Operation-parallel adaptation of a co-simulation for discrete manufacturing plants

The role of simulation in the engineering and operation of modular machines and plants is becoming increasingly important. Still, both the creation and the maintenance of the models require a considerable manual effort. To address this problem, previous publications introduced an assistance system for the automatic composition and configuration of co-simulations, focusing on models for the commissioning phase. In this paper, the authors extend the concept of the assistance system to the operation phase. The goal is to optimize the simulation with respect to accuracy and runtime. Evolutionary algorithms are used to generate different model configurations, which fulfill both goals to different degrees, and suggest a suitable variant to the operator. The concept has been prototypically validated using an industrial application example.