Computational modelling of quenching and grinding processes for bearing steels

Abstract

The manufacturing of roller bearing components from bearing steels such as 100Cr6 requires the processing of the workpieces in several successive production steps. Each of these processes modifies the initial state of the product at the beginning of the subsequent process. As a result, complex interdependencies arise between the process parameters of the individual processes and the properties of the finished product. In this work, approaches for the numerical modelling of the quenching and grinding processes in such production chains are developed.

For quenching after austenitisation, a fully thermomechanically coupled model is developed which accounts for the occurring phase transformations to martensite and bainite. The athermal transformation of austenite to martensite is modelled by a variant of the classical Koistinen–Marburger ansatz, while the diffusive transformation of austenite to bainite is modelled using a JMAK approach. It is taken into account that the resulting bainite may exhibit a different carbon content than the austenite by solving the mass balance for carbon. To this end, the parameters of the transformation models are parametrised based on the current carbon content of the austenite. The coupling to the mechanical and thermal problems arises from transformation strains and the release of latent heat during the transformations.

Due to the process forces in internal traverse grinding with electroplated cBN tools, elastic deflection occurs between the tool and the workpiece, which can lead to dimensional errors and shape deviations. In this work, a geometric simulation of the grinding process is used, where the grain geometries on the digital grinding wheel are based on measurements of real grains. To account for elastic deflection during the grinding process, two surrogate models are developed: a model for the process force contribution of a single grinding grain engaged in mesoscopic cutting, and a model for the total compliance between the tool and the workpiece. The force model is calibrated using finite element simulations of an idealised single-grain cutting process, while a static analogy test is used to calibrate the compliance model.

For the simulation of single-grain engagement, a new thermodynamically consistent material model is also developed. This model is inspired by the widely used Johnson–Cook model for cutting simulations but is adapted in several aspects to the behaviour of 100Cr6. To account for the large deformations occurring during mesoscopic material removal, a new simulation framework in the finite element software Abaqus, based on the Coupled Eulerian Lagrangian method, is presented and compared with an existing framework based on periodic remeshing.

The developed models improve the simulation capabilities for the manufacturing of roller bearing components from 100Cr6 and thereby contribute to better coordination of the process parameters across different manufacturing steps.