Haynes ® 282 ® , a relatively new superalloy is used in gas turbines in form of sheets, plates and forgings. Forgings undergo a series of deformation steps at high temperatures to form complex shapes of components. The deformation on forgings, changes the microstructural features and their distribution, and any change in distribution of microstructural features can affect the mechanical properties of the material. The present study is undertaken to investigate the possible causes of anisotropy in mechanical properties of a Haynes ® 282 ® forging through optical and electron microscopy. Microscopic investigations show that ductility is anisotropic and changes from 15% to 21%. The electron backscattered diffraction (EBSD) investigation reveals that the presence of carbide stringers (banding of MC and M6C carbides) is associated with fine grains, thereby giving a bimodal distribution of grain size. Carbide stringers follow the complexity of forgings and are identified as the primary cause for the anisotropic behavior in ductility. Furthermore, micromechanical simulations of carbide stringers in association with a bimodal grain structure was seen to qualitatively correspond to the experimental observation indicating improved ductility with banding along the tensile axis

a b s t r a c tFor low-cycle fatigue tests with smooth bars the number of cycles to initiation is commonly defined from a measured relative drop in maximum load. This criterion cannot be directly related to the crack length, which is the actual measure of interest. In order to establish a relation between load drop and crack length for the high strength titanium alloy Ti-6242, this investigation compares data from controlled low-cycle fatigue crack growth tests and numerical simulations of these tests. To achieve sufficient accuracy in this relation, focus is given to modelling of mean stress relaxation. Three constitutive models, the Chaboche, the Ohno-Wang and the Chaboche with threshold, are evaluated with respect to experiments. Furthermore, a straightforward method with cycle-scaling of the material parameters are used to efficiently reduce calculation cost. It is shown that it is possible to determine the relationship between load drop and crack length from numerical simulations, provided that care is taken to relevant aspects of the materials stress-strain response. These results are also used to numerically evaluate the effect on load drop of the extensometer position relative to the crack.

No abstract

In this paper, the cyclic modelling of Haynes 282 is investigated for a range of elevated temperatures. A calibration of a plastic model including cyclic hardening/softening has been performed and a temperature dependence for the material parameters has been determined. The calibration and validation of the temperature dependence are based on uniaxial low cycle fatigue experiments for four different temperatures. A strategy for the calibration with regard to the temperature dependence is proposed and discussed. The proposed strategy aims at minimising the complexity of the material model without losing any significant accuracy. The calibrated model response is able to capture the experimental results with good accuracy. Furthermore, a finite element example is used to illustrate the influence of the slow cyclic softening when applying many loading cycles to a structure. Owing to the many loading cycles, a technique for cycle extrapolation is incorporated in the finite element analyses and the efficiency and accuracy of this technique are briefly discussed.

Abstract. During low-cycle fatigue test with smooth bars the number of cycles to initiation is commonly defined from a measured relative drop in maximum load. This criterion cannot be directly related to the actual measure of interest -the crack length. By relating data from controlled crack growth tests under low-cycle fatigue conditions of a high strength Titanium alloy at 350• C and numerical simulation of these tests, it is shown that it is possible to determine the relationship between load drop and crack length, provided that care is taken to consider all relevant aspects of the materials stress-strain response. IntroductionTraditional low-cycle fatigue (LCF) testing is performed using either smooth round bars or test specimens with notched geometries to resemble relevant features in components. During LCF tests microcracks are initiated relatively early in the fatigue life, and the majority of cycles are spent propagating the dominating crack to failure. For many applications, the number of cycles to failure (N f ) is an inadequate measure since the aim is to determine when a macroscopic crack has formed, so that a subsequent crack propagation analysis may provide a more accurate estimation of the total life in a component of arbitrary geometry. The number of cycles to initiation (N i ) is usually defined as the point where a given metric, which is either continuously evaluated during testing or examined postmortem, exceeds some pre-determined limit. In strain controlled testing of smooth specimens, the most common metric is the X% load drop criteria [1], where initiation is defined as the point where the maximum load has dropped below X% of the saturated stable value (or in the case of continuously increasing or decreasing maximum stress the load drop compared to the extrapolated values of stress is used).Although practical, the inherent drawback of the criterion is that it is not related to the size of the crack, which is the measure that is of actual interest. If the LCF data is to be used to directly assess the life of a component, the crack size at the point of initiation must be known and compared to the dimensions of the critical location in the component. If the definition of initiation results in large crack sizes before the load drop criterion is used, the data may give un-conservative life predictions for e.g. thin-walled components. Furthermore, to allow subsequent crack propagation analysis the crack size at the point of initiation must be known. It is therefore of interest to investigate how the drop in maximum load and crack length are related during LCF testing. In addition, for round bars where the crack initiation may occur randomly around the length and circumference of the gage section, the location of the extensometer relative to the point of initiation can play a dominating part in the correspondence between crack length and load drop.In this work, we aim to demonstrate that a Chaboche-type of constitutive model can be used to assess the relationship between crack length and load dro...