This conference paper was submitted for presentation at the NAFEMS World Congress 2025, held in Salzburg, Austria from May 19–22, 2025.

Abstract

In a landscape where market expectations and competitive offerings are constantly evolving, systems that integrate and contextualize multiple data sources and deliver actionable insights are highly valuable. Closing the gap between external market data and internal data provides companies in every industry company with a strategic asset that enables them to bring products to market faster. Exploiting this potential is also the idea behind the Data Context Hub (DCH), which was developed as a research project for over six years at Virtual Vehicle Research (Graz) together with automotive and rail OEMs. The platform brings information from R&D and production data sources together as well as from telemetry data streams or storage locations. The DCH creates an explorable context map in the form of a knowledge graph from area-specific data models. These are essential for streamlining processes, reducing risks and identifying new opportunities in data-driven development. The use of state-of-the-art AI models also supports developers in gaining deeper insights from data, predicting trends and automating tasks. Especially in the virtual product development process the use of contextual graph databases opens an important approach for the implementation of artificial intelligence methods in technical use cases. As an example, we consider a crash simulation, where the develop solution creates the necessary link between internal crash results and the standards like NCAP or IIHS on which the assessment of crash safety is based. The approach here is that by constantly comparing the internal used crash values with the official values at all times, it is ensured that no outdated values from outdated protocols are used. Changes in the officially defined values and standards are already recorded before the crash tests. They are already included in the test simulations for the structure, materials and restraint systems of innovative vehicles, for example, at an early stage before the evaluation of crash safety. In the concrete example of crash simulations, the contextual graph databases can be used to immediately recognize which specific values and standards form the basis for the evaluation of crash safety in the event of changes to components on a vehicle. The contextual graph databases are done via an internal engine for context creation, which creates factually linked data points and visualizes them via a comprehensible path within the graph. To interpret these results, generative AI models such as Large Language Models (LLM) can be used to provide users with more precise answers about the data sources and relationships. In the example mentioned, the contextual graph databases can be used to quickly determine which crash test simulations need to be rerun to check the system response and reevaluate crash safety performance. As a result, only these necessary crash test simulations can be carried out on the available resources, which saves companies development time and costs in the long term. Linking internal data sources with external data sources enables companies to gain differentiated insights that were previously inaccessible. DCH's contextual graph databases can provide precise, relevant answers to specific user queries in specific data contexts through the clever use of state-of-the-art AI models. This enables users to better understand the answers and make faster decisions.