Reverse engineering in process automation

Abstract

This paper provides a method for the reverse engineering process of automation engineering data and a suitable data format for storage and further information enrichment of the gathered data. A method is presented that analyzes process automation flowcharts. Artificial neural networks recognize apparatus and text in the flowchart. Synthetic training data for another artificial neural network is created to detect connections between the recognized apparatus. The neutral data format AutomationML is used to store the gathered data and make it available for further enrichment with plant data. With the method presented in this paper it is possible to overcome the lack of compatible interfaces in the heterogeneous tool landscape of process automation engineering and to reuse analog engineering data for further development of a plant.