Monte Carlo Simulation of the Mechanical Processing of Bulk Materials with Fluctuating Compositions - Compositional Probability Density

Compositions are omnipresent in process engineering, and often directly interact with process performance. In this work, a Monte Carlo based approach for considering their impact in the case of fluctuating compositions is presented. Based on experimental data, the first component – a model for the compositional probability density is targeted. A variety of descriptive approaches that involve parametric and non-parametric density estimations is discussed. Finally, for a D-dimensional composition, a multivariate Gaussian kernel density estimation for a bijective projection on D-1 linearly independent statically constrained coordinates in combination with boundary reflection is presented as a suitable approach.