NLP for Ontology Development - A Use Case in Spacecraft Design Domain

Abstract

Ontologies can help in the digitalization of engineering design processes. We focus on the design and development of spacecraft, in particular on archiving and discovery of spacecraft part specifications. We created an ontology that is used as a skeleton for creating database schemas and data exchange APIs as well as populating a knowledge graph. However, the development of an ontology from scratch is a rather laborious task. We have to consider various sources of information, such as existing space-system standards, product specifications provided by manufacturers, and actual consumers, i.e. engineers involved in the design process - each with their own perspective. Furthermore, these documents and demands are changing more and more frequently due to new disruptive technologies and an increasing competition in the space business. We aim to close the gap between the tedious process of manual ontology development and fast-paced spacecraft parts development. Due to the lack of machine readable descriptions in this area, we apply NLP techniques to extract technical information from these documents. This information can be used in two ways: first, to create and improve our spacecraft parts ontology and second, to populate a knowledge graph.