Regelbasierte Modelltransformation in prozessleittechnischen Laufzeitumgebungen

Mersch, Tina; Epple, Ulrich (Thesis advisor); Schürr, Andy (Thesis advisor)

Düsseldorf : VDI Verlag GmbH (2017, 2018)
Book, Dissertation / PhD Thesis, Report

In: Fortschritt-Berichte VDI : Reihe 8, Mess-, Steuerungs- und Regelungstechnik 1261
Page(s)/Article-Nr.: 1 Online-Ressource (VIII, 152 Seiten) : IllustrationenReport number: Reihe 8 . Nr. 1261

Dissertation, RWTH Aachen University, 2017


A survey of 1800 members of the German engineering association VDMA [Sch12] showed that 61% of development and design engineers surveyed don’t feel up to the challenges in the development of new machines and equipment. More than one half of the respondents expect that the development of efficient, powerful and flexible machines will become even more complex in future. In addition, they predict that users and operators will become less qualified. This concurrence of increasing complexity and decreasing knowledge demands new methods in the engineering of plants. Anne Schneller, author of the VDI article about the survey, suggests that automation technology has to progress toward "parameterization instead implementation" to cope with these challenges. The work at hand aims to contribute to make this paradigma applicable for complex automation functions. The approach presented in this work takes advantage of the strong correlation between different models of plants. It uses model transformation as the basis of a system-neutral development of automation functions. Those automation functions can be parametrized with the plant specific planning data without further coding costs. The knowledge about the model correlations is stored once as rules, available for a large set of plants. The method of rule-based model transformation has its roots in computer science, where the approaches already developed achieved significant results in the model worlds created for them. Despite of many years of progress in the field of model transformation, automation technology presents special challenges for adoption. In particular, semiformal model descriptions and the variances in modelling as well as multiple source models make the use of standard methods impossible. Not only are the high demands of computer science on formalization and the fit of the participating models a hurdle in merging the two disciplines but the conservative attitude of automation technology brings particular challenges with it as well. One requirement for the acceptance of these approaches in automation technology is the smooth integration of the concepts in the application domain without ignoring the domain experts. The concept presented realizes one of the most promising approaches from computer science and integrates it seamlessly into automation programming. With this approach the installation technician can combine model transformation and standard programming in accordance with his purposes.