Our experience with the latest materials science and materials engineering findings enables us to optimize forming tools and processes, while numerical simulations facilitate the avoidance of trial and error loops. We develop material models that describe material behavior and determine relevant material and process parameters.
The derivation of processing control actions for the production of materials with certain, desired properties is the "inverse problem" of the causal chain "process control" - "microstructure instantiation" - "material properties". The main goal of the proposed project is the creation of a new basis for the solution of this problem by using modern approaches from the field of Machine Learning.
The inversion will be composed of two explicitely separated parts: "Property-Structure-Mapping" and "structured-guided optimal process control".
The focus of the project lies on the investigation and development of methods which allow an inversion of the structure-property-relations of materials, which are relevant in the industry. This inversion is the basis for the design of microstructures and for the optimal control of the related production processes. Another goal is the development of optimal control methods yielding exactly those structures which have the desired properties. The developed methods will be applied to sheet metals within the frame of the project as a proof of concept.
The goals include the development of methods for inverting technologically relevant "Structure-Property-Mappings" and methods for efficient microstructure representation by supervised and unsupervised machine learning.
Adaptive processing path-optimization methods, based on reinforcement learning, will be developed for adaptive optimal control of manufacturing processes.
We expect that the results of this work will lead to an increasing insight into technological relevant process-structure-property-relationships of materials. The instruments resulting from the project will also promote the economically efficient development of new materials and process controls.
The main issue of project Grey-Box models is to improve machine learning methods (black-box models) by integrating domain knowledge, using for example deterministic models (white-box models). Grey-box models aim especially at industrial applications, as there are typically only few significant (distributed over the space of process parameters) data available, whereas the processes themselves are relatively well known. During the project, we will work on three use-cases focusing on material science: automatic crack surface detection in microstructural images, surrogate modeling of deep drawing processes and the identification for material model parameters using neural networks.
Within the Priority Program 1713 of the German Research Foundation (DFG), we are working on a novel simulation tool for hot forming and heat treatment of metallic materials. Our approach links the thermomechanical material behavior and microstructure evolution using a comprehensive thermodynamic framework. This allows us to efficiently represent elastic-plastic material behavior, recovery, recrystallization, grain coarsening, texture evolution and precipitation as well as the related hardening and softening processes. After a successful first project phase, a second phase of three years duration was granted and started in December 2017.
Telephone +49 761 5142-479
TWIP-steel features a tensile strength of approx. 1000 MPa with a breaking elongation of 40-50 %. By using TWIP steels, both the energy absorption of components and the structural safety of the vehicle can be significantly improved. The strength of this material allows for a reduction of the sheet thickness used in components and contributes to a more efficient use of resources. Scientists at the Fraunhofer IWM have developed an appropriate material model so that the mechanical properties of TWIP steel can be accurately described. An essential characteristic of this model is the physically based description of microstructural properties and especially the development of the twin volume fractions depending on deformation and state of stress.
The entire process chain can be virtually described through the linking of either very similar or quite different simulation methodologies. At the Fraunhofer IWM we develop methods for the linking of subsequent process steps: cold rolling simulations realized via the finite element method (FEM) are combined with heat treatment descriptions. The ensuing results are then used in microstructure simulations to predict macroscopic, mechanical characteristics which are incorporated into material models for component forming simulations. This enables us to test the influence of individual process parameters on material properties.
The "Virtual Lab" is a simulation tool for the numerical determination of macroscopic material properties which takes the microstructure into account. Data produced through our "Virtual Lab" can be used in exactly the same manner as experimental data and is especially applicable to the complex material models which are required when working with modern, high-strength sheet metal materials. These complex material models utilize many parameters which can be identified by the additional data obtained from the "Virtual Lab".