Combined Geometric and Kinetic Data Model in Model-Based Systems Engineering of Robotic Cells

摘要

The main intention of the presented research is the development of an engineering framework that allows the automated generation of a combined geometric and kinetic digital twin of a robotic cell. The engineering framework is based on graph-based design languages and an appropriate compiler for their translation in order to realize a novel form of a machine-executable V-model for Model-Based Systems Engineering (MBSE). In this novel machine-executable MBSE process, the primary process management objects are no longer some documents but abstract process descriptions that can be automatically compiled into concrete product models. For such a holistic MBSE process, the models have to be enriched with physical behavior and performance information. The method is illustrated with a use-case of the engineering process of robotic cells. The starting point of the automated synthesis of the robotic cell with its resources is the abstract modelling by means of object-oriented programming. The use-case illustrates a pick-and-place operation that requires a specific geometry of a monolithic gripper. The gripper geometry synthesis is based on design automation with topology optimization, which is also realized inside the executable V-model. A large number of synthesis results from the geometry generating processes can be included in automated kinetic simulations, thus generating digital twins for a large solution spectrum to achieve an overall holistic optimization. Further, a digital twin created from geometrical data and kinetic simulation enables process monitoring based on physical values (joint forces, pressure, etc.) which then can be used to predict failure or evaluate new/optimized motion sequences.