Project Details
Projekt Print View

Development of a procedure for the determination of flow curves by explicit pointwise inverse modelling

Applicant Dr.-Ing. Emad Scharifi Arab, since 7/2024
Subject Area Primary Shaping and Reshaping Technology, Additive Manufacturing
Mechanical Properties of Metallic Materials and their Microstructural Origins
Term from 2018 to 2025
Project identifier Deutsche Forschungsgemeinschaft (DFG) - Project number 393221806
 
Final Report Year 2025

Final Report Abstract

The accurate determination of flow curves is essential for modeling material behavior in forming simulations. However, conventional laboratory methods such as tensile, compression, and torsion tests become increasingly inhomogeneous beyond certain deformation levels due to phenomena like necking, frictional effects, or temperature gradients in hot forming. This inhomogeneity undermines the accuracy of purely analytical evaluation methods that rely on simplifying assumptions. Inverse modeling using finite element (FE) simulations offers a solution by accounting for the real test conditions. However, conventional inverse methods are dependent on a mathematical equation whose parameters are determined using, e.g., a gradient-based optimization method. For this, an equation must be predefined and the resulting flow curves from this method are limited to the selected flow curve equation. Within this project, a novel flow curve determination approach - FepiM (Flow curve determination by explicit point-wise Inverse Modelling) - was developed, which is independent of a specific mathematical equation. A point-wise method for flow curve determination was implemented in which the data pairs of flow stress and corresponding plastic strain are calculated directly in a single simulation based on the experimental data. This approach does not require a predefined mathematical equation and allows the analysis of complex phenomena such as dynamic strain aging or multiple cycles of dynamic recrystallization, where oscillations in the flow curve may occur. Since the flow curve is determined without gradientbased optimization, problems with convergence, local minima and long calculation times due to numerous iterations with a high number of FE simulations are also excluded. The results show that the FepiM method can predict flow curves with an accuracy of less than 5% deviation from the experimental data.

Publications

 
 

Additional Information

Textvergrößerung und Kontrastanpassung