Sitting on a gold mine: the story of the process industry's automatic formation of a digital twin

TL;DR

This paper discusses a semi-automatic approach to generate digital twins of process systems by extracting and transforming factory floor data, such as P&IDs, into adaptable graph models to accelerate digitization.

Contribution

It introduces a method for semi-automatic digital twin creation from factory data, utilizing image, pattern, and text recognition to generate and modify intermediate graph models.

Findings

Effective extraction of raw model data from P&IDs.

Generation of adaptable intermediate graph models.

Enhanced system reliability and flexibility.

Abstract

The use of a software tool chain to generate Digital Twins (DTs) automatically can speed up digitization and lower development costs. Engineering documents and system data are just two examples of source information that can be used to generate a DT. After proposing a general plan for semi-automatic generation of a DT for a process system, this work describe our efforts to extract necessary information for the generation of a DT of a process system from existing information in a factory floor like piping and instrumentation diagrams (P&IDs). To extract initial raw model data, techniques such as image, pattern, and text recognition can be used, and then an intermediate graph model can be generated and modified based on requirements. In order to increase the system's adaptability and reliability, this research will delve deeper into the steps involved in creating and manipulating an intermediate graph model.