated on a basic academic problem inspired by industrial interests.
展开▼
