Laying the First Brick of the Digital Twin: Legacy Data Integration for Industrial Turbines

Building the digital twin of long lifed complex equipments which are in
use at a remote customer site is a challenging task. We report on a
project at Siemens Corporate Technology which addresses this problem for
industrial turbines using a knowledge graph.

The first goal is computing the current bill of materials by combining
the informations from a variety of sources like engineering data,
maintenance reports and replacement parts orders. The results of this
integration is then loaded in the system where the business processes
are implemented.

Using a knowledge graph as the heart of the ETL pipeline proved to be
very adequate as it provided the flexibility required for the successive
refinements of the integration as the understanding of the data
progresses. A further benefit is that the data can be analyzed using
generic graph exploration and querying tools.

In this project, an ontology defined on top of the knowledge graph data
provides an easily tunable inference layer which lifts the data at the
desired abstraction level, adding classification and path shortcuts for
efficient facetted browsing and querying.