Even the smallest fluctuations in production are enough to change the material properties of workpieces and therefore also the component quality and function. To prevent this, samples are repeatedly taken and examined during production. However, such a sample component has to be disassembled into individual parts and measured for tests, which takes a lot of time.

Researchers at the Fraunhofer Institute for Mechanics of Materials IWM tackled this problem and have now demonstrated the feasibility of digitally mapping many such material histories for the first time: they developed a sample material data space for additively manufactured test specimens and have thus created a basis for the digital material twin. "With the data room concept, we can integrate material information of any kind into digital networks - which is important with regard to Industry 4.0, among other things," explains Dr. Christoph Schweizer, Head of the Materials Evaluation, Lifetime Concepts business unit at Fraunhofer IWM in Freiburg. "From the material data space, we want to automatically generate a digital twin for each material that describes the current state of the material object under consideration."

The material data room offers the advantage that all relevant parameters are available at a glance; previously, the information on various material parameters was usually scattered across numerous data repositories and in different data formats, making a quick comparison impossible. "The material data room could become the production brain of the coming years. Whenever the component quality is not as desired, you could compare it in the material data room with components from the past and find out whether the current one can still be used or has to be discarded," says Schweizer.

Material ontologies as a basis

In order to create the material data space and manage the heterogeneous material data, a suitable information model is required. "This model mirrors the natural world of materials, in which the material states and properties are divided into specific categories," explains Dr. Adham Hashibon, a scientist in the Manufacturing Processes business unit. The researchers rely on ontologies - i.e. a logical, hierarchical structure. The researchers explain the principle using Facebook as an example: In the network, individual people represent nodes that have links to their preferences, for example their taste in music. "We create semantic links between the individual material objects and the associated processing operations," explains Hashibon. The relationships between them, i.e. friendships on Facebook, are information on the chronological sequence of production or work steps in the material data space, such as "comes out of the additive manufacturing process" or "this laser takes part in the 3D printing process".

The aforementioned demonstrator for additively manufactured metal covers sample production, material characterization and the subsequent data analysis and determination of material properties. Due to the logic of the underlying structural model, very complex queries can be made to the data space that would not be possible with this level of flexibility using conventional databases.

With its work on the digitized materials data room, the Fraunhofer IWM is contributing to European materials modeling topics within the framework of the European Materials Modelling Council and to Baden-Württemberg's digitization strategy. In the medium term, it is planned to convert the entire data management at the Fraunhofer IWM to the data room system. Cooperation partners and pilot users who can help shape the developments are welcome for this and other applications.

According to documents from the Fraunhofer IWM / ag