Heat Treatment Simulation and Experimental Investigation of Thermal Distortion with a Special Focus on Fineblanked Parts

Abstract

Most fineblanked components undergo some sort of heat treatment in order to achieve desired mechanical properties. At the same time it causes unwanted changes in size and shape of the workpiece leading to subsequent working steps on tools or parts. The extent of this thermal distortion depends on many factors such as material, geometry, type of treatment and treatment conditions. Thus, accurate numerical distortion prediction has a high complexity but can potentially save time and costs of reworks on manufacturing tools. Although the present work has a special focus on through-hardening processes, developed methods are applicable to other heat treatments, such as case hardening, with some adjustments. The goal is to bring thermal distortion prediction for fineblanked parts to a level where it can be used to reduce the amount of reworks on tools and parts. In a first step, mechanical and thermal material properties of through-hardenable C60E steel are characterised. The yield curve, anisotropy and continuous cooling temperature (CCT) diagram are experimentally determined. Additionally, a newly developed end quench experiment for sheet metal, a modification of the well-known Jominy-experiment, is used to characterise hardenability. By recording the temperature history during quenching with an infrared camera, the modified Jominy approach can also be used to validated thermal material data in finite element simulations. Subsequently, several experiment series are carried out to find the major influencing factors on thermal distortion, such as residual stresses from the forming or bending step, geometry and batching. All parts are measured before and after their treatment with digital image correlation technique and are then compared with their target geometry. In order to differentiate between stochastic fluctuations and deterministic relationships, the results are statistically analysed. Opposed to bending operations, which influence distortion results, pure fineblanking shows no significant change in distortion. As a consequence the blanking operation may be neglected for distortion prediction. The observable orientation dependency during quenching with a flat, disk-like specimen could not be reproduced with another similar geometry. An additional experiment indicates the distortion reduction capabilities of batching during quenching. Two of the investigated geometries develop multiple different distortion modi although boundary conditions stay the same. Hence, their occurrence frequency seems random. All necessary material input data for heat treatment simulations are assessed and either taken from literature, measured or modelled according to current state of the art methods. Transformation kinetics, describing phase transformation behaviour during heating and quenching, prove to be a key factor in successfully describing the emergence of thermal distortion. As a widely accepted approach in estimating transformation kinetics based on the material’s chemical composition fails to reproduce results of the previously measured CCT diagram, an optimisation procedure is developed. Hardness values and transformation behaviour are optimised consecutively based on controlled quenching experiments. An additional validation of the modified material model, including the modified Jominy experiment, emphasises the significantly improved prediction capabilities for phase structure and hardness values. FE-simulations of the previously conducted experiments are set up and carried out. All simulations are realised with the commercially available FE-software Forge by Transvalor. The simulations enable the tracing of residual stresses, temperature, microstructure, and hardness during heat treatment. The computational time of heat treatment simulations is drastically reduced by neglecting the blanking step as remeshing and element deletion can be avoided while maintaining the same quality of predictions. Distortion and hardness of specimens are compared with measured data. Although numerical simulation results are in line with the experiment series, not all experimentally observed deformation modi can be identified. A buckling analysis is carried out which successfully anticipates the additional modi that are not covered by the regular simulation.