An Algorithm for Identifying Texture Components in the Framework of Statistical Crystal Plasticity Models

An approach based on involving clustering techniques to identify texture components for a given polycrystalline aggregate is formulated. A heuristic iterative clustering algorithm operating only with a sample of lattice orientations and the concept of closeness between them is proposed for this task. The mentioned closeness concept is introduced via a special pseudometric distance, which is induced by the natural Riemann metrics and takes the lattice symmetry into consideration. The developed procedure allows representing an inhomogeneous orientation distribution by localized clusters in the space of orientations and thus can be used to reduce the dimension of orientation distribution data in statistical crystal plasticity models. Two examples of such an application of the clustering procedure are provided.