The need for geometrical variations management is an important issue in design, manufacturing and all other phases of product development. Two main axioms cover geometrical variations, namely the axiom of manufacturing imprecision and the axiom of measurement uncertainty. Therefore, this paper reviews common models for the description of non-ideal geometry (shape with geometric deviations) and shows how the random field theory can be applied to create more realistic skin models (a model which comprises these geometric deviations).
Furthermore, methods to estimate and to express the underlying random field from a sample population are shown. These can be used to create and simulate random shapes considering systematic and random deviations observed through measurement or gathered from manufacturing process simulations.
The proposed approach incorporates given information from manufacturing process simulations or prototypes. Based on these information, skin model samples are created which can represent the “realistic” part in assembly simulations or other geometrical analyses. This can help to identify the optimal tolerance sets within every stage of the product development process. The efficiency of the introduced approaches is shown in a case study.