Validation of Multi-scale Model Describing Microstructure Evolution in Steels

Abstract

Properties of deformed steels depend on various microstructure parameters such as distribution of grain size and precipitates. Strain, strain rate and temperature inhomogeneities make quantitative prediction of microstructure difficult but the Finite Element method is able to model these inhomogeneities. Different scales of phenomena occurring in deformed materials are another difficulty in modelling. Microstructure evolution can be described by more realistic methods (e.g. Cellular Automata CA, Monte Carlo), which, on the other hand, are unable to simulate larger samples. Therefore, development of the methods capable of spanning multiple scales became a current challenge. CAFE modelling, which couples FE and CA methods, is the objective of the paper. The model consists of two layers. The micro-scale layer, simulated by CA, represents microstructure evolution including nucleation and growth of the grains. Evolution of a dislocation density is described for every grain separately by solving differential equation. The FE thermal-mechanical model is used as a macro-scale part. Multistage plane strain compression tests for niobium steel are considered. Distributions of initial and final grain size are measured during the tests. The results from the CAFE model are compared to the measurements and to the predictions by a conventional model. The comparisons confirm the capability of the CAFE method to predict flow stress, recrystallized fraction and grain size distribution. Conventional approach gives a good agreement with experiments for an average grain size only.