SidyW - Simulation of dynamical resistances during the operation of construction and conveying machines

点击翻译按钮获取中文摘要

The handling of a manifold of goods is the task of many construction and conveying machines. The term 'good' stands here generally for granular materials that are moved during the actual working process of the construction machine or the regular transport on continuous conveyors. The interaction of machine and good characterizes the working and transport process.The increasing need for more energy efficiency, longer product life cycles and raised cost pressure initiates and drives the development of efficient cross-domain simulation methods. This desire for such simulation tools is reinforced by the fact that construction and conveying machines are often produced in small series or as unique specimen so that producing and testing prototypes is very uneconomical.For these reasons, the use of machine-process-simulations, as a tool of product development, has increased significantly in recent years. Particle-mechanical processes are thereby usually simulated with the Discrete Element Method (DEM), whose material parameters have to be determined in extensive calibration procedures. Calibration is done by comparing the results of experimental studies with those of a simulation series with varying parameters. This is called inverse parameter determination. So far most calibration is done with the target to consider a realistic flow behavior of granular materials. This is not sufficient for a realistic prospective load analysis.The aim of this project is therefore the development of a method for the automated parameterization of DEM models which can be applied for the simulation of dynamical resistances in mobile construction equipment and conveying machines. This is the missing link for a realistic assessment of process loads and motion resistances and enables simulation-based prospective analysis of such machines.The main part of research is required by developing an efficient optimization system on the one side and tools and methods for result extraction and post-processing on the other side. The so called optimization system includes all components which are needed to solve an optimization problem (initialization, optimization algorithm, objective function value calculation, etc.). Therefore, new approaches of intelligent initialization as well as the evaluation of multi-criteria objectives are studied. As optimization algorithms artificial neural networks and metamodel-based methods will be used. The calibration will be done by help of relatively simple calibration tests, while the validation of the calibrated models will be done by comparing simulation results and experimental data from large-scale experiments.