Abstract

Virtual commissioning (VC) is used to test control code deployed on Programmable Logical Controllers. Simulation models of a plant are the core of any VC approach. Simula-tion models should represent the plant in a way so that the correct process execution can be tested under customers’ conditions. Simulation models of a plant are usually not built monolithically, but by many partial simulation models that represent the modules or components of the investi-gated plant. To ensure that the VC is efficient and provides helpful results, these partial simulation models can be implemented at different levels of detail, depending on the current test scenario. Usually, the definition of the mod-ules’ and components’ level of detail is fixed. However, situations exist where more than one level of detail can be adequate. A dynamically adaptable level of detail seems beneficial to e. g. keep computing time at a reasonable level and to ensure meaningful results of the plants simula-tion model.
However, no method or approach exists so far to handle a dynamically adaptable level of detail. The paper presents the research results of the authors on virtual commissioning and focuses on a simulation point of view and is organized as follows: In Section 1, a brief description is given on how to define the right granularity of simulation models used for virtual commissioning. Based on these results, several levels of detail and model types that can be used for a VC approach are introduced in Section 2. In Section 3, situations are described where more than one level of detail is suitable. In Section 4 and 5, potentials and challenges of a dynamically adaptable level of detail are dicussed and possible solution contributions that could yield benefits for a VC approach are shown.