Multi-level simulation concept for multidisciplinary analysis and optimization of production systems

Delbrügger, T.; Meißner, M.; Wirtz, A.1, a; Biermann, D.1, b; Myrzik, J.; Rossmann, J.; Wiederkehr, P.2, c

Institute of Machining Technology, TU Dortmund University, Baroper Str. 303, D-44227 Dortmund, Germany
Chair 14 for Software Engineering – Virtual Machining, TU Dortmund University, Otto-Hahn-Straße 12, D-44227 Dortmund, Germany

a); b); c)


In the context of digitization and industry 4.0, the production-related disciplines developed powerful simulation models withdifferent scopes and varying levels of detail. As these simulation systems are usually not built in a compatible way, themodels cannot be combined easily. Co-simulation techniques provide a promising basis for combining these models into onesuperordinate model and utilizing it for planning new factories, adapting existing ones or for production planning. However,today’s co-simulation systems do not benefit from the inherent flexibility of the represented production systems. Simulation-based optimization is carried out inside each discipline’s simulation system, which means that interdisciplinary, globaloptima are often impossible to reach. Additionally, the aspect of human interaction with such complex co-simulation systemsis often disregarded. Addressing these two issues, this paper presents a concept for combining different simulation modelsto interdisciplinary multi-level simulations of production systems. In this concept, the inherent flexibilities are capitalized toenhance the flexibility and performance of production systems. The concept includes three hierarchical levels of productionsystems and allows human interaction with the simulation system. These three levels are theProcess Simulationlevel,theFactory Simulationlevel, and theHuman Interactionlevel, but the concept is easily extendable to support additionallevels. Within the multi-level structure, each simulation system carries out a multi-objective optimization. Pareto-optimalsolutions are forwarded to simulations on higher hierarchical levels in order to combine them and meet flexibly adaptableobjectives of the entire production system. The concept is tested by means of a simplified production system, to optimize itin terms of throughput time and electric energy consumption. Results show that the presented interdisciplinary combinationof heterogeneous simulation models in multi-level simulations has the potential to optimize the productivity and efficiencyof production systems.


Virtual Testbeds, Process simulation, Factory modeling, Multi-level simulation, Co-simulation


The International Journal of Advanced Manufacturing Technology, 103 (2019) 2, pp. 3993-4012, doi: 10.1007/s00170-019-03722-1