Data-Science

Poster

Capabilities of a data-based framework for assisting the planning process from atoms to components

Tobias Wagner, Institute of Machining Technology / TU Dortmund, Dortmund, Germany
Dirk Biermann, Institute of Machining Technology / TU Dortmund, Dortmund, Germany
Stefan Hess, Institute of Machining Technology / TU Dortmund, Dortmund, Germany

Within the Materials Chain, different disciplines aim at considering all phases from the atom to the component for designing novel types of products in an integrated way. Statistical and empirical methods may process and merge the data of the different research areas for providing planning tools, which guide the product planner through the novel possibilities provided by the Materials Chain. This approach holds huge potential, but the heterogeneity of data and scales poses a big challenge to the underlying methods.

Within the transregional collaborative research center SFB/TR 30, complex thermomechanically coupled process chains composed of multiple subsystems, such as machines, process components, or resources, have been modelled using empirical data. The dependencies between the subsystems are considered by defining so-called technological interfaces, such as material temperature or stress distributions, which are addressed as inputs of the individual models in addition to their actual process or system parameters. The output of each process step includes the property distribution of the current state of the component. Due to phase transformations or other local transformations of the microstructure, these property distributions often show complex gradients.

This flexible hierarchical modelling approach allows process chains to be synthesized with regard to a specification of the properties of the final component. Based on the resulting tree structure of interdepending models of the subsystems, the parameter values of the subsystems can be optimized. Hence, the optimization algorithm can exploit structural information. The robustness with regard to the propagation of uncertainties in the subsystems can be evaluated.

In this talk, the capabilities of the methods within the described planning procedure are presented. It will be discussed whether they can be adjusted with regard to the requirements of the Materials Chain.

« back