Tempered martensitic steels are used for turbine components in power plants under high temperatures $(T\approx 873\mathrm{K})$ . In addition, the components are subjected to temperature changes during start-up and shutdown procedures. In order to model the mechanical behavior of tempered martensitic steels in the high-temperature range, this article presents a phase mixture model accounting for hardening and softening effects. The alloy is described as an iso-strain mixture of a hard phase and a soft phase. Thereby, the hard phase is related to the subgrain boundaries, that is, regions with high dislocation density, while the soft phase is related to the subgrain interior, that is, regions with low dislocation density. The volume fraction of the hard phase is assumed to decrease toward a saturation value, such that softening effects due to the coarsening of subgrains are taken into account. The introduction of two internal variables, a backstress and a softening variable, allows for the calibration based on macroscopic material tests, that is, creep and high-temperature tensile tests. Special emphasis is placed on the detailed description of this non-trivial calibration procedure. The calibration is based on the distinction of stresses and strain rates in initial and steady states. Starting from three differential equations with respect to the inelastic overall strain rate, the backstress, and the softening variable, all required functions and parameters are determined. The calibrated model is verified by the simulation of additional material tests at high temperature, highlighting the wide application range with respect to stresses $(100-700\mathrm{MPa})$ and temperature $(673-923\mathrm{K})$ .

This article aims to analyse the influence of temperature and strain rate on the mechanical behaviour of the high-chromium martensitic steel X20CrMoV12-1. The analysis is based on two series of high-temperature uniaxial tensile tests. In a first series, the tensile tests are conducted until rupture, and temperature as well as strain rate are varied systematically. The corresponding stress–strain curves show an extended softening stage. In order to examine softening, it is crucial to distinguish between microstructural changes and strain localisation due to necking. For this reason, a tensile test at low strain rate is performed several times, while the test is terminated at different strain levels in order to examine the onset of necking. Based on the test results and surface measurements of the deformed specimens, the strain level at which necking starts is determined, and possible interactions between softening and necking are discussed. The tensile tests have been conducted in order to calibrate a mechanical model which supplies reliable predictions on the material behaviour under different loading scenarios at elevated temperatures. For this reason, a framework based on microstructural processes is presented in the second part of the article. The model applies a binary mixture approach in conjunction with an iso-strain concept. Furthermore, backstress and softening variables are introduced to consider hardening and softening effects. This procedure results in a system of three differential equations describing the mechanical behaviour.