Dimensional variation propagation analysis in straight-build mechanical assemblies using a probabilistic approach

Product quality in mechanical assemblies is determined by controlling the propagation of manufacturing variations as the structure is built. This paper focuses on straight-build assembly and uses a probabilistic approach to analyse the influence of component variation on the eccentricity of the build. Connective models are used to predict assembly variations arising from individual component variations, and a probabilistic approach is used to calculate the probability density function (pdf) for the eccentricity of the build. The probabilistic approach considers three different straight-build scenarios: (i) direct build; (ii) best build; and (iii) worst build, for two-dimensional “axi-symmetric” assemblies. The probabilistic approach is much more efficient than Monte Carlo simulation. The paper also uses numerical examples to investigate the accuracy of the probabilistic approach in comparison to Monte Carlo simulation.

► A probabilistic approach is used to calculate the probability density function (pdf). ► The probabilistic approach considers three different straight-build scenarios. ► The proposed probabilistic approach yields accurate and efficient results. ► The proposed approach provides a valuable tool in assembly process design.