Automatic Generation of Simulation Models for Digital Twins from State-of-the-Art Simulation Frameworks

Abstract

Simulations are an important part of Digital Twins due to their ability to forecast behavior of a physical asset or perform what-if analysis. Especially for products with long life cycles, there is a lack of sufficient data for the generation of such simulations and information has to be digitized in tedious manual work and adapted to state-of-the-art frameworks. Thus, there is the urgent need for automation of this process including the digitization and generation of simulations using these frameworks. In this work, we propose a workflow that is able to completely automate these steps by combining different digitization methods for in situ measurement data and engineering diagrams into one generalized system description form. From this system description a set of multiple simulations from different frameworks can be generated automatically. The workflow is successfully demonstrated with the example of two different frameworks for pedestrian flow simulations, one agent-based and one using finite element methods.