Model-Based Validation of Industrial Control Systems

TLDR

Industrial applications increasingly require complex, safe, trustworthy control systems with high flexibility and reusability. The paper seeks to improve engineering processes by enabling collaborative design through a model‑based approach that defines domain languages and validates designs for industrial control systems. It introduces domain‑specific models—CAEX, PLCopen, and MathML—created by experts and integrated via model transformations to support tool collaboration. Model checking guarantees specification correctness, while model transformations uncover design errors early in the development cycle.

Abstract

Current industrial applications demand the design of more and more complex, safe and trustworthy control systems exhibiting a high degree of flexibility and reutilization. To achieve this, the engineering process should be improved by making the engineering tools involved in the development process to collaborate during the design. This paper presents a model-based approach for designing complex automation applications. The core of the approach is constituted by a set of domain specific models that depend on the application field and whose elements, syntax and semantics are defined from the point of view of the experts that participate in the design of the system. The domain models are defined using engineering tools as the design progresses and they can be used to achieve tool integration through model collaboration. This can be achieved following the Model Driven Engineering approach by means of model transformations. This paper specifically focuses on the first step of this paradigm: the definition of domain languages, in this case for industrial control systems, as well as validation mechanisms of application designs coming from different domain tools. Three well known and widely used industrial standards have been used: Computer Aided Engineering eXchange (CAEX), PLCopen (a representation format for the IEC 61131-3 standard) and MathML (a language for defining mathematical constraints). Using model checking it is possible to assure the correctness of the control system specification and using model transformation it is possible to detect design errors in early stages of the design.