Description
Simulations are an important part of modern product development. These lead to shorter and more efficient development cycles and can reduce high costs for complex tests. One of the biggest challenges for good results in simulation applications is modelling. The ideal CAD model is always used for geometric analyses. However, due to unavoidable inaccuracies in the manufacturing process, this model is subject to deviations which falsify the results of the simulations. This thesis deals with the topic of how geometric manufacturing deviations can be integrated into the simulation process in order to obtain models that are as accurate and fit for simulation as possible and thus improve the quality of the results.
As a methodical foundation, a layer-model is presented, which guides the product developer to evaluate a comparison of nominal and actual geometry with regard to its significance for a simulation. The actual state is generated with high-resolution digitization techniques. Based on this comparison, a decision is forced on the user as to whether a simulation carried out with an ideal CAD model is sufficient. The knowledge elements provided at each layer can be used to answer specific questions in order to assess the significance of the deviations that occur and to evaluate their relevance to simulation results.
If a simulation with real geometric elements is recommended on the basis of this process, four methods for the generation of simulation-capable models have been developed and presented for this purpose: Parametric correction of the CAD model for simple deviations; the complete but complex process of Reverse Engineering (RE); generation of hybrid models consisting of CAD geometry and scan inserts from RE; adaptation of FE meshes to the scanned data sets (point clouds and polygonal models). All process models shown are carried out for verification with application examples from industry and practice. The simulation results were refined and brought closer to the real conditions.
The presented model offers an approach to the improvement of simulation models by integrating production deviations. An even earlier use in the development cycle should be the goal in order to reduce costs and can be achieved by using process simulations if these deliver reliable results. Since the data structures of the scans and process simulations are similar, the methods presented here can be used in the same way. The approach presented here for corrected, more realistic geometries can also make an important contribution outside the simulation application, for example within the context of a digital twin.
Reviews
There are no reviews yet.