Modelling of high strain rate plasticity and metal cutting
Sammanfattning: Metal cutting is one of the most frequently used forming processes in the manufacturing industry. Extensive effort is made to improve its process and simulation has become an integrated part, not only in the product development process but also in the customer relations. However, simulation of metal cutting is complex both from numerical as well as physical point of view. Furthermore, modelling the material behaviour has shown to be crucial. Errors in the material model cannot be reduced by the numerical procedures. The magnitudes of strain and strain rate involved in metal cutting may reach values of 1-10 and 103-106 s-1. The dissipative plastic work together with the chip tool friction also leads to locally high temperatures. These extreme ranges of conditions imply that a diversity of physical phenomena is involved and it is a challenge to develop a material model with adequate accuracy over the whole loading range. Furthermore, this intense and severe deformation represents thermo-mechanical behaviour farfrom what is generated from conventional material compression and tension testing. A highly desirable feature is also a material model that can be extrapolated outside the calibration range. This is not trivial since materials exhibit different strain hardening and softening characteristics at different strains, strain rates and temperatures. Models based on modelling some aspects of the underlying physical process, e.g. the generation of dislocations, are expected to have a larger range of validity than engineering models. Though, engineering models are the far most common models used in metal cutting simulations. The scope of this work includes development of validated models for metal cutting simulations of AISI 316L stainless steel. Particular emphasis is placed on the material modeling and high strain rate plasticity phenomena. The focus has been on a physically based material model. The approach has been to review the literature about flow stress models and phenomena and particularly at high strain rates. A previous variant of a dislocation density model has thereafter been extended into high strain rate regimes by applying different mechanisms. Some of the models have been implemented in commercial finite element software for orthogonal cutting simulations. Experimental measurements and evaluations that include SHPB-measurements, cutting force measurements, quick-stop measurements and some microstructural examinations has been conducted for calibration and validation. The compression tests, within a temperature and strain rate range of 20-950 °C and 0.01-9000 s-1 respectively, showed that the flow stress increased much more rapidly within the dynamic loading range and hence depends on the strain rate. The dynamic strain aging (DSA) that has been observed at lower strain rates is non-existent at higher strain rates. The temperature and strain-rate evolution is such that the DSA is not necessary to include when modelling this process. Furthermore the magnetic balance measurements indicate that the martensite transformation-strengthening effect is insignificant within the dynamic loading range. In the present work the concept of motion of dislocations, their resistance to motion and substructure evolution are used as underlying motivation for description of the flow stress. A coupled set of evolution equations for dislocation density and mono vacancy concentration is used rendering a formulation of a rate-dependent yield limit in context of rate-independent plasticity. Dislocation drag due to phonon and electron drag, a strain rate dependent model of the subcell formation, a strain rate and temperature dependent recovery function and a structural dependent thermally activated stress component have among others been considered. Best predictability was obtained with a strain rate dependent subcell formulation. Dislocation drag did not improve the predictability within the measured testing range. Although showed to has a greater influence outside the range of calibration when extrapolated. It has been shown that extrapolation is uncertain. Results from experiments and modelling of material behaviour and metal cutting together with the literature indicate that the predictability of the material behaviour within and outside the measured testing range can be further tuned by implementing models of the phenomenon mechanical twining and recrystallization.
Denna avhandling är EVENTUELLT nedladdningsbar som PDF. Kolla denna länk för att se om den går att ladda ner.