Cold forming processes: some examples of predictions and design optimization using numerical simulations

Predicting the behaviour of steel during a deformation process, and then under service conditions, is one of the main challenges in cold forming. The design of optimized forging schedules, by means of classical trial+errors procedures, has become increasingly heavy in terms of time and cost in a competitive environment. Simultaneously, the improvement of steel qualities requires the microstructure, constitutive behaviour and deformability to be known a priori regarding a targeted application. During the last few years, numerical simulations have become a very efficient tool to reach these goals. In this paper, we give examples of innovating forging sequences developed by numerical simulations, including the investigation of damage in tools and forged parts. In case of specific processes with very determined geometry — such as wire drawing — we show how systematic numerical studies may lead to predictive models of force, local strains and residual stress… However, reliable predictions from numerical simulations require reliable input data, including constitutive laws, friction conditions and propensity to ductile damage. These data must be characterized under realistic sollicitations. Typical cold forging loadings are indeed very severe: local strains up to 600%, strain rates locally greater than 1000 s −1, and subsequently, plastic heating over 500°C. To characterize the constitutive behaviour, the standard upset test between grooved dies is used along with an original methodology to derive the strain hardening curve from the experimental force-displacement recording. Tool elastic deformations, specimen strain heterogeneity… are taken into account. This enables a precise determination of the strain hardening curve up to about 100% of strain. The extrapolation of the flow stress to greater deformations is then very easy and reliable. Such a test can be performed under quasi-static and isothermal conditions (0.1 s −1 but also adiabatic and rapid conditions (up to 10 s −1). This procedure was adapted to a Pellini hammer, which enables very simple characterization at 800 s −1. The comparison of all these flow curves lead to the formulation of an original constitutive model, which accounts for the effects of plastic heating, strain rate, dynamic aging… In order to predict ductile fracture during the forging process, the most classical criteria were tested over a wide range of experimental loading conditions. None of them were general enough to solve all the col