Influence of material inhomogeneity on the mechanical response of a tempered martensite steel

Abstract

Failure in steel weldments operating at high temperatures often occurs in the heat affected zone (HAZ) adjacent to the weld. Such failures can be a result of material inhomogeneity within the HAZ and in the case of tempered martensite steels has been linked with regions of untransformed (ferrite) phase or over-tempered martensite within the intercritical region of the HAZ (IC-HAZ). In this work, two-dimensional Voronoi tessellation is used to construct polygonal Voronoi cells to represent the microstructure of the HAZ of a weld in a tempered martensite steel. The Voronoi construction is treated as a representative volume element (RVE) of the material and is discretized by 8-node linear brick elements, with periodic boundary conditions. The lattice orientation at each material point is specified by three Euler angles, which are assumed to be randomly distributed, to represent the initial lack of texture in the IC-HAZ. The constitutive response is represented by a nonlinear, rate-dependent, finite-strain crystal plasticity model. The results indicate that small amounts of ferrite can induce significant enhancements in stress and inelastic deformation at the interface of the ferrite and martensite grains. This localisation of stress and strain may be critical for microcrack and/or void formation and may be a contributory factor to Type IV cracking.

The effect of ferrite phases on the micromechanical response and crack initiation in the intercritical heat‐affected zone of a welded 9Cr martensitic steel

“…A methodology to obtain the viscoplastic constitutive response in heterogeneous (welded P91) material by using digital image correlation is described in Touboul et al The material parameters for the constitutive model described in Section 2 were determined by fitting to the experimental data in Touboul et al for the IC‐HAZ at room temperature. The representative volume element (RVE), constructed as described in Section 2.1, has been checked to establish a sufficient number of blocks to give a consistent macroscopic response by using the polycrystalline FE model . Two RVEs were first generated: an all‐martensite RVE and an all‐ferrite RVE as shown in Figure .…”

“…The model contains about 30,000 elements. The all‐ferrite RVE (see Figure B, inset) consists of 200 PAG grains with 35,000 elements, following our previous work . In both cases, periodic boundary conditions (see Figure C) are used to represent the overall material microstructure, as in Li and Sun et al and Li and Golden et al…”

“…These are further classified, according to their characteristic microstructure into coarse‐grained, fine‐grained, and intercritical (IC‐HAZ). Under complex operational loading conditions, failures can often initiate within the IC‐HAZ, a phenomenon known as type‐IV cracking …”

“…Figure (adapted from Abe et al) shows the relationship between peak temperature, T p , and the key phase transformation temperatures for a P91 weld. For the coarse‐grained HAZ, T p is significantly greater than A c3 (the upper austenite transition temperature) so that significant austenite grain growth occurs, whereas in the fine‐grained HAZ region, recrystallization dominates leading to grain refinement; in the IC‐HAZ region, A c 1 < T p < A c3 , leading to partial transformation to austenite and the development of a mixture of ferrite and tempered martensite following cooling . The temperature in this region is not sufficiently high to cause dissolution of carbides, leading to the formation of carbide‐free ferrite in the IC‐HAZ region …”

“…Small amounts of ferrite can induce significant stress concentrations and associated strain gradients at the ferrite‐martensite grain boundaries, as shown in our previous work . This may lead to premature microcrack initiation and influence the fatigue life of welded joints.…”