Efficient multi-scale modeling by means of FE-FFT and model order reduction techniques

Stefanie Reese, Christian Gierden, Johanna Waimann, Bob Svendsen, RWTH Aachen University, Germany

In many cases within material mechanics, the microstructural material behavior is of very large interest. An example is production technology, where one seeks to choose the process parameters in such a way that a desired material modification comes out. This requires to find a relation between the material loading and the material modification, the so-called process signature. For this purpose, two-scale simulation methods, such as the FE² or the FE-FFT-based method, are very suitable (FE finite element, FFT fast Fourier transforms). A major disadvantage of such two-scale simulations is the extremely large computational effort. To reduce this computational effort particularly for the FFT-based microstructure simulation, we present a model order reduction (MOR) technique, which is based on a reduced set of Fourier modes. Since the accuracy of this MOR technique strongly depends on the choice of Fourier modes, we investigate various definitions of sampling patterns. To achieve a further reduction of the computational effort, the MOR technique can additionally be coupled with a clustering analysis. The method is applied to the process of deep rolling.

Reducing wear to prolong tool life is a key challenge in metal machining. To facilitate that, the active wear mechanisms need to be understood, and in our research group we are working to contribute to this increased understanding. Our methodology comprises tribological testing combined with careful surface analysis, to identify the wear mechanisms and observe any changes to the surfaces. The metal machining process will not only result in tool material removal but also in tool surface modifications, vastly important for the wear process. After machining, the tool surfaces are studied in detail, using high resolution scanning electron microscopy, elemental analysis, and surface profiling as our key methods. The machining tests are complemented by sliding tests, where the chip formation process is omitted in favour of successive detailed analysis of the tool surface, enabling studies of the wear initiation and observations of minute wear. Further, the tool wear is strongly dependent on the work material, influencing wear mechanisms, surface modifications and acceptable cutting parameters. This talk will be focused on machining of titanium alloys, known to be difficult to machine materials causing rapid wear of the tools, but also case hardening and carbon steels will be touched upon.