A Datum Temperature Calibration Approach for Long-Term Minimum-Creep-Strain-Rate and Stress-Rupture Prediction Using Sine-Hyperbolic Creep-Damage Model

Abstract

This study introduces a datum temperature (DT) calibration approach for improved extrapolation of minimum-creep-strain-rate (MCSR) and stress-rupture (SR) data. The ASME B&PV code III outlines stringent requirements for the approval of materials where each heat is to be tested to 10,000+ hours to be qualified for service. Additionally, components operating at a range of service conditions require tests to be performed at many combinations of stress and temperature. Subsequently, it takes years to decades for new creep-resistant alloys to be implemented due to the number, duration, and costs of tests involved. The increasing demand for new alloys for IGT applications and the desire to reduce qualification time has driven the urge for rapid qualification testing, calibration, and modeling techniques. To that end, a datum temperature (DT) calibration approach is applied to a contemporary creep-damage model for improved long-term extrapolation of creep data. In the DT approach, data across multiple temperatures are mathematically transferred to a datum temperature creating a master curve. This collapse of the data to a single isotherm (i.e. master curve) increases the amount of data available for model calibration. Next, the model is calibrated to the master curve; afterward, the model is transferred back to original temperatures. A DT approach can significantly reduce: the overall duration of creep testing; effort required for model calibration; and eliminate the requirement for temperature-dependent material constants. In this study, the DT calibration method is applied to the continuum-damage-mechanics (CDM)-based Sine-hyperbolic (Sinh) model to extrapolate the MCSR and SR for 18Cr-8Ni (304SS) stainless steel. The MCSR and SR data across multiple isotherms are gathered from the National Institute for Material Science (NIMS) database. Mathematical rules to transfer data to a datum temperature are developed for the Sinh MCSR and SR equations. The Sinh material constants are obtained by creating and fitting the DT master curve. The model is shifted back to the original temperatures and extrapolation credibility is assessed. The normalized mean square error (NMSE), coefficient of determination (R2), and mean square percentage error (MSPE) statistics are employed to analyze the prediction quality. The NMSE at datum temperature is observed to be 2.044 and 0.233 for MCSR and SR, respectively. The corresponding MSPE statistics is low at 0.296 and 0.191. The extrapolation at low stress and high temperature and vice versa is observed to be devoid of any inflection point. The DT approach for Sinh is further verified and validated by comparing against additional MCSR and SR data for 18Cr-12Ni-Mo (316SS) stainless steel that were not used for calibration. It is observed that the Sinh extrapolated MCSR and SR are free of inflection points. Based on the goodness-of-fit of the extrapolations, a recommendation to use DT approach for past and modern creep-damage model is provided.