Parametric, asymmetric and stochastic-based 3D CAD model of Tonda Gentile Trilobata hazelnut variety

•A hazelnut shell-kernel model in a feature-based 3D CAD environment is presented.•The aim for these models is to improve production processes machinery design.•Three different approaches with increasing complexity and accuracy are analysed.•A conceptual model and two alternative symmetric and asymmetric models are used.•The resulting models are compared to actual hazelnut samples.

A family of parametric 3D CAD models based on stochastic distribution of not-engineering real life object has been studied. The subject considered was the Tonda Gentile Trilobata (TGT) hazelnut variety, which is the most important variety in the Langhe region in Piedmont, Italy. The aim was to obtain a master model which measured, or stochastically-obtained, values of parameters given as input, to generate a number of samples which have the same statistical distributions of the real world ones. These generated models could then be used in virtual simulations of machinery and equipment used by industries to process the hazelnut, making possible studies of modifications to improve the production process performance. A single parametric CAD model was considered with respect to several scanned models from real hazelnuts, with the possibility of generating families of solutions suitable for the simulation aims. The work was carried out in steps of increasing complexity checking accuracy using different types of comparisons to validate the different models.