UrWerk – Material Data Rooms for Product Development

Project UrWerk: Company-specific material (system) data spaces for accelerated product development

In times of digitalization, companies rely on modern methods of data analysis and machine learning also in product development. Well-known software tools (for CAD and CAE) are thus further developed and expanded. However, virtual testing of product properties requires high-quality material data in sufficiently large amounts. This includes, among other things, the mechanical properties of individual - but also combined - material states and components as well as information on their manufacturing. 

In the Fraunhofer UrWerk project, we are collaborating with two other Fraunhofer Institutes (IWM and IAIS) to develop data spaces that clearly map the history of materials and their interactions in complex material systems in the form of graphs and enable the coupling of analysis tools. These data spaces are adapted to the specific needs of each company and thus offer easy access to material data.

The distribution of tasks and the cooperation of the Fraunhofer Institutes involved in the UrWerk project.
© Fraunhofer IWM / Flavicon
The distribution of tasks and the cooperation of the Fraunhofer Institutes involved in the UrWerk project.

Tailor-Made Data Spaces and Ontologies Create Digital Context

We would like to design our project as practically as possible and, in the orientation phase, first record customer specific requirements for a respective data space for materials. In this way, we already take individual wishes into account and include them later in the development process. At the same time, we research suitable methods for the storage, analysis and use of the data as well as ontologies for the construction of the data room and further develop these approaches for our purposes.

The development and implementation of the concept will focus on the following areas:

  • Software framework for user interaction in the data space
  • Data Analysis Tools - such as Design of Experiments (DoE) or Machine Learning (ML)
  • Interactive user interface for data analysis 

Depending on the application and the company, we adapt the data spaces individually and set them up on site at the customer's company.

Our Tasks and Competencies in the Project

The distribution of tasks and the cooperation of the three Fraunhofer Institutes involved are based on their core competencies. We concentrate on the following focal points, methods and areas:

  • Modelling, simulation and experiments on cables and hoses (software IPS Cable Simulation, measuring machine MeSOMICS®)
  • Methods for statistical analysis of existing data: Extension of classical methods from the field of statistical data analysis and data mining
  • Methods for data-based prediction of material system properties (greybox modelling), i.e: Adaptation and extension of classical model structures for dynamic system identification (machine learning, neural networks, etc.) and their extension in combination with physical model descriptions


Project Partner

With our software IPS Cable Simulation, one can for example simulate the assembly and disassembly of the wiring harness system of cars consisting of many different cables.
© Fraunhofer-Chalmers Research Centre
With our software IPS Cable Simulation, one can for example simulate the assembly and disassembly of the wiring harness system of cars consisting of many different cables. This simulation requires knowledge of the mass, the geometric characteristics and the effective stiffness properties of the cables.

Application Examples of the Data Space

We would like to demonstrate the added value of material (system) data spaces in practice and illustrate the benefits using two particular application examples. The prediction in both demonstrators is based on a large and complex amount of data. The focus here is not only on classical material properties, but also on material systems in which complex, internal interactions of the individual components significantly influence the properties and behavior of the components.

Components Made of High-Strength Steel

Predictions of material fatigue are of great importance for bearing and gear components made of high-strength steel that are operated without inspection (such as in wind turbines). The prediction is based on a lot of data and due to the complexity no rules and regulations are available yet. Our data space allows for reliable predictions, and we demonstrate that material states with optimal fatigue properties can be proposed.

High-Voltage Cables for Electric Mobility

Several hundred different cable types are already installed in a car today. Cables and cable systems are mainly developed using CAE software tools, but information such as geometric data and mechanical properties is required for the individual cables. Especially for properties such as stiffness, friction or damping there is often not enough information available and the experimental determination is costly due to the variety of types. We are therefore developing a methodology that predicts properties relevant to the material system using a suitable data space.

Example High-Voltage Cable

In contrast to conventionally powered vehicles, electric cars generate much higher voltages. This requires high-voltage cables, as shown below: the two conductor cross-sections belong to cables with copper or aluminium strands. Although the power conduction properties of the two cables are the same, they differ considerably in their mechanical properties. In the copper cable (left) there are many copper strands with a very small diameter in the cross-section, in the aluminium cable (right) there are considerably fewer strands with a larger diameter. The braiding is also different in both cables. With the same deformation of the cable due to bending and/or torsion, a different stiffness behaviour of the two cables can be observed due to the different material properties of copper and aluminium and also due to the different intensive contact interaction of the strands. With the methods developed in UrWerk it is possible to predict such effects and to consider them afterwards e.g. in a simulation of the cable assembly.

© Fraunhofer ITWM
Der Antrieb von Hybrid- und Elektrofahrzeugen ist verbunden mit Hochvolt-Leitungen und Hochvolt-Verbindungstechniken. Die Verbindungstechniken müssen auf hohe Stromstärken ausgelegt und zugleich einfach zu handhaben sein. Die Systeme sind komplex und Anforderungen an die Zuverlässigkeit der einzelnen Komponenten und deren Schnittstellen sind hoch.
© Fraunhofer ITWM
Im Gegensatz zu konventionell angetriebenen Fahrzeugen fließen weitaus höhere Spannungen durch E-Autos. Dazu nötig sind Hochvoltkabel.