Methods for Localization in Multi-Scale Material Flow Simulation for Virtual Commissioning

While the importance of virtual commissioning is increasing and not only individual machines but also production lines are being simulated, the simulation of material flow for numerous piece goods is still a challenge. Combining models in a multi-scale simulation can exploit their advantages. The macroscopic flow model excels at modeling the behavior of large quantities of piece goods, however, it cannot provide information about individual piece goods. This led to the approach of the combination with a microscopic physics-based model in a multi-scale material flow simulation. To switch between models, the coordinates of the positions (microscopic) have to be translated in one direction into the material description of the macroscopic flow model (the density distribution). In the other direction, the density distribution has to be translated into the coordinates of the positions, leading to a computationally intensive optimization problem. This paper develops various methods for finding a start solution and compares them to a method of the state of the art. The evaluation is based on their accuracy with reference to various parameters of the data set. The results show the best accuracies for the method which traverses the density distribution and calculates the sum of the densities until it matches a single piece good. Depending on the accuracy required, it may be possible to use the start solution directly without further optimization.