Deterministische, dynamische Systemstrukturen in der Automatisierungstechnik

  • Deterministic, dynamic system structures for automation technology

Mersch, Henning; Epple, Ulrich (Thesis advisor); Wollschlaeger, Martin (Thesis advisor)

Als Manuskript gedruckt. - Düsseldorf : VDI Verlag (2016)
Book, Dissertation / PhD Thesis

In: Fortschritt-Berichte VDI. Reihe 8, Mess-, Steuerungs- und Regelungstechnik 1245
Page(s)/Article-Nr.: 1 Online-Ressource (XI, 131 Seiten) : 77 Bilder, 3 Tabellen

Dissertation, RWTH Aachen, 2015


For the future development of automation technology the enhanced collaboration of automation devices is important. This is true for all phases of a plant, from planning to production to maintenance, as well as for the horizontal and vertical integration during the production. A lot of current topics like "Industry 4.0" or "Cyber Physical Systems" act on the assumption that engineering information is available during the production phase, which is not the case today. The dissertation addresses this topic. Nowadays information is often stored not electronic analyzable (e.g. as graphic) or accessibleonly by single programs. So existing information is not accessible as far as it could be. Models are acting as an important part: They describe characteristics of a plant. Most of today’s models are defined in an electronically representable and evaluable format by their specification. So a computer could host and provide these models as well as evaluate and use them. If they are provided and used during the production phase, dynamical changes are made possible, which is not usually the case nowadays. Therefore, in automation technology models are used for the description of different topics. Topic-specific models are developed, that are independent from each other and describe different aspects of the domain of automation technology. In contrast attempts to describe the whole domain of automation technology in one model were not successful. Those models could not be widely established since their adoption could not be achieved. This could be due to the fact that models describing a whole domain ("World-Models") are not detailed enough to be used for specific cases. So specific models were required to be defined, which contradicts the purpose of whole domain model. This work describes instruments without a centralized model: Existing, heterogeneous models can be used in all phases and levels of a distributed system like a plant in a homogenous way. Therefore a distributed, dynamic, model-driven execution environment is described. This could be seen as a further development of existing technologies, but is described independent of those. In this model-driven execution environment parts of a model could be provided by distributed devices. In a common way models are discover- and query-able. Therefore, information could be stored in a judicious place, that is defined by most frequent usage, highest reliability or fastest availability. At the same time such a model-driven execution environment provides for the possibility of dynamic changes: Changes are established in a collaborative way from different applications. For that purpose transaction security as well as comprehensibility (determinism) needs to be assured. Additionally, this work describes the concept of interconnections of models: Parts of models, which are designed independent of each other, are put in relation. Interconnections are a special type of relations not having dependencies at start- or endpoint. This enables modeling of additional aspects, so interconnections of models are models again. This combination of instruments represents a distributed, deterministic and dynamic model execution environment of system structures. A requirement for this is a common used metamodel as well as a complete understanding of the topic. Models can be specified independent of each other. A model execution environment will provide interfaces for querying information and for making changes to all models. Applications will be based on this. The overall result is a solution, which makes the use of models feasible during all phases of a plant - especially during production time. The integration of the distributed models provides an information-technology foundation for dynamic changes on system architecture. This covers changes of plants due to rebuilding as well as production orders.