A novel tolerance analysis method for three-dimensional assembly

Abstract

Three-dimensional tolerance analysis is increasingly becoming an innovative method for computer-aided tolerancing. Its aim is to support the design, manufacturing, and inspection by providing a quantitative analysis of the effects of multi-tolerances on final functional key characteristics and predict the quality level. This article proposes a novel approach for three-dimensional assembly analysis—a hybridization of vector loop and quasi-Monte Carlo method. The former is used to establish the three-dimensional assembly chain and obtain the assembly function. The latter is adopted to generate n sets of dimensional values according to the distribution of each dimension in chain. The new method is shown to inherit many of the best features of classical vector loop and quasi-Monte Carlo, combining easy-to-obtain assembly function with accurate statistical analysis. For every set of dimensional values, one sample value of a functional requirement can be computed with the Newton–Raphson iterative procedure. A crank slider mechanical assembly is shown as an example to illustrate the proposed method.