Adaption of new strategies for parameterization of a microstructure-based flow stress model (Paperback)

One of the most important tools in the metal forming technology is the integrated process and microstructual simulation using Finite Element Methods (FEM). It has become more and more popular in recent years, especially in the segment of hot metal forming, e.g. open die forging of large scaled and hardly deformable materials, such as nickel-based super alloys for turbine shaft. Theoretically using this method it is possible to calculate the microstructual evolution along the whole process chain in the numerical simulation of the considered metal forming process. Based on this knowledge a series of benefits can be achieved for the practice, such as to optimize a present metal forming process, to predict the mechanical properties of the final products under the given forming conditions, to detect the possible product failures prematurely, to assist the design of a new production chain and so on. In the face of these trends of the scientific research as well as the industrial demands, different material models have been released in the market to combine the commercial FEM programs specialized in the numerical simulation of metal forming. Among others microstructure-based ow stress models show outstanding performance. Through this kind of material model not only the microstructure such as recrystallization and grain size, but also the interaction between the microstructual evolution and the working hardening, effectively ow stress, can be quantitatively represented.

Towards accurate and efficient material modeling, the model parameters have to be determined conveniently and reliably. For this propose a new Hybrid strategy combining the advantages both of direct and indirect methods has been proposed using the example of StrucSim, which is a very good representative of a mircostructure-based ow stress model. At first different aspects, which lead to the disadvantages of the conventional method, i.e. direct method, were discussed. In doing so a high manganese steel was characterized as an example by stepwise graphical and regression analysis. It was found that, the precondition of direct methods, namely recording ow curves under constant Zener-hollomon-parameter conditions, are basically not possible due to both limitations of test equipment and unconquerable physical mechanism like dissipation heating. The common solution to compensate these factors may lead to further inaccuracies, uncertainties and complexities despite large testing and evaluating efforts. Further in order to improve the model quality calibrated by the conventional direct method an efficient hybrid strategy has been derived by combining inverse analysis with offline calculation of ow stress and microstructure. Three different variations of the hybrid strategy were introduced to deal with different available experimental databases, such as isothermal and non-isothermal ow curves. To demonstrate the developed routines of these three hybrid possibilities two kinds of materials including a nickel-based super alloy and a high manganese steel have been taken into account. The investigation has shown that through the introduced hybrid methods better model quality can be achieved even with less experimental data. Owing to the convenience of the inverse technique much experimental and evaluating effort and complexities can be avoided. Finally, another inverse analysis based on inhomogeneous deformation has been proposed, in which hot compression tests with double cone specimen were employed. Thanks to the inhomogeneity of strain and microstructure distribution within the specimen, it becomes possible to get sufficient relevant information as constraints for the inverse parameterization through even fewer experiments. In addition, the established routine of a hybrid strategy as well as the inverse analysis based on non-uniform deformation enhances the transferability of determined model coefficients between two materials within the same group or among similar materials.