Evaluation of Power-Law Creep in Bending

Abstract

Materials for engineering components operating at high temperatures during their service like rockets, aero-turbines, automobile engines, nuclear power plants, steam power plants and micro-electronics are often designed for creep resistance. Creep studies for such systems are performed either as a material selection routine or for investigating the residual-life, integrity and reliability of the components. Conventionally, uniaxial tests, which are the most reliable creep testing method, are adopted to study the dominant creep mechanism of materials systems at a particular stress and temperature. However, with increased demand for improving system efficiency, high throughput testing and miniaturization, a need for alternate testing methodology is realised. This work analyses the structure-property correlation of bending creep in the power-law regime and aims to establish it as one such alternative testing method with the prospects of obtaining reliable creep data from a small sample volume and at a yield rate that is 10 times faster than the conventional testing routes. In this work, the possibility of an uniaxially equivalent deformation state in bending was investigated. Digital image correlation was adopted to obtain strain and strain-rates at multiple locations in a cantilever. This coupled with the stress information obtained using developed analytical equations and numerical methods enables gathering hundreds of “stress-strain-rate” pairs from a single cantilever. Furthermore, the strain-time information obtained from digital image correlation was used to calculate primary and steady-state creep parameters from a single test performed on a cantilever. The developed method was applied to various material systems, such as commercially pure Pb and Al, and high purity poly- and single-crystal Al. An excellent match was observed between the creep parameters, including those relevant for describing primary and steady-state stages, obtained from bending and uniaxial creep tests. In addition, the effect of sample size on the creep properties obtained from cantilevers was investigated. It was realised that the creep properties obtained from cantilevers of small thickness show a linear hardening in the size range of 3 to 0.5 mm which is inversely proportional to the cantilever thickness. On the other hand, samples tested in uniaxial compression show a softening in creep below a size range below 2 mm diameter with comparable sample thickness and substructure size. A possible interplay between the hardening in creep behaviour due to straingradients and softening due to small sample size is discussed in the context of dislocation substructures. Experimental evidence for the same from EBSD patterns is presented. These are used to discuss the differential creep behaviour along the length and thickness of a cantilever where regimes of hardening, bulk behaviour and softening with respect to uniaxial bulk EE

behaviour are observed in a single cantilever. Overall, this study aims at establishing creep of cantilevers as a high throughput creep testing methodology. Testing a cantilever is shown to accurately reproduce all relevant creep parameters as observed in uniaxial tests in the powerlaw regime and in a significantly shorter time. This is possible as long as the sample thickness is such that the strain gradients are small and allows development of equilibrium substructures as per the operable creep mechanism.