An advanced creep law for large stress and temperature ranges, derived from the Larson-Miller master curve concept

Abstract

Accurate determination of creep properties is crucial for predicting the long-term performance of components in high-temperature environments. Most reliable material properties can be obtained by conducting creep experiments under the specific conditions experienced by the component in service. However, these experiments often require significantly more time than is typically available for component design. An alternative approach involves conducting short-term creep experiments and extrapolating the results to estimate the component’s lifetime, albeit with reduced precision. Although the widely used Norton law effectively predicts creep behavior within narrow stress ranges, it overestimates the creep life when extrapolating from short-term data. To overcome this limitation, the Larson-Miller time-temperature-parameter technique (LM-TTP) is revisited, offering improved extrapolation accuracy by trading time with temperature, so that long-term creep behavior can be predicted by use of short-term data obtained at higher temperatures. However, to formulate a creep law in terms of stress vs. creep strain rate, an additional relation of the Larson-Miller parameter (LMP) to the applied stress is needed. This paper presents the derivation of a creep law by introducing a novel stress function to the LM-TTP. Building on the work of Boček and Wolf (1983), the proposed creep law demonstrates its superior extrapolation capabilities, aligning with fundamental experimental observations. Furthermore, the implementation of the proposed creep law into the Finite Element Analysis (FEA) software Abaqus is discussed, providing a ready-to-use CREEP subroutine for general application. Two parameter identification strategies are suggested, with one requiring no simulations for inter- or extrapolation of rupture times, making it particularly appealing for preliminary studies and industrial applications. The alternative strategy utilizes nonlinear optimization techniques for inverse parameter identification. The results of the identification procedures are compared based on uniaxial experiments of the heat-resistant steel P91 (1.4903). Finally, the extrapolation capabilities of the proposed creep law are compared to the Norton law using a creep dataset of the heat-resistant steel 1.4948. Even with experimental data restricted to rupture times less than 100 h, the proposed model exhibits the potential for accurate extrapolation to about 3.8 years, with an error margin of less than 1%, showcasing its reliability in predicting long-term creep behavior. • Novel viscoplastic law for secondary creep behavior that is built on the Larson-Miller time-temperature-parameter. • The material model has been specifically developed for applications covering large stress and temperature ranges. • Demonstration of two parameter-identification procedures, including statistical uncertainty evaluations. • Good extrapolation behavior demonstrated over two orders of magnitude. • Detailed explanation on how to implement the creep model in commercial FEA software such as Abaqus .

Topics & Concepts