Using Explicit and Machine-Understandable Engineering Knowledge for Defect Detection in Automation Systems Engineering

Today the costs of a failure in operation of huge industrial complexes are very high. Traditional approaches for defect detection in automation systems engineering in principle work, but generally don’t take into account the semantic heterogeneity of tools and data models which are used within the engineering of industrial automation systems. Thus, some defects can remain undetected. Also, such systems have to be implemented anew for each concrete case. In this paper we present our ongoing and planned research aimed to improve the defect detection processes. Our approach is based on using explicit knowledge about industrial system stored in a set of ontologies which integrate information from different heterogeneous data sources and present it in machine-understandable form. Another important part of the approach is rules describing system’s logic. Such rules can, through the use of integrated engineering knowledge stored in ontologies, detect faults which otherwise are hard to identify using traditional methods. Major expected results are the more efficient and effective defect detection and the potential reuse of the created ontologies in other projects.