Fatigue crack propagation in a ductile superalloy at room temperature and extensive cyclic plastic flow. (November 2015)

• Record Type: Journal Article
• Authors: Ewest, D.; Almroth, P.; Leidermark, D.; Simonsson, K.; Sjödin, B.
• Is Part Of: International journal of fatigue. Volume 80(2015)
• Page Start: 40
• Page End: 49
• Publication Date: 2015-11
• Subjects: Fatigue crack propagation -- Haynes 230 -- Large scale yielding -- Cyclic J-integral or ΔJ -- Crack closure; Materials -- Fatigue -- Periodicals; Materials -- Fatigue; Periodicals; 620.1122
• DOI: 10.1016/j.ijfatigue.2015.04.006
• Languages: English
• ISSNs: 0142-1123

Abstract

Highlights: Results from experimental TMF tests with long hold-times were presented. TMF stress relaxation in a single-crystal superalloy was analysed and evaluated. Creep was modelled using a basic Norton type of expression on each slip system. Creep parameters were obtained in an optimisation step matching the experiments. Good correlation of the stress relaxation compared to experiments was achieved. Abstract: Fatigue crack propagation experiments under both force and displacement control have been performed on the wrought superalloy Haynes 230 at room temperature, using a single edge notched specimen. The force controlled tests are nominally elastic, and the displacement controlled tests have nominally large plastic hysteresis at the beginning of the tests, but saturates towards linear elastic conditions as the crack grows. As some tests are in the large scale yielding regime, a non-linear fracture mechanics approach is used to correlate crack growth rates versus the fracture parameter Δ J . It is shown that crack closure must be accounted for, to correctly model the crack growth seen in all the tests in a unified manner. For the force controlled small scale yielding tests the Newman crack closure model was used. The Newman equation is however not valid for large nominal cyclic plasticity, instead the crack closure in the displacement controlled tests is extracted from the test data. A good agreement between all tests is shown, when closure is accounted for and effective values of Δ J are used.