Gradient Temperature Creep Testing: A Novel Approach to Parallelized Deformation Analysis

This study explores the gradient temperature creep test (ΔT) as a parallelized method for acquiring hundreds of creep curves across a range of temperatures using a single specimen. In conventional creep testing (CCT), a single creep curve is acquired per specimen at a set temperature. Many CCTs must be performed to determine the creep resistance of an alloy at service-like conditions. In the ΔT test, a quasi-static temperature gradient is produced using induction heating. Strain and temperature are mapped using Infrared 3D Digital Image Correlation, allowing for the extraction of creep deformation curves throughout the temperature gradient. This innovative approach enables the determination of the minimum-creep-strain-rate, activation energy, and the construction of partial creep curves at various isotherms along the specimen's gauge section. In this study, 4130 steel is subject to conventional creep tests at 4 isotherms (525, 550, 575, 600°C) and a single ΔT test is performed spanning approximately 245 to 600°C. Tests are performed on flat dogbone specimen and are interrupted at 7.5 hours. The creep curves, minimum-creep-strain-rate, and activation energy of the two methods are compared. A Python extraction algorithm is developed to filter the 13,462 raw creep curves measured during the ΔT. Datapoints below the creep activation threshold (525°C), outside the ASTM limits (± 2°C), and susceptible to stress concentrations are culled. The remaining datapoints are collated into equally spaced isotherms (5°C increment) within the ASTM tolerance (± 2°C). Minimum-creep-strain-rates are extracted from each curve and the Arrhenius equation plotted to furnish creep activation energy. Using this processed dataset, the ΔT test reproduces CCT behaviour, with the creep‑activation‑energy only 1.32% higher and the minimum‑creep‑strain‑rate differing by < 3% across the 525–600°C range. A total of 474 creep curves meet/or exceed ASTM standards. This outcome is equivalent to 474 independently run CCTs, representing over 3,547.5 total hours of testing time, while cutting both material usage and its associated cost by about 99.78%. A discussion on how to further optimize this test method is proposed.